Diversity in the Work Force, Volume 14 (Research in the Sociology of Work) (Research in the Sociology of Work)

CONTENTS LIST OF CONTRIBUTORS

ix

WORKFORCE DIVERSITY: WHY, WHEN, AND HOW Corinne Post and Nancy DiTomaso

1

SECTION I: SOCIAL NETWORKS WHO BENEFITS? GENDER DIFFERENCES IN RETURNS TO SOCIAL NETWORK DIVERSITY Lisa Torres and Matt L. Huffman

17

NETWORKS OF INCLUSION AND EXCLUSION IN THE ECONOMIC CONCENTRATIONS OF ASIAN INDIAN IMMIGRANTS IN NEW YORK AND LONDON Maritsa V. Poros

35

IF I KNOW IT, WILL I SHARE IT?: THE POTENTIAL EFFECTS OF GROUP COMPOSITION ON THE CREATION AND STABILITY OF ORGANIZATIONAL KNOWLEDGE Melvin L. Smith

63

THE IMPACT OF SOCIAL CAPITAL ON AFRICAN-AMERICAN AND WOMEN SURVIVORS OF ORGANIZATIONAL DOWNSIZING Rochelle Parks-Yancy

87

v

vi

SECTION II: MANAGING DIVERSITY FOR THE BENEFIT OF ALL PREDICTING AFFIRMATIVE ACTION ATTITUDES: INTERACTIONS OF THE EFFECTS OF INDIVIDUAL DIFFERENCES WITH THE STRENGTH OF THE AFFIRMATIVE ACTION PLAN David A. Kravitz and Stephen L. Klineberg

107

BETWEEN SOLIDARITY AND INDIVIDUALISM: COLLECTIVE EFFORTS FOR SOCIAL REFORM IN THE HETEROGENEOUS WORKPLACE Orly Lobel

131

HIERARCHY-ENHANCING VERSUS HIERARCHY-LEVELING PERSPECTIVES IN ORGANIZATIONS: VALUING VALUE DIVERSITY IN THE NEW ECONOMY Wayne Eastman

165

SECTION III: PROFESSIONALS AND WORK-FAMILY LINKS WOMEN, MEN, CAREER AND FAMILY IN THE U.S. YOUNG PHYSICIAN LABOR FORCE Susan W. Hinze

185

MOTHERHOOD AND CAREER COMMITMENT TO THE LEGAL PROFESSION Jean E. Wallace

219

THE MARRIAGE ADVANTAGE FOR MEN IN SCIENCE AND ENGINEERING ORGANIZATIONS Corinne Post, George F. Farris and Rene Cordero

247

vii

SECTION IV: CONCLUSION: A FRAMEWORK REGARDING THE EFFECTS OF WORKFORCE DIVERSITY UNDERSTANDING THE EFFECTS OF WORKFORCE DIVERSITY ON EMPLOYMENT OUTCOMES: A MULTIDISCIPLINARY AND COMPREHENSIVE FRAMEWORK Sheryl Skaggs and Nancy DiTomaso

279

LIST OF CONTRIBUTORS Rene Cordero

School of Management, New Jersey Institute of Technology, USA

Nancy DiTomaso

Department of Management and Global Business, Rutgers Business School – Newark and New Brunswick, USA

Wayne Eastman

Department of Business Environment, Rutgers Business School – Newark and New Brunswick, Rutgers University, USA

George F. Farris

Department of Management and Global Business, Rutgers Business School – Newark and New Brunswick, Rutgers University, USA

Susan W. Hinze

Case Western Reserve University, USA

Matt L. Huffman

Department of Sociology, University of California, Irvine, USA

Stephen L. Klineberg

Department of Sociology, Rice University, USA

David A. Kravitz

School of Management, George Mason University, USA

Orly Lobel

Visiting Lecturer, Yale Law School, USA

Rochelle Parks-Yancy

Department of Business Administration, Jesse H. Jones School of Business, Texas Southern University, USA

Maritsa V. Poros

Department of Social Sciences, Illinois Institute of Technology, USA

Corinne Post

Lubin School of Business, Pace University, USA ix

x

Sheryl Skaggs

School of Social Sciences, The University of Texas at Dallas, USA

Melvin L. Smith

Weatherhead School of Management, Case Western Reserve University, USA

Lisa Torres

Department of Sociology, George Washington University, USA

Jean E. Wallace

Department of Sociology, The University of Calgary, Canada

WORKFORCE DIVERSITY: WHY, WHEN, AND HOW Corinne Post and Nancy DiTomaso ABSTRACT In this introduction to the volume on Workforce Diversity, we review some of the critical issues that need to be addressed in thinking about and doing research on workforce diversity, and then we discuss where the topics addressed in this volume fit within the larger literature on diversity at work and how the contributions in the volume advance our understanding of workforce diversity. By necessity this is a brief overview. The papers in this volume elaborate some of the issues, but a great deal more needs to be done to fully understand how diversity both affects and contributes to workplace outcomes. We conclude with the observation that the issue of diversity in the workforce is inherently a political issue, and hence one where interests are often in conflict and sometimes contradictory. The challenge for managing diversity in the workforce is the ability to bridge such political differences, and so far, few organizations have met that goal. While social scientists have always been concerned with inequality in the labor market, the attention to “workforce diversity” emerged only when literature aimed at a management audience argued that the changing composition of the workforce portended grave consequences for employers who ignored such issues. As one popular piece told managers, “demography is destiny” (Johnston & Packer, 1987). Such hyped appeals clearly overstated both the extent and impact of demographic Diversity in the Workforce Research in the Sociology of Work, Volume 14, 1–14 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0277-2833/doi:10.1016/S0277-2833(04)14013-2

1

2

CORINNE POST AND NANCY DITOMASO

diversity in the workplace (see critique by Friedman & DiTomaso, 1996), but even so, claims about the inevitability of workforce diversity created a new policy agenda for many large employers, if not for the workforce as a whole. There is no doubt that the U.S. labor force has become more diverse in terms of race, ethnicity, gender, and other demographic or cultural dimensions. Such diversity, however, is neither sudden nor unprecedented, and it may not have the constraints or the outcomes that some have claimed are likely. The proportion of adult women in the paid labor force has been increasing for at least a century, with a slight acceleration in the post-World War II period, but there is not much expectation for continued growth in the future (Bergmann, 1986; Friedman & DiTomaso, 1996; Padavic & Reskin, 2002). The growth in the population among African Americans has been slightly higher than that for Whites, primarily because of a declining birth rate among Whites, but Black labor force participation rates and employment continue to trail that of Whites. And, immigration, especially among Hispanics and Asians, has accelerated since the 1965 changes in U.S. immigration laws reduced the barriers to entry. Although the rate of increase among Asians is faster than that of any other group, the small proportion of Asians in the population at the beginning of this period means that their proportion in the labor force is still small and likely to be concentrated on both coasts and in larger cities. There has also been a rapid increase in the proportion of Hispanics entering the U.S., both legally and illegally, such that the proportion of Hispanics in the U.S. population is now slightly higher than the proportion of Blacks. Hispanics, however, have often entered the U.S. labor force in low skilled and low paid jobs, and their vulnerability in the labor force because of questions about their immigration status has tended to dampen any efforts for them to mobilize politically. All of these changes in the U.S. population and the labor force, with perhaps the exception of the labor force participation rates of adult women, have been marginal changes, and hence, White Americans are still and will continue to be for many years to come the largest group within the labor force (Friedman & DiTomaso, 1996). Unless something dramatic changes, White Americans are also likely to be found disproportionately in high status and higher paid jobs, with more advancement opportunities and more access to benefits and favorable working conditions than others. The favorable economic circumstances of Whites compared to others in the labor force has contributed to an improvement in their standard of living, a decline in the birth rate among Whites, and consequently, extremely slow population growth among Whites that has led employers to look for new sources of labor, both in the U.S. and abroad. Women were enticed into the paid labor force because of a rise in the opportunity costs of staying home (Bergmann, 1986). Immigrants were enticed into the labor force by reducing the barriers to entry that had existed in U.S. policy. In both cases, these factors could shift, and indeed, they

Workforce Diversity

3

have shifted in the last several decades as employers have sought an available and lower paid labor force by contracting out both in the U.S. and abroad. The structure of the labor force and the role of diversity have also been affected by the changing nature of work and organizations (DiTomaso, 2001). As work has become more specialized, interdependent, and complex, and as more and more companies depend on the ability to create new and innovative products in order to compete in global markets, new demands have been made on employees to learn to work together in shifting and diverse coalitions of people who may work inside or outside the company and across both geographical and legal boundaries. The growing need for complex problem solving in project based work requires a type of skill in managing diversity that was not necessary when the economy was dominated by large manufacturers producing for a mass market. That sort of workplace required careful planning, effective controls, and creative marketing. The current, innovative workplace requires good people skills, effective information systems, and often requires cross-cultural competence. Thus, many work organizations have accepted that there are high stakes in understanding and managing the diversity in their workforce, given the changes in both labor force composition and the nature of work. The “business case” for diversity, as the argument regarding improved organizational performance has been called, can be conceptualized broadly or narrowly. It can be defined narrowly in terms of the benefits associated with well-managed diversity: higher stock price, profit levels, growth, or innovation, and research suggests that those organizations that are better at managing diversity are also more profitable (Huselid, 1995; Huselid et al., 1997). At the same time, there are costs associated with not managing diversity well. Most important in this regard is the withdrawal of effort that is likely by employees who feel that they are not fully able to participate in an organization. Hence, those organizations that are not effective in managing diversity well are not only likely to face higher costs in production of goods or services, but they are also likely to lose out on attracting a diverse customer base, to have difficulty tapping into a wider pool of talent, and to fail in generating the kind of creativity that is necessary to stay ahead of the competition. Further, diversity efforts may generate conflict in the workplace, because some employees may resist the efforts of their organizations to address diversity issues (Lynch, 1989, 1997). In this case, organizations have to weigh the relative costs and benefits of attention to diversity. While for some this is purely a business issue, for others it is also a broader, ethical issue (it’s the right thing to do). As some corporate executives have become aware of how the actions of their own companies have contributed to inequality outside of their walls, they have come to believe that they have a responsibility for how their own policies and practices contribute to social outcomes (DiTomaso & Hooijberg, 1996).

4

CORINNE POST AND NANCY DITOMASO

Consideration of such issues, however, raises questions about what the goals of diversity management are within organizations. Is the intent to reduce inequality and to create a more equitable distribution of opportunities and rewards (changing the rules of the game), or is the goal to simply open up the competition to more players. In the former case, the broader distribution of rewards may mean that some people have to take less. In the latter case, competition should become more intense, rather than less, if more people are allowed to compete and the rewards themselves are not expanded in scope. While the answers to these questions may be quite varied, it is important to acknowledge that there are differences of opinion and that not everyone supports the goal of greater equality (even of opportunity). Hence, what some people claim as the results of managing diversity well may be difficult to obtain, in part, because what key actors had in mind when they lent their support to managing diversity was not what others were trying to accomplish. Thus, one of the reasons that diversity may be so hard to manage well is because different people have different assumptions about what it is they are trying to accomplish and for whom. Further, there may be contradictions in the management of diversity in the workplace such that what is needed to accomplish a goal of greater equality may be incompatible with what is necessary to accomplish a goal of greater equity. Addressing issues of diversity in the workforce, therefore, involves an examination of the composition of the workforce (opening the game to everyone), both in the country and in a given organization, as well as attention to the power and status differences that exists between and among groups (how the rules of the game advantage some rather than others). Those concerned primarily with composition have given attention to issues of how people get along with each other, learn to cooperate, share information, effectively communicate, and develop a sense of commonality and purpose. Those concerned primarily with issues of power and status differences have given attention, instead, to perceptions of fairness or justice, to the distribution of rewards and opportunities, and to the patterns of inclusion or exclusion that have taken place among groups within organizations or in the society at large. Those who want to manage diversity well must learn skills for effectively communicating, working with teams, managing processes, and building coalitions. For companies, it involves managing interdependence, while encouraging creativity and innovation, and successfully working across boundaries (e.g. in the development of joint ventures, alliances, mergers and acquisitions, and networked organizations). Further, managing diversity well requires the ability to foster “swift trust” (Meyerson et al., 1996). One of the complexities of managing diversity well is that creativity and innovation require diversity among organizational participants. Increasing diversity, almost by definition, will increase differences of opinion and outlook, and these are what contribute to creativity in ideas, but if organizational participants

Workforce Diversity

5

are not able to resolve the conflict sufficiently that participants are willing to listen to each other and draw from and use the ideas of those who are different from them, then the result can be creativity that is never implemented, lost opportunity, time, and money. Without diversity in the composition of the workforce, organizations may reduce conflict, but in doing so, they will also reduce innovation. Hence, organizations that depend on innovation, of necessity must recruit a more diverse workforce, but they also must provide the organizational participants with the skills for managing diversity in order for them to work through the conflict and, thus, reap the benefits of creativity and increased productivity. Figure 1 provides a graphic illustration of these seemingly contradictory effects of diversity in organizations. When should organizations seek the benefits and be willing to incur the costs of having a diverse workforce? There are several criteria that are supported by the research on diversity in the labor force. First, organizations should seek diversity when they are dependent on innovation, and that implies when work tasks are complex and require complex problem solving among interdependent organizational participants. Organizations should also seek diversity when the pool of available workers from which they have been hiring is insufficient for them to fill necessary jobs with qualified people at reasonable costs. And, organizations should seek diversity when there is evidence of untapped potential (for example, when

Fig. 1. How to think about the Trade-Offs Among Productivity, Conflict, and Innovation in Organizations with Variation in Diversity and in Skills for Managing Diversity Well.

6

CORINNE POST AND NANCY DITOMASO

their competitors are doing better than they are). Organizations also should incur the costs of diversity and seek the benefits when organizational leaders understand the demands of work and organizations in the 21st century. Hence, diversity is not only a product of government regulation and appeals to good will. There hasn’t been much government enforcement of anti-discrimination policies since the early days of the Reagan administration, although there has been an occasional and notorious lawsuit that has shown up in the news for especially egregious acts of discrimination and hostility within a company toward minority or female workers. Nor is the need to give attention to diversity only a result of demographic change. The changes have been rather marginal, and Whites are still a majority in most parts of the country and in most companies, especially in good jobs. Companies should, though, welcome and learn to manage diversity because of their need to do business in a global economy, and especially because of the changing nature of work and organizations toward flatter, more flexible, diverse, global, and networked organizations (Ancona et al., 1998). Ironically, though, at the same time that employers were being told that they had to accept and develop policies regarding workforce diversity, there were new and organized efforts to undo the policies of the Civil Rights Movement that were intended to increase access and provide greater rewards for those who had been previously excluded from good jobs. For example, as early as 1978, the use of race in admissions decisions of selective universities was challenged in the Bakke case, and many subsequent cases in both university admissions and employment have been brought since (http://www.acenet.edu/bookstore/ descriptions/making the case/legal/digest.cfm, http://www.acenet.edu/bookstore/ descriptions/making the case/legal/legal overview.cfm). Of course, the renewed attention to workforce diversity in the context of political efforts to roll back or restrict the policies of the Civil Rights Movement contributed to increased controversy among those studying workforce diversity about the political ramifications of such efforts. For example, Thomas (1990) argued that “affirming diversity” was a distinct and more business savvy policy than affirmative action, which he argued was neither necessary nor effective in most workplaces. Others argued that processes of exclusion still characterized most workplaces and that continued vigilance and support from public policies such as affirmative action were both necessary and fair when substantial inequality in wages, authority level, and access to jobs still exist (Bergmann, 1996). This sort of political controversy about the role of diversity in the workplace contributed to the kinds of questions that have subsequently been addressed in research, including the extent of representation in the labor force of workers of different kinds (who holds what kinds of jobs, at what levels, and with what rewards); how people get their jobs and the extent to which social networks come

Workforce Diversity

7

into play in recruitment and hiring; how people move through the hierarchy in organizations and in their careers and whether there are biases associated with being selected for moving up and through organizations; what contributes to successful performance on the job and how people position themselves for performing well; factors that either hinder or facilitate success on the job; and the impact of diversity in interpersonal relations, group process, and organizational performance. In many cases, these are empirical questions that lend themselves well to research. In other cases, however, these are political questions that depend on the vantage point and the assumptions that are brought to bear, as is evidenced perhaps most markedly, by the research on affirmative action. For example, have the competitive pressures on companies eroded any tendency toward discrimination in hiring, as Thomas (1990) claims, or are there other factors that contribute to unequal access and unequal rewards (Bielby & Baron, 1986; Tomaskovic-Devey & Skaggs, 1999)? Further, do such competitive pressures force companies to offer employees fair treatment and decent working conditions, or on the contrary, have employers faced with competitive pressures undertaken a race to the bottom that entails outsourcing, reduced security, and lower wages (Gordon, 1996)? It may be that, depending on the circumstances, companies do both. They endeavor to recruit employees they believe can serve their needs, and sometimes this means providing rich compensation or benefits, and sometimes even hiring bonuses or incentive pay, but at the same time, companies often adopt technologies that eliminate jobs or they contract out for those who can provide the same skills at lower cost. The mix and the trends depend a great deal on what the socio-economic environment is in which the company is doing business, and the extent of political organization and support in the legal system for redistributive policies that force companies to share the wealth. The set of articles in this volume address a range of diversity issues that capture some of this complexity. One of the volume’s strengths’ is that it includes both papers that explore the outcomes for individuals of having specific demographic characteristics and papers that consider group- and organizational-level consequences of diversity. In this volume diversity is conceptualized in the broadest sense to include individual differences in race, gender, ethnicity, family characteristics, and political ideology. Another positive feature of this volume are the contributions pertaining to everyday mechanisms through which inequality is reproduced. Some of these everyday mechanisms are embedded within the structure of group-based relational ties. Several papers in the volume, for example, examine the role of social networks in explaining patterns of inequality in workplace opportunities, such as access to information about job openings, and challenges, for example in the access to instrumental and emotional support following layoffs. The articles in this volume are grouped into four sections. The first section brings together papers that discuss a critical aspect of how people get jobs, specifically

8

CORINNE POST AND NANCY DITOMASO

their access to social networks or social capital. The articles in this section explore the effects of diversity on differential access to social networks or social capital in terms of access to job information; the implications of the make-up, transformation, and break-up of occupational niches; knowledge exchange; and how access to social networks are affected for women and minorities in the context of organizational downsizing. The second section explores the more substantive issues of what happens when employees are already on the job, and it broadens the discussion by considering whether and how the fragmentation of the workplace into an increasing number of distinct special interest groups constitutes threats or opportunities for others. The third section is comprised of research articles on the links between work and family among three groups of professionals: lawyers, medical doctors, and scientists and engineers. To conclude the volume, a review paper lays out the various bodies of research that have addressed the issues of workplace composition and diversity and then integrates them into a coherent model. In the following, we briefly describe the content of each of the articles in the volume.

SECTION 1 – SOCIAL NETWORKS The first section discusses the impact of networks in finding a job and doing well in the workplace. With the growing interest in social capital (Burt, 1992; Lin, 2001; Mouw, 2003), the issues of networks, their composition, and their impact have been of keen interest within both management and sociology. The section includes results from a study by Torres and Huffman on the returns to network diversity. They argue that job information is selectively transmitted and affected not only by the sex of the informant, who may have access to different types of job information, but also by the sex of the recipient. Drawing on social closure theory, the authors argue that because women tend to be excluded from circles of influence and information sharing that men are part of, they are less likely than men to obtain information on good jobs. Further, drawing on status expectancy theory, the authors argue that better job opportunities are reserved for men by both men and women because men constitute a high-status group in the workplace. This process of opportunity hoarding suggests that men are not hurt by having diversity in their networks (i.e. by having women in their networks), and that women benefit from including men in their networks. In the second article in this section, Poros examines how different types of social ties among Asian Indian immigrants in both London and New York help explain the extent to which economic behavior is rooted in co-ethnic social networks or develops outside of them. She identifies three types of social ties: interpersonal (among family members and close social relations), organizational (primarily developed through ethnic organizations), and composite (ties that are

Workforce Diversity

9

both organizational and family related). The paper outlines implications for co-ethnic based workforce diversity in industry or occupational niches. Poros concludes that there is diversity within immigrant niches and that depending on where and how immigrants enter the immigrant community and tap into social resources within it, their labor market experiences and outcomes may vary. Smith’s article then examines what happens once employees are on the job. He specifically discusses how demographic dissimilarity and demographic imbalance in work groups can be detrimental to knowledge sharing. His analysis highlights the role of numerical majorities/minorities in groups, the role of categoricallybased status dissimilarity between group members in information sharing, and the effect of composition on different types of knowledge that are likely to be shared in groups. Implications are drawn as to how diverse teams can better be managed to increase the likelihood of information sharing by and with minorities and women in diverse teams. Concluding this section, Parks-Yancy’s paper considers the importance of diversity in workplace social networks for women and African Americans, who find emotional support in relationships with people like themselves and find instrumental support (i.e. opportunities for advancement, learning about new jobs, etc) through Whites and men. The patterns for women and minorities contrast with those of White men who do not need diverse networks because they can access emotional and instrumental support from other White men (who are more likely to have valuable social resources at their command). Given the importance for advancement and mobility of having strong relationships with Whites and men, Parks-Yancy discusses how layoffs will disrupt the social networks of African Americans and women and, thus, how losing a job will change the types of support available to them. There has been a flurry of recent research on social capital and social networks and the effects of both on job outcomes such as the status of the job obtained and income. So far the research has found limited effects, perhaps, in part, because the use of social contacts is so prevalent for everyone.

SECTION 2 – MANAGING DIVERSITY FOR THE BENEFIT OF ALL The second section of the volume discusses the implications and opportunities of diversity management in the relationships among groups and between employees and employers. Diversity management, especially under the name of affirmative action, continues to be controversial in the public opinion and unavoidably, in the workplace. Kravitz and Klineberg’s paper considers how attitudes toward affirmative action are shaped simultaneously by the intensity of the affirmative action program under consideration (the strength of the preferential treatment

10

CORINNE POST AND NANCY DITOMASO

toward minorities), and by individual differences in racism, political orientation, and self-interest, defined as the negative impact people feel they will personally, or collectively incur as a result of affirmative action programs. Hence, this paper draws attention to the importance of also considering how diversity management programs (e.g. affirmative action programs) are perceived. The section also includes an analysis by Lobel, in which she urges us to take a step back and think of some of the consequences of taking an “individualistic” approach to fairness in the workplace. She raises the question of whether efforts directed at ensuring fairness and equal opportunity for a number of minority groups, have diluted more basic efforts of fairness in employee-manager relationships in general. She cites the example of childcare, which has not been able to garner much attention in workers’ bargaining positions specifically because it has been framed as a women’s issue, despite having larger implications for others in the workforce as well. She then develops an alternative framework for thinking about diversity issues. While the management of diversity has had the detrimental effect, according to Lobel, of drawing attention and energy away from concerns with “universal” employee rights, Eastman’s paper suggests that one way to bring the employeemanagement debate back into the picture is by openly allowing for debate on contradictory values held between groups he labels as “traditionalists” and “innovationists.” These groups hold different values when it comes to acceptable levels of hierarchy, inequality, and flexibility in work routines. Eastman argues that both managers and employees hold a mix of values that may highlight hierarchical differences on the one hand or reduce such differences on the other. He further argues that organizations that allow such differences to come to the fore and to be debated within and across groups in the organization can actually suppress the sort of conflict that may create a sense of difference or hostility and can at the same time reduce an unhealthy opportunism that may undermine workplace productivity and harmony.

SECTION 3 – PROFESSIONALS AND WORK-FAMILY LINKS Given the importance of the increase of adult women in the paid labor force and the association of women with childcare, the issue of work-family balance is an inevitably important issue in discussions of workforce diversity. This section brings together research articles that explore the links between the work and family spheres for work in high technology and high knowledge occupations. There are papers that examine the balance of work and family in three groups of professionals: lawyers, physicians, and scientists and engineers. The section starts

Workforce Diversity

11

with Hinze’s analysis of the changing work patterns and family patterns of male and female physicians in the U.S. (i.e. one of the most highly paid occupations) in light of the slowly closing gender gap in earnings among physicians. The paper draws on and analyzes census data from three time periods: 1980, 1990, and 2000. Hinze’s paper also documents the work and family patterns among dual-physician families, a phenomenon that is on the rise. Implications are discussed for organizations that employ physicians. Wallace’s paper explores whether career commitment is contingent on motherhood status of female lawyers. Starting with the largely unverified claims that motherhood explains women’s lag in career success among lawyers, in part, because it presumably decreases their professional commitment, Wallace compares mother and non-mother lawyers with regard to their career commitment. Using interviews and survey data, she considers the role of time norms and career norms in the law profession and examines to what extent adherence to these norms by both mothers and non-mothers can help explain differences in their level of commitment to the profession. Wallace finds that contrary to common wisdom, it is women without children who are more likely to take time off from their jobs and who appear to be less committed to their careers. Implications are drawn for the heavy reliance of the law profession on adherence to time and career norms for evaluating the commitment of female lawyers. This section concludes with Post, Farris, and Cordero’s analysis of the contingency of married men’s advantages at work. Taking as a starting point the well-documented advantages of married men in science and engineering work settings, the authors explore whether married men’s advantages in promotability are contingent on the organizational characteristics of the research and development (R&D) organizations in which they work. Using data from a survey of 3163 scientists and engineers and 690 managers employed in the R&D laboratories of 24 U.S.-based private, multi-billion dollar firms, this paper also takes into account the intermediary processes of advantage in everyday practices such as the allocation of favorable work experiences that help explain promotability for married men compared to others.

SECTION 4 – CONCLUSION: A FRAMEWORK REGARDING THE EFFECTS OF WORKFORCE DIVERSITY Skaggs and DiTomaso’s review article develops a conceptual framework for understanding the effects of workforce diversity on employment outcomes. The article draws on literatures from sociology, psychology and management, including

12

CORINNE POST AND NANCY DITOMASO

research on such topics as on stratification, cognitive processes in intergroup relations, and diversity and organizational demography. The authors argue that we need to consider the effects of power (access to valued and scarce resources), status (relationships of deference among groups) and numbers (the composition and average characteristics of groups) in order to understand the effects of diversity in the workplace. Their framework links the structure of groups to the mechanisms that contribute to both everyday and long term inequality. The analysis considers, for example, the role of exploitation and opportunity hoarding, legitimating myths and the process of social identity formation, and competition and threat in intergroup relations.

SUMMARY The efforts of some organizations to address issues of workforce diversity are often quite varied, given the broad issues that are encompassed by diversity as a concept. Some have limited their efforts to announcing their commitment to follow the law and to not discriminate. Some announce their intention of being “an affirmative action employer.” And some go just one step further in engaging in what Lynch (1989) calls “folkloric multiculturalism,” that is, attention to variations in customs and holidays, but little substantive attention to programs and policies that may affect group outcomes. Many organizations over the last several decades have undertaken ways to expand the pool from which they recruit, and many have even encouraged students to choose certain kinds of career goals and have helped provide scholarship money and opportunities for internships and exposure to the kinds of work done in their companies. All of these efforts are laudatory and undoubtedly contribute to an atmosphere of inclusiveness that provides some comfort to those who otherwise might feel excluded. But very few organizations address issues of power and inequality, and virtually none have made a commitment to equality of outcomes or to substantial changes in the way business is conducted in the U.S. Thus, there has been a great deal of attention to awareness of diversity, some training to improve the acceptance of a more diverse workforce, calls for greater cooperation and intergroup harmony, and some training in practical skills such as intercultural communication, team building, and managing change. By addressing these sorts of issues, companies have assumed that such efforts are going to have an effect on motivation, productivity, and profits, but the evidence is still mixed (Kochan et al., 2002) and depends, it seems, on providing skills to effectively draw from the resources that diversity provides while minimizing the costs. Perhaps one of the reasons that diversity programs have been so disappointing in practice, though, is because so few of them address the issues of power and inequality in

Workforce Diversity

13

the workforce. To the extent that programs to manage diversity empower people to make effective demands to serve their needs, then those who have withheld effort may begin to trust the organization and extend themselves both with their effort and with their ideas. Similarly, to the extent that attention to diversity changes how resources are distributed and creates an environment where those who had been favored come to see the benefits of redistribution, then diversity management may lead to both greater harmony and greater creativity and productivity. But, these are rarely what are meant when workforce diversity is addressed either as a problem or as a challenge. Hence, the topic still engenders controversy and for some, failed expectations. Managing diversity well requires political skills and political outcomes. It requires attention to people and to groups, to how they interact, and how they command and use resources. It requires attention to inclusion and exclusion and to sharing information, ideas, and inspiration. And, it requires consideration of such issues at different levels of analysis, across different types of diversity, and at different time periods and settings. In other words, diversity is diverse. In a volume such as this one, only a small number of issues can be addressed, but we hope that these selections contribute to our understanding of how and when diversity in the workplace is an issue that needs to be considered.

REFERENCES Ancona, D., Kochan, T., Van Maanen, J., Scully, M., & Westney, E. (1998). Managing for the future: Organizational behavior and procedures. Cincinnati, OH: South-Western College Publishing. Bergmann, B. R. (1986). The economic emergence of women. New York: Basic Books. Bergmann, B. R. (1996). In defense of affirmative action. New York: Basic Books. Bielby, W. T., & Baron, J. N. (1986). Men and women at work: Sex segregation and statistical discrimination. American Journal of Sociology, 91, 759–799. Burt, R. (1992). Structural holes. Cambridge, MA: Harvard University Press. DiTomaso, N. (2001). The loose coupling of jobs: The subcontracting of everyone? In: I. Berg & A. L. Kalleberg (Eds), Sourcebook on Labor Markets: Evolving Structures and Processes (pp. 247–270). New York: Plenum Press. DiTomaso, N., & Hooijberg, R. (1996). Diversity and the demands of leadership. Leadership Quarterly, 7(2), 165–190. Friedman, J., & DiTomaso, N. (1996). Myths about diversity: What managers need to know about changes in the U.S. labor force. California Management Review, 38(4), 54–77. Gordon, D. M. (1996). Fat and mean: The corporate squeeze of working Americans and the myth of managerial downsizing. New York: Martin Kessler Books. Huselid, M. A. (1995). The impact of human resources management practices on turnover, productivity, and corporate financial performance. The Academy of Management Journal, 38, 635–672. Huselid, M. A., Jackson, S. E., & Schuler, R. S. (1997). Technical and strategic human resource management effectiveness as determinants of firm performance. The Academy of Management Journal, 40, 171–188.

14

CORINNE POST AND NANCY DITOMASO

Johnston, W. B., & Packer, A. H. (1987). Workforce 2000: Work and workers for the 21st century. Indianapolis: Hudson Institute. Kochan, T., Bezukova, K., Ely, R., Jackson, S., Joshi, A., Jehn, K., et al. (2002). The effects of diversity on business performance: Report of the Diversity Research Network. Unpublished manuscript. Lin, N. (2001). Social capital: A theory of social structure and action. New York: Cambridge University Press. Lynch, F. R. (1989). Invisible victims: White males and the crisis of affirmative action. New York: Greenwood Press. Lynch, F. R. (1997). The diversity machine: The drive to change the “White male workplace”. New York: Free Press. Meyerson, D., Weick, K., & Kramer, R. M. (1996). Swift trust and temporary groups. In: R. Kramer & T. Tyler (Eds), Trust in Organizations: Frontiers of Theory and Research (pp. 166–195). Thousand Oaks, CA: Sage. Mouw, T. (2003). Social capital and finding a job: Do contacts matter? American Sociological Review, 68, 868–898. Padavic, I., & Reskin, B. R. (2002). Women and men at work. Thousand Oaks, CA: Pine Forge Press. Thomas, R. R., Jr. (1990). From affirmative action to affirming diversity. Harvard Business Review (March–April), 107–117. Tomaskovic-Devey, D. D., & Skaggs, S. (1999). An establishment level test of the statistical discrimination hypothesis. Work and Occupations, 26, 422–445.

WHO BENEFITS? GENDER DIFFERENCES IN RETURNS TO SOCIAL NETWORK DIVERSITY Lisa Torres and Matt L. Huffman ABSTRACT Previous research has shown that social networks can be an important source of employment information, such as job leads, for both men and women. However, few studies have considered whether women and men benefit equally from having a diverse set of personal contacts. We argue that because previous research has not considered both the sex of the person providing job-related information and the sex of the person receiving the information, one cannot determine whether returns to network diversity differ by gender. We describe a unique data set that permits such an analysis, and report findings suggesting sex differences in returns to network diversity.

INTRODUCTION Since Regents of the University of California v. Bakke (1978), legal experts, policymakers, and institutional leaders have framed the benefits to “diversity” in terms of either strategy or resource. As a strategy, diversity is a means of ameliorating race and gender inequality within institutions and organizations, particularly as formal affirmative action programs lose their legal footing (Kelly & Dobbin, 1998). As a resource, diversity contributes to organizational goals such as expanding Diversity in the Workforce Research in the Sociology of Work, Volume 14, 17–33 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0277-2833/doi:10.1016/S0277-2833(04)14001-6

17

18

LISA TORRES AND MATT L. HUFFMAN

into new markets or increasing worker productivity (Gilbert & Ivancevich, 2000; Jehn et al., 1999). Diversity is also thought to benefit organizational members, who accrue new sources of information through social interactions with those of dissimilar backgrounds and experiences (Robinson & Dechant, 1997). In this view, diversity is a source of capital from which individuals can draw. In this paper, we step back from these general claims to ask: who benefits from diversity? If, as advocates suggest, diversity serves as a resource, is this value experienced similarly among individuals? As a long history of social science research has shown, social resources and their associated benefits, are often distributed unequally across social groups (Burt, 1998). We approach our question of “who benefits” by examining the returns to diversity in one specific social context: the social networks that connect men and women to employment information. By social networks we mean that set of individuals from whom one may receive information about jobs. Such individuals include, but are not limited to, co-workers, friends, business associates, and social acquaintances. We also use the term personal contacts to refer to one’s network members. Our strategy is to link the argument of “diversity as resource” to research on social networks as a source of job leads. A substantial body of research and theory shows that social networks channel crucial information about jobs (e.g. Marsden & Hurlbert, 1988; Silliker, 1993). Additionally, superior job information is gleaned from network members or personal contacts who are less socially similar to us (Granovetter, 1974). If dissimilar others are resources that provide new, non-redundant sources of information, then it is reasonable to assume that diverse social networks are better sources of job lead information than networks that are more homogeneous (Burt, 1998). And, as implied in the “diversity as resource” argument, the benefits (job leads) from network diversity will be experienced in similar ways by both men and women. We proceed as follows: First, we discuss the overall differences in women’s and men’s social networks that relay job information, and their implications for gender inequality in the workplace. We then describe our data set and how it allows us to directly test whether diversity in network composition affects men and women differently in terms of the quality and quantity of job information. This analysis will give us some leverage in answering our specific question: whether diverse social networks are a resource that benefits men and women equally.

DIVERSITY IN SOCIAL NETWORKS Social networks are valuable resources for learning about job opportunities not always available through public channels (Granovetter, 1974), and are also useful

Who Benefits?

19

for upward mobility and other job rewards (Brass, 1985). A considerable literature addresses the link between the characteristics of an individual’s pool of contacts (sex/race composition), type of contact (such as co-worker or friend), and employment outcomes including occupational sex segregation and earnings inequality (Drentea, 1998; Huffman & Torres, 2001; Moore, 1990; Rankin, 2003; Torres & Huffman, 2002). Findings from this literature speak to the issue of how diversity of networks reduces workplace gender inequality.

How Diverse are Men’s and Women’s Social Networks? Research documents important differences between men’s and women’s social networks with respect to contact type and the quality of information provided by one’s contacts. Men’s networks tend to include mostly non-intimates such as business acquaintances and co-workers (Marsden, 1987). In contrast, women’s networks tend to include mostly family and close friends (Moore, 1990). Women’s contacts also represent a smaller subset of occupations than do men’s (Campbell, 1988). However, both men’s and women’s networks are similar in terms of sex composition; that is, both are sex-homophilous (Marsden, 1987; Straits, 1998; Wellman, 1990). Regardless of the causes of the lack of gender diversity in women’s and men’s networks, it may have consequences for gender inequality.1 Because jobs, occupations, and even organizations are highly sex segregated (Baunach, 2002; Wells, 1999), when women turn to their contacts for information about job opportunities, they are most likely to hear about female-dominated positions that are low-paying and with few opportunities for advancement. The paucity of men in women’s networks may be both a cause, and a result of, occupational and job-level sex segregation (Ibarra, 1992; Straits, 1998). For example, Hanson and Pratt (1991) report that women employed in female-dominated jobs are more likely to use female contacts to gather job-related information. Additionally, Drentea (1998) argues that when women use personal contacts in a job search, their odds of being tracked into low-wage, female-dominated jobs increase significantly.2 Since women tend to have female-dominated networks, acting on job information provided by those contacts may sustain occupational sex segregation and, consequently, associated forms of inequality such as wage disparities. If women are disadvantaged by the sex characteristics of their networks, then it is reasonable to assume that women whose networks are more sex diverse will have access to a greater variety of employment opportunities. If same-sex contacts result in negative outcomes for women, then diversity should be a resource to women, alerting them to opportunities for better-paying and more sex-balanced jobs. Indirectly, sex diversification of women’s networks will also serve as a

20

LISA TORRES AND MATT L. HUFFMAN

strategy for minimizing gender inequality. Mencken and Winfield (1999) report some evidence supporting this hypothesis – they find that women are less likely to be employed in female-dominated jobs when their networks include men. This bolsters the idea that one proximate cause of gender inequality is sex-segregated networks (Drentea, 1998; Reskin & Roos, 1990). Importantly, this type of explanation does not address the possibility that how much one benefits from diversity in one’s networks might depend on one’s sex, over-and-above the sex composition of one’s network. This is the key empirical question targeted by our analysis.

Returns to Network Diversity: Consequences for Gender Inequality In one view, men hear of more desirable job information than women simply because of network homophily – men tend to make connections with other men, who pass along information about the most desirable jobs. This male advantage could result from gender differences in attachment to the labor market – because men, on average, work longer hours than women and are employed in a wider variety of jobs, they hear about more job opportunities than women (Petersen et al., 1998). In terms of job lead quality – rather than quantity – men may also be better sources of job opportunities because of their advantaged position in the occupational structure. Thus, occupational and job-level segregation foster same-sex networks, which result in men’s networks being more useful than women’s. Therefore, the lack of diversity in both women’s and men’s networks may reproduce sex segregation on the job, thereby sustaining various forms of inequality. However, this kind of structural explanation assumes no sex difference in the returns to network diversity. In other words, a man and a woman who receive job leads from contacts of the same sex will receive leads of equal quality. This means that any gender gap in the quality of job leads is accounted for solely by who provides the leads. In this way, the influence of one’s gender on the quality of the job-related information received is indirect through factors such as the sex and status composition of social networks. Sex and other ascribed characteristics of the job lead recipient are only important because they condition the structural factors that ultimately influence the quality of job information transmitted through social networks. Gender can be viewed primarily as the basis for exclusion of women from men’s networks; however, those women who gain access to men’s networks will benefit from them just as men do. In contrast, status construction theory highlights the direct effect of ascriptive characteristics by positing that individuals always take sex (and other visible characteristics) into account when forming personal contacts and passing along information about jobs. For example, Ridgeway (1997) and Ridgeway et al.

Who Benefits?

21

(1998) describe the salience of cultural beliefs and gender stereotypes that lead to the devaluation of women’s attributes and the assumption that women are less competent than men (see also Deaux, 1985; Eagly, 1987). Ridgeway (1997) argues that the relationship between gendered organizational arrangements and inequality is mediated largely by interactions among workers, interactions that are guided by pervasive gendered status beliefs. As McGuire (2000, p. 504) suggests, because individual network resources differ by gender, others may assume that gender accurately reflects their ability and competence to perform certain work roles. Importantly, these beliefs constructed around gender may disadvantage women with respect to their social contacts. For example, women may be excluded from important work-related personal networks. Worse yet, the network rewards accruing to women may be lower than those benefiting men in similar positions in organizational hierarchies and with comparable constellations of skills, credentials, and resources (McGuire, 2000). As a result, it may be that network diversity has different implications for women and men with regard to employment. For example, even when men and women have similar networks, women may benefit less if men “hold out” on women when passing on job information. This could occur if men’s decisions about what kind of job information to give to whom are based on sex stereotype-driven perceptions about appropriate work roles for women and men. The fact that gender stereotypes about work roles are widespread and stable across time and place (Heilman, 1995; Williams & Best, 1990) suggests that any positive effects of diversity may be counteracted by other – in this case, cognitive – factors. Such factors may not be ameliorated simply by increasing the diversity of one’s network members. In sum, there are two non mutually-exclusive ways that diversity in networks could lead to unequal outcomes with respect to job leads. In one scenario, the role of gender may be indirect, primarily affecting the diversity of one’s employment networks. Due to causes that are both dispositional and structural, women and men may end up in segregated networks. This would be a net disadvantage to women, because they would have less, on average, men in their pool of contacts from which they could glean useful job information. It is important to note that in this case women would still be disadvantaged if men always pass on job information independently of the sex of the recipient of the information. Men, however, would experience no disadvantage from same-sex networks. In a second scenario, it may be that men (and women) do not ignore the sex of the contact recipient when they pass along job information. This suggests a sex difference in returns to network diversity. If men reserve their best information for men and pass along less useful information to women, any positive effects of increasing the diversity of women’s networks may be weakened. This suggests that diversity may not be sufficient for reducing inequality in the absence of outcome-based organizational policies and

22

LISA TORRES AND MATT L. HUFFMAN

practices (Konrad & Linnehan, 1995). In other words, although the importance of the sex segregation of networks is clear (Hanson & Pratt, 1991; Reskin & Roos, 1990), reducing inequality may require more than just desegregating them, given the existence of other powerful gender-based processes such as sex stereotyping. The two views of how diversity in networks impacts gender inequality described above pose a challenge for researchers. Below we describe a unique data set that we constructed which gives us insight on the issue of network diversity and its effect on gender inequality in employment outcomes. We do not purport to offer a definitive study that allows us to say with certainty which conception of the role of network diversity – additive or multiplicative – best describes reality. Rather, we hope that the analysis we present below will spur researchers to think about novel data collection methods and more carefully specify how diversity in one important context influences workplace inequality. We describe our data set and our analysis below.

DATA AND ANALYSIS Our Sample Unfortunately, researchers interested in questions regarding gender and the effects of network diversity confront a scarcity of data. Data sets used to link gender differences in important features of workers’ networks to outcomes such as sex segregation and earnings inequality (for example, see Drentea, 1998; Green et al., 1999; Hanson & Pratt, 1991) lack information necessary to address the fundamental question we have outlined: gender differences in returns to network diversity. One notable exception is recent work by McGuire (2000), who collected data with these questions in mind. However, her data describe a single organizational setting, prompting questions about generalizability. To address questions about the effects of network diversity, we collected original survey data from members of the Experience Unlimited (EU) job clubs located in California. These clubs operate under the auspices of the California Employment Development Department (EDD) and aid white-collar workers in finding employment. When we collected our data in 1997 there were 26 EU job clubs in California, representing workers from a broad range of occupations and industries.3 Individual club membership ranged from under twenty to approximately two hundred.4 Membership in EU job clubs is voluntary, free-ofcharge, and contingent upon unemployment from a layoff or other involuntary means.

Who Benefits?

23

We contacted the designated EDD coordinator of each club soliciting participation. Subsequently, each coordinator was mailed enough questionnaires to cover their estimated membership, along with instructions for completing the surveys to be administered at their next club meeting.5 Of the 26 clubs contacted, 23 participated in the study by providing us with completed surveys.6 At the individual level, we obtained surveys from approximately 60% of the 23 participating clubs’ average weekly attendance, as estimated by the coordinators. Because club membership is fluid, the denominator for calculating the individual-level response rate must be estimated. The job club coordinators reported to us that “nearly all” of those present at the meeting when the surveys were distributed completed the questionnaire. Thus, if the number present at the time of survey distribution is used, the response rate is actually higher than 60%. Our sample comprises 524 respondents, of which 324 (62%) are men, and 200 (38%) are women.7 Unfortunately, our sample includes too few racial/ethnic minorities to permit comparisons across racial lines. Although our sample is not representative of all white-collar workers in California, it includes a relatively large and diverse cross-section of male and female workers. This variety of work backgrounds overcomes a serious limitation of research on network structures and processes that often relies on data collected from a single organization or industry (e.g. Ibarra, 1992; Leicht & Marx, 1997; McGuire, 2000). In contrast, our data span diverse organizational contexts and work settings.8 One key question regarding our data is how likely is it that the self-selection into the EU job clubs produced a sample that differs substantially from the population of white-collar workers in California. Comparisons with California workers taken from the Current Population Survey (March 1997 file) suggest that those in our sample are comparable on several key variables to those in that nationally representative sample. On average, our respondents are only slightly older than comparable workers in California. In addition, our respondents report working several hours longer per week, on average, then those in the CPS. They are also less likely to be married, and more likely to have a graduate degree. Finally, although our sample is more heavily male-dominated than one would expect in a random sample, the distribution of race (white vs. non-white) is similar to that found in the CPS data set. Nonetheless, we think inferences from our sample should be made somewhat cautiously.

Measures One strength of our data set is that we asked respondents not only about their own characteristics, but also about the overall quality of information that passes to them

24

LISA TORRES AND MATT L. HUFFMAN

through their personal contacts. We asked respondents to describe up to three job leads they recently received, even if they did not act upon them. Specifically, we asked, “Over the last few years while you were employed in your last job, did a contact ever alert you to a job opening in your field, whether you pursued the job lead or not?” Although respondents were unemployed at the time of the survey, in order to exclude those leads that might have been received while participating in the EDD job clubs, we asked only about leads received while employed. This is important because the gender dynamics that occur in actual work settings are likely to differ from those found in the course of the job club activities, such as sharing job leads with other members. Respondents were also instructed to exclude job leads offered by family members, following prior research (e.g. Granovetter, 1974). Respondents were then asked to provide the annual salary of each job lead as a measure of its quality. Limiting our questions to three job leads minimizes potential recall-based measurement error. Brewer (2000) has shown that subjects generally have poor recall of weak ties, a problem that is magnified when they are asked about a large number of contacts. We designed our survey to balance two competing goals: maximizing the amount of information about job leads while minimizing recall bias. The distribution of the salary associated with each lead is important because this variable is our outcome measure in our analysis. The untransformed distribution is slightly right skewed (mean = $50,241, median = $49,500, skewness = 1.24). However, a logarithmic transformation did not substantially affect our results; therefore, for ease of interpretation, we present results based on the dependent variable in its original metric. Additionally, we asked respondents to describe the sex of each person who provided a job lead. Our data set is unique because we know not only the sex of the job seeker, but also the sex composition of her or his network. Knowing this, we are able to directly measure the amount of sex diversity in the respondent’s network. Furthermore, we are able to link network diversity to a specific outcome such as the quality of job leads received. This makes our data set well-suited for assessing the degree to which gender differences in the quality of job leads received is due to the additive component (which stresses only “who you know”) vs. the interactional component (which highlights both “who you know” and “who you are”).

Analyses From our data, we constructed a job-lead level file where the unit of analysis is the individual job lead. As noted above, we began with a sample of 524 individuals. Cases were excluded due to missing values on the recipient-level control variables

Who Benefits?

25

(described below), or because they reported receiving no job lead information. This reduced our sample to 175 individuals. Our final data file contains complete information on 274 separate job leads; we know the yearly salary for each lead as well as the sex of the provider and recipient of the lead. Of the 175 individuals in the final data set, approximately 37% reported receiving one job lead, about 26% reported receiving two job leads, and approximately 37% reported receiving three leads. We coded the four combinations of job lead characteristics as follows: (1) male job lead provider to male recipient; (2) male job lead provider to female recipient; (3) female job lead provider to male recipient; and (4) female job lead provider to female recipient. We denote these four combinations as follows: male → male, male → female, female → male, and female → female. Of the 274 leads, just over half (51.5%) were male → male. The remaining leads were distributed across the other three types nearly uniformly: 13.5% were male → female, 16.4% were female → male, and 18.6% were female → female. Men, on average, have larger pools of individuals who provide them with job leads – men are the recipients of just over twice as many job leads as women (men receive 67.9% of job leads compared to 32.1% received by women). Men also offered more leads than did women in similar proportions: 64.6% of all job leads originated with men vs. 35.1% originating from women. Our analysis is straightforward. Regressing the job lead salary on dummy variables representing the four job lead profiles described above will indicate the extent that job lead quality depends simultaneously on the sex of the provider and the recipient. However, isolating the gendered nature of the transmission of job information requires controls for worker characteristics that are associated both with job lead quality and respondents’ sex. Therefore, we control for marital status (1 = married; 0 = not married), and the presence of children under 18 years old in the household (1 = yes; 0 = no). These controls are important because previous research has demonstrated that marital status and the presence of children affect the amount of time workers spend with co-workers, which in turn influences network size and composition (McGuire, 2000; Munch et al., 1997; Wellman, 1985). Two dummy variables capture differences across three educational categories: (1) whether the respondent has a graduate degree (the omitted category); (2) whether the respondent has a 4-year college degree; and (3) whether the respondent has less than a 4-year degree (i.e. “some college”). Occupational experience is a continuous measure describing the number of years the respondent has worked in his or her present occupation (this measures general human capital). Finally, respondents were asked about the number of hours typically worked per week at their last job.9 Descriptive statistics for these individual-level the variables appear in the Appendix.

26

LISA TORRES AND MATT L. HUFFMAN

RESULTS Gross Differences in Job Lead Quality Results from our job lead-level analysis appear in Table 1. The first model does not include any respondent-level controls; therefore, this model tests the unadjusted mean differences in job lead quality across the four groups (sex of recipient × sex of lead provider). The intercept in this model ($54,409) is the average salary of male → male leads – the coefficients on the other three job lead combinations reflect mean deviations from this group in job lead quality. In terms of the overall rank ordering of the job leads, male → male leads are the highest paying, followed by female → male. Male → female leads rank third, and female → female leads rank last, offering information about the lowest-paying jobs. In short, with no respondent characteristics controlled, men always get the best leads – accrued from both men and women contacts. In contrast, women always get lower quality job leads; however, they receive better leads from men than they do from women. In terms of contrasts between male → male leads and the other lead profiles, a statistically significant difference exists between male → male job leads and Table 1. Ordinary Least Squares Regression Models Predicting Salary of Job Leads, by Sex of Recipient and Lead Provider. Variable

Model 1

Model 2

Intercept Job lead characteristicsa female → male job lead male → female job lead female → female job lead

54,409**

31,428**

−4,620 −5,363 −14,311**

−3,868 −122 −8,034**

– – – – – –

542** −9,680** −8,256** 6,507* 341** −855

0.05 5.19**

0.17 5.62**

Respondent characteristics Occupational experience Education: Some college Education: College degree Marital status (1 = married) Hours worked Children <18 years old (1 = yes) R2 F

Note: Standard errors available upon request. N = 274 job leads in both models. a Male provider, male recipient (male → male) job lead is the omitted category. ∗ p < 0.05. ∗∗ p < 0.01.

Who Benefits?

27

female → female job leads, which are associated with significantly lower-paying jobs (−$14,311, p < 0.01). Additional statistical tests, not shown in the table, provide an assessment of whether other pairs of job lead profiles are significantly different from one another. These tests indicate that two remaining contrasts: female → male leads, and male → female leads are both associated with significantly higher salaries than female → female leads (p < 0.05 for both differences). Thus, for male recipients, the sex of the lead provider does not substantially affect the quality of the leads they receive. However, women receive significantly better job leads from men than they do from women. Of course, these differences could be misstated because respondent characteristics are not controlled in model 1. We discuss the differences adjusting for those factors below.

Adjusted Differences in Job Lead Quality The quality of job leads received by individuals reflects, at least in part, characteristics such as education, occupational experience, and other factors that correlate both with gender and job lead quality. The second model in Table 1 includes controls for such respondent characteristics. Predictably, those with more than a college degree receive significantly higher quality job leads than those with some college only or a college degree. Receiving higher quality job leads is also associated with being married, working more hours, and high levels of occupational experience. Although being childless increases the quality of job leads received, this effect is not significantly different from zero (p > 0.05) when the other respondent characteristics are controlled. When individual-level variables are controlled, female → female job leads average $8,034 (p < 0.01) less than their male → male counterparts. The individual-level variables in model 2 account for 43% of the gross difference between male → male leads and female → female leads observed in model 1 (the difference is reduced from $14, 311 to $8,034). Also, net of individual-level characteristics, male → female leads are virtually identical to male → male leads (−$122, p > 0.10). Model 2 also indicates no significant difference between female → male leads and male → male leads (−$3,868 p > 0.10). Of the remaining contrasts, additional tests confirm that only the difference between female → female and male → female is statistically significant at any conventional level. Net of differences in job lead recipients’ measured human capital and other characteristics, when women receive job leads from men, they are of significantly higher quality (p < 0.05) than those received from women. Thus, all else equal, women give better leads to men than they give to women.

28

LISA TORRES AND MATT L. HUFFMAN

In sum, women benefit from having male contacts; for men, the quality of the job leads they receive are largely unaffected by who is providing them with leads.

The Effect of Network Diversity: Sex Invariant? The job-lead level analysis presented above was motivated by the question of whether the effects of network diversity are the same for women and men. In other words, do men and women benefit equally from diverse networks? The results indicate that such a difference may exist – for men, it does not matter who their leads come from. For women, network diversity does matter. Specifically, women’s job leads are substantially better when those leads come from men than when they come from women. Importantly, that difference holds in the presence of controls for individual-level factors that could account for the difference. Another way to conceptualize variability in the effect of network diversity is to examine the relationship between aggregate characteristics of one’s pool of job lead providers and overall job lead quality. In another analysis, originally reported in Huffman and Torres (2002), we report results that also speak directly to the issue of variability in returns to network diversity. In that analysis, we compute the average job lead salary across the leads provided to each respondent. Additionally, we computed the gender composition of each respondent’s pool of job lead providers, thereby measuring diversity. By modeling variation in the average salary of one’s job leads as a function of the sex composition of one’s pool of job lead providers, we are able to address whether women and men accrue similar benefits (or disadvantages) from network diversity. Figure 1 shows the results of this analysis, which is discussed fully in Huffman and Torres (2002). The plot depicts the relationship between network diversity and average lead salary, separately for men and women. The sex composition of each respondent’s job lead pool is on the x-axis; importantly, each end of the x-axis represents the lowest level of diversity – that is, all-male and all-female job lead pools. The predicted values in the plot come from a regression model that includes the individual-level controls used in our job lead-level analysis. The plot reveals a similar pattern as the job lead-level analysis. Increasing the number of men in men’s job-related networks has little effect on the quality of leads they receive. In fact, the effect of network sex composition for men is not different from zero at any standard significance level. On the other hand, the effect is more pronounced among women. Net of relevant individual-level factors, women receive significantly better quality job leads from men than they do from women. It is also interesting that in the middle of the x-axis, where network diversity is highest, the gender difference in lead quality is smallest.

Who Benefits?

29

Fig. 1. Association Between Sex Composition of Pool of Job Lead Providers and Mean Salary of Leads Received, by Sex. Note: Plot is based on a regression model that includes controls for respondent characteristics (see Table 1, model 2). Predicted values are for married college degree holders with mean values on all continuous control variables. Source: Recalculated from Huffman and Torres (2002, Fig. 1).

DISCUSSION AND CONCLUSION We began this paper by conceptualizing diversity in networks as a potentially important resource for providing job-related information, and asking who benefits from it. We addressed this question by bringing in theory and research on social networks. We argue that this research is highly salient for discussions about diversity as a means of reducing gender inequality in the workplace. Concomitantly, we have suggested that because prevailing gender stereotypes about appropriate work roles for women vs. men permeate even “diverse” organizational settings (and have remarkable staying power), increasing the numbers of women may not change patterns of gender inequality as profoundly as diversity advocates might wish. Therefore, while diversity efforts may be a laudable practice of a fair and equitable workplace, in terms of reducing social inequity, they may not be sufficient in achieving this aim. A substantial body of research shows that diversity in work groups and other contexts minimizes in-group preference, outgroup derogation, and other consequences of gender stereotypes (see Reskin, 2000 for a review). To the extent that

30

LISA TORRES AND MATT L. HUFFMAN

our findings indicate that diversity may have different effects for women and men, a more equitable workplace may require not only increasing diversity but also more proactive steps such as holding employers accountable for personnel decisions and instituting “identity conscious” human resource policies and practices (Konrad & Linnehan, 1995). Indeed, these kinds of interventions are explicitly advocated by Ridgeway (1997, p. 232), in light of the fact that modification of gender status beliefs may significantly lag behind changes in the demographic composition of workplaces. Interestingly, our findings also suggest that diversity – at least with respect to social networks – does not come at the expense of men, even though research shows that men are less likely to support diversity efforts than are women and minorities (Kossek & Zonia, 1993). In terms of the quality of job information that passes through their networks, men benefit both when their networks are same-sex and when they are sex mixed. Therefore, while men have always benefited from the gendered and segregated nature of work organizations (Acker, 1990), they also benefit from the effects of desegregation – such as in our example when networks are sex-mixed. Women, on the other hand, benefit when their contacts are diverse with respect to sex. These findings are complicated, and deserve more attention than is possible within the scope of this paper. Although preliminary, we hope that our research will inspire future data collection efforts aimed at connecting diversity efforts to measurable outcomes. With respect to social networks, we would benefit substantially from data collection efforts that also address the networks of racial minorities. Those kinds of data collection efforts would expand our understanding of the role played by diversity in personal networks for both matching workers to jobs and reproducing both gender and racial inequality in paid employment.

NOTES 1. One explanation for this lack of diversity may be the sex-segregated nature of jobs and workplace settings. Men and women have few opportunities to form mixed-sex relationships. However, it could also reflect culturally accepted notions about the appropriateness of such relationships. 2. Personal contacts include all individuals known to the job seeker, and are not limited only to “strong ties” – that is, close friends and family members (Granovetter, 1974). Following Granovetter, we exclude family members from our analysis. 3. In 1997, the unemployment rate in California ranged between 7 and 8%. Major layoffs in aerospace and other high-technology sectors were common. 4. Because memberships in EU job clubs is not static but fluctuates as members continually find new employment, it was not possible to develop a sampling frame of current members.

Who Benefits?

31

5. Although attendance at each club’s weekly or monthly meeting is strongly encouraged – in fact, it is mandatory for members – each member cannot attend every meeting. 6. When first contacted, one EU club was in the stage of disbanding and merging with another club. Therefore, from a pool of 25 viable clubs, 23 agreed to participate. 7. Respondents represent a variety of industries and services including banking, healthcare, aerospace, manufacturing, utilities, government, education, agricultural, and advertising. Our sample includes corporate vice-presidents, consultants, administrators, managers, director of sales/marketing, project managers, technical writers, program analysts, engineers of various types, environmental scientists, and several chief executive and chief financial officers. 8. A few other shortcomings of our sample should also be noted. First, because respondents are members of a voluntary organization, they may systematically differ from non-members along some dimensions. For example, some might join a job club because they have limited network connections; however, in personal interviews with club members we found many to have well-established social networks. It is also possible that people who join job clubs may in fact be better connected than the average individual as a result of the increased likelihood of belonging to other clubs as well. 9. In a supplementary analysis, we also included a variable measuring the percentage female in each respondent’s 3-digit occupational category. This variable was taken from the 1997 Current Population Survey (March file), and appended to each respondent’s record. This variable was neither significant in the multivariate models, nor did it alter the coefficients on any of the other independent variables. Therefore, we did not include it in the analysis presented here.

ACKNOWLEDGMENTS The authors contributed equally to this paper. We thank three anonymous reviewers for their helpful suggestions on an earlier version of this paper.

REFERENCES Acker, J. (1990). Hierarchies, jobs, and bodies: A theory of gendered organizations. Gender & Society, 4, 139–158. Baunach, D. M. (2002). Trends in occupational segregation and inequality, 1950–1990. Social Science Research, 31, 77–98. Brass, D. J. (1985). Men’s and women’s networks: A study of interaction patterns and influence in an organization. Academy of Management Journal, 28, 327–343. Brewer, D. D. (2000). Forgetting in the recall-based elicitation of personal and social networks. Social Networks, 22, 29–43. Burt, R. S. (1998). The gender of social capital. Rationality and Society, 10, 5–46. Campbell, K. (1988). Gender differences in job-related networks. Work and Occupations, 15, 179–200. Deaux, K. (1985). Sex and gender. Annual Review of Psychology, 36, 49–81. Drentea, P. (1998). Consequences of women’s formal and informal job search methods for employment in female-dominated jobs. Gender & Society, 12, 321–338.

32

LISA TORRES AND MATT L. HUFFMAN

Eagly, A. (1987). Sex differences in social behavior: A social role interpretation. Hillsdale, NJ: Erlbaum. Gilbert, J. A., & Ivancevich, J. M. (2000). Valuing diversity: A tale of two organizations. The Academy of Management Executive, 14, 93–105. Granovetter, M. (1974). Getting a job: A study of contacts and careers. Cambridge, MA: Harvard University Press. Green, G. P., Tigges, L. M., & Diaz, D. (1999). Racial and ethnic differences in job-search strategies in Atlanta, Boston, and Los Angeles. Social Science Quarterly, 80, 263–278. Hanson, S., & Pratt, P. (1991). Job search and the occupational segregation of women. Annals of the Association of American Geographers, 81, 229–253. Heilman, M. E. (1995). Sex stereotypes and their effects in the workplace: What we know and what we don’t know. Journal of Social Behavior and Personality, 10, 3–26. Huffman, M. L., & Torres, L. (2001). Job search methods: Consequences for gender-based earnings inequality. Journal of Vocational Behavior, 58, 127–141. Huffman, M. L., & Torres, L. (2002). It’s not only “who you know” that matters: Gender, personal contacts, and job lead quality. Gender & Society, 16, 793–813. Ibarra, H. (1992). Homophily and differential returns: Sex differences in network structure and access in an advertising firm. Administrative Science Quarterly, 37, 422–447. Jehn, K. A., Northcraft, G. B., & Neale, M. A. (1999). Why differences make a difference: A field study of diversity, conflict, and performance in workgroups. Administrative Science Quarterly, 44, 741–763. Kelly, E., & Dobbin, F. (1998). How affirmative action became diversity management. American Behavioral Scientist, 41, 960–984. Konrad, A. M., & Linnehan, F. (1995). Formalized HRM structures: Coordinating equal employment opportunity or concealing organizational practices? Academy of Management Journal, 38, 787–820. Kossek, E. E., & Zonia, S. C. (1993). Assessing diversity climate: A field study of reactions to employer efforts to promote diversity. Journal of Organizational Behavior, 14, 61–81. Leicht, K. T., & Marx, J. (1997). The consequences of informal job finding for men and women. Academy of Management Journal, 40, 967–987. Marsden, P. V. (1987). Core discussion networks of Americans. American Sociological Review, 52, 122–131. Marsden, P. V., & Hurlbert, J. S. (1988). Social resources and mobility outcomes: A replication and extension. Social Forces, 66, 1038–1059. McGuire, G. M. (2000). Gender, race, ethnicity, and networks: The factors affecting the status of employees’ network members. Work and Occupations, 27, 500–523. Mencken, C. F., & Winfield, I. (1999). Employer recruiting and the gender composition of jobs. Sociological Focus, 32, 201–220. Moore, G. (1990). Structural determinants of men’s and women’s personal networks. American Sociological Review, 55, 726–735. Munch, A., McPherson, J. M., & Smith-Lovin, L. (1997). Gender, children, and social contact: The effects of childrearing for men and women. American Sociological Review, 62, 509–520. Petersen, T., Saporta, I., & Seidel, M. L. (1998). Offering a job: Meritocracy and social networks. American Journal of Sociology, 106, 763–816. Rankin, B. (2003). How low-income women find jobs and its effects on earnings. Work and Occupations, 30, 281–301. Regents of the University of California v. Bakke. 438 U.S. 265, 318–19, 98 S. Ct. 2733, 2763, 57 L. Ed. 2d 750 (1978).

Who Benefits?

33

Reskin, B. F. (2000). The proximate causes of employment discrimination. Contemporary Sociology, 29, 319–328. Reskin, B. F., & Roos, P. (1990). Job queues, gender queues: Explaining women’s inroads into male occupations. Philadelphia: Temple University Press. Ridgeway, C. (1997). Interaction and the conservation of gender inequality: Considering employment. American Sociological Review, 62, 219–235. Ridgeway, C., Boyle, E. H., Kuipers, K. J., & Robinson, D. T. (1998). How do status beliefs develop? The role of resources and interactional experience. American Sociological Review, 63, 331–350. Robinson, G., & Dechant, K. (1997). Building a business case for diversity. The Academy of Management Executive, 11, 21–31. Silliker, S. (1993). The role of social contacts in the successful job search. Journal of Employment Counseling, 30, 25–33. Straits, B. C. (1998). Occupational sex segregation: The role of personal ties. Journal of Vocational Behavior, 52, 191–207. Torres, L., & Huffman, M. L. (2002). Social networks and job search outcomes among male and female professional, technical, and managerial workers. Sociological Focus, 35, 25–42. Wellman, B. (1985). Domestic work, paid work, and net work. In: S. Duck & D. Perkman (Eds), Understanding Personal Relationships (pp. 151–191). Beverly Hills, CA: Sage. Wellman, B. (1990). The place of kinfolk in personal community networks. Marriage and Family Review, 15, 195–228. Wells, T. (1999). Changes in occupational sex segregation during the 1980s and 1990s. Social Science Quarterly, 80, 370–380. Williams, J. E., & Best, D. L. (1990). Measuring sex stereotypes: A multination study. Thousand Oaks, CA: Sage.

APPENDIX Descriptive Statistics for Individual-Level Variables Used in the Analysis Variable

Range

Mean

Std. Dev.

Occupational experience Education: Some college Education: College degree Education: Graduate degree Marital status (1 = married) Hours worked Children < 18 years old (1 = yes)

2–30 0–1 0–1 0–1 0–1 10–75 0–1

13.60 0.32 0.41 0.27 0.31** 44.40* 0.03

8.90 0.47 0.49 0.45 0.46 13.70 0.18

Notes: N = 175 individuals who provided information about 274 job leads. ∗ p < 0.05 for two-tailed test of mean gender difference. ∗∗ p < 0.01 for two-tailed test of mean gender difference.

NETWORKS OF INCLUSION AND EXCLUSION IN THE ECONOMIC CONCENTRATIONS OF ASIAN INDIAN IMMIGRANTS IN NEW YORK AND LONDON Maritsa V. Poros ABSTRACT The role of social networks is central to the phenomenon of employment and ownership in ethnic businesses, ethnic enclaves, and more generally ethnic economies. Social capital within migrant or co-ethnic social networks is generally viewed as an aid to niche employment, in other words as processes of network inclusion. This article examines both processes of inclusion and exclusion in the social networks of Asian Indian migrants in and outside of ethnic economies. Evidence from the life histories of these migrants in New York and London allows us to see the role of social networks in producing cooperation and conflict within modes of economic inclusion and exclusion.

INTRODUCTION Cities today and historically have been magnets for immigrants from many different ethnic origins. Immigrant groups become noticeable especially when Diversity in the Workforce Research in the Sociology of Work, Volume 14, 35–61 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0277-2833/doi:10.1016/S0277-2833(04)14002-8

35

36

MARITSA V. POROS

they populate places in large numbers. However, even smaller populations of immigrant or ethnic groups become noticeable when they are concentrated geographically, socially, politically, or economically. The economic concentration of immigrant groups is apparent in a variety of occupations, industries and sectors. One group, Gujaratis, whose origins lie in the Western Indian state of Gujarat, provide an illuminating case for examining these economic concentrations and their implications for workforce diversity in the host societies of these immigrants. Gujaratis have a centuries long history of migration and diasporic settlement through trade, indentured labor, and colonial relations. Today, walking along the High Streets of London or New York’s Broadway, Gujarati Indian newsstand agents can’t escape one’s notice. The “Patel motel” in the United States has become another well-known phenomenon as Gujaratis have dominated the mid-size motel industry in the U.S. Apart from these ethnic businesses, Indians (including Gujaratis) become concentrated in particular professional occupations, including among physicians in New York and London, or software engineers in Silicon Valley and other high-tech regions. They can also be found in the public sector, concentrated in civil service positions in teaching, transportation, health, and engineering. How do these economic concentrations form? Who gains entry into these niches and who is left out of them? How do niches maintain stability over time? How do they change? Why might they break up or transform into new niches?1 A vast and growing literature has developed around the phenomena of ethnic businesses, ethnic enclaves, and more generally ethnic economies representing a variety of industries and economic sectors in the United States and Europe (see Bonacich & Light, 1988; Kloosterman & Rath, 2001; Light & Gold, 2000; Model, 1997; Waldinger, 1990, 1996). Central to this literature is the role of social networks in niche employment or ownership, and in the consequences of niche employment both for co-ethnic employees in those jobs and industries where niches exist and for the wider society. This literature has done much to demonstrate the importance of social capital within migrant or co-ethnic social networks as an aid to niche employment and ownership, in other words, as processes of inclusion. Few studies, but no less significant, have noted conflict or the negative effects of those employed in ethnic economies, but not necessarily exclusion from those economies (see Portes & Sensenbrenner, 1993). This article builds on these earlier studies by examining both processes of inclusion and exclusion in the social networks of Asian Indians in and outside of ethnic economies. Evidence from the life histories of migrants in New York and London, including their work histories, allows us to see the role of social networks in producing cooperation and conflict within modes of economic inclusion and

Networks of Inclusion and Exclusion in the Economic Concentrations

37

exclusion. The advantage of the life history data is that they show change in migrant social networks and in the outcomes they produce for ethnic economic niches. In brief, the data show that the extent and mode of embeddedness of migrants in social networks produce different consequences for niche employment or ownership through three aspects of networks: different social tie configurations, social network location and transactions within ties, and the formation of new social ties and brokers. Networks dominated by particular configurations of social tie types, specifically organizational, interpersonal, and composite ties, and the location of actors within those networks, structure the conditions of transactions that produce inclusion and exclusion from economic opportunities. These networks also change through the formation of new ties and the role of brokers in providing new opportunities. These processes result in variation in the extent and mode of embeddedness of migrants. Embeddedness here refers to the way in which economic behavior is affected by networks of social relations (Granovetter, 1985) – in this case, the way that immigrant economic behavior is embedded in social networks or relations. The inclusion and exclusion of immigrants embedded in different configurations of social ties, or more simply distinct types of networks, unevenly penetrate economic sectors. Following a methodological discussion, the above argument is developed in a theoretical discussion of network inclusion and exclusion. Then, narrative evidence from the life history interviews is presented to illustrate three different modes of embeddedness in economic niches and their consequences.

METHODOLOGY The collection of data was not specifically designed to examine particular occupational or industry niches prevalent among Asian Indians, but rather, more broadly, to understand the role of social networks in the migration flows of Indians and their subsequent social and economic incorporation into American and British society. That incorporation sometimes included niche employment and sometimes did not, and in some cases included movement in and out of niches over the life course of individuals. The sample included 56 life history interviews with Gujarati Indian immigrants and 24 interviews with association leaders and members within the broader Indian populations in New York and London. The interviews were collected in 1998 and 1999 during the author’s fieldwork. Those interviews addressed the experiences of migrants’ natal and marital family life, education, migration, work, cultural life, and economic activities. The sample focused on Indians whose origins lie in the western Indian state of Gujarat because high

38

MARITSA V. POROS

proportions of Gujaratis live in New York and London. Although Gujaratis may be considered an exceptional immigrant group because of their centuries-long history of migration owing to trade, indentured labor, and colonial labor recruitment, the narrow origins of the sample provide a strong test of whether social origins or social processes found within the transactions of network ties drive migration flows and opportunities for incorporation. The cross-national comparison of Gujaratis also demonstrates to what extent the socioeconomic context of New York and London affected Gujaratis’ migration flows and incorporation as opposed to network processes. In order to sample Gujaratis and their networks, four common surnames were used to produce a random sample recruited from public telephone directories in both cities. A short-string snowball sample was then obtained from the random sample in order to provide primary data on the social networks of the migrants in the random sample. The initial sample corresponded to the top residential areas of settlement of Indians according to 1990 U.S. Census Bureau data for the New York Consolidated Metropolitan Statistical Area (CMSA) and the Greater London area of the 1991 British Census. The interviews provided a retrospective narrative account of how these Gujaratis migrated, how their social networks changed across the course of their lives, and what kinds of transactions passed through their ties affecting their access to and use of crucial social and economic resources, such as loans, information about jobs and housing, and educational opportunities.

THEORIZING NETWORK INCLUSION AND EXCLUSION IN ECONOMIC NICHES Network studies of migration have contributed to our understanding of how migrants’ social capital – that is “the ability of actors to secure benefits by virtue of membership in social networks or other social structures” (Portes, 1998, p. 6) – or social networks as the embodiment of social capital (Faist, 2000; Massey et al., 1998), provides the necessary means to migrate and command the use of network resources in their home and host countries.2 In these studies, social networks are typically identified as relationships to kin and community. For instance, Massey et al. (1998, p. 42) define migrant networks as “sets of interpersonal ties that connect migrants, former migrants, and non-migrants in origin and destination areas through ties of kinship, friendship, and shared community of origin.” The importance of co-ethnic interpersonal ties on the subsequent economic incorporation of immigrants into their host societies has also been well investigated in studies of ethnic economies (see Light & Gold, 2000; Waldinger,

Networks of Inclusion and Exclusion in the Economic Concentrations

39

1990, 1996; Zimmer & Aldrich, 1987). Other types of social ties, however, such as to organizations or mixed ties of friends and family in organizations raise several questions about the possibility of differential effects of the embeddedness of migrant economic behavior in networks in terms of their opportunities for niche employment. This possibility leads us to question: Do distinct types of network ties affect the inclusion or exclusion of migrants in occupational or industry niches? This question is taken up below in a theoretical elaboration of how social tie configurations, network location and transactions, and the formation of new ties and brokers between ties can affect the uneven embeddedness of immigrants in different niches, thus including some and excluding similar others from these niches.

Social Network Tie Configurations: Interpersonal, Organizational, and Composite Ties The migration and social networks literatures provide somewhat different but important ways of conceptualizing social ties, which combined, allow us to better define migrant networks and examine their relationship to economic behavior. As described above, the migration literature tends to emphasize the importance of interpersonal ties for migrants’ ability to migrate and become incorporated into a new society. Family and friends provide the necessary contacts and resources for migration across borders. Similarly, the social network literature identifies friends, relatives, and neighbors (see Krackhardt, 1992) as the types of interpersonal ties that are important for gaining access to crucial resources, such as getting a job (Granovetter, 1974). While the network literature often measures interpersonal ties in terms of the amount and quality of interaction, affection and time within a tie (Krackhardt, 1992), these measures become problematic for examining migrant ties, which often span great physical distances. Nevertheless, kin, friends, neighbors, and those who share the same community are all important types of interpersonal social ties. Examples of organizational ties are ties to others in firms, schools or universities, businesses, cultural institutes, voluntary associations, and government or state agencies. Organizations actively recruit both their own members and outsiders to work, live, or study abroad. Individuals pursue overseas opportunities that become known to them through organizations. Formal aids to labor recruitment, which exist within organizations, such as interviewing for overseas jobs or taking professional certification exams in immigrant sending countries are also common. Once in the host country, ties to organizations also become important for economic and other opportunities. Some of those ties are to native institutions

40

MARITSA V. POROS

and organizations while others are to organizations created by the immigrant groups themselves. Interpersonal ties of friendship and family also exist within organizations (Heimer, 1992). However, identifying ties based on the channel through which two people are related, i.e. identifying the way two people meet or the organizational context of the relationship, can help us to distinguish how social ties operate differently as interpersonal and organizational ties. The channels through which two people know and relate to each other, at least initially, organize social ties in terms of mutual relations or interests, for example, a mother and daughter, a foreman and workers, or a university department faculty. The organizational context of a relationship mediates it in different ways, such that colleagues, co-workers, and supervisors are different sorts of friends than are people who associate as neighbors or fellow church goers. The former relationships are shaped more by their differential status in an organizational hierarchy and can act as weak ties, while the latter are strong ties (see Granovetter, 1973). In some cases, such as large or transnational family businesses, these interpersonal and organizational relations overlap in such a way that they are indistinguishable or inseparable from each other. For example, the president of a family firm might have his sons and nephews each manage separate branches of the firm. A married couple may own and operate a business together. I refer to these as composite types of social ties. Composite ties are distinct from interpersonal and organizational ties because the relations within those ties are mediated by the organization and the family or community together as a single entity. In other words, if my uncle is also my boss, these relations are indistinguishable when both roles are prominent and act in concert in transactions within the tie of nephew-employee and uncle-boss. These social tie distinctions have significant and differential consequences for migrant incorporation because interpersonal, organizational, and composite ties regulate network relations differently. Migrant networks, therefore, are composed of different configurations of interpersonal, organizational, and composite ties, which often span great geographical distances, and which link particular migrants to specific destinations and opportunities at those destinations. These different configurations of ties and the transactions that take place within them drive migration flows and social and economic opportunities in the host countries of migrants (see Poros, 2001). Shared ethnicity in these migrant networks often appears to be the glue that binds immigrants or co-ethnics together. However, one’s location in a social network dominated by interpersonal, organizational, or composite ties, which conditions transactions within ties, is more important for defining access to network resources such as niche employment.

Networks of Inclusion and Exclusion in the Economic Concentrations

41

Network Location and Transactions Niche employment involves a process Charles Tilly calls “opportunity hoarding.” The hoarding of opportunities takes place when valued resources are confined to members of an in-group (Tilly, 1998). In-group members hoard economic opportunities and niches by relying on interpersonal, organizational, and composite ties in networks that span both the home and host societies. Various resources pass through these social ties to kin, community, and colleagues, which help to sustain their occupational niches. Information about jobs, recommendations to potential employers, vouchers regarding reputation and status, and financial loans and credit are among the most important. How that “in-group” becomes defined, however, is crucial to understanding who is included and excluded from a particular niche. Inclusion and exclusion of individuals in niche employment, or hoarding opportunities for some but not for others, does not simply follow co-ethnic affinities, community membership, and other similar group constructs. We cannot take for granted that one’s membership in a social network grants that individual access to coveted resources. The immigration literature and research on ethnic economies have gone far in documenting the role of social networks in generating advantages and disadvantages for ethnic groups’ ability to locate jobs, or in explaining rates of entrepreneurship among other forms of economic activity. However, we know little about how inequality is generated within a network. The resources of a network are generally treated in a homogeneous way – equally available to its members based on shared co-ethnicity. For example, in Light’s and Gold’s (2000) critique of class theory as an explanation for variation in rates of entrepreneurship, they highlight the importance of ethnic group membership as a structure for generating economic resources. They argue that ethnic group membership both promotes an “ideology of solidarity” and the passing of specific forms of knowledge, skills and norms particular to an ethnic group, which “transcend class boundaries, uniting groups marked by different class positions and interests” (Light & Gold, 2000, p. 108). Waldinger (1995) uses a similar explanation in his examination of immigrant niches in the public sector. He finds that co-ethnic affinities are the basis for sharing information and resources that build ethnic niches in certain occupations in New York City government. Shared ethnicity or community acts as proxies for trust and predictability. These explanations leave us to wonder about which co-ethnics were excluded from this co-ethnic network mechanism. When one looks at the structure of ethnic networks, a transcendence of class and other boundaries in favor of ethnic ones is far from clear. There are significant inequalities within a co-ethnic network in which class,

42

MARITSA V. POROS

gender, caste, and other markers of identity may be played out. An individual’s location in a network is further marked by combinations of such attributes and is manifested in the context of exchange relations or transactions with other members. That is, network location affects what types of exchange occur between two members, with influence and access to resources being two of the most indicative sources of power and inequality (Marsden, 1982; see Stinchcombe, 1989, on persuasion as power). Niche employment, therefore, depends, in part, on the social location of an individual in a network, which can exclude people who have the same ascriptive characteristics of most of the network members. One’s location in a network affects the exchange of resources that can take place between two or more of its members. Put another way, network location demonstrates relative positions of power or inequality, which affect in whose favor/disfavor occupational (and other) opportunities will be hoarded. This hoarding, in turn, affects the formation of niches, who obtains access to niche employment, and to particular positions within a niche. Niches also change. The conditions of opportunity hoarding might also change as a result of the formation of new social ties and of brokers present in networks in the host society.

The Formation of New Social Ties and the Role of Brokers The formation of new social ties in the host country occurs in a variety of contexts including work, neighborhood, philanthropic activity, voluntary associations, and places of religious worship. The immigrants in this study forged new network ties in two main ways. First, immigrants sponsor family members from India and other parts of the diaspora to join them in the host country, a process known as chain migration. Second, they developed relations with other members of the Indian immigrant population through participation in the many and diverse Indian voluntary associations in New York and London. There is no doubt that Indians also formed friendships and ties with co-workers, neighbors, and others of non-Indian origin. However, sponsoring family members for immigration and participating in voluntary associations were predominant paths for new social ties because they simultaneously reinforced immigrants’ ethnic identity with other identities related to occupation, religion, residence, politics, the arts, etc. (Less typical pathways are discussed in particular narrative descriptions when relevant.) The priority of family reunification in the immigration policies of the U.S. and Britain has also greatly encouraged chain migration. Kin are, thus, an integral part of migrant networks in both the home and host countries. However, as most people know from personal experience, not all kin serve a family or network in

Networks of Inclusion and Exclusion in the Economic Concentrations

43

equal or similar ways. Siblings can drop out of favor with each other or with their elders when risky undertakings arise as often happens within the composite ties of transnational entrepreneurs. Fights can ensue over inheritance rights, dividing families. Birth order and gender can be instrumental in decisions concerning who migrates or who will assume ownership of the family business. Thus, the content or quality of kin relations affects the network location of individual family members. It cannot be taken for granted that all kin relations produce inclusive and strong relations of exchange within a network. Therefore, the family members that constitute an interpersonal network are not always structurally equivalent to each other. Some kin are favored over others for different types of economic activity. Relations of inequality often become clear on this basis. However, generally, kin play a significant if not always positive role in bridging networks between the home and host societies and in forming new networks in the host society. Apart from kin and co-ethnic relations, voluntary associations are an important conduit for establishing new ties in the host country. Hundreds of Indian associations exist in the U.S. and Britain, including numerous regional chapters within many of them.3 The sheer variety of associations in both sites is also remarkable. Many associations are culturally specific, for people who share a common regional/linguistic or caste background. These associations are especially conducive to strengthening and developing new interpersonal networks. General cultural associations are open to all Indians (pan-Indian). Typically, these associations hold cultural performances and events related to Indian art, literature, music, and dance. Also included in this category are neighborhood Indian associations, which address the local concerns of Indians living in the same neighborhood or local jurisdiction. Religious, political, philanthropic, occupational, and umbrella organizations are also numerous. While membership and participation in these types of associations are not new experiences (many such organizations exist in India, and some were transplanted to New York and London or other parts of the diaspora), the content of these associations differs by addressing newly formed interests in the host countries. For purposes of analysis, I consider new network ties formed through these associations as organizational ties. Although it is clear that some associations foster interpersonal ties, while others promote organizational or composite ties, the path to their formation is organizational. The medium of the organization shapes the kinds of transactions that take place within these newly formed ties. For example, business contacts are made in organizations as diverse as temples and industry associations. Ties and transactions between businesspeople, who meet and know each other (and others in their networks) as members of the same temple, will be hard pressed to adhere to the ethical strictures of their religious organization, such as not selling alcohol,

44

MARITSA V. POROS

in order to uphold their status and reputation. Similar business ties within other types of associations incur different kinds of constraints on the transactions and practices of their members. One example of associations as social ties mediated by organizations comes from religious associations. Temples can serve as sites where religious and economic networks meet. Recently, the South Asia Development Partnership (SADP), a philanthropic organization, which focuses on the promotion and support of Asian business in London, held an event at the large and influential Swaminarayan temple in London. In an interview, SADPs founder recalls: I held my first event there involving the business community, talking about ethics in a global context – 300 businesses turned up. Now, I’ve not seen that number of people turn up in any other network from one community group, the Gujaratis in Neasden4 . . . Now in a global business context, it was a business issue in a temple, but the temple was being used as a community hall, now that could be tremendous. . . . It’s not [yet] happening in a concerted, coordinated way. And how through any of these initiatives – how do you really convince people in a sense to trust outside sources? I think this is the point . . . I think that by building trust with the temple structure and the church structure and if the leaders establish the trust level . . . if they go to the key movers and shakers and you know identify 100 and build trust with 10, those 10 will influence. We made contact with one guy, Mr. Lal, who came to a lecture I organized with the House of Lords. He was so encouraged by that, he said I want to do something at my temple and he brought 300 people. And it’s that kind of multiple effect that’s needed. So in a sense, it’s creating new networks. Correct, it’s identifying existing networks, but it’s bridging network to network . . . moving networks and linking them.5

The language of networks is by no means foreign to this leader of the South Asian community. He fully understands and utilizes his position as a broker between Indian organizational networks. Furthermore, he realizes how the act of bridging or brokering between networks can lead to more and better economic opportunities. In many ways, however, this is a top-down strategy emanating from elite members of the Indian community. Those elites are primarily concerned with raising the economic and social status of Indians through the development of mid-size and large businesses rather than helping the family-run corner shop. Mr. Lal, incidentally, runs a “small” business and was seeking help to take the business public. To be fair, the founder of SADP recognizes the weaknesses of the top-down strategy emanating from the elite. As he expressed it: “At the moment I suspect sometimes there’s not enough pull [from the top], or there is pull at the wrong places, and there’s not enough push.”6 Most Indian voluntary associations serve multiple functions by incorporating cultural, business, and even religious and political interests on the same platform. Caste-based and regional/linguistic/religious-based cultural associations are especially conducive to this blending of interests that promotes the formation of new interpersonal ties. Occupational (especially professional) organizations,

Networks of Inclusion and Exclusion in the Economic Concentrations

45

on the other hand, tend to be pan-Indian and encourage a broader form of ethnic identity in favor of shared class and occupational interests. Occupational associations also encourage the formation of organizational ties that can lead to new and better economic opportunities. Industry-specific associations are often comprised of individuals, who work in the same industry and rely on interpersonal and composite ties to support their niche in these industries. Therefore, the structure of their networks is reinforced by their participation in industry-related associations that are ostensibly open to all Indians. Brokers, who link networks at the organizational level, such as the SADP founder above, are typically elite members of the community and play an important role in helping to direct the opportunities of association members. Brokers further play a crucial role in directing new opportunities both inside and outside of niches. In the network literature, brokers are described as weak ties linking networks which otherwise do not or would not interact with each other (Granovetter, 1973). Thus, brokers are a mechanism by which certain members of a network are conferred advantages or disadvantages (Marsden, 1982). Those who control resources in economic niches hoard opportunities for a defined group and as such provide advantages to those with access to the group. Brokers are instrumental in conferring these advantages on specific individuals or in linking networks to broaden a niche. Different types of brokers are also prevalent in occupational niches dominated by interpersonal, organizational, and composite ties. Brokers tend to correspond to the different configurations of ties prevalent in migrant networks, but also specific ties are more dominant than others. For instance, brokers mediate many of the employment opportunities of organizational networks. They are supervisors, teachers, or colleagues, who can link an individual from one organizational network to another, usually in the same field or industry. In composite networks, uncles are key brokers, but also siblings, in-laws, parents, and spouses are often the owners, managers, suppliers, and buyers, who make transnational trade happen. Niche entry and new economic opportunities in the niche are mediated by brokers who are organizationally and interpersonally linked to different segments of their trade. Brokers in niches dominated by interpersonal networks include a wider array of interpersonal ties, such as co-ethnics, friends, siblings, spouses, and parents.

The Irregularity of Embeddedness Variation in network location, changing configurations of ties, and transactions within ties in interpersonal, organizational, and composite networks suggest

46

MARITSA V. POROS

that differences in network structure affect economic sectors differently. In Granovetter’s (1985, p. 491) theoretical elaboration of embeddedness he has stressed that “networks of social relations penetrate irregularly and in differing degrees in different sectors of economic life thus allowing for what we already know: distrust, opportunism, and disorder are by no means absent.” His statement suggests that we need to identify which areas of economic life are penetrated irregularly and what are the consequences of embeddedness in different types of social networks. For instance, do some networks generate more trust or more malfeasance than others? The following narratives of Gujarati immigrants demonstrate how the economic niches represented by interpersonal, organizational, and composite networks are penetrated irregularly by social relations. This unevenness results in different sets of economic (and other) opportunities and trajectories for those immigrants embedded in these three types of networks.

NARRATIVES OF INCLUSION AND EXCLUSION Interpersonal Networks and Constrained Embeddedness Gujaratis have been quite successful at establishing themselves in industries in New York and London in which they had no prior experience either in India or East Africa (where many of London’s Indians were born or first migrated). Two economic sectors with predominant Gujarati niches in New York and London are retail trade and services. In retail trade, common industry niches in both cities include confectionery/tobacco/newsstand (CTN), groceries/food, and stationery. Small retail trade in London is overwhelmingly dominated by South Asians. Hindus make up approximately 70% of newsagents in London, most of whom belong to the Patidar caste community (Lyon & West, 1995). Of the 576 Patel (usually Patidar)7 small businesses surveyed by Lyon and West, 67% were in the CTN industry. A further 23% owned groceries, general stores, off-license (liquor) stores, and post office outlets. Fewer were self-employed professionals (12%); dentists, doctors, and pharmacists topped the list. Retail businesses are geographically dispersed across central and Greater London, thus catering to a broad mix of clients in order to maximize profits. Many small retail businesses in commercial and residential Indian enclaves (in both New York and London) sell goods that cater to a co-ethnic and larger South Asian community. However, most Indian retailers are able to pursue the economic advantages of serving the mainstream market. In New York, Logan et al. (2000) report that 22% of Indians were owners in retail and wholesale (mostly retail) in 1990. They also show that new niches in

Networks of Inclusion and Exclusion in the Economic Concentrations

47

New York have been growing in food stores, eating places, and social services. For instance, Gujaratis in the U.S. have been forming a niche in particular food franchise outlets, notably in Dunkin’ Donuts franchises (see Sun, 1998). In the service sector, Gujaratis are a significant part of the hospitality industry in the U.S., owning close to 50% of mid-size motels according to the Asian American Hotel Owners Association, but have no such niche in the U.K. (see McDowell, 1996; Varadarajan, 1999). These motel franchises are concentrated in the southern and midwestern states of the U.S. In these entrepreneurial niches, owners often employ kin and others from their community. Therefore, the owners and employees in these niches form an ethnic ownership economy, but not an ethnic enclave economy (see Light & Gold, 2000). The clientele that they serve are part of the general population. Thus, Indian businesses are located in predominantly native white, commercial and residential areas. Typically, Gujaratis cite cultural reasons for pioneering or entering a new industry in the host country. Industries that are not polluting (e.g. do not require handling food) are especially desirable, as are those that offer immediate cash flow and housing. As vegetarians, Gujaratis are keen not to enter industries where they must handle meat, fish, or alcohol. Thus, the ubiquitous curry house in the U.K. is actually a Bangladeshi niche, among whom meat, fish, and alcohol restrictions are not always as strong. Vegetarianism, immediate cash flow, and housing, often located above one’s business, are the same reasons Gujaratis give for starting CTN shops, postal service outlets, stationery stores and small groceries which only sell fruits, vegetables, and packaged frozen and dry goods. The overarching reason for starting small businesses in these industries is that they require a small amount of capital and these owners already have interpersonal ties that can help them to navigate the industry. Furthermore, Gujaratis are able to quickly mobilize this necessary capital and labor from within their interpersonal networks. In addition to the cultural aspects of pioneer niches, the socioeconomic and institutional context of the host country must be open to such entrepreneurial activities for immigrants (see Kloosterman & Rath, 2001). In the case of Gujarati niches in both New York and London, the strategy was the same. Gujaratis looked for shops to lease or buy in declining commercial areas and sometimes in declining industries. Many relied on interpersonal ties to find these empty shops and neighborhoods. Zimmer and Aldrich (1987) found that choosing a geographical site for a business depended either on formal methods or interpersonal ties. For example, a little over half of the shopkeepers in the Ealing borough of West London used formal channels to locate available shops for sale. The other 50% used information from family or they knew the owner. Newspapers such as Exchange and Mart and Dalton’s Weekly are two of the most popular sources of formal advertisement

48

MARITSA V. POROS

for the sale of London shops. These advertisements also describe the location and turnover of shops.8 Gujarati small business owners have also introduced innovative practices to their businesses, thus giving them a competitive advantage over native owners in the same industries. One notable example is parallel trading. A leader of the Asian Business Association at the London Chamber of Commerce explained this phenomenon to me. In parallel trading, a retailer may buy products manufactured in the U.K., but that are destined for sale in Kenya or some other country. Indications of this might be that a tube of toothpaste, such as the one I bought in the South Asian neighborhood of Wembley, has authorization from the Dental Associations of several African countries. Often, though, those products are exported to foreign countries for sale at a much cheaper price and then re-imported back into the original country for sale at the more expensive price.9 Thus, even small, local retailers can operate with transnational advantages. Other advantages were related to how Indian families ran their shops, and the relations or transactions within these interpersonal ties composed of shopowners and employees. One leader in the South Asian community described how his family first entered the retail business in London in the 1970s: Oh it wasn’t difficult because we had cash, even though it was small cash, a few, a few hundred pounds. We pooled it because these shops were really being underutilized. They were opening at 7AM and closing at 5 or 6 in the evening so the volume of turnover was, and there was a formula . . . So we then went in and a) we added firepower of people for free, so we had an edge. We made our margins. We opened longer, right. We brought in niche goods like Irish goods . . . and that immediately boosted our turnover . . . Technically, we were three families. But, we said look we’ll all be together, we’ve got no choice, we are brothers, sisters, nephews, all together, so 15 [people staying in] 4 bedrooms, you buy your first shop. You really get a go and six months later you can buy another shop because you’ve made the money and you’ve really not spent as much . . . You’re generating capital, which spawns the second shop. You’ve now got two . . . so literally one became two, became three, became six. And in the second shop, for example, there, the brother wanted to outdo this brother. So, he said look I’ve got a clever idea. I’ll go to Covent Garden Market, buy fresh fruits and vegetables or ethnic goods, which come in, and bring them here. And that’s got even more margin, but it means getting up at 4 in the morning. It means going to the market at 5, and coming back at 6, and offering everybody cassava and plantain and coriander that you couldn’t get otherwise.10

Many aspects of this story were echoed by other shopkeepers I interviewed in New York and London: the introduction of ethnic goods, the long, hard hours family members worked behind cash registers and stocking shelves, waking up before dawn to buy goods at wholesale cash and carry stores or farmer’s markets, and the subordination of one’s personal motives and desires for the collective motive of making the shop profitable. Most shopkeepers, however, did not have such an enormous amount of labor power to put to work in these shops as in

Networks of Inclusion and Exclusion in the Economic Concentrations

49

the story of Ram’s family above. Many shopkeepers sufficed with only three or four family members in one shop, over the course of many years, even decades. Less labor power to exploit, “firepower of people for free” as Ram put it, has kept these shops from growing and kept the workers within them from opportunities to exit the niche. More importantly, interpersonal ties, which on the one hand made these niches possible and profitable, simultaneously constrained mobility. In other words, both the networks of actors within these niches and the set of businesses representing the niche were constrained in their embeddedness in the niche and in the wider economy respectively. Even when new ties were made in temple organizations or industry associations, these organizational ties reinforced the occupational status of owners and employees in the niche. In some cases, one’s status in a temple organization, such as the Swaminarayan Hindu sect, could become intimately linked with one’s status as a shopkeeper. Therefore, exiting the niche would be akin to exiting one’s community, or at least risking giving up the practices associated with representing one’s social status. This intertwined relationship between business and community keeps some college-educated shopkeepers from pursuing other lines of employment. While these network processes include in-group members albeit in a constrained way, out-group members will be excluded from the niche because of a lack of old and new interpersonal ties to small businesses such as CTN shops. This constraint is less than the extreme embeddedness of composite networks in much riskier economic activities such as the global diamond trade, which will be discussed below. Thus, interpersonal networks behave differently in different economic sectors and industries, allowing easier but constrained niche entry and exit in small, less risky businesses such as CTN shops.

Organizational Networks and Partial Embeddedness During the first half of the twentieth century, the British began recruiting civil servants, including teachers, administrators, clerks, and lawyers, from India to the East African Protectorates of Uganda, Kenya, and Tanzania (see Gregory, 1993). In addition to their development of transport and commerce, the British had begun developing an administrative infrastructure and a formal educational system for the native African populations and the large Asian (mostly Indian) population. To fill these positions, representatives of the Ugandan, Kenyan, and Tanzanian High Commissions made regular trips to India to recruit directly from a small, educated middle class. Recruitment took the form of placing advertisements in Indian schools and newspapers, soliciting recommendations of students by Indian university educators, and conducting interviews in India. Formed from colonial

50

MARITSA V. POROS

relations, these organizational ties to personnel in recruiting agencies and formal means of recruitment such as advertisement aided young, educated, middle-class Indians, many from particular caste groups, such as the Anavil Brahmins who are prevalent in Gujarat, to migrate and form a teaching niche in Asian schools in East Africa. Gregory (1993, p. 228) notes that Asian teachers far outnumbered European and African teachers by a ratio of five to one, even though the proportion of Asians to Europeans in the population was only three to one. Some Indian teachers were also recruited from within East Africa, as in the case of Radhika. Radhika is not from the Anavil caste community. Although she belongs to a much lower caste community, where teaching is not a traditional occupation, her family comes from the same region in Gujarat as the Anavils. Radhika was born in Kampala, Uganda in 1934 and was educated there. After finishing her A-levels, she qualified to be a teacher and was hired to teach the primary grades at an Asian school in Kampala. She held this position until the Asian expulsion by Idi Amin in 1972, at which time she left for London. Most civil servants were permitted to stay in Uganda beyond the expulsion date in order to keep the country running. Nevertheless, Radhika and her husband, Suresh, a lawyer in Kampala, strategically split up during that time. Radhika migrated to London as a refugee and Suresh stayed in Uganda. At the time, they thought that the political situation might improve so that Radhika could eventually re-join Suresh in Uganda. Instead, as the political environment for Asians worsened, Suresh was forced to join her in London in 1973. Migration to London for Radhika and other Gujarati teachers resulted in degraded employment opportunities or underemployment. Their human capital qualifications were re-evaluated in this new socio-economic environment. It was precisely because the British held lower standards for the Asian teachers and their schools that they would not accept their qualifications in British schools that taught British children. For many teachers in the British colonies and protectorates good A-level scores, bachelor’s degrees, or teaching certifications obtained in India or Africa were sufficient to build a career as a professional teacher. After all, the British had imposed their own educational system on these countries. These qualifications were deemed sufficient by the British because they were just beginning to develop formal education for mostly African and Asian children. The segregation of children into African, Asian and European schools promoted different standards in the curriculum and in teacher qualifications. The European schools were, of course, held to a higher standard closer to their origin countries. Therefore, the transfer of qualifications from the East African to the British context was not recognized unless Indian teachers had a post-graduate degree. Few had such a degree because it was not required for teaching in British India or East Africa. Furthermore,

Networks of Inclusion and Exclusion in the Economic Concentrations

51

these Indian civil servants did not migrate to London voluntarily. Especially in the case of Uganda, they were politically and economically forced out of the country. These difficult circumstances surrounding their migration hindered their ability to form strong organizational ties with British educational institutions or organizations that might have helped them retain their teaching niche. In fact, most of their organizational ties were broken in East Africa by the shift in power from the British to the Africans. During and after this political upheaval, Britain only reluctantly accepted the Indian economic and political refugees. Therefore, the breakup of the teaching niche resulted from a changed socio-economic environment in East Africa and in the receiving context of these migrants in Britain. But, that’s not all. These circumstances also led to a break in organizational ties among teachers and other civil servants. The following profiles from two networks of teachers provide evidence for the effect that changes in migration and socio-economic environments have on sustaining niches. This evidence also has wider implications for what happens to organizational networks, generally, when organizational ties become weak or non-existent in a new socio-economic context. Pravin, Uma, and Kanti migrated to East Africa in the 1950s after being hired to become teachers there. Pravin taught in a small school in Tanzania until 1968 when the political situation of the country pushed him to migrate to London with his British passport. He migrated to London alone. His wife and three children had already returned to India so that his son could pursue a college education there. By 1970, when Pravin had found suitable accommodations for his wife and children, they joined him in London. Pravin went to the employment department to find his first job in London. Upon learning that his college education and experience as a teacher were not sufficient qualifications to continue at that job in Britain, he took his first job as an estimator in the sanitary department of a builder merchants firm. Pravin survived in this same position for 20 years under five different managing companies until his retirement. Pravin, Uma and Kanti are all members of the Gujarat Welfare Association in their working-class London neighborhood, where they met. Uma and Kanti worked in the same jobs in London for about fifteen years until they retired. Kanti was formerly a headmaster at an Asian school in Uganda and his wife, Uma, was a teacher at that same school. In London, Kanti took a job as a clerk. Uma worked as a machinist in a garment factory. They, too, could not transfer their teaching experience and secondary- and bachelors-level education in order to teach in London. Pravin, Uma and Kanti were all nearing middle age when they were politically forced to migrate to London. Their financial capacity was also very limited. In the case of Uma and Kanti they were only allowed to take £150 per person when they left Uganda, even though they had been a modestly wealthy

52

MARITSA V. POROS

middle-class family. Both families also had children to support. Thus, with almost no capital to invest in obtaining British teaching qualifications, they had little choice but to accept inferior employment opportunities in their new country of immigration. Radhika and Suresh had similar experiences. As noted earlier, Radhika was a teacher and Suresh was a lawyer and a government magistrate in Uganda. He had a Bachelor’s degree in Law from Bombay and she had good A-levels and teaching certification from Kampala. They were nearing middle age, when they were forced from Kampala to London as refugees. Although they did not have children to support, they both felt that it was too late in their lives to pursue further education. As refugees, they also lacked the capital to do so. Thus, Radhika became a clerk in a motor company and later took a low administrative post in government. Suresh was downgraded to the post of a magistrate’s clerk in government. Radhika and Suresh’s work and educational qualifications were also not sufficient in their new socioeconomic environment. An important counterexample suggests that the field of teaching was not closed to Indians as a result of individual racial discrimination per se, but rather a changed socio-economic context and changes in their social ties. Bharat, from the Anavil network of teachers, held a postgraduate degree from a college in Bombay. He was the headmaster of a prestigious European school in Kampala. Through this position, he met an official of the Ugandan High Commission, who recommended him for an educational post in London, which involved the development of multi-racial schools. Bharat was given the choice to teach or to become an officer for the project. He chose the latter position. Fortunately, he arrived in London the day before Idi Amin issued his “fatwah” on the Asians in Uganda. Bharat continued in this position until retirement. Not only did Bharat have the educational qualifications required in the British socioeconomic context, but also he was strategically located in a high position at a European school in Uganda, where he developed the organizational ties that enabled his recruitment. Even if Radhika, Pravin, Uma, and Kanti had developed similar kinds of organizational ties, these probably would have been of little use to them since their educational qualifications were still too low to become teachers in Britain. Nevertheless, they had little opportunity to develop those types of organizational ties with British educational officers and institutions because of their network location. They were teaching in small Asian schools during a time of declining British power. Neither did they seek Bharat’s help to find jobs once they befriended him in London. The breakup of the Indian teaching niche in East Africa, thus, occurred in the context of an involuntary migration to a new socioeconomic environment that did not recognize the experience and qualifications of migrants from the sending

Networks of Inclusion and Exclusion in the Economic Concentrations

53

region. A break in organizational ties, which first aided Indians’ employment in this niche, also occurred as a result of the political circumstances surrounding their migration. This was common for many other civil servants and professionals in East Africa, who relied on the host governments for jobs. Degraded employment opportunities for immigrants in a new host society are not original findings in migration research. However, the irony in this situation is significant. The British held different and lower standards of education for different racial groups of people in the overseas colonial context. This, in fact, helped to create the Indian occupational niche in East Africa that would later be broken in Britain using the same logic. The predominance of newly-formed interpersonal ties in the teachers’ networks also constrained access to the necessary resources to overcome the poorer employment opportunities they faced in London. Capital to pursue further education and ties to network members in organizations and better positions were sorely missing. Composite Networks and Extreme Embeddedness The structure of social networks in the transnational diamond trade is exceptional in many ways. In such a high-stakes business, social ties are closely guarded and controlled more than most other businesses. The trade is well known for its insularity, which guards against risk and the possibility of malfeasance (Westwood, 2000). The high risk of malfeasance in the trade stems from the transfer of such tiny and valuable goods across the world’s diamond trading centers on what seems like no more than a handshake or a phonecall. One diamond dealer reflected on her role selling diamonds on the American market: It has become a very risky, very competitive trade . . . It’s a lot of sleepless nights and a lot of money you lend out . . . your life is on the line all the time because you’re taking millions of dollars of goods with you in your hands and selling to people, walking from door to door. And you’re a target once they identify you.11

That handshake or phonecall is buttressed by huge portions of enforceable trust, which lubricate the trade. These relations of trust are unsurprisingly found in the dense network ties of Gujarati Jains in the trade.12 Their composite ties to family members, who have been trained and have worked in the trade across many generations, offer real and perceived security. I was told a story of malfeasance in the community: [Some time ago], one of the Indian merchants . . . borrowed goods from other diamond, jewelry Indians, you know. About $25 million worth of it. Went and sold it for $11 million and pocketed the money. Till last week somebody caught him when he was not paying his bills back. And you know, that’s how the business has become. Luckily, we were not hurt by it because we didn’t

54

MARITSA V. POROS sell to this guy. But there are at least another 40 Indians who had been selling to this guy, who are now fighting for their money. And that’s how risky it has become, the business. And that’s how business is done. Like once I know you, and I’ve dealt with you for a while, I don’t mind giving you any amount of goods, the truth. It’s like a handshake. I have to just know you and I’ll come give you goods. But they had never sold to this guy before? They all had been selling to him before. And he had a good reputation. And he used that . . . You know, it’s all from the community, people, he’s this one’s this, or that one’s brother, or that one’s nephew, and . . . it’s quite frightening. And how can you get lawyers involved, and you know, the lawyers get half the money, and there are 40 people, and you know, it’s tough . . . It’s a very good community. As I was saying this man went bankrupt. But, you know, in a year or two, they were forgiven. They’ll again give him goods and he’ll again be accepted back into the community. It’s not [as] if he did all this and so many people were hurt, and that he’s totally ostracized, no. That doesn’t happen. People forgive him. They’ll again get him back. That’s the nice part.13

How smooth the process of regaining trust in a dense network such as these diamond traders is arguable. However, the idea(l) that the community always envelops its members even when they do harm attests to the extreme embeddedness of those networks. This kind of embeddedness enables those within the network, many of whom are born into it, to become employed in the niche very easily. Even the dealer who betrayed his business partners and customers will be contained within the community because it would be too risky to ostracize him completely, given the secretive nature of business information. This extreme embeddedness can also prevent network members from exiting the niche when they find it in their interest to leave. It also keeps outsiders, those without generations of family linked to the trade, far away from niche opportunities and activities. One diamond dealing couple, who were pioneers on the American market migrated to the U.S. in 1970. Their immigration visas were sponsored by the headquarters of their firm in Antwerp. They had few contacts, but they joined the India Diamond and Colorstone Association and forged new ties with other Indian pioneers on the American diamond market. Being among the first Indian families to break into the American diamond market, they had to make contacts with buyers from scratch. Soon after these initial forays into the American market, and with the help of their sponsor firm in Antwerp, their business made great profits and they expanded into the manufacturing and sale of jewelry. They had a series of partnerships with other Jain traders. Finally, their firm became independent while still selling diamonds and jewelry polished and manufactured in India. Their composite ties to other traders in New York and Antwerp and to their exporters in India helped the firm to grow, which fed back into economic expansion in India. When we came into the country all we did was sell diamonds, no jewelry . . . loose diamonds to the Jewish people and the retailers . . . Then as we learnt the trade, we saw how jewelry was being made over here, and it was very expensive and labor back home was very cheap. That’s

Networks of Inclusion and Exclusion in the Economic Concentrations

55

when we started developing factories. And now we’ve got one of the largest factories in India, many factories in India . . . Our laborers, I mean artisans in India, they are so good at copying anything. You know, because they don’t have machines and all that, and they just know how to copy it with their hands . . . 14

Their first partner ran the Indian side of the business, exporting diamonds to their New York branch, which Lakshmi and her husband managed. Later, they became partners with two other firms in order to enter the jewelry trade, which required greater amounts of capital. Conflict with one of those partners, a leading diamond merchant with a $150–200 million business, led to their decision to become independent. By most standards, their business has been a success and Lakshmi proclaims that she loves the diamond trade. However, Lakshmi’s extreme embeddedness in their network of Jain diamond traders and particularly her position as a female member produces severe constraints on her ability to pursue new opportunities. Lakshmi’s generation was the third generation of Jains in the trade, but the first for women. Many women from diamond families were never taught the trade. They were instead encouraged to marry within the diamond community, and to become mothers and caretakers. Although Lakshmi was born into the trade and has been learning the trade ever since she was young, she holds a subordinate position to her husband in their firm. Her husband, somewhat ironically, was an outsider from a prominent Hindu business family. He was “adopted” and taught the trade by Lakshmi’s uncle. Here, Lakshmi describes the bounds of their relationship: Though my husband is modern . . . if I’m there in the office, he wants to make sure that I never speak back to him. If I do, we will have major fights, if I say anything in front of anybody else in the office . . . Everything I can say, we’re modern and everything, but if today he says you can’t do this, I cannot do it.15

Lakshmi’s husband controls their firm and he controls her role in it. As Lakshmi later explained to me, she wanted some capital to start a jewelry firm of her own. Her husband, however, would not provide her with the necessary start-up capital. To consider other employment options was out of the question. The high prestige, status, and extreme embeddedness of the diamond trade have an almost prohibitive effect on exiting the niche. Not only did she feel that her husband would not permit her to leave the trade for another industry, but that the community would also ridicule them. Social pressures within the trade are onerous. People come in from India, visit us all the time. So when they come to visit us we have to be with them. We take them out shopping, we’ll do everything, whether we like it or not. Because if we don’t, we’ll be talked about, mocked upon. So, we’re very much all traditions still.16

Niche exit is not impossible for differently located members of this network of diamond traders. Chandrika, a friend of Lakshmi, was also born and raised

56

MARITSA V. POROS

in a diamond family. Like most other women in their community, she was not expected to work in the trade. Instead, she was expected to marry and to become a mother and caretaker. Her marriage was arranged to Madhu, a Jain from their community, who lived in New York. Although he also comes from a diamond family, his father had entered the garment manufacturing industry in Bombay and Madhu ran the New York branch of this transnational firm. After their marriage, Chandrika moved from Antwerp to New York where she took an informal job in the New York office of her parent’s diamond firm. Well, when I got here, I wasn’t allowed to work so I didn’t have a valid working permit so I worked, my family had an office here at that time, so I worked there as a sorting person for diamonds . . . [then] I switched from being a sorting person to being a counting person.17

Afterwards, Chandrika attended college and received an associate degree in business management. She worked for another diamond company, which left her dissatisfied. Later she had a child and began working in her husband’s firm because she did not want to stay at home all day, but she needed flexible hours. Although she was raised in a diamond family, Chandrika had never been extensively trained in the diamond business. She held low positions as a counter or a sorter. Her marginal network location in the niche allowed her more flexibility to exit niche employment, especially because her husband was not active in the trade. His involvement in the garment industry, instead, created an easy link into the possibility for employment in his firm. As a result of this link, exit from the diamond niche did not result in ridicule or social pressure by other members of their network. Thus, Chandrika was able to switch her employment from one niche industry to another with some ease. However, it should be noted that because Chandrika and Madhu are from diamond families, they are still closely tied to the extremely embedded networks of Jain diamond traders active in this industry niche.

CONCLUSIONS AND IMPLICATIONS FOR CONSTRAINED, PARTIAL, AND EXTREME EMBEDDEDNESS The economic niches composed of interpersonal, organizational, and composite networks described above illustrate how the extent and mode of embeddedness of economic behavior in distinct networks of relations varies. The location of individuals in particular configurations of ties, the formation of new ties, and the transactions or relations of exchange within those ties are part of the network processes that make up these differences in embeddedness.

Networks of Inclusion and Exclusion in the Economic Concentrations

57

Variation in the network location in interpersonal, organizational, and composite configurations of ties can cause very different outcomes in embeddedness. Constrained, partial, and extreme embeddedness roughly correspond to interpersonal, organizational, and composite network processes, where network location, new social ties and brokers, and the relations of exchange within ties contribute to processes of inclusion and exclusion in economic niches. For example, Lakshmi was extremely central to her network of diamond traders, but in a relationally subordinate position to her husband. She is severely constrained in her ability to work independently even though she and her husband are co-owners of a very successful firm within a lucrative industry niche dominated by members of their co-ethnic community. Contrast her position with Chandrika. Chandrika’s low position in the diamond trade and legitimate access to the garment trade allowed her to easily exit the diamond trade and enter the garment trade, where her husband is an owner of a profitable wholesale business. Thus, the network location and the consequent transactions within those ties – Lakshmi’s inability to act with authority in business matters, and Chandrika’s ability to avoid ridicule and the social pressures of their network – led to quite different outcomes for each of them. This, even though both could be considered upwardly mobile (surely an understatement) in lucrative niches dominated by members of their Gujarati Jain communities. These kinds of transactions are less evident in the stories of organizational networks. Niches formed through organizational networks are monitored by external organizations, such as educational and certifying institutions or states. These organizations are very exclusive and selective in terms of human capital and organizational requirements. Once organizational criteria are met, transactions within these ties are less risky than within interpersonal and composite tie networks. This was evident in the story of Bharat, who met the human capital requirements to continue as a teacher in Britain, but who was also strategically located in a network that included British brokers in the education industry, who could connect him to a job in education in Britain. This partial embeddedness of economic behavior is shaped and regulated by organizations and network ties allowing for niche change on this basis. In composite networks, however, the dense network ties based on trust do the job of regulating and sorting individual and network level economic behavior such that embeddedness is extreme and niche exit and entry are more difficult. Embeddedness in interpersonal networks is typically constrained because interpersonal ties offer few new opportunities for exiting a niche or for mobility within it. Where there is a lack of brokers in one’s interpersonal networks, as is the case with many small shopkeepers, niche employment can constrain individual mobility. Similarly, the absence of interpersonal ties or brokers who can connect co-ethnics to industry niches exclude some individuals from these opportunities altogether.

58

MARITSA V. POROS

In the above narratives, we also see that different configurations of ties and their transactions penetrate economic sectors irregularly (in Granovetter’s terms). Despite heaps of trust, which are inherent in the composite ties needed to run the diamond trade, the risk of malfeasance always looms, quite unlike niches dominated by organizational networks. Interpersonal networks also penetrate economic sectors such as retail trade and services through interpersonal network ties that select labor and monitor behavior. There are broader implications for these findings that are related to the role of social origins of migrants in niche employment and ownership, and to the broader diversity of workers in different industries or economic sectors. The narrative evidence from this study shows that social processes, that is, the network processes that include and exclude immigrants from particular economic opportunities, are more important than social origins. In other words, co-ethnicity operates on many different levels and very unevenly in networks such that it is not the overriding mechanism for the cooperation necessary to sustain niche employment. Individuals with very similar attributes, such as gender, caste, ethnicity, and class status are presented with very different kinds of opportunities for niche employment as a result of their network location and transactions within their social ties, rather than based on those individual attributes. Thus, examining processes of inequality within (and even between) different immigrant groups by focusing on co-ethnicity seems likely to be a misguided approach. The variation in embeddedness of these industry niches and the individuals who work within them also have implications for how diverse workforces can be in certain industries. In industries, such as diamonds and garments, where economic behavior is extremely embedded in relations of composite ties, workforce diversity is likely to be low or at least very segmented. In industries where economic behavior is less extreme but constrained by the penetration of interpersonal networks of relations, workforce diversity is similarly likely to be low, but subject to change or turnover to other groups more easily. Niche entry, but more importantly, niche exit is more possible in these industries, which can provide room for different groups to populate ownership and employee positions within them. In industries where economic behavior is partially embedded, such as in professions or civil service, organizational requirements provide openings to many different ethnic groups as long as they meet the human capital and organizational requirements of those positions. Examining variation in embeddedness, therefore, is important for understanding how social processes inherent in networks of relations vary in the way that they penetrate different economic sectors, which can make a workforce appear more or less ethnically diverse.

Networks of Inclusion and Exclusion in the Economic Concentrations

59

NOTES 1. The terms concentration and niche are used interchangeably throughout this article. Some scholars define ethnic niches as ethnic groups that dominate 50% of ownership or employment in a particular industry. Here, an ethnic economic niche or concentration will refer less strictly to ownership or employment of a particular immigrant or ethnic group, which exceeds their proportion of the general population. 2. Such migrant networks are often based on co-ethnicity. Thus, membership by virtue of co-ethnicity is thought to grant one access to the tangible and intangible benefits of network-mediated social capital, which can aid socio-economic incorporation in the host country. In particular, social capital can be converted into other (positive) forms of capital, such as financial capital in the form of migrant remittances (Massey et al., 1998). Some migration scholars have also acknowledged and documented the constraints of belonging to such networks (see Faist, 2000 on “negative” social capital; Kyle, 2000; Levitt, 2001; Menj´ıvar, 2000; Portes, 1998; Portes & Sensenbrenner, 1993). 3. Membership or some participation in events sponsored by Indian voluntary associations is very common. Although it is difficult to establish just what proportion of Indians engage in association activities, the great majority of respondents in this study (86%) were paid or listed members of at least one association and some attended the activities of others. Prior studies also support the claim that Indian immigrants participate heavily in Indian voluntary associations (see Desai, 1963; Fisher, 1980; Helweg & Helweg, 1990; Lessinger, 1995). 4. Not only is the Neasden Swaminarayan temple dominated by Gujaratis, but the Patidar community, a caste group in Gujarat, is also very prominent within it. The Neasden temple is located in a sorely unattractive Northwest London neighborhood. Surrounded by dilapidated terraced housing, enormous shopping centers, and the M25 beltway, which heavily pollutes the area, the temple seems to rise from the ashes as one approaches it. Built of marble and limestone, it was carved by artisans in India and shipped to London in approximately 26,000 pieces. The temple has become such a strong status symbol, that it has been dubbed the “face of Hinduism in Britain.” See Williams (2001) on Swaminarayan Hinduism, generally. 5. From interview on 6/2/99, London. Pseudonyms have been used to protect the anonymity of all respondents and persons mentioned in the interviews. 6. From interview on 6/2/99, London (italics are mine). 7. Patel is a very common surname most often, but not always, belonging to the Patidar subcaste in Gujarat. Caste and subcaste in Gujarat is a highly complex system of stratification. See Shah and Desai (1988) and Pocock (1972) for a more detailed discussion. 8. From interview on 6/2/99, London. 9. From interview on 4/27/99, London. 10. From interview on 6/2/99, London. 11. From interview on 11/9/98, New York. 12. Jainism is one of several world religions represented in India’s population. 13. From interview on 11/9/98, New York. 14. From interview on 11/9/98, New York. 15. From interview on 11/9/98, New York. 16. From interview on 11/9/98, New York. 17. From interview 9/21/98, New York.

60

MARITSA V. POROS

REFERENCES Bonacich, E., & Light, I. (1988). Immigrant entrepreneurs: The Koreans in Los Angeles, 1965–1982. Berkeley: University of California Press. Desai, R. (1963). Indian immigrants in Britain. London: Oxford University Press. Faist, T. (2000). The volume and dynamics of international migration and transnational social spaces. Oxford: Clarendon Press. Fisher, M. (1980). The Indians of New York City: A study of immigrants from India. Columbia, MO: South Asia Books. Granovetter, M. (1973). The strength of weak ties. American Journal of Sociology, 78, 1360–1380. Granovetter, M. (1974). Getting a job: A study of contacts and careers. Cambridge: Harvard University Press. Granovetter, M. (1985). Economic action and social structure: The problem of embeddedness. American Journal of Sociology, 91, 481–510. Gregory, R. (1993). The South Asians of East Africa: A social and economic history. Boulder, CO: Westview Press. Heimer, C. (1992). Doing your job and helping your friends: Universalistic norms about obligations to particular others in networks. In: N. Nohria & R. Eccles (Eds), Networks and Organizations: Structure, Form, and Action (pp. 143–164). Boston: Harvard Business School Press. Helweg, A., & Helweg, U. (1990). An immigrant success story: East Indians in America. Philadelphia: University of Pennsylvania Press. Kloosterman, R., & Rath, J. (2001). Immigrant entrepreneurs in advanced economies: Mixed embeddedness further explored. Journal of Ethnic and Migration Studies, 27, 189–202. Krackhardt, D. (1992). The strength of strong ties. In: N. Nohria & R. Eccles (Eds), Networks and Organizations: Structure, Form, and Action (pp. 216–239). Boston: Harvard Business School Press. Kyle, D. (2000). Transnational peasants: Migrations, networks, and ethnicity in Andean Ecuador. Baltimore and London: Johns Hopkins University Press. Lessinger, J. (1995). From the Ganges to the Hudson: Indian immigrants in New York City. Boston: Allyn & Bacon. Levitt, P. (2001). The transnational villagers. Berkeley: University of California Press. Light, I., & Gold, S. (2000). Ethnic economies. San Diego: Academic Press. Logan, J., Alba, R., Dill, M., & Zhou, M. (2000). Ethnic segmentation in the American metropolis: Increasing divergence in economic incorporation, 1980–1990. International Migration Review, 34, 98–132. Lyon, M. H., & West, B. J. M. (1995). London Patels: Caste and commerce. New Community, 21, 399–419. Marsden, P. (1982). Brokerage behavior in restricted exchange networks. In: P. Marsden & N. Lin (Eds), Social Structure and Network Analysis. Beverly Hills: Sage. Massey, D., Arango, J., Hugo, G., Kouaouci, A., Pellegrino, A., & Taylor, J. E. (1998). Worlds in motion: Understanding international migration at the end of the millennium. Oxford: Clarendon Press. McDowell, E. (1996, March 21). Hospitality is their business: One ethnic group’s rooms-to-riches story, The New York Times, p. D1. Menj´ıvar, C. (2000). Fragmented ties: Salvadoran immigrant networks in America. Berkeley: University of California Press. Model, S. (1997). An occupational tale of two cities: Minorities in London and New York. Demography, 34, 539–550.

Networks of Inclusion and Exclusion in the Economic Concentrations

61

Pocock, D. F. (1972). Kanbi and Patidar: A study of the Patidar community of Gujarat. Oxford: Clarendon Press. Poros, M. V. (2001). The role of networks in linking local labour markets: The case of Asian Indian migration to New York and London. Global Networks, 1, 243–259. Portes, A. (1998). Social capital: Its origins and applications in modern sociology. Annual Review of Sociology, 24, 1–24. Portes, A., & Sensenbrenner, J. (1993). Embeddedness and immigration: Notes on the social determinants of economic action. American Journal of Sociology, 98, 1320–1350. Shah, A. M., & Desai, I. P. (1988). Division and hierarchy: An overview of caste in Gujarat. Delhi: Hindustan Publishing Corporation. Stinchcombe, A. (1989). An outsider’s view of network analyses of power. In: R. Perrucci & H. Potter (Eds), Networks of Power: Organizational Actors at the National, Corporate, and Community Levels. New York: Aldine de Gruyter. Sun, L. (1998, November 3). Merchants share heritage, hard work, tragedy. The Washington Post, p. B1. Tilly, C. (1998). Durable inequality. Berkeley: University of California Press. Varadarajan, T. (1999, July 4). A Patel motel cartel? The New York Times Magazine, p. 36. Waldinger, R. (1990). The making of an immigrant niche. International Migration Review, 28, 3–30. Waldinger, R. (1995). The other side of embeddedness: A case study in the interplay of economy and ethnicity. Ethnic and Racial Studies, 18, 555–580. Waldinger, R. (1996). Still the promised city?: African-Americans and new immigrants in postindustrial New York. Cambridge: Harvard University Press. Westwood, S. (2000). A real romance: Gender, ethnicity, trust and risk in the Indian diamond trade. Ethnic and Racial Studies, 23, 857–870. Williams, R. B. (2001). An introduction to Swaminarayan Hinduism. Cambridge: Cambridge University Press. Zimmer, C., & Aldrich, H. (1987). Resource mobilization through ethnic networks: Kinship and friendship ties of shopkeepers in England. Sociological Perspectives, 30, 422–445.

IF I KNOW IT, WILL I SHARE IT?: THE POTENTIAL EFFECTS OF GROUP COMPOSITION ON THE CREATION AND STABILITY OF ORGANIZATIONAL KNOWLEDGE Melvin L. Smith ABSTRACT This article presents arguments regarding the importance of information sharing to the growth and stability of organizational knowledge. In addition, the article discusses the expected effects of group composition on the nature and degree of information sharing that takes place within groups. While group composition may vary along a number of dimensions, this article focuses primarily on differences in group membership represented by various race and gender combinations. The specific research question explored is, to what extent does group composition affect the likelihood that individually held (unique) information will be shared with group members during group discussion? A conceptual model is presented and its implications for both research and practice are discussed.

Diversity in the Workforce Research in the Sociology of Work, Volume 14, 63–86 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0277-2833/doi:10.1016/S0277-2833(04)14003-X

63

64

MELVIN L. SMITH

INTRODUCTION Whether performing a routine task or making a strategic decision regarding a future course of action, organizations (and the groups within them) often draw upon what they have learned from past experiences. This requires that the knowledge from past experiences be retained in some fashion within the organization in order that it may be retrieved as needed. Researchers have referred to this process as organizational learning and the resulting stores of information have been said to represent organizational memory (Levitt & March, 1988). Most discussions of organizational memory cite individuals as a key repository of organizational knowledge (Argote, 1999). It thus seems apparent that unless the knowledge possessed by an individual is somehow transferred to other individuals within the organization and/or other knowledge repositories, the organization is very susceptible to losing it. Such knowledge can, therefore, be considered to be fragile. If the individual forgets the information or leaves the organization, then that knowledge will no longer be available for the organization to draw upon. It can further be inferred that in order for it to be represented as a stable component of organizational memory, knowledge possessed by individuals must first be shared. Once the organization is aware of the knowledge possessed by an individual, decisions can be made as to if, and if so how, that knowledge can be transferred to other organizational repositories. Given the suggested importance of the sharing of individual knowledge to creating stable elements of organizational memory, this article develops specific propositions that are intended to enhance our understanding of information sharing in groups. The article builds upon previous work in the area of collective information sampling (CIS) by considering the potential impact of group composition (and the corresponding degree of demographic similarity/dissimilarity experienced by individual group members) on the level of information sharing that takes place within a group. While group composition may vary along a number of dimensions, this article focuses primarily on the expected differences in the level of information sharing resulting from differences in group membership represented by various race and gender combinations. The specific research question to be explored is, to what extent does group composition affect the likelihood that individually held (unique) information will be shared with group members during group discussion? In the following sections, arguments are presented relative to the importance of information sharing to the creation of a stable base of knowledge in organizational memory. The expected influence of group composition on the nature and degree of information sharing that takes place within a group is then discussed. This is followed by the consideration of several factors suggested to have a moderating

If I Know It, Will I Share It?

65

effect on the relationship between an individual’s demographic dissimilarity and their willingness to share uniquely held information during group discussion. Specific propositions are offered and a theoretical model is presented. Finally, the article concludes with a discussion of implications for both research and practice.

SHARED INFORMATION AND ORGANIZATIONAL KNOWLEDGE A fundamental activity of groups is the integration of individual knowledge into collective knowledge (Okhuysen & Eisenhardt, 2002). Additionally, in order to perform organizational tasks, teams or workgroups must often perform the knowledge processing activities of acquisition and creation (Anand et al., 2003). With knowledge integration, complementary knowledge separately held by group members is combined to form new knowledge (Grant, 1996). Knowledge acquisition involves bringing in knowledge from external sources (Huber, 1991). Knowledge creation in workgroups entails sharing information to become aware of the various members’ expertise, constant reinterpretation of each other’s perspectives until new ideas emerge, and testing the newly created knowledge for validity (Anand et al., 2003). Whether acquired from external sources or created internally, in order for knowledge to be utilized for future organizational purposes it must be stored in organizational memory so that it can be retrieved by the organization when needed. Walsh and Ungson (1991), in an effort to develop a coherent theory of organizational memory, contended that once acquired, information could be stored in a variety of “retention bins” within an organization. They described five such information retention facilities – individuals, culture, transformations, structures, and ecology. Although information might conceivably be stored in any or all of these retention facilities, the focus of this article is on individuals (and ultimately groups of individuals) as information/knowledge repositories. According to Wexler (2002), collective remembering entails locating information or experiences that either one has directly dealt with earlier or more indirectly has been left in a repository by others, including earlier generations, for future retrieval (p. 393).

Unshared knowledge that resides strictly within the memory of an individual, however, and thus is not stored in any other organizational retention facility, represents what can be thought of as a fragile component of organizational memory. In cases such as this, an organization is particularly susceptible to

66

MELVIN L. SMITH

memory loss due to the fact that if the individual were to leave the organization for any reason, then that unshared knowledge would leave with them. Several empirical studies have, in fact, shown that turnover can have a significant impact on organizational memory (see Argote, 1999 for examples). The intent here is not to suggest that organizational memory should ideally be comprised completely of shared information. To the contrary, researchers have discussed the benefits of transactive memory systems (Moreland et al., 1996; Wegner, 1987, 1995), where group members retain only certain items in individual memory, beyond which they simply need to remember who in the group knows what, or who is good at what. This enables the group as a whole to store more information than any group member alone could retain. Therefore, it appears that there are certain benefits inherent in not limiting organizational memory solely to shared information, where all relevant knowledge that any group member possesses in individual memory is also known/retained by another (other) group member(s). What I am suggesting, however, is that there are several distinct elements of which the individual component of organizational memory is comprised, and that these elements can be placed on a continuum ranging from fragile sources of organizational memory at one extreme to stable sources at the other. This continuum is illustrated in Fig. 1. Toward the right end of the continuum (position 3a) is shared knowledge. This represents specific knowledge possessed by more than one individual in the group. For example, in an application-programming group of an organization, John and Sally might both have detailed knowledge of the software code underlying the company’s accounts payable and accounts receivable applications. Thus, if either John or Sally were to leave the group, then that knowledge would still be maintained in organizational memory due to the redundancy of retention

Fig. 1. Organizational Knowledge Continuum.

If I Know It, Will I Share It?

67

facilities (multiple individual repositories). Shared knowledge thus represents the most stable element of organizational memory residing in individual repositories since it should be relatively unsusceptible to personnel turnover. The greater the redundancy of the knowledge the more stable organizational memory will be relative to that knowledge. On the left-center section of the continuum (position 2a) is known, unshared knowledge. This, in essence, represents transactive memory. In other words, the group is aware that an individual possesses particular knowledge, yet that individual may be the only member of the group that possesses that knowledge. Continuing with the previous example, in this case the group may be aware that John has uniquely detailed knowledge of the accounts payable application and that Sally has uniquely detailed knowledge of the accounts receivable application. Organizational memory in this scenario is less stable than was the case with shared knowledge since if either John or Sally were to leave the group, then their specific knowledge regarding either the accounts payable or the accounts receivable application will no longer be accessible to the group. There is some degree of stability, however, since the group is at least aware of who knows what. Given this awareness, the group has the option of putting a plan in place to transfer John and/or Sally’s specific application knowledge to one or more additional group members, which would then shift it into the shared knowledge category. To the extent that the group chooses not to do this, it represents what I refer to as conscious vulnerability. This is analogous to making a conscious decision not to carry insurance and hoping that no unforeseen, loss-creating event occurs. At the far left of the continuum (position 1a) is unknown, unshared knowledge. This represents knowledge possessed by individual members that may be relevant to the effective functioning of the group, yet the group (with the exception of the individual possessing the knowledge) is unaware that the knowledge is available to them. Again considering the previous example, this case might be represented by a situation where Sally may have, through the course of performing her job, become aware of a particular problem with the interface between two software applications. The problem may require manual intervention each month to avoid compromising data integrity, which would result in erroneous information being fed into the payroll application. Rather than initiating a major re-coding effort to correct the problem, she makes a decision to perform the manual procedure each month. As long as Sally is a member of the group and she retains this knowledge in her individual memory, then it would also be classified as organizational memory according to the definition provided by Walsh and Ungson (1991). They define organizational memory as, “. . . stored information from an organization’s history that can be brought to bear on present decisions.” Similarly, Anand et al. (1998)

68

MELVIN L. SMITH

suggest that information is incorporated in an organization’s memory when at least one organizational member possesses the knowledge or can retrieve it. I contend, however, that this knowledge represents an extremely fragile element of organizational memory in that it is not only unshared, but its existence is also unknown to the group as a whole. Therefore, the group cannot in this case make a conscious decision as to whether or not they should attempt to transfer the knowledge to one or more additional members of the group. As a result, if Sally were to leave the group without transferring this knowledge to another group member, then the group (and ultimately the organization) is likely to experience negative consequences. I thus refer to this as unconscious vulnerability, which I suggest is characteristic of the most fragile forms of organizational memory residing in individual repositories.

SOCIAL NETWORK IMPLICATIONS In addition to possessing knowledge that may be brought to bear in solving a particular group problem, an individual may also possess network ties, or social relationships that provide them with access to relevant information/resources that may not necessarily be available within the group. These network ties can be classified in the same fashion as the three elements of the previously described continuum. First, toward the right end of the continuum (position 3b) are shared network ties, representing redundant contacts, meaning that more than one individual in the group has a relationship with a particular contact. Next, on the right-center section of the continuum (position 2b), are network ties that are known but unshared. Ties of this type represent known, non-redundant contacts. In other words, although the group may be aware of a network contact that provides the focal individual with relevant information/resources, no one else in the group has a relationship with that particular contact. Similar to the previously discussed example of known but unshared knowledge, these ties represent a degree of conscious vulnerability. Given that only one individual has a relationship with the contact in question, if that individual were to leave the organization, then the organization would be susceptible to losing access to resources associated with that relationship. Finally, toward the left end of the continuum (position 1b) are network ties that are both unshared and unknown. These represent unknown, non-redundant contacts and are similar to the unknown, unshared knowledge previously discussed in that they represent to the group a source of unconscious vulnerability. The group may be currently benefiting from these contacts, yet since it is not aware of their existence it is unable to make provisions to protect these particular sources

If I Know It, Will I Share It?

69

of information/resources. Thus, if the individuals with these network contacts terminate their membership with the group, then the access to these resources will be terminated as well. To recap, I suggest that information sharing is essential to creating and maintaining organizational memory, not only due to the transfer of specific knowledge that occurs, but also due to the fact that information sharing can play an important role in simply making the group aware of who knows what. Or, in some cases, who has access to relevant knowledge/resources that exist outside of the group. This awareness can serve to increase the stability of organizational memory, as represented by group knowledge.

GROUP DISCUSSION DYNAMICS Since no single individual typically has all of the relevant information necessary for the performance of a group task, the ability of a group to successfully integrate the knowledge of its members and create new knowledge is central to the effectiveness of both the group and the organization. In order for this integration to occur, group members must share their relevant knowledge. The process of information sharing during group discussion, however, is not always effective. Critical information held by individual group members is often not used by the group (Okhuysen & Eisenhardt, 2002; Stasser & Stewart, 1992; Wittenbaum & Stasser, 1996). One explanation that has been offered for this failure to fully share critical information is a phenomenon referred to as collective information bias, which is the tendency for groups to discuss shared rather than unshared information (Wittenbaum et al., 1999). Research on collective information sampling (CIS) examines the process by which groups sample items for discussion when information is not evenly distributed among members. Previous CIS research has found that groups often omit unshared information from discussion and focus on information commonly held by members (Wittenbaum & Stasser, 1996). Stasser et al. (1989), for example, found that when three- and six-person groups were provided with both shared (commonly held) and unshared (uniquely held) information, 46% of the shared information was discussed, whereas only 18% of the unshared information was brought up in discussion. Such behavior suggests that groups may often forego the benefit of having diverse knowledge. In addition to the suppressive effects of collective information bias on the discussion of uniquely held information during group discussion, dysfunctional group processes such as conflict, miscommunication, and marginalization of minority group members (i.e. women and people of color), may prevent groups from taking advantage of their diverse knowledge (Elsass & Graves, 1997). Not

70

MELVIN L. SMITH

only may this lead to groups missing the opportunity to capitalize on their own diverse knowledge, but it may also result in groups foregoing the opportunity to take advantage of the varied network contacts represented by the webs of social relationships in which group members are involved. The effect of this potential missed opportunity for leveraging the social ties of group members is likely to be particularly pronounced in the case of minority group members since they have been shown to be more likely to develop networks of relationships outside of their workgroup (often for emotional support not available within the group), as well as within the group (Ibarra, 1995). These external social ties conceivably represent non-redundant sources of information and other resources which can be of value to the group. Combining what we know from previous research on information sharing with our knowledge of the effects of group composition and intergroup relations on group processes, several interesting propositions may be derived regarding the anticipated effects of group composition on the nature of information sharing within groups. The propositions presented here stem from the baseline CIS model developed by Stasser and Titus (1987). According to the model, the probability that a group will discuss an item, p(D), is a function of the number of members who have knowledge of the item, n, and the likelihood that any one of these members will mention it, p(M). The function can be represented by the following equation: p(D) = 1 − [1 − p(M)]n This baseline model considers the probability that a member will mention an item, p(M), to be the same for all group members. However, recognizing that heterogeneous groups are likely to be composed of members with different experiences, values, attitudes, and cognitive approaches (Elsass & Graves, 1997), there are reasons to believe that the probability of mentioning an item of information during group discussion might not necessarily be the same for all group members. Thus, drawing on both theory and empirical research on a variety of social psychological dynamics suggested to be evoked in the presence of demographic heterogeneity in groups, the following section will consider how group composition might possibly influence the willingness of individual group members to share uniquely held information during group discussion.

GROUP COMPOSITION AND INFORMATION SHARING An increasingly diverse U.S. labor pool and the trend toward the use of teams to coordinate work has led to more heterogeneous workgroups in organizations

If I Know It, Will I Share It?

71

(Milliken & Martins, 1996). While it has been suggested that there is value in this increased diversity, much empirical work is still needed to test this assertion (Cox & Blake, 1991; Cox & Nkomo, 1990; McLeod et al., 1996). Furthermore, results of the empirical work that has been conducted to date are somewhat inconclusive (see Williams & O’Reilly, 1998 for a review). Harrison et al. (2002) state that, as research accumulates, there has been a growing recognition that paths linking work team diversity to team functioning and outcomes are complex (p. 1029).

On the one hand, diverse groups have been found to be more creative (Jackson, 1992; Nemeth, 1992), more willing to adopt innovative solutions (Bantel & Jackson, 1989; Wiersema & Bantel, 1992), and more apt to make higher quality decisions (Hoffman & Maier, 1961). On the other hand, group heterogeneity has been shown to be associated with lower levels of social integration and higher levels of turnover (O’Reilly et al., 1989), and greater conflict (Pelled, 1996; Pelled et al., 1999). Given the increased use of teams in organizations, the increasing diversification of the workforce, and mixed results to date from research examining the benefits of diversity in groups, it is becoming increasingly important that we develop a better understanding of the effects of group composition on workgroup processes and outcomes. The process of information sharing within groups is one area in which such an understanding could be of significant benefit. Of particular interest here is the potential effect of group composition (more specifically demographic differences) on the collective information sampling that takes place during group discussion in the context of decision making.

DEMOGRAPHIC DISSIMILARITY EFFECTS Research on the effects of group composition focuses on the relationship between the demographic profile of a group and various outcomes such as group performance, collective attitudes and/or collective behaviors. While the compositional approach to examining demographic differences in groups adds considerably to our understanding of a variety of group processes and outcomes, it treats the effect of the group’s demographic distribution as being the same for all group members, which in essence ignores the experiences of specific individuals within the group based on their demographic makeup as compared to the demographic profile of the group (Tsui & Gutek, 1999). A relational demography approach to examining demographic differences in groups, however, allows not only for the consideration of how certain demographic characteristics impact the outcomes, attitudes, and behaviors of individuals, but also considers the effects on the individual of being

72

MELVIN L. SMITH

relatively different from (or similar to) others in a social unit on a variety of demographic characteristics (Tsui & O’Reilly, 1989). Relational demography research is based on the premise that demographic factors are important sources of self-identity for individuals and that salience of the various social categories that serve as bases for an individual’s identity can be made more or less salient by the relative presence of other similar or dissimilar individuals (Tsui et al., 1992; Tsui & Gutek, 1999). One of the underlying theories for this research is the similarity-attraction paradigm (Byrne, 1971), which suggests that individuals are attracted to others that share their basic attitudes and beliefs, and that various physical, social, and status attributes are used to infer similarity or dissimilarity. Another theory that serves as a conceptual foundation for relational demography research is social identity theory (Tajfel, 1974), which maintains that individual self-evaluation is based in part on one’s psychological group memberships (i.e. social categories). As an extension of social identity theory, self-categorization theory (Turner, 1987) further posits that members of a group are likely to use salient features of their social identities to categorize one another. This categorization results in the creation of an “in-group” (those who are similar to ourselves on the attribute upon which the categorization is based), and an “out-group” (those from whom we are different on that same attribute). Research has shown that even the creation of categories based on arbitrary criteria (e.g. the flip of a coin) can lead to favoritism of one’s in-group and bias against the out-group (Brewer, 1979). These categorizations can, in turn, evoke cognitive processes such as status judgments, stereotyping, and similarity bias, all of which are likely to lead to negative expectations for the focal individual (Ashforth & Mael, 1989; Chatman et al., 1998). DiTomaso and Hooijberg (1996) state that, [w]ithin organizations, these negative stereotypes and attributions can severely hamper cooperation and communication and negatively affect production (p. 167).

In the context of group communications, these processes have been shown to lead to exclusive communicative behaviors, whereby those who are observably different from others are essentially excluded from the in-group conversation through the use of various divergence or distancing communication strategies (Akoyo et al., 2002; Larkey, 1996). This may result in those who are demographically different from the majority of the group being less able (and eventually less willing) to share their views with the group, especially when their potential contributions are different from, or even opposed to information already shared during the discussion. While there are a variety of demographic differences that could possibly have a negative impact on an individual’s contributions to a group, the demographic

If I Know It, Will I Share It?

73

attributes of race and gender are likely to be more salient than other attributes. This is due to the relative ease with which they may be distinguished, as well as the importance that each of these categories has had on the historical interactions between individuals both in and out of the workplace. Therefore, in the context of an organizational workgroup, minority group members (i.e. women and racial minorities) may be less willing to share information with the group in part due to long existing stereotypes and corresponding negative expectations. DiTomaso and Hooijberg (1996) describe these potential effects in the context of inequality that they suggest emerges and is reinforced through a variety of mechanisms. They add that group differences are ‘embedded,’ in the sense that they are reproduced throughout the culture . . . they are not only perceived, but also experienced in ways that make them appear justified to those who are privileged (p. 174).

The majority group may thus be generally less receptive to the contributions of minority group members regardless of their individual knowledge, skills, and abilities. Subsequently, minority group member awareness (or even perception) of this potential lack of receptivity could further lessen the likelihood that they would share information with the group. This leads to the following proposition. Proposition 1. An individual’s demographic dissimilarity in terms of (a) race and (b) gender relative to other group members will be negatively related to that individual’s willingness to share uniquely held information during group discussion. Previous studies have also shown that as the proportion of individuals who share a particular demographic attribute grows smaller, people in the group who possess that characteristic will become increasingly aware of their social identity (Kanter, 1977; Mullen, 1983; Riordan & McFarlane Shore, 1997). In other words, the salience of that particular characteristic is likely to increase as it becomes more distinctive relative to the characteristics of others in the group. This effect is likely to be even more pronounced when the individual is the only group member of his or her race or gender (McGuire & McGuire, 1981). Thus, based on the arguments presented above relative to in-group/out-group communicative behaviors, I propose the following: Proposition 2. An individual’s willingness to share uniquely held information during group discussion will be significantly less when that individual is the only group member of his or her (a) race or (b) gender than when there is at least one other group member of the same race or gender.

74

MELVIN L. SMITH

While demographic dissimilarity is proposed to negatively influence individual willingness to share uniquely held information during group discussion, there are several individual-, as well as group-level factors that are likely to influence that relationship. These factors are discussed in the following sections.

PROPOSED MODERATING FACTORS – INDIVIDUAL LEVEL As discussed in the previous section, group members’ race or gender, as well as their demographic dissimilarity on either of these attributes, is likely to influence their contributions to group discussions. Previous research has examined the effects of race and/or gender on the attitudes and behaviors of individuals within groups (see Tsui & Gutek, 1999 and Phillips & O’Reilly, 1998 for reviews). For instance, empirical evidence shows that men in task groups express more opinions, make more suggestions, and take the lead more often than do women (Johnson et al., 1996). One explanation that has been offered for this observed pattern of behavior is provided by status characteristics theory, which suggests that gender differences in interaction in mixed-sex groups are a result of women’s lower status in the larger society (Johnson et al., 1996). I suggest that this argument can be extended to racial minorities as well as females. Alderfer and Smith (1982) described a similar phenomenon, which they referred to as hierarchical intergroup effects, which result from embedded intergroups. Status, however, is not derived simply from race or gender. Instead, both (as well as other factors) are likely to contribute to the status attributions made of individuals within groups. Status incongruency theory, in fact, suggests that there is a “pecking order” that is assigned to individuals based on their demographic characteristics. Further, the lower one’s status in the hierarchy, the more negative the stereotypes and the greater the disadvantages one faces (Fernandez, 1981; Gilbert et al., 2003; Pettigrew, 1979). Lau and Murnighan (1998) similarly suggest that the consideration of a single demographic attribute can cause researchers to miss the potential impact of other attributes and/or their interactions. In addition, several researchers have stated a need for consideration of the joint effects of race and gender noting that, particularly in the case of African-American females (and potentially females of other racial minority groups as well), these individuals tend to get grouped into the category of females or racial minorities in research studies. The fact of the matter is, however, that their differences from the dominant majority (i.e. White males) in terms of both race and gender create a unique organizational

If I Know It, Will I Share It?

75

experience for minority females, which is not fully captured by simply examining race or gender effects alone (see for instance Bell, 1990; Lau & Murnighan, 1998; Parker, 2001; Parker & Ogilvie, 1996). Additionally, Cox and Nkomo (1990) make a specific call for future research of this nature in their status report of race as a variable in organizational behavior research suggesting that future theory constructions, should give more attention to the intersection of race and gender . . . research that appears in other sources suggests that organizational experiences of minority males and minority females may differ in significant ways (p. 428).

It thus stands to reason that the interaction of race and gender might have a significant influence on the attributions made toward members of a demographically diverse group and that these attributions and associated stereotypes might, in turn, influence individual attitudes and behaviors. Based on status incongruency theory, it could be expected that the negative relationship between race dissimilarity and the willingness to share uniquely held information during group discussion would be stronger for females than for males due to their lower perceived status. Status incongruency theory might also lead to the prediction that the strength of the negative relationship between gender dissimilarity and the willingness to share uniquely held information during group discussion would be greater the lower an individual is on the perceived status “pecking order” based on their race. However, as previously discussed, the interaction between race and gender is likely to be more complex than this due to the distinct cultural history and socialization experiences of certain race/gender groups (see for instance, Parker and Ogilvie, 1996 for a discussion of such distinctions in the case of African-American females). Thus, in an effort to avoid overgeneralization regarding the expected nature of the interaction between race and gender, yet to also avoid articulating specific expectations for each possible race/gender combination of interest, I offer the following proposition. Proposition 3. An individual’s specific race and gender will significantly influence the strength of the relationship between their demographic dissimilarity in terms of (a) race and (b) gender and their willingness to share uniquely held information during group discussion. Elsass and Graves (1997) suggest that the cognitive processes discussed above (status judgments, stereotyping, and similarity bias) negatively impact the willingness of minority group members to share information partly due to feelings of self-doubt stemming from the internalization of the low status and negative stereotypes associated with their category memberships. This suggests

76

MELVIN L. SMITH

that the level of a minority group member’s self-esteem may play a role in their willingness to share information in a White male dominated group. In a broader consideration of how self-esteem might impact the behavior of individuals within a group, Brown (1997) draws on the theory of narcissism, which suggests that individuals have a need to maintain a positive sense of self, and will thus engage in ego-defensive behaviors in order to preserve self-esteem. While high-esteem individuals are said to have a preference for ego defenses that repress, deny, or ignore challenges, it is believed that low-esteem individuals are more open to social influence. In addition, Kernis et al. (1989) found that low self-esteem individuals had a greater tendency than did high self-esteem individuals to over-generalize the implications of negative feedback to other aspects of their identities. Therefore, it is likely that self-esteem may influence an individual’s contributions to the discussion of a diverse group when they find themselves in the numerical minority regardless of whether they are also in a socially defined minority category. In addition to self-esteem, there are other personality traits concerned with how an individual views himself or herself that might also influence their contributions during group discussion within a diverse workgroup. Judge et al. (1997) proposed the existence of a higher order construct that they termed core self-evaluations or positive self-concept. Core self-evaluations represent “basic conclusions or bottom-line evaluations that individuals hold about themselves” (Judge & Bono, 2001, p. 80). The construct consists of four specific personality traits. The four traits are: (1) self esteem (the overall value one places on oneself as a person); (2) generalized self-efficacy (one’s estimate of one’s fundamental ability to cope, perform, and be successful); (3) internal locus of control (one’s belief that they can control a broad array of factors in their life); and (4) emotional stability or low neuroticism (the tendency to be confident, secure, and steady). Judge et al. (1998) concluded from their study of job and life satisfaction that core self-evaluations as a construct have consistent effects on how people view their jobs as well as their lives. Having a positive self-concept can thus potentially offset the negative stereotypes associated with being demographically dissimilar from other group members and the associated effect that it has been proposed to have on an individual’s sharing of uniquely held information during group discussion. This suggests the following proposition. Proposition 4. An individual’s positive self-concept (i.e. self-esteem, generalized self-efficacy, locus of control, and emotional stability) will significantly decrease the strength of relationship between their demographic dissimilarity in terms of (a) race and (b) gender and their willingness to share uniquely held information during group discussion.

If I Know It, Will I Share It?

77

PROPOSED MODERATING FACTORS – GROUP-LEVEL In addition to the proposed individual-level moderating factors discussed above, there are several group-level factors also likely to moderate the relationship between an individual’s demographic dissimilarity and their contributions to group discussions. One such factor is the nature of the group’s task. With more routine tasks, groups have less of a need to exchange ideas or to challenge one another (Pelled et al., 1999). When the task is more uncertain, however, given the task requirements, group members are likely to be expected to share their diverse opinions and ideas with the group. Therefore, when the group task lacks certainty, individuals who are demographically dissimilar may feel less constrained and less compelled to share only information that is consistent with what has already been discussed (shared information) and, thus, may be more likely to share uniquely held information that might facilitate the completion of the group’s task. Therefore, I propose the following: Proposition 5. Task uncertainty will significantly weaken the negative relationship between an individual’s demographic dissimilarity in terms of (a) race and (b) gender and their willingness to share uniquely held information during group discussion. Other group-level factors that might possibly influence the relationship between demographic dissimilarity and unique contributions to the group are group characteristics such as group longevity and the nature of the group’s leadership. First considering group longevity, when individuals initially join a group, levels of familiarity with other group members is likely to be low, as are trust levels. Anand et al. (2003) contend that, individuals who lack familiarity and trust are likely to carefully assess the norms of the group and are unlikely to present information that may be unusual, competing, or different from what they presume the core knowledge of the group to be (p. 25).

They further suggest that under these circumstances the benefits of external knowledge may never be fully realized. As an individuals’ tenure with the group increases, however, and the group gains experience working together, corresponding increases in familiarity and trust are likely to make group members feel more comfortable offering information that is not commonly held by other group members during group discussion. Pelled et al. (1999) contend that the effects of increased group longevity are partly due to a blurring of social categories based on demographic attributes over time. They suggest that, over time, the “in-group” becomes the entire workgroup rather than the dominant majority. Additionally, as

78

MELVIN L. SMITH

group longevity increases group members are less likely to resort to stereotypes and negative attributions based strictly on demographic characteristics. Instead, any categorizations that are made are more likely to be based on “deep-level” information such as values, beliefs, and attitudes rather than “surface-level” characteristics such as race and gender (Harrison et al., 1998, 2002). These things taken together suggest that increased group longevity should decrease the strength of the negative relationship between demographic dissimilarity and individual willingness to share uniquely held information during group discussion. I, therefore, propose the following: Proposition 6. The strength of the negative relationship between an individual’s demographic dissimilarity in terms of (a) race and (b) gender and their willingness to share uniquely held information during group discussion will be significantly decreased by group longevity. The other group characteristic considered here as a potential moderating factor of the relationship between demographic dissimilarity and the sharing of uniquely held information is the nature of the group’s leadership. While some groups are self-managed and, therefore, function without a formal group leader, the presence of a leader can play an important role in the creation and use of knowledge within groups and organizations (Bryant, 2003). Leaders may facilitate the creation and sharing of knowledge in a variety of ways, including their ability to recognize and reward information sharing behavior, or simply by creating a climate that is receptive to new ideas (Bartol & Srivastava, 2002; Bryant, 2003). Using these and other techniques, a group leader has the opportunity to promote the open sharing of information on the part of all group members. However, not all leadership styles are conducive to facilitating the sharing of information within groups. Drawing on the work of Conger (1999) and Yukl and Howell (1999), Bryant (2003) suggests that a transactional leadership style encourages specific exchanges of information closely tied to expected or contracted behavior. While this may facilitate the effective use of existing knowledge, it is not likely to lead to the creation and sharing of new knowledge. Bryant contends that transformational leadership, on the other hand, is conducive to the creation and sharing of new knowledge due to the effect that it has on individual motivation and commitment to the organization. Extending this argument to the context of information sharing within demographically diverse workgroups, it is likely that the presence of a transformational leader will decrease the strength of the negative relationship between demographic dissimilarity and group member contribution, thus having a positive influence on the individual sharing of uniquely held information. This leads to the following proposition.

If I Know It, Will I Share It?

Fig. 2. Meso-level Model of Group Composition and Information Sharing within Groups.

79

80

MELVIN L. SMITH

Proposition 7. The strength of the negative relationship between an individual’s demographic dissimilarity in terms of (a) race and (b) gender and their willingness to share uniquely held information during group discussion will be significantly decreased by the presence of a transformational group leader. A final group-level factor suggested to moderate the relationship between group member demographic dissimilarity and their unique contributions to the group during group discussion is the presence of organizational incentives that reward information sharing behavior. An individual’s reluctance to share uniquely held information may partially stem from a lack of motivation to do so (Szulanski, 1996). This lack of motivation can stem from a desire to hoard knowledge for reasons of power, or from the belief that information sharing behavior will not be adequately rewarded (Bartol & Srivastava, 2002; Szulanski, 1996). Thus, one way to encourage group members to share uniquely held information would be to reward that behavior via group and/or organizational incentives designed to reward that type of behavior. The presence of such incentives should serve to decrease the negative effects of demographic dissimilarity on the sharing of uniquely held information, thus leading to the following proposition. Proposition 8. The strength of the relationship between an individual’s demographic dissimilarity in terms of (a) race and (b) gender and their willingness to share uniquely held information during group discussion will be significantly decreased by the presence of group or organizational incentives designed to reward information sharing behavior. A conceptual framework representing each of the proposed relationships is presented in Fig. 2.

IMPLICATIONS FOR RESEARCH AND PRACTICE The propositions developed in this article build upon the work of Stasser and Titus (1987) as well as Stasser et al. (1989). Both of these studies examined factors influencing the discussion of unshared information in groups. However, neither study considered the potential effects of group composition on the information sharing behavior exhibited by group members. The consideration of potential group composition effects and, more specifically, the effects of individual group member demographic dissimilarity on individual willingness to share uniquely held information during group discussion is the primary contribution of the present article. Several individual- and group-level factors proposed to moderate this

If I Know It, Will I Share It?

81

relationship were also considered. Future research could broaden this contribution by further developing and then testing the propositions presented. Additionally, future research would benefit from the exploration of other factors not considered here, but that could also conceivably influence information sharing behavior in diverse groups. For instance, it is possible that in addition to the task and group characteristics considered in this article, occupations or industries within which individuals in decision-making groups are embedded may have an influence on their contributions to the group. Further, while group longevity was proposed to have an influencing effect, the specific stage of the group in its development cycle based on Tuckman’s (1965) forming, storming, norming, and performing framework could also be considered. Researchers could examine whether the sharing of uniquely held information might be more or less likely during one stage of the group’s development vs. another. Another potentially interesting avenue for future research might be the examination of situations where the members of each socially defined subgroup possess information that is not shared with the “other” group. This information would thus represent partially shared information, as discussed by Cruz et al. (1997). This would allow the testing of whether information that is shared only among “like” members of a group affects their willingness to mention that information during a discussion involving the entire group. Yet another interesting extension regarding the nature of unshared information in a group might be the examination of whether minority group members would be more likely to share uniquely held information with the group if they know that the information is crucial to the accomplishment of the group’s task. These are just a few of the many possibilities for expanding our understanding of information sharing in demographically diverse groups. The ideas developed in this article also represent potentially significant implications for practitioners. First, the concept of fragile vs. stable organizational memory suggests that organizations might want to attempt to consciously manage their degree of vulnerability regarding certain elements of organizational memory. To the extent that this is the case, processes may be put into place to encourage increased sharing of information. Organizations could proactively balance the desire for greater efficiency (focusing on minimizing redundant knowledge) and less vulnerability (focusing on creating redundant knowledge sources). In doing so, judgments would have to be made as to whether the potential costs to the organization resulting from the loss of unshared information justify the additional costs associated with creating redundant stores of that information/knowledge. Second, understanding the potential effects of group composition on the level of information sharing within groups might suggest to organizations a need to

82

MELVIN L. SMITH

more actively manage the composition of its various groups (teams) and/or to facilitate information sharing through interventions aimed at reducing potential process losses often associated with heterogeneous groups. This may mean consciously considering the demographic composition of teams put in place to accomplish various tasks as opposed to creating the teams strictly based on functional representation or other commonly used criteria. Finally, organizations might want to ensure that when group tasks depend upon the contributions of individuals with a wide variety of perspectives, there are sufficient sources of motivation in place to encourage the desired information sharing behavior. This might entail the use of organizational incentives and/or placing leaders in place that can create a climate that promotes the active sharing of information from all group members regardless of their specific demographic attributes. In sum, this article makes a contribution by suggesting several propositions regarding the effects of group composition relative to specific social categories on the degree of information sharing that takes place within the group. It also stresses the potential importance of information sharing on the stability (or instability) of organizational memory, which has potentially significant implications for both research and practice. Rather than considering simple group composition effects on the level of information sharing within groups (a macro level approach), this article proposed a variety of micro- and macro-level factors that might influence the willingness of individual group members to share uniquely held information during group discussion.

REFERENCES Alderfer, C. P., & Smith, K. K. (1982). Studying intergroup relations embedded in organizations. Administrative Science Quarterly, 27, 35–65. Akoyo, O. B., Hartel, C. E. J., & Callan, V. J. (2002). Resolving the puzzle of productive and destructive conflict in culturally heterogeneous workgroups. International Journal of Conflict Management, 13, 165–196. Anand, V., Clark, M. A., & Zellmer-Bruhn, M. (2003). Team knowledge structures: Matching task to information environment. Journal of Managerial Issues, 15, 15–33. Anand, V., Manz, C. C., & Glick, W. H. (1998). An organizational memory approach to information management. Academy of Management Review, 23, 796–809. Argote, L. (1999). Organizational learning: Creating, retaining, and transferring knowledge. Boston: Kluwer. Ashforth, B. E., & Mael, F. (1989). Social identity theory and the organization. Academy of Management Review, 14, 20–39. Bantel, K. A., & Jackson, S. E. (1989). Top management innovations in banking: Does the composition of the top team make a difference? Strategic Management Journal, 10, 107–124.

If I Know It, Will I Share It?

83

Bartol, K. M., & Srivastava, A. (2002). Encouraging knowledge sharing: The role of reward systems, Journal of Leadership. Journal of Leadership & Organizational Studies, 9, 64–77. Bell, E. L. (1990). The bicultural life experience of career-oriented Black women. Journal of Organizational Behavior, 11, 459–477. Brewer, M. B. (1979). In-group bias in the minimal intergroup situation: A cognitive-motivational analysis. Psychological Bulletin, 86, 307–324. Brown, A. D. (1997). Narcissism, identity, and legitimacy. Academy of Management Review, 22, 643–686. Bryant, S. (2003). The role of transformational and transactional leadership in creating, sharing, and exploiting organizational knowledge. Journal of Leadership & Organizational Studies, 9(4), 32–47. Byrne, D. (1971). The attraction paradigm. New York: Academic Press. Chatman, J. A., Polzer, J. T., Barsade, S. G., & Neale, M. A. (1998). Being different yet feeling similar: The influence of demographic composition and organizational culture on work processes and outcomes. Administrative Science Quarterly, 43, 749–780. Conger, J. A. (1999). Charismatic and transformational leadership in organizations: An insider’s perspective on these developing streams of research. Leadership Quarterly, 10, 145–169. Cox, T. H., & Blake, S. (1991). Managing cultural diversity: Implications for organizational competitiveness. Academy of Management Executive, 5, 45–56. Cox, T. H., & Nkomo, S. M. (1990). Invisible men and women: A status report on race as a variable in organization behavior research. Journal of Organizational Behavior, 11, 419–431. Cruz, M. G., Boster, F. J., & Rodriguez, J. I. (1997). The impact of group size and proportion of shared information on the exchange and integration of information in groups. Communication Research, 24, 291–313. DiTomaso, N., & Hooijberg, R. (1996). Diversity and the demands of leadership. Leadership Quarterly, 7, 163–187. Elsass, P. M., & Graves, L. M. (1997). Demographic diversity in decision-making groups: The experiences of women and people of color. Academy of Management Review, 22, 946–973. Fernandez, J. P. (1981). Managing a diverse workforce. Lexington, MA: Lexington Books. Gilbert, J., Carr-Ruffino, N., Ivancevich, J. M., & Lownes-Jackson, M. (2003). An empirical examination of inter-ethnic stereotypes: Comparing Asian American and African American employees. Public Personnel Management, 32, 251–265. Grant, R. M. (1996). Prospering in dynamically-competitive environments: Organizational capability as knowledge integration. Organization Science, 7, 375–387. Harrison, D. A., Price, K. H., & Bell, M. P. (1998). Beyond relational demography: Time and the effects of surface- and deep-level diversity on work group cohesion. Academy of Management Journal, 41, 96–107. Harrison, D. A., Price, K. H., Gavin, J. H., & Florey, A. T. (2002). Time, teams, and task performance: Changing effects of surface- and deep-level diversity on group functioning. Academy of Management Journal, 5, 1029–1045. Hoffman, L. R., & Maier, J. R. F. (1961). Quality and acceptance of problem-solving. Journal of Abnormal and Social Psychology, 58, 27–32. Huber, G. (1991). Organizational learning: The contributing processes and the literatures. Organization Science, 2, 88–115. Ibarra, H. (1995). Race, opportunity, and diversity of social circles in managerial networks. Academy of Management Journal, 38, 673–703.

84

MELVIN L. SMITH

Jackson, S. E. (1992). Team composition in organizational settings: Issues in managing the increasingly diverse work force. In: S. Worchel, W. Wood & J. A. Simpson (Eds), Group Process and Productivity (pp. 138–173). Newbury Park, CA: Sage. Johnson, C., Clay-Warner, J., & Funk, S. J. (1996). Effects of authority structures and gender on interaction in same-sex task groups. Social Psychology Quarterly, 59, 221–236. Judge, T. A., & Bono, J. E. (2001). Relationship of core self-evaluations traits – Self-esteem, generalized self-efficacy, locus of control, and emotional stability – With job satisfaction and job performance: A meta-analysis. Journal of Applied Psychology, 86, 80–92. Judge, T. A., Locke, E. A., & Durham, C. C. (1997). The dispositional causes of job satisfaction: A core evaluations approach. Research in Organizational Behavior, 19, 151–188. Judge, T. A., Locke, E. A., Durham, C. C., & Kluger, A. N. (1998). Dispositional effects on job and life satisfaction: The role of core evaluations. Journal of Applied Psychology, 83, 17–34. Kanter, R. M. (1977). Some effects of proportions on group life: Skewed sex ratios and responses to token women. American Journal of Sociology, 82, 965–990. Kernis, M. H., Frankel, B. S., & Brockner, J. (1989). Self-esteem and reactions to failure: The mediating role of overgeneralization. Journal of Personality and Social Psychology, 57, 707–727. Larkey, L. K. (1996). Toward a theory of communicative interactions in culturally diverse workgroups. Academy of Management Review, 21, 463–491. Lau, D. C., & Murnighan, J. K. (1998). Demographic diversity and faultlines: The compositional dynamics of organizational groups. Academy of Management Review, 23, 325–340. Levitt, B., & March, J. G. (1988). Organizational learning. Annual Review of Sociology, 14, 319–340. McGuire, W. J., & McGuire, C. (1981). The spontaneous self-concept as affected by personal distinctiveness. In: M. D. Lynch, A. A. Norem-Hebeisen & K. J. Gergen (Eds), Self-Concept: Advances in Theory and Research. (pp. 147–171). Cambridge, MA: Ballinger. McLeod, P. L., Lobel, S. A., & Cox, T. H. (1996). Ethnic diversity and creativity in small groups. Small Group Research, 27, 248–264. Milliken, F. J., & Martins, L. L. (1996). Searching for common threads: Understanding the multiple effects of diversity in organizational groups. Academy of Management Review, 21, 402–433. Moreland, R. L., Argote, L., & Krishnan, R. (1996). Socially shared cognition at work: Transactive memory and group performance. In: J. L. Nye & A. M. Brower (Eds), What’s Social About Social Cognition? Research on Socially Shared Cognition in Small Groups (pp. 57–84). Thousand Oaks, CA: Sage. Mullen, B. (1983). Operationalizing the effect of the group on the individual: A self-attention perspective. Journal of Experimental Social Psychology, 19, 295–322. Nemeth, C. J. (1992). Minority dissent as a stimulant to group performance. In: S. Worchel, W. Wood & J. A. Simpson (Eds), Group Process and Productivity (pp. 95–111). Newbury Park, CA: Sage. Okhuysen, G. A., & Eisenhardt, K. M. (2002). Integrating knowledge in groups: How formal interventions enable flexibility. Organization Science, 13, 370–387. O’Reilly, C. A., Caldwell, D. F., & Barnett, W. P. (1989). Work group demography, social integration, and turnover. Administrative Science Quarterly, 34, 21–37. Parker, P. S. (2001). African-American women executives’ leadership communication within dominant-culture organizations: (Re)conceptualizing notions of collaboration and instrumentality. Management Communication Quarterly, 15, 42–82. Parker, P. S., & Ogilvie, D. T. (1996). Gender, culture, and leadership: Toward a culturally distinct model of African-American women executives’ leadership strategies. Leadership Quarterly, 7, 189–214.

If I Know It, Will I Share It?

85

Pelled, L. H. (1996). Demographic diversity, conflict, and work group outcomes: An intervening process theory. Organization Science, 6, 615–631. Pelled, L. H., Eisenhardt, K. M., & Xin, K. R. (1999). Exploring the black box: An analysis of work group diversity, conflict, and performance. Administrative Science Quarterly, 44, 1–28. Pettigrew, T. F. (1979). The ultimate attribution error: Extending Allport’s analysis of prejudice. Psychological Bulletin, 5, 461–476. Riordan, C. M., & McFarlane Shore, L. (1997). Demographic diversity and employee attitudes: An empirical examination of relational demography within work units. Journal of Applied Psychology, 82, 342–358. Stasser, G., & Stewart, D. (1992). Discovery of hidden profiles by decision-making groups: Solving a problem vs. making a judgment. Journal of Personality and Social Psychology, 63, 426–434. Stasser, G., Taylor, L. A., & Hanna, C. (1989). Information sampling in structured and unstructured discussions of three- and six-person groups. Journal of Personality and Social Psychology, 57, 67–78. Stasser, G., & Titus, W. (1987). Effects of information load and percentage of shared information on the dissemination of unshared information during group discussion. Journal of Personality and Social Psychology, 53, 81–93. Szulanski, G. (1996). Exploring internal stickiness: Impediments to the transfer of best practice within the firm. Strategic Management Journal, 17, 27–43. Tajfel, J. (1974). Social categorization, social identity and social comparison. Social Science Information, 12, 61–76. Tsui, A. S., Egan, T. D., & O’Reilly, C. A. (1992). Being different: Relational demography and organizational attachment. Administrative Science Quarterly, 37, 549–579. Tsui, A. S., & Gutek, B. A. (1999). Demographic differences in organizations: Current research and future directions. Lanham, MD: Lexington Books. Tsui, A. S., & O’Reilly, C. A. (1989). Beyond simple demographic effects: The importance of relational demography in superior-subordinate dyads. Academy of Management Journal, 32, 402–423. Tuckman, B. W. (1965). Developmental sequence in small groups. Psychological Bulletin, 63, 384–399. Turner, J. C. (1987). Rediscovering the social group: A self-categorization theory, New York: Basil Blackwell. Walsh, J. P., & Ungson, G. R. (1991). Organizational memory. Academy of Management Review, 16, 57–91. Wegner, D. M. (1987). Transactive memory: A contemporary analysis of the group mind. In: B. Mullen & G. R. Goethals (Eds), Theories of Group Behavior (pp. 185–208). New York: Springer-Verlag. Wegner, D. M. (1995). A computer network model of human transactive memory. Social Cognition, 13, 319–339. Wexler, M. N. (2002). Organizational memory and intellectual capital. Journal of Intellectual Capital, 3, 393–415. Wiersema, M. F., & Bantel, K. A. (1992). Top management team demography and corporate strategic change. Academy of Management Journal, 35, 91–121. Williams, K. Y., & O’Reilly, C. A., III (1998). Demography and diversity in organizations: A review of 40 years of research. In: Research in Organizational Behavior (Vol. 20, pp. 77–140). Greenwich, CT: JAI Press. Wittenbaum, G. M., Hubbell, A. P., & Zuckerman, C. (1999). Mutual enhancement: Toward an understanding of the collective preference for shared information. Journal of Personality and Social Psychology, 77, 967–978.

86

MELVIN L. SMITH

Wittenbaum, G. M., & Stasser, G. (1996). Management of information in small groups. In: J. L. Nye & A. M. Brower (Eds), What’s Social About Social Cognition? Research on Socially Shared Cognition in Small Groups (pp. 3–28). Thousand Oaks, CA: Sage. Yukl, G., & Howell, J. M. (1999). Organizational and contextual influences on the emergence and effectiveness of charismatic leadership. Leadership Quarterly, 10, 257–283.

THE IMPACT OF SOCIAL CAPITAL ON AFRICAN-AMERICAN AND WOMEN SURVIVORS OF ORGANIZATIONAL DOWNSIZING Rochelle Parks-Yancy ABSTRACT Research has found that African-Americans and women have opportunities for advancement in the traditional corporate environment through resources embedded in their social networks. However, layoffs can affect the composition of their social networks, their positions in the networks, and rewards from those networks. I suggest that the racial, cultural, and gender differences between African-American and women layoff survivors and White and male layoff survivors will negatively affect their access to and benefits from social capital resources. Yet, strong tie relationships with White and male layoff survivors in key strategic positions can help African-American and women survivors maintain their existing job position because they can then borrow the social capital resources of the White and male survivors. Thus, while research has found that weak ties help individuals advance in their workplaces, strong tie relationships with majority groups may be more beneficial to minority groups in maintaining their position after a layoff.

Diversity in the Workforce Research in the Sociology of Work, Volume 14, 87–103 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0277-2833/doi:10.1016/S0277-2833(04)14004-1

87

88

ROCHELLE PARKS-YANCY

This paper examines the relationship between social capital and downsizing for two traditionally disadvantaged groups: African-Americans and women. Social capital is the term given to the information, influence, solidarity, and opportunities deriving from the relationships among people in a social group or network (family, community, society or group based on other forms of affiliation). The behaviors among people in the group are governed by norms (accepted behaviors) that define reciprocal obligations (give-and-take behaviors) and the sanctions that are likely for violating group norms (Lin, 2000; Putnam, 2000). In his seminal piece on interpersonal ties, Granovetter (1973) suggested that weak ties provided greater mobility opportunities than strong ties. He defined the strength of a tie as “a combination of the amount of time, emotional intensity, the intimacy, and reciprocal services which characterize a tie” (p. 1361). Ties denote the type and amount of social capital resources an entity may have access to, as well as the entities’ position in a social network (Lin, 2000). Strong ties refer to close emotional connections (e.g. family), whereas weak ties are weaker connections (e.g. acquaintances). Granovetter’s theory was supported by his empirical study on how people find jobs (1995). Jobs that were considered to be a “step up” from previously held jobs were often found through weak ties. Furthermore, his research has been extended by other researchers, such as Burt (1998) and Ibarra (1992, 1995). Burt (1998) and Ibarra (1992) found that gender affected the ways in which social networks facilitated upward mobility in organizations for men, women, and non-Whites, respectively. While women obtained social support from relationships with other women, they utilized relationships with men for instrumental support because the women’s networks were not as rich in information, influence, or opportunity as the men’s. In contrast, men tended to utilize other men for both social and instrumental support. Social support refers to emotional encouragement, whereas instrumental support refers to support needed for moving ahead in one’s career. This type of support includes access to key information, influence, opportunities and solidarity. Ibarra (1995) found that the experiences of minorities were parallel to those of women: due to similar deficiencies in their same-race social networks, they tended to use Whites for instrumental support and other minorities for social support. Also, the women had strong tie relationships with other women and relatively weaker ties with the men (Ibarra, 1992). However in Burt’s study (1998), in order for women to be able to utilize their relationships with their male colleagues for instrumental support, they had to develop strong tie relationships with men in key positions of power. For women, strong ties with these colleagues enabled them to access the weak ties of these colleagues (thereby developing more weak ties of their own), which facilitated their career advancement.

The Impact of Social Capital

89

These authors’ findings suggest that ideal combinations of strong and weak ties for women and minorities in organizations entail having relationships with other women and minorities for social support (strong tie) and utilizing relationships with Whites and males for instrumental support (strong and weak ties). However, organizational changes can disrupt the pattern of strong and weak ties to the extent that the combinations and benefits of the ties may be altered. One such change, which can have a profound effect on an organization and its employees, is downsizing (Brockner, 1992). Downsizing can affect the productivity, morale, and organizational commitment of even those employees who remain with the organization, hereafter referred to as “survivors” (Brockner, 1992). Survivors may experience a change in relationships with supervisors, subordinates, and peers and a change in access to important organizational resources (Susskind et al., 1998). They may also experience mental and emotional trauma associated with the layoffs. Several researchers have examined downsizing in conjunction with other organizational phenomena. Shah (2000) studied downsizing and its effect on network structures, Armstrong-Stassen (1998) studied survivors’ perceptions of the procedural justice of the downsizing, and Susskind et al. (1998) examined downsizing and its effect on structural holes within the networks of survivors. Shah (2000) examined the changes in survivors’ structural position in their firm after the downsizing of friends and co-workers. He found that survivors were dissatisfied with the loss of friendship networks after the layoffs, but that the loss of co-workers (who were not friends) resulted in less perceived competition for promotions. Armstrong-Stassen (1998) found that gender and organizational level affected survivor perceptions of the procedural injustice of the downsizing. Female technicians who survived the layoff felt more vulnerable than male technicians to the possibility of a layoff happening to them, due to perceptions of gender discrimination. Susskind et al. (1998) found that a structural hole index affected the survivors’ perceptions of the organization. Structural holes are disconnects between people within different networks. The holes provide opportunities for people, who have relations with the disconnected parties, to “broker” information flows between the disconnected parties (Burt, 1997) and to benefit from a broader array of network access. In Susskind et al.’s study (1998), survivors whose networks increased in structural holes after the layoff were less satisfied than those whose networks’ structural holes decreased. These studies did not examine the nature of the weak and strong ties of survivors, nor did they address how these processes may differentially affect particular groups. Although Armstrong-Stasssen (1998) recommended that strategies to help survivors adjust to downsizing should be tailored specifically to the differing concerns of various survivor groups, these results may differ for

90

ROCHELLE PARKS-YANCY

survivors who are also members of historically disadvantaged groups. Thus, this paper examines the changes in social capital following a downsizing for women and African-Americans.

STRONG AND WEAK TIES Ties are the social relations that facilitate exchanges of information, influence, opportunity, and solidarity among individuals, groups, and organizations. Strong ties include familial relationships, close friendships and close co-workers, whereas weak ties include relationships with acquaintances and distant co-workers. The nature of strong ties relates to the concept of “bonding social capital.” Bonding social capital facilitates the maintenance of current positions in a network. Weak ties are related to the concept of “bridging social capital.” Bridging social capital are linkages between distant networks that facilitate mobility between those networks (Putnam, 2000). “Bonding social capital is good for getting by, bridging social capital is good for getting ahead” (Putnam, 2000, p. 23). These two types of social capital concur with Granovetter’s (1973) characterization of strong and weak ties. Granovetter (1995) studied how people obtain jobs in professional, managerial, and technical careers. He noted that structure has a large impact on individuals’ size, type, and access to strong and weak ties. Structure refers to one’s social relationships, which shape the size and content of one’s networks. One usually does not have much control over the ties that one has. Educational level and income are dimensions that can shape the size and content of a network. The greater one’s educational level, usually the larger one’s network of contacts that can provide job information; the same holds true for income level. Granovetter found that 80% of higher income respondents found jobs through work contacts (weak ties), as opposed to family-social contacts (strong ties). Lower income respondents were likely to find jobs equally through family-social contacts and work contacts. As race often corresponds to income, Granovetter’s findings suggest that racial minorities, particularly African-Americans and Hispanics, who are often at a lower income and educational level than Whites or Asians, may have a smaller network of work contacts from which to draw to facilitate job mobility, and the contacts they have may be less helpful. Because members of these groups may have less education and lower income levels than their Asian or White counterparts, their strong ties may also not provide the means to obtain jobs that facilitate upward mobility. In this regard, Blacks and Hispanics may be more likely to use formal means such as employment agencies and advertisements to obtain jobs. Rosenbaum et al. (1999) found that young females and

The Impact of Social Capital

91

African-Americans who had recently graduated from high school received more help from institutional contacts, as opposed to relatives, in finding initial jobs. Nine years after the initial study, the effects were still the same: for African-Americans and young women school placement still had more effect on current earnings than did relatives. In Granovetter’s study (1995), the respondents were predominantly White males. In his study, strong tie relationships were most beneficial in obtaining first jobs, as the respondents had not yet developed the contacts that could be instrumental in obtaining employment or better employment. Strong tie relationships were also beneficial to those in dire need of a new job, such as the unemployed. However, career advances were primarily obtained through weak ties over successive jobs. Each job usually provided access to additional occupational contacts that had the potential to offer information about subsequent jobs. There are several reasons why weak ties, as opposed to strong ties, are more beneficial in facilitating job mobility over a career span. As strong ties are, by definition, close relations, little information is known by one member of the network that is not known by other members of the network. That is, job information will tend to be redundant, thereby limiting knowledge of new opportunities. Weak ties are relations in which individuals do not necessarily have knowledge of or access to the same people. These ties can encompass one’s acquaintances, distant co-workers, and former co-workers, all of whom will have relations with their own networks. Hence, the prospect of learning about new job opportunities is much greater with weak ties (Granovetter, 1973). While Granovetter’s study was in the context of how people found jobs in a labor market, Burt (1997) examined social capital and career mobility within organizations. He concurs with Granovetter that social capital depends on “a person’s location in the social structure of a market or hierarchy” (1997, p. 339). His study of the contingent value of social capital suggests that the information flow and control benefits of bridging structural holes are more valuable to managers with few peers than managers with many peers. The ability to bridge structural holes allows the information “brokers” timely access to knowledge and referrals (information) and to control who receives information and in what manner (control). Information “brokers” gain information from both sides of the structural hole by serving as a bridge between disconnected parties. As the “broker” receives information from both sides of the structural hole, the information tends to be richer in content, detail, and usefulness. The “brokers” can also disseminate information about opportunities, and, in doing so, incur obligations from members of various networks. Bridging structural holes is a form of social capital, in that it defines social capital in terms of the benefits of information and control in brokering interactions

92

ROCHELLE PARKS-YANCY

between people whose networks are otherwise disconnected within a social structure (Burt, 1997, p. 340). It is of more value to managers with few peers than managers with many peers for two reasons: (1) because managers with few peers have a limited number of people from which to model “appropriate” behavior; and (2) because their work does not have the legitimacy that would be provided if there were more people doing the same job. Burt also found that new managers and non-White managers valued social capital more than other managers. Because both new and non-White managers are usually very visible, it tends to highlight any errors they may make. Bridging structural holes provided them with the necessary information access and control so they could monitor their behaviors in accordance with explicit and tacit expectations.

MINORITIES, WOMEN, AND CAREER MOBILITY THROUGH INTERPERSONAL TIES The major premise of this paper is that career paths can be interrupted by the loss of social ties from whom individuals draw information, influence, and opportunity to affect their career outcomes. When individuals lose access to social connections who can, for example, help them obtain a job, obtain better jobs, or get promoted, it may be difficult for them to continue the presumably upward direction of their career trajectory. This is because developing the trust and reciprocal obligations that inhere in social relations takes time (Burt, 1998). Hence, having to spend time developing new social ties suggests that one’s career outcomes will be forestalled, at least until the trust and obligations with the new social ties are sufficiently developed to warrant them helping individuals with their career aspirations. We know that most jobs are found through social capital resources and that individuals in jobs that pay well and offer high prestige are often embedded in a web of social relations that facilitate obtaining and maintaining those jobs (Granovetter, 1973; Lin, 2000). As such, it may be very difficult for individuals who lose those social ties to obtain and keep those kinds of jobs, because the ties can no longer offer information, influence, opportunity, or solidarity to help them with their careers. This may be especially true for Blacks and women, most of whom have lesser access to social capital resources anyway, relative to Whites and men. Certainly women and Blacks who want to improve their careers can benefit from developing more social ties with men and Whites, because such ties can provide better quality and quantity of social capital resources than would be available within their own group (Burt, 1998; Hanson & Pratt, 1991). However, social capital resources from

The Impact of Social Capital

93

men and Whites to women and Blacks may be limited because the ability for women and Blacks to reciprocate may also be limited. Social capital resource exchanges between groups that differ by gender or race/ethnicity do take place from time to time because of family ties, friendship, or occasionally general good will. Yet, because of the existing differences in social capital resources (Lin, 2001) and people’s homophilic tendencies (Ibarra, 1992), exchanges between groups that differ on these dimensions are less likely than they are within groups that are more homogeneous. For example, studies of the interpersonal ties of minorities and women have been undertaken by researchers such as Ibarra (1992, 1995), Burt (1998), and Lin (2000). These researchers have generally found that, in professional and managerial careers, women and minorities have less access to instrumental ties that would put them on an equal or greater footing than White males. There are two reasons for these findings: (1) there are often fewer minorities and women in such positions to provide instrumental ties; and (2) Whites and men often have homophilous (same race and gender) preferences for instrumental networks, so they are less likely to share their opportunities with women and minorities (Ibarra, 1992). Because they lack legitimacy as members, women and minorities may have difficulty gaining entrance to the “inner circles” shared by White males (Burt, 1998). Ibarra (1995) empirically examined differences between the informal networks of minority managers and White managers. She found that minority managers had a smaller proportion of homophilous (same race) ties than White managers, but high-potential minority managers had a wider range of informal networks than White managers or lower-potential minorities. Even so, these networks included a higher proportion of minority contacts, and they were less likely to provide instrumental utility for high-potential minorities, compared to their ties with White managers. While the minority networks were beneficial for sharing common experiences and gaining grapevine information (for social support), network ties to Whites were still necessary for career advancement. Both White and minority middle level managers utilized ties to upper level White managers for advancement. However, the ties to White upper level managers did not necessarily yield the same results for high-potential middle level minorities as they did for White middle level managers. The White middle managers still advanced faster and further than minorities of similar rank. This was attributed to the racial and the positional differences between the minorities and the White upper level managers, while for White middle level managers, the positional differences between them and White upper level managers did not create barriers, presumably because of the shared race/ethnic background.

94

ROCHELLE PARKS-YANCY

In her study of gender differences in network structure and strategy, Ibarra (1992) found that “gender inequalities in organizational power distributions were due to sex differences in homophily and the conversion of personal attributes and resources into network advantages” (p. 422). Men tended to use other men for instrumental and social support, whereas women utilized other women for social support and men for instrumental support. Men also tended to have more central network positions than women, which suggests that men’s credibility and legitimacy in the network was more likely to be assured than that of women. Tenure could provide women with more network centrality, but tenure paid off, even for men, only when it led to higher authority jobs. Although men tended to use other men for both instrumental and support networks, some were also likely to include women in their support networks. Therefore, women who restricted their networks to other women may have overlooked a possible “opening” in the dominant group’s support network. This “opening” could conceivably have led to a position in men’s instrumental networks. Burt (1998) addressed the issue of how women gain access to and acceptance in men’s instrumental networks. He examined the “borrowing” of social capital for legitimacy and career advancement. His study found that those women and young men who developed strong ties with a strategic partner who had extensive weak ties advanced faster than those who did not develop such ties. The strategic partner was usually a senior male whose networks had many structural holes. The partner was a bridge between the structural holes, and he provided access to those holes for women and young men with whom he had developed close relationships (strong ties). Burt (1998) argued that women and entry-level men suffer from legitimacy issues in the firm, therefore the strong ties to senior managers with weak ties extended their own ties and, hence, both provided legitimacy and facilitated their acceptance by other senior members. This “borrowing” of social capital from a strategic partner enabled women and young men to enter senior managers’ instrumental networks, be accepted as legitimate members of those networks, and, as a result, facilitated their own career advancement. Although “borrowing” social capital worked well for women and young men, it had a negative effect on senior men’s promotion opportunities. Senior men were expected to take direct advantage of structural holes for networking and promotions, so for them indirect ties were not helpful. Burt (1998) also proposed that the differences in the networks available to men and women related more to legitimacy issues than to attributes such as race or gender. Acceptance into a network, he argued, is dependent on sponsorship by an established member of the network, rather than occurring only through homophilic (same race or gender) ties within the network. While

The Impact of Social Capital

95

his own study supported the notion that legitimacy is more important than race or gender, other researchers have found that acceptance in a group is affected by the type and number of differences between the network members and the individual desiring acceptance into the network (Tsui et al., 1992). In their study of work group attachment, Tsui et al. (1992) found that as groups became more diverse in terms of age, gender, and racial differences, the psychological attachment to the group decreased. This effect was particularly pronounced for Whites and men. The greater the differences in appearance between a dominant group and a minority group, the lower the status of the minority group and the less chance of acceptance by the dominant group. Multiple differences in appearance, especially when the differences are pronounced, often result in unfavorable comparisons and correspondingly negative attributions between groups (Tilly, 1999). Furthermore, legitimacy may not be equally available to all network members. For example, although women can be accepted into networks dominated by White males, their legitimacy is not likely to be equal to that of the male network members. Ibarra’s (1992) work supports this assumption. She found that men, in general, tended to occupy a central network position, whereas women occupied a more peripheral position. Lin (2000) argues that women and minorities have less career-enhancing social capital than men. In an earlier study, Martineau (1977) found that AfricanAmericans had fewer strong or weak ties than Whites. However, Lin (2000, p. 760) stated that “we should not assume that social capital deficiency translates directly into return deficit.” Return deficits are differentials in returns from similar levels of investment. Although two groups may have differing returns, they can still have the same level of social capital deficits or vice versa. For example, two middle managers, one White and one Black, may form strong ties with an upper level White manager. Although both parties may expend an equal amount of time in developing the strong tie with the upper manager, the returns on that strong tie may not be equal between them. Specifically, the White middle manager may gain a greater return on the strong tie by the likelihood of having a central network location and, therefore, he may gain greater access to career enhancing information compared to the Black middle manager. Active mobilization of ties refers to purposely seeking and activating relationships with individuals or groups for either instrumental or social purposes. The mobilization of ties, however, does not necessarily equate to greater job opportunities or career outcomes, because it is those groups and individuals who have resource-poor embedded ties that are more likely to need to activate ties for job searches and career advancement. Since their networks have limited resources for instrumental support, they have to purposefully reach outside of their

96

ROCHELLE PARKS-YANCY

networks for that support. As minorities often have less resource-rich ties than Whites, they may have to work harder to locate and activate instrumental ties than Whites do (Lin, 2000). Mobilization of ties highlights the value of a high level of social capital. Groups who have low levels of social capital are likely to find it more difficult than groups with a higher level of social capital to advance in social standing and in their careers. Lin et al. (1981) found that educational and occupational achievements, along with weak ties to high level contacts enabled the job seeker to advance to more prestigious jobs. As they advanced to more prestigious jobs, they gained more social capital in networking with individuals of similar or higher rank. If individuals or groups have few weak ties to higher status contacts, i.e. have resource-poor networks, they may find it more difficult to gain a higher social and career status. Given that Whites not only have more social capital, but also get higher returns on their social capital, it is likely that their networks are more resource-rich and, therefore, provide greater advantages, even if their numbers of network connections are the same as those of minorities. Hence, even if a historically disadvantaged group does have ties to high status contacts within an organization, it may be difficult for them to regain access to networks of an equal level if downsizing interrupts the nature of those ties. This may be because of their lack of access to others of a different race or gender, who could provide either social or instrumental support, because they lack knowledge of how to regain access to and acceptance by other higher level people. Also, majority group members are more likely to help others like themselves given individuals’ homophilic tendencies anyway (Ibarra, 1995).

THE EFFECTS OF DOWNSIZING ON THE SOCIAL CAPITAL OF WOMEN AND MINORITY MANAGERS Firms can experience a change in the productivity, morale, and organizational commitment of survivors after a layoff. Factors that affect survivors’ reactions are their perceptions of justice in the layoff, the relationship of the survivors to the laid off employees, and how working conditions are changed by the layoff (Brockner, 1992). Issues that affect perceptions of procedural justice are the justification given for the layoff, the length and nature of the advance notification given about the layoff, and the implementation process of the layoff. If the survivors are close to the layoff victims, that can also affect perceptions of the layoff (Brockner, 1992). However, as Shah (2000) found, perceptions of layoffs are also related to changes in the survivors’ network position. He examined layoff survivors’ reactions to the downsizing of friends and coworkers in similar

The Impact of Social Capital

97

structural positions. The dismissal of friends negatively affected survivors’ emotional responses to the layoff and their location in the firm’s friendship networks. The loss of friends was associated with a decrease in work satisfaction and it also decreased the survivors’ access to informal information networks in the firm. However, the loss of friends increased the survivors’ advice networks with supervisors. Shah (2000) found that employees needed advice to perform their work effectively, and survivors compensated for the loss of friends from whom they could gain advice by forming new ties with other survivors, including with supervisors. Finally, the loss of structurally equivalent, but non-friend, co-workers was positively related to the survivors’ satisfaction with their promotion opportunities within the firm. Since there were fewer co-workers with whom they would have to compete in similar functional and hierarchical positions, survivors recognized that they were in a better position to get promoted. Although their networks were smaller as a result of the downsizing, it resulted in a potentially greater return on social capital than the survivors had before the downsizing. Armstrong-Stassen (1998) explored the effect of gender and organizational level on survivors’ perceptions of a layoff. She found that female technicians perceived a greater sense of procedural injustice from the layoff than male technicians of equal rank. She suggested that the female technicians may have felt more vulnerable to a layoff happening to them, and therefore identified more with the layoff victims. “Survivors who identify with the layoff victims are likely to view the layoff as more unfair” (Brockner, 1990, p. 99). The female technicians may have also felt more vulnerable after the layoff due to a perception, as Burt (1998) suggests, of having less legitimacy in their positions, compared to the male technicians. The changes in working conditions after a layoff also affect survivors’ perceptions of the layoff. Some of the concerns about those changes are the possibility of future layoffs, how survivors’ jobs compare to their job before the layoff, and reactions of fellow survivors (Brockner, 1992). Susskind et al. (1998) studied the effect of downsizing on structural holes within survivors’ networks, with regard to survivors’ receptiveness to organizational change and their perceptions of organizational chaos. While some employees’ access to structural holes were negatively affected by the layoffs, others were positively affected. Survivors, whose structural holes increased after the downsizing were less receptive to organizational change and perceived greater organizational chaos than survivors whose structural holes decreased due to the layoffs. Survivors, whose network structural holes were decreased by the layoffs, tended to gain closer network contacts. This gain facilitated their willingness to support additional organizational changes.

98

ROCHELLE PARKS-YANCY

Burt (1998) found that when women and young males “borrowed” social capital from a senior strategic partner who had several structural holes – and who was a broker across those holes – the advancement opportunities of women and young males and the pace of their advancements increased. He suggested that the difficulty in accessing varied networks was due to legitimacy issues, as opposed to differences in race, gender, or age. “Legitimacy is key to the social situation of the person, not to the person’s attributes” (Burt, 1998, p. 35). In other words, sponsors matter in organizations, and sponsorship can overcome disadvantages of gender or race/ethnicity. However, in traditional corporations, most senior managers tend to be White and male. As Mills (1997) suggests and Ibarra (1992, 1995) found, Whites and males tend to exhibit homophilic preferences in instrumental networks. Therefore, it may be more difficult for African-Americans and women to gain legitimacy in those networks dominated by Whites and men than for other Whites and males. If downsizing removed the strategic partner from whom legitimacy was obtained, African-American and women survivors would most likely lose their legitimate status in the instrumental networks dominated by Whites and men. The fact that their legitimacy was obtained through a relationship to a strategic partner, who is no longer in the network, combined with the physical and cultural differences from the dominant group, could probably negatively affect their network position. Proposition 1a. Compared to White males, the loss due to downsizing of a senior strategic partner from whom social capital was “borrowed” will negatively affect African-American and women survivors’ position in instrumental networks that are dominated by Whites and men. The ability to borrow social capital lends legitimacy to otherwise “illegitimate” individuals and facilitates their mobility across structural holes. The senior strategic partner serves as a broker between those holes. If the senior strategic partner leaves because of downsizing, the subordinate partners would probably have lessened access to the structural holes previously accessed by the senior partner, because the senior partner would no longer be available to serve as broker, and hence, to provide legitimacy. Under circumstances of downsizing, the senior strategic partner’s focus would probably be on maximizing resources for himself from individuals with whom he shares instrumental and social ties. Research has shown that such sharing would most likely be with other Whites and males. Proposition 1b. The downsizing of a senior strategic partner will be associated with a stagnation or decrease in mobility between previously accessible structural holes for African-American and women survivors.

The Impact of Social Capital

99

Furthermore, since layoffs generate fears about job security and about future opportunities within the firm, potential senior strategic partners may be less inclined to “loan” social capital resources to individuals who are not perceived as being able to reciprocate equally. Under circumstances of downsizing, the senior strategic partner’s focus would probably be on maximizing resources from individuals with whom she/he shares instrumental and social ties for themselves, which research has shown would most likely be to fellow Whites and males. Consequently, if a senior strategic partner were downsized, African-American and women survivors would probably have more difficulty than Whites and men in developing a relationship with another senior strategic partner from whom to “borrow” social capital. Proposition 1c. Downsizing will negatively affect African-American and women survivors’ access to a new senior strategic partner from whom to “borrow” social capital resources, should their original senior strategic partner be downsized. The quality and quantity of network structural holes are probably not, necessarily, the same for every potential senior strategic partner. Therefore, it may be difficult for African-American and women survivors’ to access a new senior strategic partner who had similar or better networks as the downsized partner’s. Furthermore, each new network requires an investment of time to cultivate ties. Even if African-American and women survivors gained access to new structural holes by “borrowing” the social capital of a new senior strategic partner, it may take them a while to develop the relationships that pay off compared to Whites and males in the same social networks. Proposition 1d. Even if African-American and women survivors gain a new strategic partner, their access to social capital resources in instrumental networks will be negatively affected by the downsizing of a senior strategic partner because a new network requires building new relationships. Since high-potential minorities and women tend to utilize homophilic ties for social support and for informal information networks (Ibarra, 1992, 1995), this would most likely apply to African-American and women survivors in traditional corporate environments as well. As strong ties are beneficial in uncertain situations (Granovetter, 1995), survivors would presumably gain emotional support from co-workers and supervisors with whom they have close relationships. Such close relationships would probably come from homophilic ties, because the individual and group status of survivors would be highly salient to them in this circumstance of vulnerability. African-Americans and women tend to have lower group status than White males in traditional corporations. Also, surviving the individual and

100

ROCHELLE PARKS-YANCY

group crisis of a layoff could develop or enhance strong tie relationships among survivors left in similar network positions after a layoff. Survivors would most likely share “grapevine” information with fellow survivors with whom they have close relationships and who are members of the same groups. Since they would perceive the layoff as affecting fellow group members similarly, they would be more likely to share with them information that was not disseminated through formal corporate networks. Proposition 2. Homophilic, strong tie relationships of African-American and women survivors will be positively associated with social support and grapevine information after a layoff. The survivor research indicates that concerns about future layoffs and changes in workload relate to career stability issues. Survivors are greatly concerned about their continued employment and position in the firm following a layoff. While career mobility would probably still be of concern, survivors would most likely be more concerned about holding on to their current position, especially after having seen other employees lose their jobs. One of the means by which African-American and women survivors could conceivably maintain their position would be through strong ties with majority members in the firm. Ibarra (1992) found that some women used ties to males for instrumental purposes and proposed that, among other types of ties, high-potential minorities should also use heterophilic ties for instrumental purposes (Ibarra, 1993). Strong tie relationships to the majority members of a firm, which are beneficial in times of uncertainty, could increase survivors’ likelihood of keeping their position during a layoff. Thus, because majority members in traditional corporations tend to be Whites and males: Proposition 3. Strong tie relationships with White male survivors will be positively related to the ability of African-American and women employees to keep their jobs during a layoff.

IMPLICATIONS AND DISCUSSION African-Americans and women have opportunities for advancement in the traditional corporate environment through their access to social capital. However, layoffs can affect the composition of those networks, the positions of African-American and women in the networks, and rewards likely from those networks. The racial, cultural, and gender differences of African-American and women layoff survivors will probably negatively affect their access to and benefits from social capital compared to White and male layoff survivors.

The Impact of Social Capital

101

Yet, strong tie relationships with White and male layoff survivors can also help African-American and women layoff survivors maintain their existing job position after a layoff, because African-Americans and women can borrow the social capital resources of senior strategic partners (who are often Whites and males in traditional corporations). Thus, in downsizing situations the weak tie relationships, which researchers suggest are beneficial in getting ahead, may not be as useful for Black and women survivors for whom strong tie relationships with dominant groups may be more beneficial. In fact, irrespective of the cause of unemployment, Blacks and women are still more likely to be unemployed and to have difficulty finding another job relative to Whites and men. For example, women may leave the labor force due to childcare responsibilities (among other reasons). This may severely reduce the extent of their social ties, especially their access to weak ties (Munch et al., 1997), because more time spent at home in child-rearing is more likely to foster relations among kin-folk than among acquaintances. Blacks are less likely to be employed at all and, once employed, are more likely to lose a job than are Whites, irrespective of human capital attainments (education, training, etc.). Given their lesser access to social ties who can help them get a job, when they lose access to useful social ties or lose their jobs altogether, they can have a very difficult time finding useful social contacts and often have difficulty finding another job. Indeed, some research has found that during periods of downsizing, White men are the first to get hired to other jobs, White women are second, Black men are third, and Black women are fourth (Spalter-Roth & Deitch, 1999). And, Whites who have lost their jobs are more likely to be re-employed in similar occupations, whereas Blacks who lose their jobs are more likely to experience downward occupational mobility. Thus, not only do Blacks lose access to social ties when they lose a job (or when a key social contact is downsized), even when they find a job (through social ties or not), the job is not likely to pay as well nor be of equal prestige as their previous job. Downsizing is just one of many issues that continue to warrant attention from social capital researchers. While the relationship of race, gender, and social capital within corporations has been studied (Burt, 1998; Ibarra, 1992, 1993, 1995), there is limited research on how downsizing affects the social capital of survivors who are racial minorities and women. Research in this area is theoretically and practically warranted for two reasons: First, social capital and downsizing both can have a tremendous impact on individuals’ and groups’ existing and future socioeconomic opportunities. Hence, examining the relationship between them can assist with greater theoretical rigor in deriving propositions and hypotheses about career trajectories. Second, studying the relationship between social capital, layoffs, and race and gender can provide industry leaders and government officials

102

ROCHELLE PARKS-YANCY

with a greater understanding of why and how different groups are affected in different ways by layoffs. Such understanding could facilitate the development of or changes to policies to better manage those differential effects. This paper is not suggesting that White and male survivors will not experience career challenges during a layoff. However, due to the great access to networks and social status for many males and Whites, they will often be in a more favorable position during a layoff than women or minorities. While this paper examines the social capital of African-American and women survivors in corporate environments, the analysis may differ in other environments such as non-profit or government organizations where promotions are more structured and bureaucratic. Future research should ascertain if and how the type of environment relates to the social capital of African-American and women survivors.

REFERENCES Armstrong-Stassen, M. (1998). The effect of gender and organizational level on how survivors appraise and cope with organizational downsizing. Journal of Applied Behavioral Science, 34, 125–141. Brockner, J. (1990). Scope of justice in the workplace: How survivors react to coworker layoffs. Journal of Social Issues, 46, 95–106. Brockner, J. (1992). Managing the effects of layoffs on survivors. California Management Review, Winter, 9–27. Burt, R. S. (1997). The contingent value of social capital. Administrative Science Quarterly, 42, 339– 365. Burt, R. S. (1998). The gender of social capital. Relationality and Society, 10, 5–36. Granovetter, M. (1973). The strength of weak ties. American Journal of Sociology, 78, 1360–1380. Granovetter, M. (1995). Getting a job: A study of contacts and careers (2nd ed.). Chicago, IL: University of Chicago Press. Hanson, S., & Pratt, G. (1991). Job search and the occupational segregation of women. Annals of the Association of American Geographers, 81, 229–253. Ibarra, H. (1992). Homophily and differential returns: Sex differences in network structure and access in an advertising firm. Administrative Science Quarterly, 37, 422–438. Ibarra, H. (1993). Personal networks of women and minorities in management: A conceptual framework. Academy of Management Review, 18, 57–87. Ibarra, H. (1995). Race, opportunity, and diversity of social circles in managerial networks. Academy of Management Journal, 18, 673–703. Lin, N. (2000). Inequality in social capital. Contemporary Sociology, 29, 785–795. Lin, N., Ensel, W. M., & Vaughn, J. C. (1981). Social resources and strength of ties: Structural factors in occupational status attainment. American Sociological Review, 46, 393–405. Martineau, W. H. (1977). Informal social ties among urban Black Americans: Some new data and a review of the problem. Journal of Black Studies, 8, 83–104. Mills, C. W. (1997). The racial contract. Ithaca, NY: Cornell University Press. Munch, A., McPherson, J. M., & Smith-Lovin, L. (1997). Gender, children, and social contact: The effects of childrearing on men and women. American Sociological Review, 55, 509–520.

The Impact of Social Capital

103

Putnam, R. D. (2000). Bowling alone. New York, NY: Simon & Schuster. Rosenbaum, J. E., DeLuca, S., Miller, S. R., & Roy, K. (1999). Pathways into work: Short-and long-term effects of personal and institutional ties. Sociology of Education, 72, 179–196. Shah, P. P. (2000). Network destruction: The structural implications of downsizing. Academy of Management Journal, 43, 101–112. Spalter-Roth, R., & Deitch, C. (1999). “I don’t feel right-sized I feel out-of-work sized”: Gender, race, ethnicity and the unequal costs of displacement. Work and Occupations, 26, 446–482. Susskind, A. M., Miller, V. D., & Johnson, J. D. (1998). Downsizing and structural holes. Communication Research, 25, 30–64. Tilly, C. (1999). Durable inequality. Los Angeles: University of California Press. Tsui, A. S., Egan, T. D., & O’Reilly, C. A., III (1992). Being different: Relational Demography and organizational attachment. Administrative Science Quarterly, 37, 549–574.

PREDICTING AFFIRMATIVE ACTION ATTITUDES: INTERACTIONS OF THE EFFECTS OF INDIVIDUAL DIFFERENCES WITH THE STRENGTH OF THE AFFIRMATIVE ACTION PLAN David A. Kravitz and Stephen L. Klineberg ABSTRACT Affirmative action and diversity management are complementary efforts to achieve an inclusive workforce. Research on attitudes toward affirmative action can therefore contribute to an understanding of reactions to diversity programs. Using data from two studies, we assess the extent to which the strength of the affirmative action plan (AAP) influences the relationship between attitudes and individual difference predictors. The relation of attitudes with the respondents’ race and perceived self-interest increased monotonically with AAP strength, whereas measures of racial prejudice and political orientation best predicted attitudes toward AAPs of intermediate strength. We explore the implications of these findings for the theory and practice of diversity management.

Diversity in the Workforce Research in the Sociology of Work, Volume 14, 107–130 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0277-2833/doi:10.1016/S0277-2833(04)14005-3

107

108

DAVID A. KRAVITZ AND STEPHEN L. KLINEBERG

INTRODUCTION Contemporary interest in workforce diversity evolved in part from the need to satisfy the requirements of equal employment opportunity and affirmative action regulations (Kelly & Dobbin, 1998). Affirmative action refers to measures that are legally required to correct for past or present discrimination. Diversity management entails a voluntary effort on the part of companies to maximize “the ability of all employees to contribute to organizational goals and to achieve their full potential unhindered by group identities such as gender, race, nationality, age, and departmental affiliation” (Cox, 1993, p. 11). Affirmative action and diversity management are complementary approaches to the creation of an inclusive workforce. Although the primary reasons for their existence differ, affirmative action and diversity management share many operational procedures (Kelly & Dobbin, 1998). Konrad and Linnehan (2003) point out that affirmative action contributes to workforce diversity by increasing the number of employees who belong to traditionally underrepresented groups. Conversely, an appropriate diversity management program will advance affirmative action goals by increasing the employment success of individuals from affirmative action target groups. Affirmative action and diversity management are closely interconnected – historically, conceptually, and operationally. The survival and expansion of workplace diversity programs depend on their acceptance by organizational decision makers and by the general public. It is therefore important to uncover the bases of attitudes toward diversity policies. Although some empirical research has been published on attitudes specifically with regard to workplace diversity management policies (e.g. Hopkins et al., 2001; Kossek et al., 2003; Richard et al., 2000), this literature is dwarfed by research on attitudes toward affirmative action. In light of their interconnections, it seems clear that research on the factors that predict attitudes toward affirmative action will also be helpful to scholars and practitioners concerned about reactions to diversity management practices. Unfortunately, much of the social science research erroneously equates affirmative action with strong racial or gender preferences. This approach not only fails to assess attitudes toward affirmative action policies as they are actually implemented, but also neglects the complex interactions between implementation procedures and the attitudinal predictors. Those interactions are the focus of the present chapter. After reviewing research on the predictors of attitudes toward affirmative action, we argue that the effects of several important predictors are likely to vary systematically with the strength of the affirmative action plans (AAP), and we

Predicting Affirmative Action Attitudes

109

make use of two disparate studies to test these claims. The results have implications for the development of a comprehensive theory of attitudes toward affirmative action and for the practice of workplace diversity management more generally.

AAP Implementation in Practice Despite recent political and legal retrenchment, federal affirmative action regulations still apply to approximately 22% of the private workforce (Office of Federal Contract Compliance Programs, 2002). Other affirmative action policies are required at state and municipal levels and by programs developed in all branches of the U.S. military. Affirmative action plans are complexly configured and highly variable in their requirements. Federal AAPs include organizational structures, assignments of responsibility, workforce analyses, and reports to federal regulators. When workforce analyses indicate that a protected group is underrepresented, AAPs must specify goals and associated timetables to improve the representation, along with a description of the procedures that will be taken to achieve the goals. Various actions are possible, depending on the severity and cause of the underrepresentation. In contrast to the complexity of real AAPs, the descriptions of affirmative action that are used in research are generally simplistic and potentially misleading. In many cases, respondents are simply asked to report their reactions to “affirmative action,” with no description or definition being offered. Because people construe “affirmative action” in many different ways, it is not possible to know what policy the respondents in this research believed they were evaluating. In other studies, a simple definition is provided in the phrasing of the question. In experimental research, the affirmative action plan is often described in a few sentences. Krysan (2000) found that survey questions generally ask about one of three types of racial policies: equal treatment, opportunity enhancement, or (most often) equal outcomes. The same alternatives are among those used in experimental studies as well (Bobocel et al., 1998; Slaughter et al., 2002). Equal treatment policies require no more than the elimination of discrimination. Opportunity enhancement involves additional procedures designed to improve the opportunities available to the target group. These might include such actions as targeted recruitment, additional training, or the establishment of formal mentoring plans. Equal outcome policies designate actions that are intended to equalize the outcomes (or proportional outcomes) for all the different groups rather than just to improve the opportunities of targeted groups. Such policies might entail the establishment of quotas, the selection of target group members over equally qualified majority applicants (a tiebreak procedure), or even the selection of

110

DAVID A. KRAVITZ AND STEPHEN L. KLINEBERG

minimally qualified target group applicants over more qualified majority group members (strong preferential treatment). In practice, almost all affirmative action programs involve policies and practices that fall in the first two categories – equal treatment and opportunity enhancement. The three types of racial policies designate different approaches to affirmative action (Gutman, 2000; Office of Federal Contract Compliance Programs, 2002). If no protected group is underrepresented, a non-discrimination policy is sufficient. If underrepresentation is present but there is no clear evidence of discrimination, regulations usually call for additional opportunity enhancement policies. Racial or gender preferences are legal when, but only when, they are ordered by a Court in response to evidence of egregious discrimination. Because the full range of policies can be construed as different versions of affirmative action, a full understanding of affirmative action attitudes requires attention to all three types of policies. Research has consistently found that public support for AAPs is inversely related to the strength of the plan, at least among individuals who are not members of the targeted groups (e.g. Kravitz et al., 1997; Lipset & Schneider, 1978; Matheson et al., 1994; Nosworthy et al., 1995). A key question is whether the effects of specific attitudinal predictors are equally powerful across the range of AAPs. Unfortunately, as noted earlier, many social science studies on racial policy attitudes equate affirmative action with strong racial preferences. For example, data from the National Election Studies ask about the preferential hiring and promotion of Blacks and the use of racial quotas in college admissions (e.g. Bobo & Smith, 1994; Federico & Sidanius, 2002). Williams et al. (1999) used this item: “I would not mind giving special preferences in hiring and job promotions to Blacks.” Sniderman and Carmines (1997, p. 42) analyzed questions specifying “quotas in hiring for large companies” and “special advantages in jobs and schools.” Sears et al. (1997) used attitude items that referred to the “preferential hiring and promotion of Blacks,” “quotas for admitting Black students in universities,” “special treatment for Blacks,” and “set-asides for Black contractors.” Even when scholars employ what might be described as multiple versions of affirmative action, they often restrict the term, “affirmative action,” to questions involving strong racial preferences (e.g. Sears et al., 1997). The equating of affirmative action with preferences severely limits the practical relevance of this research. Insofar as the relations between predictors and attitudes vary with AAP strength, it is dangerous to generalize the results to AAPs that do not involve preferences. This problem is particularly serious because preferences (and quotas) are almost always illegal and presumably are the least common type of affirmative action. Although many people do interpret affirmative action to mean

Predicting Affirmative Action Attitudes

111

quotas and preferences, many others do not (Crosby, 1994; Golden et al., 2001; Kravitz & Platania, 1993). The dominant approach does little to clarify the attitudes of those who do not associate affirmative action with preferential treatment. Furthermore, it perpetuates the misperception of affirmative action as requiring strong racial or gender preferences, a belief that demonstrably stimulates social conflict (Crosby & VanDeVeer, 2000; Locke, 1997), as well as the stigmatization of women and racioethnic minorities (Evans, 2003; Heilman et al., 1998; Maio & Esses, 1998). In sum, it is important to study AAPs that range in strength from the elimination of discrimination (equal treatment), through opportunity enhancement procedures, to preferences (equal outcomes). Otherwise, it is impossible to know whether the bases of support or opposition to different types of affirmative action are fundamentally the same or different. We believe that there are compelling reasons to expect the effects of some of the most important attitudinal predictors to vary systematically with AAP strength.

Major Predictors of Affirmative Action Attitudes Social science research on the determinants of racial policy attitudes has been guided by three general conceptualizations (Krysan, 2000). Some scholars emphasize the importance of racial prejudice, particularly in its contemporary forms. A second approach is grounded in assumptions about individual and group conflict and self-interest. A third focuses on nonracial perspectives, such as principled conservatism. Advocates of these different conceptualizations have engaged in a vigorous debate (e.g. Bobo, 2000; Federico & Sidanius, 2002; Kluegel & Smith, 1983; Kuklinski et al., 1997; Sniderman & Carmines, 1997; Williams et al., 1999). Racial Prejudice The decades since the Civil Rights movement of the 1950s and 1960s have seen a waning of old fashioned (or “Jim Crow”) racism, in which Blacks were widely believed to be inherently inferior, segregation was advocated, and blatant discrimination was accepted. Today, such views are far less common, but this is not to say that racial prejudice has disappeared. Many scholars claim that the old-fashioned forms of racism have been replaced by more subtle contemporary forms. Two of the most useful conceptualizations are “modern racism” (McConahay, 1986) and “aversive racism” (Gaertner & Dovidio, 1986). Despite some important differences, the two approaches agree that White Americans today generally avoid behaving in overtly discriminatory ways, either because they do not want to be seen as violating the norms of equality or because

112

DAVID A. KRAVITZ AND STEPHEN L. KLINEBERG

they genuinely oppose prejudice and discrimination. However, many will discriminate against Blacks when doing so can be attributed to some non-racial cause – when the situation enables them to deny any racist motivation for their racially tinged actions and beliefs. For example, Dovidio and Gaertner (2000) studied White students’ recommendations of White and Black candidates for a peer counselor position. The candidate’s race did not affect recommendations when qualifications were either strong or weak. But when qualifications were ambiguous, and rejection of the Black candidate could be attributed to that person’s mediocre qualifications, White candidates were significantly more likely to be recommended. According to these conceptualizations of contemporary racism, the bases of prejudice lie in cognitive biases and/or in the racist socialization still prevalent in American society. Racial prejudice, in this view, persists as a function of individual psychology – rather than as a consequence of underlying group conflict or a concomitant of other non-racist ideological perspectives. Racially prejudiced individuals are likely to oppose affirmative action because the AAPs are intended to assist the targets of their prejudice (Kravitz et al., 2000). This expectation has been confirmed in the ample evidence of a negative correlation between various measures of racial prejudice and support for affirmative action (e.g. Harrison et al., 2001; Hughes, 1997; Sears et al., 1997; Sidanius et al., 1996). Group Conflict and Self-Interest A second research tradition focuses on the role of self-interest in motivating opposition to affirmative action. Sears and Funk (1991) define self-interest in terms of its short- to medium-term impact on the individual’s own material well-being. Collective self-interest refers to the impact of the policy for the respondent’s demographic group as a whole (e.g. Tougas et al., 1991). Both personal and collective self-interest predict attitudes toward affirmative action (Aberson, 2003; Nosworthy et al., 1995). A meta-analysis found that attitudes were somewhat more closely associated with collective self-interest than with personal self-interest, perhaps because affirmative action is a collective solution to the problem of discrimination (Harrison et al., 2001). Other scholars emphasize the interconnectedness of self-interest with racism. This perspective has been developed primarily by Lawrence Bobo and colleagues (e.g. Bobo et al., 1997), whose work is grounded in a sociological theory of race prejudice as a sense of group position (Blumer, 1958) and in an historical analysis of economic and political forces. Blumer’s (1958) group position theory assumes that members of the dominant racial group believe they are entitled to their dominant position and that their position is threatened by the claims of subordinate groups. They develop racist ideologies in an effort to justify and maintain a privileged status. The specific details of such ideologies will be

Predicting Affirmative Action Attitudes

113

informed by the surrounding economic and political circumstances. Conditions in the post-Civil War South stimulated the development of Jim Crow racism. As circumstances changed in the Civil Rights era, it became more difficult to espouse the Jim Crow ideology, and a new set of beliefs emerged to justify the racial hierarchies. Bobo et al. (1997) suggest the term, “laissez faire racism,” for the current ideology. Its key components include the continued negative stereotyping of African Americans, the assumption that Blacks’ disadvantaged status is due to their own failings rather than to societal or historical forces, and opposition to government policies designed to eliminate barriers to equal opportunity. Principled Conservatism The third explanation for attitudes toward racial policies is based on the assumption that many people oppose affirmative action primarily because it is contrary to conservative principles. For the past 40 years, liberals and the Democratic Party have generally supported policies designed to enhance racial equality, whereas conservatives and the Republican Party have typically opposed such policies (Abramowitz, 1994; Gilens et al., 1998). Scholars in the principled conservatism tradition acknowledge that racial prejudice plays a role in the opposition to affirmative action, but they deny that it plays the central role. Thus, Sniderman and Carmines (1997, p. 72) report very low correlations between racial prejudice and attitudes toward affirmative action, but they define the latter “as requiring preferential treatment or racial quotas” (p. 10). Perceived Fairness and AAP Strength An important dimension of attitudes toward affirmative action is the perceived fairness of its procedures. Not surprisingly, numerous studies have found that people are more likely to support AAPs that they consider to be fair than those they believe to be unfair (Kravitz, 1995; Kravitz & Klineberg, 2000; Ozawa et al., 1996; Peterson, 1994). Furthermore, the effect of AAP strength on attitudes is mediated by the evaluations of fairness (Bobocel et al., 1998; Kravitz, 1995). That is, perceived fairness decreases with AAP strength, and this is one cause for the corresponding decrease in support.

The Interactions of AAP Strength with Individual Difference Predictors As we have seen, affirmative action attitudes are affected both by the strength of the AAP and by various individual difference variables. There are reasons to believe that the relative impact of these attitudinal predictors varies systematically with

114

DAVID A. KRAVITZ AND STEPHEN L. KLINEBERG

AAP strength. Although the importance of this possibility has been acknowledged (e.g. Bobo, 1998), it has not received much empirical attention. Predicted Monotonic1 Interactions (Self-Interest, Respondent Race) Perceptions of fairness are negatively related to AAP strength – the elimination of discrimination is considered to be the most fair, and preferential treatment the most unfair. Several studies have found that the effect on attitudes of an outcome experienced by an individual is most pronounced when the procedures determining the outcome are perceived to be unfair (Brockner & Wiesenfeld, 1996; Ehlen et al., 1999; McFarlin & Sweeney, 1992). In the affirmative action literature, ratings of self-interest correspond to the outcomes studied in the justice literature. This leads directly to our first two hypotheses. Hypothesis 1. The magnitude of the relation between personal self-interest and AAP attitudes will increase monotonically with AAP strength. Hypothesis 2. The magnitude of the relation between collective self-interest and AAP attitudes will increase monotonically with AAP strength. Respondent race, not surprisingly, is powerfully and consistently related to attitudes toward affirmative action. With few exceptions, support is strongest among African Americans and weakest among Whites, with Hispanics and Asians typically expressing intermediate levels of support (Bobo, 1998; Hughes & Tuch, 2000; Kinder & Sanders, 1996; Klineberg & Kravitz, 2003; Kravitz & Klineberg, 2000). Demographic differences, of course, represent complex bundles of variables. Blacks and Whites in America differ on average in their socioeconomic status, political orientation, experiences with discrimination, and in many other ways. They also differ fundamentally in the self-interest implications of affirmative action. At least in the short run, AAPs that target African Americans have positive implications for African Americans and negative implications for Whites. Insofar as the effect of race on affirmative action attitudes is attributable to intergroup differences in self-interest rather than to differences on other dimensions, the logic underlying Hypotheses 1 and 2 leads to an additional prediction. Hypothesis 3. The magnitude of the relation between race (being Black or White) and attitudes toward AAPs that target African Americans will increase monotonically with AAP strength.

Predicting Affirmative Action Attitudes

115

Predicted Nonmonotonic Interactions (Racism, Political Ideology) As we have seen, a key assumption in many conceptions of contemporary racism is that prejudiced individuals generally avoid revealing their racial prejudice to others and even to themselves (Gaertner & Dovidio, 1986). They will engage in prejudicial actions only if their actions can reasonably be attributed to non-racial causes. Because it is difficult to justify opposition to weak AAPs, such as equal treatment plans, even prejudiced individuals will express support for such programs. By the same token, even the non-prejudiced are likely to oppose strong preferential treatment (equal outcome plans), because such policies violate widely held norms of procedural justice (Bobocel et al., 1998). With regard to an AAP of intermediate strength (e.g. opportunity enhancement plans), however, prejudiced individuals can readily explain their opposition in justice terms – arguing that they oppose any type of special treatment for any group. Non-prejudiced individuals are more likely to support such moderate AAPs because they enhance the target group’s opportunities without harming non-target group members. It follows that the effects of racial prejudice should be most evident in attitudes toward intermediate AAPs. Hypothesis 4. Among non-Black respondents, the relation between racial prejudice and attitudes toward AAPs that target African Americans will be strongest for AAPs of intermediate strength. A key component of the distinction between the liberal and conservative political ideologies in the United States today lies in their conceptualization of the role of government, with liberals favoring a more active role (Ashbee & Ashford, 1999; Sniderman et al., 2000). Specifically, Kernell and Jacobson (2000, p. 324) state that “liberals typically favor using government to reduce economic inequalities [and] champion the rights of disadvantaged groups such as racial minorities and women.” Regarding affirmative action, conservatives (and Republicans) argue that the government should forbid racial discrimination but eschew any stronger actions to increase racial equality (Canady, 1996). Liberals (and Democrats) argue that the government should both forbid discrimination and take a more active role to enhance the opportunities available to racial minorities, though they too reject strong preferences (Clinton, 1996). Because individuals at both ends of the political spectrum support equality and oppose strong preferences, the effect of political orientation on affirmative action attitudes should be most substantial for AAPs of moderate strength. Hypothesis 5. The relation between political orientation and attitudes toward AAPs will be strongest for AAPs of intermediate strength.

116

DAVID A. KRAVITZ AND STEPHEN L. KLINEBERG

We make use of two different studies to develop preliminary tests of these hypotheses. Importantly, the studies differ in methodology (a questionnaire study and a telephone survey), in the ages of the respondents, and in the measures used to assess the variables. These differences increase confidence in the reliability of any results that are replicated across studies. Study 1 (The FIU Study) includes only non-Black respondents. As such, we can test only Hypotheses 1, 2, 4, and 5 in this study. The FIU study has the virtue of multiple item measures for most of the key concepts, although the sample size is small. Study 2 (The Houston Area Survey) reaches a larger sample that includes both Blacks and Whites. It generally uses single item measures and does not include a measure of personal self-interest. Study 2 tests Hypotheses 2, 3, 4, and 5.

THE FIU STUDY Method Data are drawn from Kravitz (1995), who manipulated AAP strength and measured personal self-interest, collective self-interest, racial prejudice, and political orientation. Participants were undergraduates at Florida International University. The sample analyzed here included 110 females and 44 males. There were 72 Hispanics, 67 non-Hispanic Whites, and 15 other non-Black respondents. Ages ranged from 17 to 44, with a mean of 23.5 (SD = 5.5). After preliminary instructions, the participants completed a variety of measures of belief systems, including a seven-item scale assessing racial prejudice (e.g. “I would be personally upset if Blacks moved into my neighborhood”). The materials went on to describe one of eight different AAPs that targeted Blacks. Multi-item scales measured the anticipated impact of the designated AAP on the respondent’s personal self-interest (e.g. “What effect would this AAP have on your chances of getting the job you want in an organization that used it?”) and collective self-interest (e.g. “What effect would this affirmative action plan have on the chances of people in your demographic group being hired?” – with demographic group being defined as “people such as yourself, of the same sex and ethnic background”). A six-item scale assessed attitudes toward the AAP (e.g. “This is a good affirmative action plan”). The respondents also identified their political orientation on a five-point scale (very conservative to very liberal). The intercorrelations among the various measures are given in Table 1, along with estimates of internal reliability (Cronbach’s alpha) for all multi-item scales. Seven of the designated AAPs can be categorized unambiguously as weak, moderate, or strong. The two weak AAPs (combined n = 43) included a condition

Predicting Affirmative Action Attitudes

117

Table 1. The FIU Study: Correlations Among Variables, Pooled Across All AAPs. Variable (Meaning of the High Value)

PSI

CSI

Racism

Politics

Attitude

Personal self-interest (positive effect) Collective self-interest (positive effect) Racism (high) Political orientation (liberal) Attitude toward AAP (positive)

0.87 0.66** −0.21** 0.08 0.68**

0.93 −0.21** −0.01 0.63**

0.78 −0.38** −0.30**

N/A 0.10

0.93

Notes: The values on the diagonal are estimates of internal reliability (Cronbach’s alpha) for the multi-item scales. ∗∗ p < 0.01 (2-tailed).

in which discrimination was forbidden and one in which it was irrelevant because demographic information was unavailable. These AAPs received similar evaluations (Ms = 4.3 and 4.1) and were preferred over all others. The three moderate AAPs (combined n = 66) included targeted recruitment, special training, and the requirement that “the proportion of Blacks hired must equal the proportion of qualified Blacks who applied for the position” (emphasis in the original). Evaluations of these three AAPs were also similar (Ms = 3.6–3.7); they were significantly less positive than for the weak AAPs but significantly more positive than for the two AAPs classified as strong. The latter (combined n = 45) included a tiebreak condition and a strong preferential treatment condition (Ms = 2.4 and 2.5). Analysis and Results We hypothesized that the correlations of the predictors with attitudes would vary systematically with AAP strength. To test these expectations, we computed the correlations between affirmative action attitudes and the four predictors across the various AAPs. The results are presented in Table 2; subscripts in each row summarize the results of significance tests (t-tests) on the differences between the correlations. Personal Self-Interest Hypothesis 1 states that the correlations of attitudes with personal self-interest will increase in size as the AAP increases in strength. The pattern of correlations presented in the first row of Table 2 is fully consistent with the hypothesis. The relationship between attitudes and personal self-interest becomes stronger as the AAP increases from weak, to moderate, to strong. The t-tests revealed that the correlation in the strong AAP condition was significantly higher than

118

DAVID A. KRAVITZ AND STEPHEN L. KLINEBERG

Table 2. Correlations Between AAP Attitudes and Predictors (FIU Study). Affirmative Action Plan Predictor*

Weak

Moderate

Strong

Personal self-interest Collective self-interest Racism Political ideology

0.28b 0.46b 0.07b 0.17a

0.56ab 0.39b –0.55a 0.05a

0.64a 0.74a −0.19b −0.07a

Notes: Correlations printed in boldface differ significantly from zero at the 0.05 level. Correlations with different subscripts differ significantly at the 0.05 level according to t-tests. ∗ High values on the predictors represent positive self-interest, strong racial prejudice, and liberal politics.

in the weak AAP condition. Given the small sample sizes and relatively low statistical power of this test, it should also be noted that the difference between the correlations in the weak and moderate AAP conditions approached statistical significance (p = 0.086). The results provide strong support for Hypothesis 1. Collective Self-Interest Hypothesis 2 states that the correlations of attitudes with collective self-interest will increase in size as the AAP increases in strength. The correlations presented in the table are only partly consistent with this hypothesis. The correlation is unexpectedly (but not significantly) smaller in the moderate AAP condition than in the weak condition. Still, as predicted by the hypothesis, the correlation of affirmative action attitudes with collective self-interest is significantly higher in the strong AAP condition than in the two weaker conditions. Racial Prejudice Hypothesis 4 states that the relation of attitudes with racial prejudice will be strongest for the AAPs of intermediate strength. The pattern of correlations presented in Table 2 is fully consistent with this expectation. Higher levels of racism are strongly associated with greater opposition to affirmative action, but only in the moderate AAP condition. The t-tests confirmed that the correlation was significantly higher for the AAPs of intermediate strength than for either the weak or strong AAPs. Political Orientation Hypothesis 5 states that the relation of attitudes with political ideology will be strongest for the AAPs of intermediate strength. This hypothesis was not confirmed. The correlations in Table 2 indicate that attitudes are unrelated to

Predicting Affirmative Action Attitudes

119

political ideology in all three AAP conditions. The significance tests confirm that there were no differences in the effect of political ideology by AAP strength. In sum, the pattern of correlations and the significance tests provide strong support for Hypotheses 1 (personal self-interest) and 4 (racial prejudice), moderate support for Hypothesis 2 (collective self-interest), and no support for Hypothesis 5 (political orientation). AAP strength clearly moderates the relationships between attitudes and three of the four predictors. The FIU study could test only four of the hypotheses, and sample sizes were small. The second study complements this research by providing data relevant to Hypothesis 3 (effects of race) and additional tests of Hypotheses 2, 4, and 5. It measures attitudes toward four AAPs, representing the three clusters of AAP strength and both levels of the strongest cluster (tiebreak and strong preferential treatment). By using telephone interviews of a large representative sample of adult residents in the Houston area, this study also makes possible some empirical triangulation. Replication of research through the use of different methodologies, respondent populations, and operationalizations of variables provides important information about the generalizability of results (cf. Mitchell & Jolley, 1996; Rosenthal, 1990).

THE HOUSTON AREA SURVEY Method Data were drawn from the 2001 Houston Area Survey, a countywide, computerassisted telephone survey conducted in February and March of that year. Respondents from the Houston metropolitan area (Harris County) were selected through a two-stage random digit dialing procedure. In each household reached by computer-generated telephone numbers, the respondent was selected randomly from all residents aged eighteen or older. After completion of the initial representative sample, additional interviews were conducted using identical selection procedures, to enlarge and equalize the samples of Whites, Blacks, and Hispanics at approximately 450 each. For this study, we limit our analyses to the 428 U.S.-born White and 407 U.S.-born Black respondents who provided data on the dependent variable. The vast majority of respondents had work experience and the distribution of work status did not vary with respondent race, Chi-square (n = 818, df = 7) = 6.83, ns. AAPs and the Dependent Variable Respondents were asked to evaluate one (randomly rotated) AAP that a company might use if minorities or women were underrepresented. The four AAPs were: the

120

DAVID A. KRAVITZ AND STEPHEN L. KLINEBERG

elimination of discrimination (NoDisc); special recruitment efforts (Recruitment); preferences for equally qualified minority and women candidates (Tiebreak); preferences for less qualified minority and women candidates (Strong Preferential Treatment or SPT). Attitudes toward the AAP were reported on a 7-point scale (1 = strongly disapprove, 7 = strongly approve). Predictors After evaluating the designated AAP, respondents rated its likely impact on “the opportunities for success of the employees in the company who are like you in terms of their ethnicity and sex.” The response scale for this measure of collective self-interest ranged from 1 (would hurt a lot) to 5 (would help a lot). The survey included one question that might serve as an indicator of contemporary racism. Respondents were asked whether they agreed or disagreed with the statement, “Black people in the U.S. are still a long way from having the same chance in life that White people have.” The question was scored 0 = agree (low contemporary racial prejudice), 1 = disagree (high contemporary racial prejudice). This item is conceptually similar to one used by McConahay (1986) to measure contemporary racism (“Discrimination against Blacks is no longer a problem in the United States.”) and to two items used by Sears et al. (1997) in their measure of symbolic racism (“If Blacks would only try harder they could be just as well off as Whites.” – “Over the past few years, Blacks have gotten less than they deserve.”). Political party identification was based on two items. Respondents were asked if they would call themselves “a Republican, a Democrat, an Independent, or something else.” Those who did not indicate an affiliation were asked if they thought of themselves as “closer to the Republican Party or to the Democratic Party.” Respondents were categorized as identifying with the party they selected on either question (0 = Republican, 1 = Democrat). Those who expressed no preference for either party were dropped from the analyses involving this variable. Self-identified political ideology was measured on a scale that ranged from 1 (very conservative) to 7 (very liberal). In addition, respondents reported their ethnicity and place of birth (0 = U.S.-born non-Hispanic White, 1 = U.S.-born African American).

Analyses and Results Collective Self-Interest As indicated in the top row of Table 3, support for each of the AAPs was significantly associated with the belief that it would improve the opportunities for

Predicting Affirmative Action Attitudes

121

Table 3. Correlations Between AAP Attitudes and Predictors (Houston Area Survey, 2001). Affirmative Action Plan Predictor* Collective self-interest Race Contemporary racial prejudice (Whites only) Political party Political ideology

ns

NoDisc

Recruitment

Tiebreak

SPT

184–214 195–221 88–126

0.31a −0.06b −0.03a

0.37a 0.16a −0.20a

0.36a 0.24a −0.03a

0.31a 0.32a −0.00a

173–193 180–207

0.05a 0.06a

0.23a 0.20a

0.21a 0.18a

0.12a 0.02a

Notes: Correlations printed in boldface differ significantly from zero at the 0.05 level. Correlations with different subscripts differ significantly at the 0.05 level according to t-tests. ∗ High values on the predictors represent positive collective self-interest, being African American, the denial of antiblack discrimination, being Democrat, and liberal politics.

the respondent’s demographic group. Contrary to Hypothesis 2, however, the size of the correlation coefficient was not related to AAP strength. Respondent Race Hypothesis 3 states that the effect of race will increase monotonically with the strength of the AAP. The pattern of correlations presented in the second row of the Table is fully consistent with this hypothesis, although the t-tests indicate that the only significant difference in the strength of the correlations is between the NoDisc AAP and the other three AAP conditions. The difference between the correlations in the Recruitment and SPT conditions approached significance, at p = 0.084. Contemporary Racial Prejudice Hypothesis 4 states that the association among Whites between racial prejudice and affirmative action attitudes will be most pronounced for AAPs of intermediate strength. The correlation coefficients shown in Table 3 are consistent with the hypothesis. The relationship between the denial of antiblack discrimination and opposition to affirmative action among Whites is larger and statistically significant only with regard to the intermediate (Recruitment) AAP. However, the t-tests did not reveal significant differences across the four conditions. Political Orientation Hypothesis 5 states that the correlation between political orientation and attitudes toward affirmative action will be strongest for the AAPs of intermediate strength.

122

DAVID A. KRAVITZ AND STEPHEN L. KLINEBERG

As indicated in Table 3, the predicted nonmonotonic effects were found in the correlations involving both measures of political orientation – party identification and political ideology. Democrats and political liberals were significantly more in favor of the AAPs involving focused recruitment or the tiebreak procedure, but they were no different from Republicans or conservatives in their attitudes toward either the weakest AAP (NoDisc) or the strongest AAP (SPT). The t-tests, however, did not reveal significant differences among the correlations, although the difference between the NoDisc and Recruitment conditions approached significance (p = 0.084) in the analysis of political party, and the difference between the Recruitment and SPT conditions approached significance (p = 0.075) in the analysis of political ideology.

DISCUSSION We have reviewed two different research efforts that explored the interactions between AAP strength and individual variables in predicting affirmative action attitudes. The studies used different research procedures, stimuli, and subject populations, and thus have complementary strengths and weaknesses. The FIU Study assessed the variables using reliable multi-item measures on a written questionnaire, but it was limited to a relatively small sample of college students and pooled results across racioethnicities (excluding African Americans). The Houston Area Survey involved telephone interviews of large samples of adult Whites and African Americans, but it was limited to single-item measures of the variables. Each study drew from a single location, but those locations (Miami and Houston) differed. It is not possible to know with certainty that the results of the two studies would generalize to the United States as a whole, but we have no reason to believe that the bases for AAP attitudes among Miami and Houston residents are atypical.

Tests of Hypotheses Table 4 summarizes the strength of the support obtained for the hypotheses in the two studies. Hypothesis 1, which predicted that the relation between affirmative action attitudes and personal self-interest would increase monotonically with AAP strength, could be tested only in the FIU Study. The pattern of correlations was fully consistent with the hypothesis. Both studies provide evidence regarding Hypothesis 2, which stated that the relation between attitudes and collective self-interest would increase monotonically with AAP strength. The Houston Area Survey yielded

Predicting Affirmative Action Attitudes

123

Table 4. Summary of Support for the Hypotheses. #

Hypotheses

Predicted monotonic interactions 1. The magnitude of the relation between personal self-interest and AAP attitudes will increase monotonically with AAP strength. 2. The magnitude of the relation between collective self-interest and AAP attitudes will increase monotonically with AAP strength. 3. The magnitude of the relation between respondent race and attitudes toward AAPs that target African Americans will increase monotonically with AAP strength. Predicted nonmonotonic interactions 4. Among non-Black respondents, the relation between racial prejudice and attitudes toward AAPs that target African Americans will be strongest for AAPs of intermediate strength. 5. The relation between political orientation and attitudes toward AAPs will be strongest for AAPs of intermediate strength.

FIU

HAS

Strong

N/A

Moderate

None

N/A

Strong

Strong

Moderate

None

Moderate

comparable effect sizes in all four AAP conditions, but the FIU Study found that the correlation of attitudes with collective self-interest was substantially larger in the strong AAP condition than in either of the weaker conditions. Hypothesis 3, which predicted that the effects of respondent race would increase with the strength of the AAP, was clearly supported by the pattern of correlations observed in the Houston Area Survey. The hypothesis could not be tested in the FIU Study. Hypothesis 4, derived from contemporary racism theory, stated that the relation between racial prejudice and affirmative action attitudes among non-Black respondents would be strongest for the AAPs of intermediate strength. The hypothesis was supported in the FIU Study, which found a strong correlation of attitudes with racial prejudice in the moderate AAP condition and no effect in either the weak or the strong conditions. The pattern of correlations in the Houston Area Survey between AAP attitudes and the denial of antiblack discrimination was also consistent with the hypothesis. It is worth noting that this confirmation was obtained despite the less-than-ideal measures of contemporary racism used in the two studies. Hypothesis 5 predicted a similar nonmonotonic effect for political orientation. The FIU Study found no significant correlations between political ideology and affirmative action attitudes in any of the AAP conditions. These nonsignificant effects are contrary to the typical finding that affirmative action attitudes are strongly associated with political orientation (Harrison et al., 2001). The sample used in the FIU Study consisted of undergraduates, whose political views may still

124

DAVID A. KRAVITZ AND STEPHEN L. KLINEBERG

be in formation and may not yet reflect the degree of variation that is typical among older respondents. In contrast, the Houston Area Survey used large samples of Harris County adults and tested the effects of both party identification and political ideology. The tests provided strong correlations with AAP attitudes, and moderate support for the expected nonmonotonic effect: The data revealed a consistent pattern of higher correlations of attitudes with political orientation for the intermediate AAPs, although the differences among the correlations were not statistically significant. In summary, two disparate studies provide strong evidence for monotonic interactions of AAP strength with personal self-interest and respondent race, and for a nonmonotonic interaction of AAP strength with racial prejudice, in predicting attitudes toward affirmative action plans. There is also moderate evidence for the hypothesized monotonic interaction of AAP strength with collective self-interest and for the nonmonotonic interaction of AAP strength with political orientation.

Theoretical Implications As knowledge about a phenomenon increases, the focus of scholarship often shifts, as it should, from simple monotonic relationships to nonmonotonic effects and interactions (Hall & Rosenthal, 1991; Reichers & Schneider, 1990). Research on affirmative action attitudes has reached that point. The results of the two studies reviewed in this chapter make it clear that any comprehensive theory of affirmative-action attitudes needs to incorporate an explicit recognition of the probable interactions between AAP strength and individual difference variables in predicting attitudes. Indeed, this general point may be more important than the specific levels of support obtained for the five hypotheses being investigated here. Other interactions are also likely. For example, Tougas et al. (1995) asked male subjects to judge the fairness of four different AAPs that targeted women. Their judgments varied with the interaction of the AAP by scores on a measure of neosexism. In a study of African Americans’ attitudes toward affirmative action, Fried et al. (1996) found an interaction of ideology (egalitarianism-individualism) by personal experiences of discrimination. As the frequency of experienced discrimination increased, the relation between ideology and attitudes became weaker. Other work has found an interaction of socioeconomic status (e.g. income, education) by race, such that the relation between social class and support for affirmative action is positive among Blacks but negative among Whites (e.g. Ludwig, 2003; Sax & Arredondo, 1999). In addition, the research reviewed in this chapter indicates that some interactions (e.g. racial prejudice, political ideology) are likely to be nonmonotonic rather than monotonic. Although it is unusual

Predicting Affirmative Action Attitudes

125

for theories to incorporate nonmonotonic terms, the results reported here clearly indicate the importance of doing so. Attitudes toward affirmative action are complexly determined, and theories designed to explain those attitudes need to be equally complex. These results are particularly important for the debate about the role of racial prejudice in determining attitudes toward racial policies. As noted above, the principled conservatism argument is that racial prejudice is a relatively unimportant predictor of opposition to affirmative action. Research in this tradition supports this supposition by reporting weak correlations between racial prejudice and racial policy attitudes (e.g. Sniderman & Carmines, 1997; Sniderman et al., 2000). However, such studies almost always describe affirmative action in terms of strong racial preferences. As we have seen, that is precisely the situation in which racial prejudice has the weakest effects. When AAPs are operationalized in terms of more moderate initiatives, consistent with affirmative action laws and regulations, the evidence indicates that racial prejudice is much more closely related to negative attitudes toward affirmative action.

Implications for Research The most important research implication of these findings is the need for studies that incorporate at least three levels of AAP strength. Ideally, research will include examples of equal treatment, opportunity enhancement, and equal outcome policies. When only two different AAPs are presented, it is obviously impossible to detect the nonmonotonic interactions. The results also indicate that the relative impact of any particular attitudinal predictor will depend on how the AAP is operationalized. The use of a moderate AAP is likely to maximize the potential impact on affirmative action attitudes of racism and perhaps of other individual difference variables such as sexism and political ideology. The impact of respondent race and self-interest, however, may be most evident for the strongest AAPs. Thus, the findings we have reviewed in this chapter underscore the need for caution when generalizing the results of research on individual difference predictors beyond the specific AAP conditions under which they were obtained.

Practical Implications An important implication of these results for practitioners concerns the use of moderate AAPs, such as opportunity enhancement. These are the kinds of AAPs that organizations are most likely to establish, because weak AAPs

126

DAVID A. KRAVITZ AND STEPHEN L. KLINEBERG

(e.g. the simple elimination of discrimination) have relatively little effect on the employment opportunities of female and minority workers (French, 1999; Konrad & Linnehan, 1995); and strong AAPs (e.g. preferential treatment) are usually illegal (Office of Federal Contract Compliance Programs, 2002). Unfortunately, the data indicate that the use of moderate AAPs, characteristic of most real-world diversity programs, will by no means eliminate conflict over their appropriateness and implementation. The research we have reviewed indicates that opposition to AAPs that seek, for example, to enhance the recruitment of minority applicants is associated with differences in racial prejudice, in the anticipated implications for personal and collective self-interest, in political ideology, and in respondent race. Diversity management programs include a variety of structures and activities, as they seek the removal of discriminatory barriers and the enhancement of opportunities for all organizational members. They also generally correspond in strength to the weak and moderate AAPs studied here. Thus, organizational leaders should anticipate some opposition to their diversity programs; such opposition will increase as programs give more explicit attention to demographic status. In the effort to build consensus, they would be well advised to develop a nuanced and moderate approach and to give clear and sustained attention to the justifications for their diversity initiatives. Despite management’s best efforts, such programs may continue to be controversial, even as they contribute importantly to the achievement of an equitable and effective multiethnic workforce.

NOTE 1. We use the term, “monotonic,” rather than “linear” because the latter term implies that the variable of interest (AAP strength) is measured on an interval scale. We have no reason to believe that the AAPs used in this research differ by consistent quantities in their relative strength (a requirement for interval scales), so it would be inappropriate to speak of linear effects. This is also the reason for our reference in Hypotheses 4 and 5 to “nonmonotonic” rather than “curvilinear” interactions.

ACKNOWLEDGMENTS The 2001 Houston Area Survey was supported by a consortium of contributors, including the SBC Corporation, the Greater Houston Community Foundation, The Houston Chronicle, the Chase Texas Foundation, KHOU-TV Channel 11, Fiesta Mart, Texas Colonial Homes, H. E. B. Grocery Company, Laredo National Bank, and Telemundo 48 KTMD. David Kravitz received summer support for preparation of this paper from the George Mason University Provost’s office and

Predicting Affirmative Action Attitudes

127

School of Management. Preliminary reports of these results were presented at the 2001 conference of the Society for Industrial and Organizational Psychology (the FIU Study) and the 2001 meeting of the Academy of Management (the Houston Area Survey). We would like to thank Steve Vallas, the editors, and two anonymous reviewers for their helpful comments on earlier drafts of this chapter. All remaining deficiencies are our responsibility.

REFERENCES Aberson, C. L. (2003). Support for race-based affirmative action: Self-interest and procedural justice. Journal of Applied Social Psychology, 33, 1212–1225. Abramowitz, A. I. (1994). Issue evolution reconsidered: Racial attitudes and partisanship in the U.S. electorate. American Journal of Political Science, 38, 1–24. Ashbee, E., & Ashford, N. (1999). US politics today. Manchester, England: Manchester University Press. Blumer, H. (1958). Race prejudice as a sense of group position. Pacific Sociological Review, 1, 3–7. Bobo, L. (1998). Race, interests, and beliefs about affirmative action: Unanswered questions and new directions. American Behavioral Scientist, 41, 985–1003. Bobo, L. (2000). Race and beliefs about affirmative action: Assessing the effects of interest, group threats, ideology, and racism. In: D. O. Sears, J. Sidanius & L. Bobo (Eds), Racialized Politics: The Debate About Racism in America (pp. 137–164). Chicago: University of Chicago Press. Bobo, L., Kluegel, J. R., & Smith, R. A. (1997). Laissez-faire racism: The crystallization of a kinder, gentler, antiblack ideology. In: S. A. Tuch & J. K. Martin (Eds), Racial Attitudes in the 1990s: Continuity and Change (pp. 15–42). Westport, CT: Praeger. Bobo, L., & Smith, R. A. (1994). Antipoverty policies, affirmative action, and racial attitudes. In: S. H. Danziger, G. D. Sandefur & D. H. Weinberg (Eds), Confronting Poverty: Prescriptions for Change (pp. 365–395). Cambridge, MA: Harvard University Press. Bobocel, D. R., Son Hing, L. S., Davey, L. M., Stanley, D. J., & Zanna, M. P. (1998). Justice-based opposition to social policies: Is it genuine? Journal of Personality and Social Psychology, 75, 653–669. Brockner, J., & Wiesenfeld, B. A. (1996). An integrative framework for explaining reactions to decisions: Interactive effects of outcomes and procedures. Psychological Bulletin, 120, 189–208. Canady, C. T. (1996). The meaning of American equality. In: G. E. Curry (Ed.), The Affirmative Action Debate (pp. 277–287). Reading, MA: Addison-Wesley. Clinton, B. (1996). Mend it, don’t end it. In: G. E. Curry (Ed.), The Affirmative Action Debate (pp. 258–276). Reading, MA: Addison-Wesley. Cox, T., Jr. (1993). Cultural diversity in organizations: Theory, research, and practice. San Francisco: Berrett-Koehler. Crosby, F. J. (1994). Understanding affirmative action. Basic and Applied Social Psychology, 15, 13–41. Crosby, F. J., & VanDeVeer, C. (2000). Sex, race, and merit: Debating affirmative action in education and employment. Ann Arbor, MI: University of Michigan Press. Dovidio, J. F., & Gaertner, S. L. (2000). Aversive racism and selection decisions: 1989 and 1999. Psychological Science, 11, 315–319.

128

DAVID A. KRAVITZ AND STEPHEN L. KLINEBERG

Ehlen, C. R., Magner, N. R., & Welker, R. B. (1999). Testing the interactive effects of outcome favorability and procedural fairness on members’ reactions toward a voluntary organization. Journal of Occupational and Organizational Psychology, 72, 147–161. Evans, D. C. (2003). A comparison of the other-directed stigmatization produced by legal and illegal forms of affirmative action. Journal of Applied Psychology, 88, 121–130. Federico, C. M., & Sidanius, J. (2002). Sophistication and the antecedents of whites’ racial policy attitudes: Racism, ideology, and affirmative action in America. Public Opinion Quarterly, 66, 145–176. French, E. L. (1999, August). Implementing equal employment opportunity through identity conscious or identity blind HR strategies: The effects on women’s advancement in management. Paper presented in: E. Yakura (Ed.), Creating Change Through Affirmative Action. Symposium presented at the annual meeting of the Academy of Management, Chicago, IL. Fried, Y., Levi, A. S., Browne, K. R., & Billings, S. W. (1996). The relation between political ideology and attitudes toward affirmative action among African-Americans: The moderating effect of racial discrimination in the workplace. Human Relations, 54, 561–584. Gaertner, S. L., & Dovidio, J. F. (1986). The aversive form of racism. In: J. F. Dovidio & S. L. Gaertner (Eds), Prejudice, Discrimination, and Racism (pp. 61–89). Orlando, FL: Academic Press. Gilens, M., Sniderman, P. M., & Kuklinski, J. H. (1998). Affirmative action and the politics of realignment. British Journal of Political Science, 28, 159–183. Golden, H., Hinkle, S., & Crosby, F. (2001). Reactions to affirmative action: Substance and semantics. Journal of Applied Social Psychology, 31, 73–88. Gutman, A. (2000). EEO law and personnel practices (2nd ed.). Thousand Oaks, CA: Sage. Hall, J. A., & Rosenthal, R. (1991). Testing for moderator variables in meta-analysis: Issues and methods. Communication Monographs, 58, 437–448. Harrison, D. A., Kravitz, D. A., & Lev-Arey, D. (2001, August). Attitudes toward affirmative action programs: A meta-analysis of 25 years of research on government-mandated approaches to reducing employment discrimination. Paper presented at the annual meeting of the Academy of Management, Washington, DC. Heilman, M. E., Battle, W. S., Keller, C. E., & Lee, R. A. (1998). Type of affirmative action policy: A determinant of reactions to sex-based preferential selection? Journal of Applied Psychology, 83, 190–205. Hopkins, W. E., Hopkins, S. A., & Mallette, P. (2001). Diversity and value commitment: A test of some proposed relationships. Journal of Managerial Issues, 13, 288–306. Hughes, M. (1997). Symbolic racism, old-fashioned racism, and whites’ opposition to affirmative action. In: S. A. Tuch & J. K. Martin (Eds), Racial Attitudes in the 1990s (pp. 45–75). Westport, CT: Praeger. Hughes, M., & Tuch, S. A. (2000). How beliefs about poverty influence racial policy attitudes: A study of whites, African Americans, hispanics, and asians in the United States. In: D. O. Sears, J. Sidanius & L. Bobo (Eds), Racialized Politics: The Debate About Racism in America (pp. 165–190). Chicago: University of Chicago Press. Kelly, E., & Dobbin, F. (1998). How affirmative action became diversity management: Employer responses to antidiscrimination law, 1961–1996. American Behavioral Scientist, 41, 960–984. Kernell, S., & Jacobson, G. C. (2000). The logic of American politics. Washington, DC: CQ Press. Kinder, D. R., & Sanders, L. M. (1996). Divided by color: Racial politics and democratic ideals. Chicago: University of Chicago Press. Klineberg, S. L., & Kravitz, D. A. (2003). Ethnic differences in predictors of support for municipal affirmative action contracting. Social Science Quarterly, 84, 425–440.

Predicting Affirmative Action Attitudes

129

Kluegel, J. R., & Smith, E. R. (1983). Affirmative action attitudes: Effects of self-interest, racial affect, and stratification beliefs on whites’ views. Social Forces, 61, 797–824. Konrad, A. M., & Linnehan, F. (1995). Race and sex differences in line managers’ reactions to equal employment opportunity and affirmative action interventions. Group and Organization Management, 20, 409–439. Konrad, A. M., & Linnehan, F. (2003). Affirmative action as a means of increasing workforce diversity. In: M. J. Davidson & S. L. Fielden (Eds), Individual Diversity and Psychology in Organizations (pp. 95–111). Chichester, England: Wiley. Kossek, E. E., Markel, K. S., & McHugh, P. P. (2003). Increasing diversity as an HRM change strategy. Journal of Organizational Change Management, 16, 328–352. Kravitz, D. A. (1995). Attitudes toward affirmative action plans directed at Blacks: Effects of plan and individual differences. Journal of Applied Social Psychology, 24, 2192–2220. Kravitz, D. A., et al. (1997). Affirmative action: A review of psychological and behavioral research. Bowling Green, OH: Society for Industrial and Organizational Psychology. Kravitz, D. A., & Klineberg, S. L. (2000). Reactions to two versions of affirmative action among Whites, Blacks and Hispanics. Journal of Applied Psychology, 85, 597–611. Kravitz, D. A., Klineberg, S. L., Avery, D. R., Nguyen, A. K., Lund, C., & Fu, E. J. (2000). . Attitudes toward affirmative action: Correlations with demographic variables and with beliefs about targets, actions, and economic effects. Journal of Applied Social Psychology, 39, 1109–1136. Kravitz, D. A., & Platania, J. (1993). Attitudes and beliefs about affirmative action: Effects of target and of respondent sex and ethnicity. Journal of Applied Psychology, 78, 928–938. Krysan, M. (2000). Prejudice, politics, and public opinion: Understanding the source of racial policy attitudes. Annual Review of Sociology, 26, 135–168. Kuklinski, J. H. et al. (1997). Racial prejudice and attitudes toward affirmative action. American Journal of Political Science, 41, 402–419. Lipset, S. M., & Schneider, W. (1978). The Bakke case: How would it be decided at the bar of public opinion? Public Opinion, 1, 38–44. Locke, E. A. (1997). Individualism: The only cure for organizational racism. The IndustrialOrganizational Psychologist, 34(4), 128–129. Ludwig, J. (2003). Public warming to affirmative action as Supreme Court hears Michigan case. Lincoln, NE: The Gallup Organization. Maio, G. R., & Esses, V. M. (1998). The social consequences of affirmative action: Deleterious effects on perceptions of groups. Personality and Social Psychology Bulletin, 24, 65–74. Matheson, K., Echenberg, A., Taylor, D. M., Rivers, D., & Chow, I. (1994). Women’s attitudes toward affirmative action: Putting actions in context. Journal of Applied Social Psychology, 24, 2075–2096. McConahay, J. B. (1986). Modern racism, ambivalence, and the modern racism scale. In: J. F. Dovidio & S. L. Gaertner (Eds), Prejudice, Discrimination, and Racism (pp. 91–125). Orlando, FL: Academic Press. McFarlin, D. B., & Sweeney, P. D. (1992). Distributive and procedural justice as predictors of satisfaction with personal and organizational outcomes. Academy of Management Journal, 35, 626–637. Mitchell, M., & Jolley, J. (1996). Research design explained (3rd ed.). Fort Worth: Harcourt Brace College Publishers. Nosworthy, G. J., Lea, J. A., & Lindsay, R. C. (1995). Opposition to affirmative action: Racial affect and traditional value predictors across four programs. Journal of Applied Social Psychology, 25, 314–337.

130

DAVID A. KRAVITZ AND STEPHEN L. KLINEBERG

Office of Federal Contract Compliance Programs (2002, January 4). Retrieved June 25, 2002, from http://www.dol.gov/esa/regs/compliance/ofccp/aa.htm. Ozawa, K., Crosby, M., & Crosby, F. (1996). Individualism and resistance to affirmative action: A comparison of Japanese and American samples. Journal of Applied Social Psychology, 26, 1138–1152. Peterson, R. S. (1994). The role of values in predicting fairness judgments and support for affirmative action. Journal of Social Issues, 50(4), 95–115. Reichers, A. E., & Schneider, B. (1990). Climate and culture: An evolution of constructs. In: B. Schneider (Ed.), Organizational Climate and Culture (pp. 5–39). San Francisco: Jossey-Bass. Richard, O. C., Fubara, E. I., & Castillo, M. N. (2000). The impact of explanations and demographic group membership: Reactions to diversity initiatives. Journal of Applied Social Psychology, 30, 1039–1055. Rosenthal, R. (1990). Replication in behavioral research. Journal of Social Behavior and Personality, 5, 1–30. Sax, L. J., & Arredondo, M. (1999). Student attitudes toward affirmative action in college admissions. Research in Higher Education, 40, 439–459. Sears, D. O., & Funk, C. L. (1991). The role of self-interest in social and political attitudes. In: M. P. Zanna (Ed.), Advances in Experimental Social Psychology (Vol. 24, pp. 1–91). San Diego: Academic Press. Sears, D. O., Van Laar, C., Carrillo, M., & Kosterman, R. (1997). Is it really racism? The origins of white Americans’ opposition to race-targeted policies. Public Opinion Quarterly, 61, 16–53. Sidanius, J., Pratto, F., & Bobo, L. (1996). Racism, conservatism, affirmative action, and intellectual sophistication: A matter of principled conservatism or group dominance? Journal of Personality and Social Psychology, 70, 476–490. Slaughter, J. E., Sinar, E. F., & Bachiochi, P. D. (2002). Black applicants’ reactions to affirmative action plans: Effects of plan content and previous experience with discrimination. Journal of Applied Psychology, 87, 333–344. Sniderman, P. M., & Carmines, E. G. (1997). Reaching beyond race. Cambridge, MA: Harvard University Press. Sniderman, P. M., Crosby, G. C., & Howell, W. G. (2000). The politics of race. In: D. O. Sears, J. Sidanius & L. Bobo (Eds), Racialized Politics: The Debate About Racism in America (pp. 236–279). Chicago: University of Chicago Press. Tougas, F., Beaton, A. M., & Veilleux, F. (1991). Why women approve of affirmative action: The study of a predictive model. International Journal of Psychology, 26, 761–776. Tougas, F., Crosby, F., Joly, S., & Pelchat, D. (1995). Men’s attitudes toward affirmative action: Justice and intergroup relations at the crossroads. Social Justice Research, 8, 57–71. Williams, D. R., Jackson, J. S., Brown, T. N., Torres, M., Forman, T. A., & Brown, K. (1999). Traditional and contemporary prejudice and urban whites’ support for affirmative action and government help. Social Problems, 46, 503–527.

BETWEEN SOLIDARITY AND INDIVIDUALISM: COLLECTIVE EFFORTS FOR SOCIAL REFORM IN THE HETEROGENEOUS WORKPLACE Orly Lobel ABSTRACT It has been argued that the workplace and the labor market in general, by processes of education, mobility and competition, have become the main forces behind the individualization and atomization in societies and in people’s lives. This paper inquires into the tensions between solidarity, identity, and individualism among workers in their efforts to organize collective struggles to improve their workplaces and their lives. Drawing on the dilemmas of increased diversity in the new workplace, the paper delineates three models of organized labor: (1) The Universalist-Individualist model of organized labor, peaking at the New Deal crisis and embedded in National Labor Relations Act, as an attempt to establish universal solidarity, which suppressed differences and presented a unified worker voice; (2) The Separatist model, which emerges as a reaction to intragroup exclusion and involves fragmentation of workers into identity groups, each representing the interests of its members; (3) The Coalitionist-Altruist model, envisioned

Diversity in the Workforce Research in the Sociology of Work, Volume 14, 131–164 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0277-2833/doi:10.1016/S0277-2833(04)14006-5

131

132

ORLY LOBEL

in the paper as a middle ground between solidarity and self-interest, through interrelated moves: a move from totalizing universal solidarity to coalitionist solidarity through continuous dialogue and “rotation of centers” and a move from rights-based identity politics and the dominance of employment antidiscrimination claims to a fuller substantive theory for social reform.

1. INTRODUCTION In the struggle for social reform, the workplace has always been a central arena. This is true both because the workplace is the locus of a tremendous proportion of human activity and because it is the main source of access to material resources (Kelman, 1996, reviewing Piore, 1995). Yet, rather than accommodating its natural tenant, the labor movement, the workplace arena has gradually become more the struggle of “identity” interest groups, which correspond to the new social movements of the 1960s. The American labor movement is facing an on-going crisis. Since their peak in the 1950s, labor unions have been constantly declining and there are predictions of their further decline in the near future (Crain & Matheny, 1999; Craver, 1998; Geman, 1997; Weiler, 1991).1 With the decline of collective bargaining, identity groups focus most of their attention on individual employee rights, and, in particular, on antidiscrimination claims. In many ways, the weakening of the labor movement and the increasing tension between workers of different identities echoes a wider crisis – that of fragmentation and self-interest in the political process, in which interest groups struggle to achieve the most for the individuals they represent rather than debating substantive ideological differences of social justice and reform. This paper inquires into the tensions between solidarity, identity, and individualism among workers in their efforts to organize collective struggles to improve their workplaces and their lives. The pattern I describe is a familiar one and appears in the histories of different social movements struggling between solidarity and individualism. Yet the pattern is perhaps most evident in the context of labor law, in which a body of law gives special recognition to the organized unit. Section 2 describes various interests and claims of different workers in the heterogeneous workplace, demonstrating the difficulties of defining mutual goals. Drawing on these dilemmas of diversity, the paper then explores three models of organized labor. Section 3 describes the first model of organized labor, peaking at the New Deal crisis and given legal recognition by the National Labor Relations Act (NLRA). This model was an attempt to establish universal solidarity, which

Between Solidarity and Individualism

133

suppressed differences and presented a unified worker voice. Section 4 describes a second model, which arises as a reaction to intragroup exclusion and involves separation into identity groups, each representing the interests of its members. I consider the reasons for such separation and the weaknesses of identity politics in the context of the workplace. Finally, in Section 5, I suggest a third model, which is an attempt to respond to the need for a middle ground between solidarity and selfinterest, by several interrelated moves: a move from totalizing universal solidarity to coalitionist solidarity through continuous dialogue and “rotation of centers” and a move from rights-based identity politics and the dominance of antidiscrimination claims to a fuller substantive theory for social reform.

2. THE HETEROGENEOUS WORKPLACE The workplace today is highly heterogeneous. In almost any workplace, workers divide along different categories of “identity” (Crain, 1994; Geman, 1997; Ho et al., 1996).2 In the following section, I briefly describe different interests and goals that are raised by these categories. This is not an attempt to give a full account of social identities that exist in the workplace, nor to provide a full list of their goals or needs. Indeed, I argue that such lists are never closed or fully pre-defined. Many of the needs and interests are yet to be articulated and they are always historically contingent and dependent on the power and the language that individuals and groups have to create and to express their desires. Rather, my aim is to provide an idea of the many different forces, needs, goals, voices and perspectives that may exist in any workplace. By describing a wide range of interests that workers might have, I hope to illuminate the complexity of workers’ experiences in a heterogeneous workplace and the difficulties of defining the goals of a diverse bargaining-unit. In a sense, voicing and problematizing the picture of identities and interests is a significant goal in itself, since the list of categories shows the impossibilities of defining clear and rigid identity boundaries (Russell, 1995).3 As Minow (1997) notes, recognizing that all workers inevitably fall into several different categories at once demonstrates the extent to which our individual and collective identities are fluid and open-ended. Noticing and thinking about paradoxes of identity can reduce the likelihood of polarized discussion while also reorienting attention from desires to fix or solve issues of identity toward attending to the circumstances that make identities seem so salient (p. 23).

Some of the interests I will describe are self-evident in the social construction of the groups with which they correspond. However, many criteria that are seemingly

134

ORLY LOBEL

neutral affect different sub-groups in very different ways. For example, childcare benefits, parental leave, day-care facilities, flexible hours and geographical stability are working conditions that seem mostly relevant to workers who have, or plan to have, young children. Moreover, while they could benefit both men and women as parents, in our society such benefits usually have a much bigger impact on the lives of women (Crain & Matheny, 1999; Finley, 1986; Frug, 1979; Gibson, 1987; Kalleberg, 1995).4 Since gender has been socially constructed in ways that women are the primary care-providers on society, women workers disproportionately suffer from the definition of working conditions as individualistic (Lobel, 2001b). But even among working mothers, attitudes toward the importance of these issues widely vary. Not only is the issue of childcare at the root of the special treatment versus equal treatment feminist debates, but also, in practice, there are many women who waive flexible hours even if these are provided, signaling their career as a priority in their lives. Yet, we could imagine that if childcare benefits would become widespread, they would influence the decisions of people regarding their family life, choosing at what stages they plan to become parents and when to develop their career. With such understandings, the questions concerning identities and interests reappear before us as circular. Thus, the issue of childcare is helpful in understanding the impossibility of defining clear categories of interests along “identity” lines, and the inevitability of the need to define substantive values in deciding among different interests. Therefore, I return to the issue of childcare in Section 5, when I discuss the need to develop a comprehensive theory of workplace justice. It should also be emphasized that the various interests of workers are not merely different from one another, but they almost always conflict. Almost any gain in workers conditions has its costs, and the resources, energy and time invested in bargaining is finite. Thus, given that resources are often of zero-sum nature, workers of different categories ultimately have to decide on allocation and reallocation based on their total interests in the workplace. Moreover, many interests are inherently designed to “jump the current queue” (as in the title of Kelman & Lester, 1997). For example, an organized group of workers in the workplace will inevitably have to make a choice between advancing the interest of promotion based on seniority, “skill” and “merit,” or identity-based affirmative action. Regarding the choice of categories, I initially tried to follow conventional categories of group identity: gender, race, ethnicity, disability, sexuality, religion, nationality, language, age, region, and rank. Yet, at the end of the list I added two categories, which, in lack of a better term, I term “class” and “personality.” The category of “class” has become less and less popular in contemporary American politics, in part because of the liberal rejection of a connection between birth and

Between Solidarity and Individualism

135

rank (Mink, 1986). Class as an identity seems unacceptable because it directly confronts the idea, albeit often illusionary, of mobility and individual change. Equality claims concerning social traits, associated with physical traits such as “immutable characteristics” or “discrete and insular minorities”5 are more likely to be heard than economic-based claims. “Personality” is conceived as an individual rather than a collective characteristic, and does not serve as a basis for categorization. However, personality, much like “identity” can be understood as social in the sense that it is often socially constructed and relational. It is ironic that one of the consequences of American identity politics is the disjunctures between “identity,” “class” and “personality.” It is even more ironic that socio-economic class and personality traits have little to do with defining the needs and interests of individuals and groups in the workplace. Gender – The rich scholarship of feminist jurisprudence has shown how women’s interests in working conditions may vary significantly from those of men. The most obvious on the list are pregnancy and maternity leaves and safety during pregnancy. Often, women workers also have a greater interest in comparable worth wages, in redefining unpaid labor, in a harassment-free workplace environment, and in affirmative action. And as discussed above, parental leave, day-care facilities, family health benefits, flexible hours and geographical stability usually have a disparate impact on the lives of women workers. These interests also correspond with another category that could be added to the list of “identities,” that of “family status,” since workers with partners or with children have different benefit interests than single/childless workers. Race/Ethnicity – Race and ethnicity theorists have identified particular interests of racial and ethnic groups in the workplace, including prevention of prejudice and racial harassment (EEOC v. Mitsubishi Motor, 1998; Marshall, 1995), redefinitions of “merit” (Crenshaw et al., 1995), affirmative action in recruitment and promotion, and elimination of systematic patterns of covert discrimination by biased requirements and seniority benefits (Griggs v. Duke Power Co., 1971).6 Moreover, cultural differences and different ethnic backgrounds can lead to different expectations and understandings of the role of “work” and “workplaces” in our lives. Disabilities – The American with Disabilities Act (ADA) states the right for reasonable accommodations for employees with disabilities. Accommodations include modification of the work site, work schedule, position restructuring, special equipment, technologies or devices to assist the employee and reassignment to a vacant position if accommodation in the employee’s current position proves unfeasible (which may sometimes conflict with seniority benefits) (Hankinson, 1998). In addition, the creation of a collegial working environment, in which workers help each other, can in certain situations be more important to workers

136

ORLY LOBEL

with disabilities. Sexual Orientation – The gay rights movement has advocated promotion of domestic partner work benefits and extended family benefits, such as leave rights for family care and health benefits. In addition, gay and lesbian groups that have formed in workplaces have among their goals the elimination of homophobia and harassment on the basis of sexuality and the encouragement of AIDS awareness (Thomson, 1993, as cited in Geman, 1997; Young, 1995). Religion – Religious affiliation can lead to various constraints regarding the workplace. The most obvious is the definition of weekdays and holidays. For example, for an observant Jew, a full workday on Friday is a burden, and shifts on Saturday are impossible (Trans World Airlines v. Hardison, 1977).7 Other difficulties may occur when dress codes, food plans and other conditions at a workplace conflict with religious requirements. Nationality/Language – Immigrants in the United States face the difficulties of cultural barriers and alienation in a highly “Americanized” workforce. They experience hardship with regard to language requisites and discrimination on the basis of foreign accents and English-only rules. In addition, many immigrants have an illegal status informally resulting in alienation from the protection of the American legal system. Age – Stability of work and pension rights may have an increasing importance according to age. Moreover, the age of the worker can undoubtedly influence her interest in promotion on the basis of seniority and in beneficial retirement programs. Yet, even “benefits” based on age have an unclear impact. For instance, the benefits of the “life cycle model” of deferred payment to older workers are unclear since they expose older worker to the risk of being fired (Issacharoff & Harris, 1997). Region – A worker’s geographical home can also be an influential factor in defining her interests. Many times, a workplace serves as an important source of income for an entire community. Conflict on the level of a national union may easily occur when a firm plans to move its base from one state to another, benefiting some workers, but risking the jobs of a community. Moreover, a geographical community is more sensitive to the relation between labor and other realms of social life, and therefore can unite members of the community not just as workers but also as consumers, voters, parents, etc. Rank – Workers of different ranks are thought to have very different interests. Higher-ranking workers have more authority and responsibility vested to them by the employer. Also, higher-ranking workers most often earn more and many times have a different educational and socio-economic background. Managerial workers are thought to have divided loyalties and different economic interests than rank and file workers (Crain, 1990; NLRB v. Bell Aerospace, 1974). They are considered so

Between Solidarity and Individualism

137

essentially different than other workers that they are not included in the union’s bargaining unit as defined by the NLRA, and thus are not allowed to take part in the process of collective bargaining (NLRA, 1994). Socio-Economic Class – The economic conditions that individuals face in their lives, including wealth, income, level of education and family background are significant in shaping their interests and needs in their vocational pursuits. In the workplace, indications of income and welfare are particularly relevant because they often serve as a link between otherwise diverse workers. However, physical and cognitive shifts from “production” to “services,” from “blue collar” to “white collar,” from “working class” to “middle class” have all contributed to the decline of worker mobilization around the category of socio-economic interests (Heckscher, 1988).8 When thinking about the categorization of the needs and interests of workers in the today’s workplace, it is significant to consider when and how did a vibrant concept of a “working class” transform from a mobilized cultural community to “the vast masses of propertyless and powerless white collar workers, in services, shops, and offices, who imagine themselves member of a ‘middle class’ to which all but the richest and poorest belong” (Unger, 1996). Personality – There is something out there called “personality.” Is it merely an aggregation of the different identity categories? Is it something so personal and individual that we cannot define nor categorize it? Can we imagine relevant characteristics for group solidarity that are unrelated to current categorizations of identity? Imagine categories or organizations of people who are ambitious, competitive, tidy, friendly, generous, greedy, active, workaholic, nervous, aware, selfish, engaged. Or people who love loud music and vibrancy in an openspace working environment vs. people who prefer quiet, separate and closed-door offices with hardly any interaction with their co-workers. People have different expectations and emotions toward their jobs. Yet, this dimension of differences is often suppressed in the American politics of identity. Indeed, the reconciliation between solidarity and individualism can also become reconciliation between identity and personality and between work and other aspects that define our lives.

3. THE UNIVERSALIST-INDIVIDUALIST MODEL The universalist model was the basis for the traditional labor union as constructed by the NLRA in the New Deal. The model is based on the idea that workers should present a unified voice to advance their common goals. Therefore, the universalizing ideal prefers solidarity to individualism and emphasizes the common while silencing differences. The model requires a well-defined form of representation, which involves strict separation between leadership and grassroots

138

ORLY LOBEL

activities, demands loyalty to the group from its members (Halley, 1998),9 and requires that representation be exclusive. Thus, the NLRA grants exclusivity to the majority union, that is, the union that wins the most votes in an election. The representatives of the majority union are the only ones who can bargain on behalf of all the workers in the bargaining unit, including non-members (Crain & Matheny, 1999; Hyde et al., 1993; NLRA, 1994).10 Paradoxically, the universal solidarity model of the NLRA is also deeply embedded in individualism and self-interest. Exclusive representation and separation between leadership and rank and file can alienate group members from each other and promote material “common” self-interest. And, indeed, the courts understood the NLRA as implying individualism by replicating the classical model of contractualism and perceiving the merits of collective action as merely a way to equalize bargaining power (Klare, 1978). Therefore, rather than interpreting the NLRA as a framework for the advancement of worker participation in ownership and management, collective bargaining was limited to be simply a means of remedying unequal bargaining power by numerical aggregation of its members and strict representation by leadership. Thus, the NLRA, as it has been interpreted by the courts, limits the scope of the bargaining unit, excluding from it “managerial employees,” limits the possibilities of workers participation, and limits the modes of labor activity and labor speech, excluding from it “political speech” (Estlund, 1992; Fischl, 1989; Hyde, 1981; Lynd, 1984; NLRB v. Peter Cailler Kohler Swiss Chocolates Co., 1942).11

3.1. Overt Exclusionist Universalism The universalist model was initially advanced in the context of overt exclusion. At the peak of the labor movement struggle, the labor movement was a relatively homogenous exclusionary movement, which made it relatively easy to universalize “the worker” as a white, male, English-speaking, able, American worker. Up till the New Deal, and in some cases even after, unions widely practiced overt exclusion of minorities, women and immigrants (Foner, 1964). During the Lochner era, almost all labor unions engaged in racial discrimination, ranging from complete exclusion to segregation (Bernstein, 1993; Gaspain, 1998; Patchen, 1998; Russell, 1994). Due to this exclusion, employers constantly used black strikebreakers during strikes (Whatley, 1993). Yet, although most black leaders opposed the labor movement, there was a complex dynamic between the civil rights movement and the labor movement (Meier, 1969, as cited in Bernstein, 1993). Many times, the NAACP refused to criticize New Deal politics that granted a monopoly power to racist

Between Solidarity and Individualism

139

unions, because it feared that criticism would harm the labor movement’s social struggle (Wolters, 1970). On the other end, in some instances, labor unions were said to be “pushed” to the Republican Party to distance themselves from the civil rights movement. Women were excluded from unions for fear that they would take up men’s work and that women’s work would decrease wages and weaken the nuclear family (Crain, 1994, p. 1942; Kessler-Harris, 1975).12 Even when women were allowed to join unions, it was partly because the American Federation of Labor (AFL) understood it would be easier to control women from within the unions (KesslerHarris, 1975, pp. 97–98). When unions finally admitted women to their ranks, many unions employed “exclusionary tactics ranging from holding late-night meetings in saloons, where women were unwelcome without escorts, to ridiculing women who spoke at the meetings” (Kessler-Harris, 1975, p. 99). As the workforce grew more and more diverse and overt exclusion from union ranks became no longer possible, the majoritarian rule of the NLRA served as a basis for covert exclusion. During the 1940s and 1950s, discriminatory collective bargaining agreements were signed by unions, including clauses that blocked the hiring of African-Americans and of married women, and clauses that forced women who married to resign (Hill, 1977; Steele v. Louisville & Nashville R. R., 1944; see also, Gabin, 1985, as cited by Crain, 1995; Hill, 1987).13

3.2. The Development of Fair Representation and Covert Discrimination The first move away from overt discrimination and from radical universalism was the increased scrutiny over the internal activities of the labor union by the development of the fair representation doctrine. Although the NLRA did not include such an obligation, the courts developed the duty of the labor union to represent fairly each of the workers that are part of the bargaining unit. The duty was first articulated in Steele v. Louisville & Nashville R. R. (1944) in regard to the Railway Labor Act.14 The court held that a union could not bargain for an agreement that discriminated against the black members of the bargaining unit. The court explained that the duty to represent fairly was an implied obligation in the idea of exclusive representation. The courts further developed the fair representation doctrine to encompass a wide prohibition, stating that a labor union may not use its discretion partially, arbitrarily or in bad faith (Vacca v. Sipes, 1971, p. 171). Failing to fairly represent its members, the union itself may be liable under Title VII or under a breach of duty (Vacca v. Sipes, 1971, p. 190). Although the doctrine of fair representation is an important development in decreasing internal discrimination, many writers have commented on the weakness

140

ORLY LOBEL

of the doctrine, its vagueness and the need for its reform (Cheit, 1982; Goldberg, 1985; Malin, 1992; VanderVelde, 1983)15 . Most problematic is the fact that, most of the time, goal definition cannot be read as discrimination, since the focus on wages and material benefits is widely understood in American culture to be the lowest common denominator of all workers. Discrimination is usually not perceived as such if it merely involves an omission to bargain on various issues that are more central to the lives of certain workers. Since all issues placed on the bargaining table are in fact competing with one another in a situation of finite time, energy and resources, it is almost impossible to detect how a bargaining process and its outcome has disparately affected different workers. Since historically, and still today, union leadership is dominated by men, even at times when they do not necessarily outnumber the women in the group,16 these questions become even more crucial. Although we could imagine union leaders committed to the representation of all members of the group, it is difficult to assume that without active participation in leadership, women’s voices are sufficiently heard and understood. The issue of childcare illuminates these problems. Traditionally, the “family” is considered essentially different from the “market,” and the two concepts are thought of as two separate spheres with rigid boundaries between them (Olsen, 1983). Thus, even though women are now a dominant part of the workforce, very few collective agreements include clauses regulating flexible hours or childcare facilities (Diktaban, 1991)17 . This is particularly salient when compared to issues of workers’ “individual care,” such as sickness, injury and pensions, that are typically included in collective bargaining agreements. Similarly, collective bargaining on equal pay for comparable worth, sexual harassment prohibitions, and issues concerning gay and lesbian workers has been scarce (Crain, 1994, pp. 1945, 1995). Even when such benefits are conferred, they are often a result of external interventions by different social movements. For example, where domestic partner benefits have been conferred on employees through collective bargaining, it has usually been under the threat of litigation, initiated by gay and lesbian identity groups such as Lambda (Briggs, 1994). Regarding disabilities, although labor unions have been strong vocal supporters of the ADA (Dubault, 1995; O’Melveny, 1993–1994), there have already been cases in which unions opposed accommodation of disabled workers that interfered with collectively bargained seniority rights (Eckles v. Consolidated Rail Corp., 1996).18 In general, in many cases, labor unions have been hesitant to represent workers in discrimination claims and grievances, and workers have turned to outside organizations of their identity group for legal and moral support (Crain, 1995, p. 33). In certain cases, union leaders have even played an active role in discrimination and have reportedly been active participants in creating hostile working environments by engaging in sexual harassment (Crain, 1995, p. 33).

Between Solidarity and Individualism

141

In sum, while the advantage of the traditional universal model of organized labor is its strong and united front against the “common adversary,” the employer, its problems are evident – silencing and exclusion of voices and interests of weaker sub-groups, coupled with pacifying workers’ struggles by a rigid separation between representatives and rank and file. The next model describes ways in which identity groups reacted to these problems.

4. THE SEPARATIST MODEL 4.1. Separate Identity Caucuses and Internal Divisions The first African-American identity caucus, Bay Area Black Employees (BABE), at Xerox, was founded by sales representatives in the San Francisco Bay area in 1969 (Hyde, 1993). BABE convinced management to increase the number of black employees at Xerox by assigning one black interviewer to each black candidate and by personal recruitment by BABE members. BABE’s goals widened when the group filed a suit against Xerox, claiming that black sales representatives were assigned inferior sales territories. The suit was settled, local management was replaced, new territories were assigned, and Xerox even provided some compensation to equalize past earnings.19 During the 1970s, black Xerox employees organized conferences to coordinate caucuses that emerged in several different regions (Hyde, 1993). Later, in the 1980s, identity caucuses of Hispanics, Asians, and women formed at Xerox (Hyde, 1993; see also, Thompson & DiTomaso, 1988). Since then, many other identity groups have formed in various workplaces, representing workers with shared “identities.”20 At AT&T, the Alliance of Black Telecommunication Employees (ABTE) was founded. Among its activities, the ABTE has provided AT&T with suggestions on how to improve diversity and racial sensitivity (Scales, 1993, as cited in Geman, 1997).21 Similarly, women’s identity-based caucuses and women’s corporate groups such as 9 to 5 have formed to advance women workers rights in various fields.22 Many of these groups are supported by Catalyst, an organization that helps businesses set up corporate women’s groups (Geman, 1997).23 Gay and lesbian identity caucuses have also formed in several workplaces. At the Walt Disney Co., the Lesbian and Gay United Employees (LEAGUE) has among its goals achieving domestic partner benefits, combating homophobia, and encouraging AIDS awareness (Thomson, 1993, as cited in Geman, 1997; Young, 1995). Another gay and lesbian workers group, Pride at Work, is currently allied with the labor movement.24

142

ORLY LOBEL

In the public sector, and particularly in fire and police departments, there have been similar moves of fragmentation. Separate organizations have formed, usually around class action anti-discrimination law suits, in order to represent the particular needs of the different groups that experienced discrimination (Marshall, 1995; McUsic & Selmi, 1997).25 In such instances of anti-discrimination suits, the courts have recognized the role of identity groups. For example, in Local Number 93 v. City of Cleveland (1986), the court allowed a group of African-American and Latino firefighters called the “Vanguards” to negotiate a consent decree with the city regarding race-conscious promotions. The court approved the decree despite the objections of the labor union, which represented the majority of the city’s firefighters. Identity caucuses have emerged both in unionized and non-unionized workplaces, yet there are significant differences among these two settings. In non-unionized workplaces, employee caucuses are informal entities that may operate in many ways and with various goals, yet the employer is not under the obligation to bargain with them. In unionized workplaces, because of the majoritarian rule, the range of activities that identity groups may engage in is more problematic. In Emporium Capwell Co. v. Western Addition Community Organization (1975), African-American workers, dissatisfied with their union’s reaction to the employer’s racial discrimination, organized and engaged in protest. The employer fired two of the protesting workers. The Supreme Court, reversing the Court of Appeals, decided that the workers’ protest should not be regarded protected union activity, because permitting separate protest and bargaining would undermine the union’s power. The court explained that allowing minority employees to bargain separately with employers could fragment and weaken unions, which would adversely affect their ability to combat discrimination in the long run. Thus, the paradoxical result of the exclusivity rule of the NLRA is that in non-unionized workplaces, identity groups can more easily protest discrimination than in unionized workplaces (Crain, 1994, p. 1976; Crain & Matheny, 1999; Iglesias, 1993; NLRB v. Downslope Industries, 1982).26

4.2. Separatism as a Remedy for Exclusion The rise of identity caucuses in the workplace has been a consequence of a complex set of developments. The first explanation is that separation is a remedy for the historical exclusion and the on-going discrimination against minorities through the traditional universalist-individualist model. Once groupwide solidarity has failed and intragroup dynamics have proved to be coercive, liberal notions of individual freedom typically call for divisions of the group

Between Solidarity and Individualism

143

into smaller interest groups, rather than advancing mechanisms to adhere to the larger group. Marion Crain has called for a separatist move of women in the workplace. She believes that in order to be effective, feminists must have separate organizations (Crain, 1995). According to Crain’s separatist model, women’s caucuses will be on a voluntary basis and will serve to inform women workers of their rights and to represent women in cases like sexual harassment. Crain recognizes the difficulties of intersections between different “identities,”27 but her solution is the voluntary basis for affiliation with these organizations, so that African-American women in a particular occupation could chose whether to join or form an identity caucus based upon their race and sex, whether to join or form a caucus organized around gender identity alone, whether to join or form a caucus organized around racial identity alone, or on some other basis (pp. 70–77).28

Elizabeth Iglesias (1993) has argued for separatist organizations for women of color workers. Drawing on a model proposed by Eileen Silverstein (1979, as cited in Iglesias), Iglesias suggests multiple representation by interest groups in a given bargaining unit, including a limited right of the groups to veto collective bargaining decisions made by the majority, subject to a super-majority override. Iglesias also offers to consider Lani Guinier’s model of proportionate interest representation through cumulative voting, in which the threshold would be reduced below 51%, so that union officials would be more representative of their electors (Guinier, 1991a, b, as cited in Iglesias, 1993). A third model Iglesias describes is that of George Schatzki (1975, as cited in Iglesias, 1993), a system of “self-determined representation,” that would produce a network of multiple-interest unions. A comparison of these models illuminates the differences between internal divisions from within and external separatism from without the group as ways of reacting to intragroup conflict. In Section 5, I consider such differences and suggest that, rather than requiring the formation of separate bargaining units, doctrinal reform of the NLRA could allow the expression of voices remaining within one union.

4.3. Separatism as a Strategy While exclusion is undoubtedly one explanation for the separatist move, it is actually only part of the story of fragmentation. Identity groups have at times been more successful in changing the workplace than traditional unions (Heckscher, 1988). The reason for this is, on the one hand, the wide hostility toward labor unions29 and toward arguments concerning class as an identity,30 and on the other

144

ORLY LOBEL

hand, the relative receptiveness of anti-discrimination and recognition claims on behalf of identity groups. In his book, Beyond Individualism, Michael Piore (1995) describes the difference between the labor movement and the new identity movements. Piore argues that American unions, because of their proximity to the economic environment, have historically been more sensitive to economic reality and more concerned with distributive justice than the new social movements. Because of its sensitivity to the economy, the labor movement has also been more ready to compromise. In contrast, the new social movements, built on the model of the civil rights movement, tend to focus on equality and recognition and less on questions of distributive justice (Halley, 1994). Traditionally, civil rights legislation bans bad treatment based on race, sex, and other specified grounds and provides remedies and thus deterrence designed to put victims of discrimination on a playing field with everyone else (p. 134).

Piore believes that the new social movements make unlimited demands, without taking economic distribution into account, because of their relative distance from the economy, and because their demands are intertwined with symbolic recognition. Since the groups are seen to represent the “whole person,” their claims “touch upon the very core of people’s self-conception.” Therefore, the demands of the new social groups are many times made more for “principle” or symbolic value than for their practical impact (Piore, 1995). In response to Piore’s account of the nature of the new social groups, Mark Kelman (1996) has argued that, rather than being victims of “cognitive” (distance from economics) and “affective” (longing for recognition) disadvantages, the new groups are more likely engaged in rational conscious instrumental rhetoric that advances their interests. The new social movements have appropriated their rhetoric from the civil rights movement “using socially salient antidiscrimination language which they have realized is simply good strategy for pursuing quite traditional material goals:” [T]he failure of identity groups to account for resource limits is driven significantly by the rhetorically motivated use of antidiscrimination imagery, for in the traditional antidiscrimination context, resource limits are presumptively illegitimate. The slogan “the cost of remedying discrimination is no defense” remains socially persuasive, at least in situations in which the defendant has engaged in “traditional” discrimination (the failure to treat the protected group member in an economically rational, impersonal fashion) (p. 993).31

I believe that separation should be understood somewhere in-between Piore’s (1995) and Kelman’s (1996) readings of workplace identity politics. Separate identity struggles in the workplace today are a product of several combined factors, including historical and on-going discrimination, the legal and social

Between Solidarity and Individualism

145

deradicalization of the labor movement’s struggles, the weakness of the NLRA, the dominance of rights discourse in American politics, and the disjuncture between group “identity” and socio-economic “class.”32 Moreover, I believe that the tendency to replicate identity groups at the workplace from other arenas, is in part, the result of the American culture of sharp separation between “work” and “play,” “family” and “market,” and “private” and “public.” These dichotomies not only result in self-alienation, creating a “double life” for workers in their different activities, but also create an image of desires and interests as constructed mostly in the sphere of the home and the community, rather than at work (Olsen, 1983).33 Hence, just as there is no unitary concept of “worker,” there is also no unitary reason for fragmentation. More important than the attempt to define the exact motivations for separation is the attempt to understand the weaknesses of separation and the ways in which alliances could be forged to remedy alienation, fragmentation and pacification of workers.

4.4. The Consequences of Separatism The cautionary call “choose your battles and your enemies” seems appropriate to describe the consequences of organized labor fragmentation. Although separation may have short-term advantages to interest groups, in the long run it contributes to a powerful employer at the expense of workers. The aggregate power of the labor union is inevitably sacrificed with the formation of smaller separate identity groups. Since antidiscrimination claims often have an underlying meaning of a zero sum distribution, much of the struggle of social movements for rights within the workplace is directed against each other rather than against the employer and the state. Put in Janet Halley’s (1998) words: Particularly where hard resources are involved, it is alarmingly easy to see that winner-take-all civil rights contests can take shape. Affirmative action programs are rife with such contests, which pit one recognized civil rights constituency against another. For instance, in minority business enterprise programs, blacks and Latinos have had ample opportunity to observe white women speed ahead of them in contests for finite resources (p. 131).

In addition, disputes among workers are used by anti-labor groups to argue that there is no workable way to involve workers in management, to bargain collectively, or to form participatory models that require cooperation between workers and ownership.34 Some social theorists even believe that working class fragmentation is a product of deliberate practice of employers, who benefit from opposition and division within groups of workers (Crain & Matheny, 1999).35 Under the “dual labor market” theory the upper classes support discrimination in

146

ORLY LOBEL

order to create a segregated dual market in which a reserve army of marginalized workers is produced, and, thus ultimately, all labor is disadvantaged (Crain & Matheny, 1999). Moreover, separation into smaller groups does not resolve the difficulties of achieving non-coercive solidarity. A comparison to other social movements reveals that there are always inner tensions and difficulties in reaching a consensus about common interests. Most identity groups are faced with the dilemma of defining mutual goals in a reality of multiple experiences and voices.36 Therefore, separation does not escape, although it may mitigate, the dangers of intragroup exclusion and silencing.37 The problem is further deepened by the fact that any participation mediated by group representatives puts into question the nature of interest representation, since those who volunteer to speak may not represent the group that they purport to represent (Chayes, 1976; Grossberg, 1989). Finally, separatism assumes a “coherentist” view about group identity, which holds that the group shares a coherent set of experiences, attitudes and political and legal interests (Halley, 1998, pp. 115–116). Yet, we have seen that in the complex context of the workplace these assumptions are rarely plausible. Separate interest groups risk undermining the differences between individuals and the complexity of human nature, and depreciating the role of work in forming identities, by replicating the same identity groups from other arenas. Therefore, it is not at all clear that the goal of organized labor should, or even could, be to form a coherent collective agenda that would encompass the self-interests of all the members of the group. Rather, goals should focus on broader ideals of workplace justice that transcend questions of identity. The workplace thus becomes an important arena to inquire into the possibility of “post-identity politics,” based on the premises that there is no preexisting “authentic” identity to which we can refer prior to the political activities of the workers (Halley, 1998, pp. 116–117).

5. THE COALITIONIST-ALTRUIST MODEL 5.1. The Importance of Collective Action The “law of work” constitutes both employment and labor laws.38 While employment law is concerned with the rights of individual employees and is monitored mostly by individual litigation,39 labor law offers a starting point of collective direct action – groups of workers acting together in order to achieve the most for themselves from their employers. While, in effect, individual protection has been replacing collective bargaining, rather than complementing it, it is important to maintain a “mixed system” of regulating the workplace by both

Between Solidarity and Individualism

147

labor and employment law. Individual legislation can aid collective bargaining and strengthen union support by removing controversial issues from the bargaining table and permitting unions to concentrate on issues that are of particular interest to the employees in a specific workplace or industry (Bales, 1997; Van Wezel Stone, 1992). Moreover, it is important to understand that although much of the conflicts of interests between union members could be settled by regulating controversial issues through protective employment legislation, there always remains further room for labor activism. First, it is almost impossible to regulate all controversies and there are always open-ended, ad hoc issues left for the bargaining table. Second, it is beneficial to allow some flexible and dynamic regulation in every workplace according to the particular characteristics of the workplace. Third, there is an inherent value in direct participation, allowing transformative empowerment and community building through decentralization and self-determination. In the context of the workplace, this is particularly significant, since the workplace is where we spend a great amount of our time and energy. Indeed, studies have found that motivation to work is higher in workplaces where there is a strong union, and that contemporary dissatisfaction of workers has been found to be expressed more passively than in the past (Heckscher, 1988; Lobel, 2001a).40 Lastly, in reality, many important issues are currently unregulated in the American market. Employment law is focused mainly on anti-discrimination litigation, while significant issues, including interests of workers that have traditionally been marginalized, such as work/family relations, cannot be argued for on the basis of discrimination. Where legislative efforts have failed, labor bargaining serves as an important medium through which the workers can achieve direct reform vis-`a-vis their employers.

5.2. Solidarity through Continuous Dialogue The current wild race-to-the-courthouse individualism is disturbing precisely because it lacks any mechanism for intracommunal dialogue, but if such a dialogue were possible, who would be invited to join in it and how could it be conducted without coercion? (Halley, 1998, p. 127).

Labor organization cannot and should not be an attempt to present a “universal worker,” nor should solidarity be viewed as possible only among workers sharing an essential “identity core.” Rather, moving toward a third “in-between” model should focus on the evolving and contingent meanings of work in the lives of different workers in different positions, places and times. Through shared political activities and constant redefinition of the workplace itself, workers can find common grounds

148

ORLY LOBEL

for social reform and the labor movement could revitalize. In her book, “Not Only for Myself,” Martha Minow (1997) urges that solidarity should be achieved by hard work of constant interaction rather than by debates over “true” or “authentic” identities: Preoccupation with what it means to be an American is no less constraining than preoccupation with being Chicano, or a woman. Both preoccupations try to summon a solidarity prior to political action; both try to skip the hard work of inventing, crafting, and often straining for understanding and connection in the very process of deliberating and judging. Solidarity does not arise from understandings prior to politics, but instead politics permits its construction, in particular moments (p. 145).41

Michael Piore (1995) has suggested that labor unions could serve as an important institution that would emphasize the existence of “borderlands,” where particular individuals are attached to more than one group. Piore borrows the notion of borderlands from anthropologist Renaldo Rosaldo (Rosaldo, 1989, as cited in Piore, 1995), who uses the term “borderland cultures” to explain that people who experience multicultural interactions relate to their cultures in ways that alter the interpretative framework of culture itself. Rosaldo thus rejects mainstream notions of anthropology, in which individuals are thought to simply reflect their culture, rather than being agents of its dynamic creation and evolvement. Piore further extends the metaphor of borderlands to describe both individuals, who are attached to different institutions and can contribute to the evolvement of crosscultural dialogue and institutions, where such cross-cultural interaction takes places. [B]orderland individuals carry the themes they encounter in each of the communities to which they are attached into the discussions and the debates of others, changing the content and character of the conversations . . . making them more cognizant of and able to talk to one another. It seems natural to think in terms of borderland individuals who perform this function but also of borderland institutions where those individuals encounter the themes they import into the communities to which they belong (p. 163).

Piore (1995) believes that universal public institutions, where the borders between groups are open, provide an opportunity to reverse the move of American society toward particularism. Therefore, he suggests that while it is important to maintain the rights of particular groups to create their own institutions, public funds should be devoted to such universal institutions as existing borderlands, and to the creation of new borderlands. Labor unions as “borderland institutions” could generate internal institutional pressure for conversations and enhance the ability of workers to expand their agenda and to force their leaders into internal dialogue. Labor unions could also serve the important role of introducing a vocabulary of economic awareness to various groups that are usually isolated from each other. Most importantly, the framework of collective bargaining has the potential of

Between Solidarity and Individualism

149

providing a process that could mediate and reconcile the conflicts among the many claims and interests of different communities and individuals. Molly McUsic and Michael Selmi (1997), building upon Piore’s idea of “borderland unions,” have offered a concept of “cosmopolitan unions,” in which unions serve as mediating institutions that emphasize both the common and the differences between their members.42 In the cosmopolitan perspective, individuals agree to engage in dialogue and to experience both understanding and conflict. Similarly, Crain and Matheny (1999) have argued that their call for separate identity groups and representation of unions could be a basis for “rainbow coalitions,” in which differences are not suppressed but voiced and coalition is not based on “principles of unity” but on multiples perspectives and experiences. Therefore, they argue that separatism is a means for meaningful coalition-building and mutual education about differences.

5.3. Developing Mechanisms for Non-Coercive Activism and Good Faith Solidarity Calls for solidarity through transformative hermeneutic dialogue have become common in light of social fragmentation, conflict and coercion. The most difficult questions arise at the steps beyond mere “calls,” in which we must elaborate on how to foster ideas of non-coercive dialogue and critical coalition formations. Yet, there is no one single model that can alone ensure that the traditional problems of exclusion and silencing will be avoided. Indeed, it is important not to neglect horizontal expressions of coercion by uncritically accepting the traditional vertical flow of power relations (Lobel, 2001a). Rather, it is important to experiment with new forms of organization and activism that could be interrogated and developed dynamically. In the following sections, I offer various focal points for such elaboration in the context of the heterogeneous workforce, both at the level of the single workplace and within the broader framework of movement struggles. Within the workplace, an adjusted model of labor organization could allow diversity without mandating complete fragmentation. Several commentators have proposed to consider abandoning the majoritarian rule and allowing multiple representation and separation into interest unions within a bargaining unit.43 Yet, critical coalitions among workers in a workplace could be formed within the forum of the union, rather than eliminating the united front of exclusive representation vis-`a-vis the employer. The default should remain a majoritarian rule and representation of all workers in a bargaining unit by a single union. However, there should also be restricted possibilities to bargain separately on

150

ORLY LOBEL

issues (rather than separate membership) in which no agreement between workers has been reached. Such mechanism could produce several important results. First, the model would mandate constant dialogue and extensive deliberation by all members in a distinct setting. Second, only when workers demonstrate that they have tried among themselves, in good faith, to reach common understandings and have failed, will a group of members be able to act separately. Third, separation will not entail permanent fragmentation in regard to union membership. Thus, bargaining for all workers on non-controversial issues would continue, and even in regard to controversial issues, separation would be on a temporary basis and would have to be periodically reconsidered. In order to make such adherence to the format of labor unions viable, there should be a renewed focus on different procedures to exchange information and opinions among workers – creating new forums for input and output regarding workers’ experiences, rights and needs, supporting regular broad meetings of workers, rather than just representatives, and investing public funds in training for union activism. Other aspects of labor law reform must also complement these changes. By decreasing current restrictions on labor activism, including the distinction between “labor” and “political” speech and the prohibition on secondary boycott, collective bargaining on issues that are less traditional (and thus more “political”) would be facilitated and the union could become more multidimensional. Finally, as I discuss in the next section, developing forums of dialogue, increasing forms of action and enhancing exchange of information are all related to the idea of shifting interests and rotation of centers among workers. While the option of separation would be provided, encouraging preliminary universal engagement has the potential to gradually reduce controversy among diverse workers.

5.4. Rotating Centers and Substantive Ideals for Social Reform Collective engagement through constructed deliberative processes has the inherent value of not only illuminating long-term mutuality and reciprocity, but also transcending the contemporary rigidity of identity politics. The creation of a model of mandatory forums for dialogue and action among diverse workers, with the provisions of “exits” if agreement has not been reached, has the potential of creating new concepts of cooperation. Such cooperation signifies the idea of a “rotation of centers,” in which all partners depart from their narrow and immediate self-interest (Valdes, 1998).44 Even prior to labor organization, co-workers are situated in the unique position of long-term relationships and direct interaction. Therefore, the forum of negotiating priorities among workers can consider not only present, but also past and future, “rounds.” In such interaction, workers can

Between Solidarity and Individualism

151

not only understand how identity dynamics lead in the long run to a race-to-thebottom, but also learn to question the vocabulary of rights and redefine the political discourse toward intersubjective understandings of selves and communities. The perspective of time, the experience of rotating centers and compromise and the fluidity of identity boundaries can show workers how, in fact, any reform transcends the particular interests of the “identity group” that advanced the reform and can change our perceptions of society as a whole. The idea of a rotation of centers can be better understood when we consider the issue of childcare. Childcare today, unlike conditions as wages and hours, is recognized as a permissive, rather than a mandatory bargaining subject.45 Compared to other countries, American workers enjoy very few childcare benefits (Kamerman, 1988; Mills, 1992; Note, 1987; Thomsen, 1996), 46 and we have seen that few labor agreements address the issues of childcare (Diktaban, 1991). We have also discussed the impossibility of defining clear identity boundaries of the interest of some workers in childcare. Yet, we know that in fact, lack of affordable childcare many times leads to the decision of women not to work or to work in part-time jobs (Diktaban, 1991). Thus, restructuring the workplace to allow both parents to share responsibilities of childcare and ensure that women can equally compete in the workforce, has been a common goal of feminists from different parts of the equal/special treatment spectrum (Kay, 1993; Williams, 1984–1985). Some workers may argue against such benefits and ask, for example, why should childless workers bear the costs of parents? The answer to such inquiries should be threefold. First, strong labor coalitions can ensure that costs are, at least mostly, directed toward the employer. Concrete progress has been made in the conditions of work over the years, and imagining further progress embodies faith in the possibilities of redistribution among owners and workers, not among workers themselves. Second, not only is it pervasively impossible to reach coherence in ideas of “identity,” but also when we understand the constructions of working conditions themselves as creating power relations and “identity” among workers, it becomes clear that such questions of who might benefit from changing welfare conditions in the workplace is, at least, partly a circular question. How we perceive our role as workers, as parents and as care providers in society is inevitably influenced by what rights we hold to define these roles. Thus, the question of who will take advantage of new rights in the workplace is not pre-defined but evolves dynamically according to how widespread the right is in a particular culture. Third, and most important, the idea of rotation of centers demonstrates how such questions of costs and “tradeoffs” are problematic. It illuminates the falsity of perceiving certain working conditions as “natural,” global and evident, while other conditions such as care giving, stereotypically an unpaid “feminine” task, are perceived as controversial, particular and marginal. Yet, in fact, all working

152

ORLY LOBEL

conditions could be depicted as burdens on other workers, including conditions as sickness leaves, retirement pensions, injury compensation, and minimum wages. In the struggle for social reform in the workplace there is an inherent value in shifting centers of power among workers themselves, de-centering individualism and utilitarianism as our privileged outlooks of human relations and changing the background rules by denaturalizing the separation between work and family. Childcare benefits serve as an important link between work and family, dissolving false private/public and personality/identity dichotomies. These shifts are precisely what can bring substantive equality among workers, change our conceptions of labor/ownership relations and broaden our imagination regarding the possibilities of workplace reform. A dynamic concept of “workers” as an identity would examine differences in economic and political power by referring not only to the roles individuals take on in their employment capacities but as consumers, parents, and citizens, and by exploring the common interests in the market, in education and in welfare. Finally, critical coalitions, rotation of centers and post-identity dynamics illuminate the need of a revitalized labor movement to develop a macro theory of justice that is not only wider than factional self-interest, but also departs from equality and rights discourse, imagining new structures of the workplace, the workforce and the market at large. The labor movement needs to develop a broader political vision for its struggle, in which collaboration with other social movements is crucial (Klare, 1982).47 Today, within the labor movement leadership there are signs that a new agenda of inclusion across gender and ethnic lines has begun to be the focus of the movement’s strategies (Crain & Matheny, 1999). On the international level, alliances between workers from different countries are also in formation (Ansley, 1992; Ho et al., 1996). A general substantive theory of workplace justice would be a dynamic platform that evolves over time and through continuous engagement. However, in its foundational principles, it can be based on the following three premises. First, workers should have more control over the conditions and policies in their workplace. Second, workers are entitled to a greater share of the material income of their production. Third, the first two goals must be achieved in ways that would advance all workers and not only certain factions. Such a theory thus includes a threefold focus: power, material distribution, and equality. It purposely does not describe ultimate detailed goals but future focal points that are defined and redefined as changes are achieved. Certain issues that currently require particular attention in a framework of rotating centers would include the expansion and enrichment of the relations between work and other political arenas and other aspects of life, including taxation, education, care, health, housing, consumption, and the formation of alliances with associations that have these issues as their agenda. It further suggests the need for greater experimentation

Between Solidarity and Individualism

153

with different forms of workplace democracy, along with the understanding that workers’ participation and unionization are not mutually exclusive.

6. CONCLUSION – FINDING A HOME IN THE WORKADAY WORLD The spirit of extended social democracy is one of radical individualism . . . This idea lends some weight of authority or, at least, inevitability to shared habits of life: in particular, to a life in which the focus of energy remains in the world of family and leisure; the most vanguardist and rebellious forms of culture take on a dream-like character as if they were festivals of the spirit incapable of finding a home in the workaday world (Unger, 1996, p. 143).

It has been argued that the workplace and the labor market in general, by processes of education, mobility and competition, have become the main forces behind the individualization and atomization in societies and in people’s lives (Beck, 1987).48 A third-way mode of collectivity could bridge current dichotomies between individualism and solidarity. Diversity does not mandate separation, because solidarity should not necessarily lead to coercion, just as individualism should not necessarily imply self-interest. An effort to unite without repressing differences is harder than falling back to one of the boxes of these traditional dichotomies. It requires dynamic experimentation, imagination and risk-taking in the difficult realm of the modern workplace. Yet the risks are worth taking since their rewards are human relationships, self-agency, continuity, and transformation. The strength of social life is precisely in the empowering experience of embodying both individual and collective empowerment, which only together can fulfill the promise of freedom.

NOTES 1. Unions represented 34.7% of the workforce in 1954, 16.4% in 1990, and approximately 14% in 1997. Of all wage and salary workers in private industry, 8.5 percent were union members in 2002. In the Public sector, 37.5 percent of government workers were union members. Bureau of Labor Statistics. http://www.bls.gov/opub/ted/2003/ feb/wk4/art03.htm. Although 1998 and 1999 have been marked with renewed labor militancy and strikes, analysts believe that union density will continue to decline in the next century. 2. There are of course many “workplaces” and sectors of labor, some more heterogeneous than others. There are genderized professions (sectors that are traditionally occupied by women workers are at times referred to as “pink collar work”) and jobs that are mostly occupied by racial minorities or immigrants. Moreover, there is a disproportionate number of women and minorities who work in unorganized sectors or in temporary jobs or who are unemployed.

154

ORLY LOBEL

3. Not only are identity categories fluid, but they also tend to change and be renamed over time, often by subgroups in order to escape stigma. 4. For example, women’s disproportionate work in part-time job is closely linked to their social role as “mothers.” 5. For a critique on the idea of immutable characteristics and on the dangers of “immutability,” see also, Halley (1994): “[T]he immutable characteristic rationale is springloaded to harm special minorities: Its hidden assumption that racial discrimination would be morally acceptable if blacks could change the color of their skin leaps into prominence when employers tells black women on their payrolls that they can’t wear braids, or Latino employees that they can’t speak Spanish” (p. 138). 6. For example, requirements of higher education for certain promotions may have a disparate impact on African-Americans. In Griggs v. Duke Power Co. the Court held that a policy requiring high school education for almost all departments in the company was discriminatory under Title VII. 7. In Hardison, the Supreme Court held that TWA did not violate Title VII when it fired the plaintiff for failing to work during his shifts on Friday and Saturday, due to his religious tenets. In general, the courts have held that accommodations required under Title VII for religious belief are limited (Willborn et al., 1998). 8. In general, white collar workers tend to view themselves closer to management than to blue-collar workers and to view union membership as associated with blue-collar work. 9. On the dangers of the “loyalty duty” in the struggle of social movements, see Halley (1998, p. 126): “Here is a dangerous point in intragroup politics. When group members promote a duty of loyalty they implicitly ask internal dissenters to fall silent.” 10. The employer is obliged to bargain exclusively with the union representing the majority of the workers. At times, workers had to pay a membership fee to the union even if they opposed its policies (Hyde, 1993). 11. Hyde (1981) notes, American labor law assumes that “employees have joint economic interests but no joint political interests” (p. 12). Lynd (1984) is critical of mutualism because “Labor solidarity is more than an updated version of the social contract through which each individual undertakes to assist others for the advancement of his or her own interest” (p. 1427). 12. On the process of “opportunity hoarding,” where insiders seek to preserve their advantages for their own use by excluding others, see Tilly (1998). 13. As noted by Hill (1977), “[N]ational labor law was, until 1964, concerned almost exclusively with the regulation of relations between management and organized labor, leaving such ‘third parties’ as blacks and women unprotected from discriminatory employment practices” (p. 26). Additionally, Hill (1987) observed that labor unions “functioned primarily to advance the interests of white workers, to guarantee them privileges in the labor market” (p. 33). 14. The rule of fair representation was further applied to the NLRA, in Ford Motor Co. v. Huffman (1953). 15. Empirically, in a study that reviewed cases regarding the duty of fair representation, it was found that plaintiffs rarely won such actions against their union. From the opinions published from 1977 to 1983, plaintiffs won in less than 5% of the cases (Goldberg, 1985, p. 96). Although this could be interpreted as showing that most cases are simply baseless, in light of the historical and on-going reality of discrimination within labor unions, I believe

Between Solidarity and Individualism

155

these figures may indicate the inefficacy of the doctrine of fair representation to provide a framework of ensuring equality and participation within unions. 16. Even in unions in which men are outnumbered by women, men tend to control leadership positions, particularly at the national level (Crain, 1994, p. 1945). 17. In effect, statistics have shown that only approximately 2% of the nation’s establishments with ten or more employees sponsor day care centers for their workers (Diktaban, 1991). See also, Crain (1994) and Crain (1995). 18. The Court held that the ADA “did not require reassignment in violation of a bona fide seniority system.” 19. Another example that occurred in the 1970s was the emergence of a group called “the Third World Caucus” in the margins of the Amalgamated Transit Union, Local 265. The group was formed by ethnic minorities who felt that their union discriminated against them. Their explicit goal was to train caucus members in union skills. What is interesting is that once the group admitted white men into their training initiatives the minority founders of the Third World Caucus soon felt they were becoming excluded and the group fell apart (Gaspain, 1998). 20. With the decline of labor unions, employee caucuses have become widespread, especially in the high-tech industry, and are not only based on identity. Such groups are formed by active employees in a certain workplace with the goal of influencing their working conditions without being a formal union. Alan Hyde (1993) has argued that labor law should protect such employee caucuses as a fundamental institution of employee voice. See also Geman (1997), proposing that states and cities focus on creating physical-local communities of employee that will focus on job development and work involvement in a certain geographical area. Also, many non-unionized companies have employed employee diversity committees as a response to complaints by minority employees (Note, 1994). 21. The company’s employee magazine “included a drawing that depicts people on various continents making telephone calls. In Africa, a gorilla makes the call.” The Alliance gave AT&T suggestions on how to improve diversity and sensitivity. 22. Nine to Five, or the Working Women’s Association, addresses the needs of women office workers and to enhance their ability to assert and enforce their rights. Although it is allied with the labor movement, it remains an independent organization (Crain, 1995, p. 72). 23. In a telephone interview with Rachel Sullivan, Catalyst Public Relations Liaison (Jan. 7, 1997), Geman (1997) learned that in a recent survey of 60 such groups from Fortune 500 firms, over 60% receive a company budget. 24. On the national level, we also witness the results of conflicts between identity groups and the labor movement. One result has been the formation of departments or divisions within the labor movement and within the national administration of large unions. There are also interest groups on the national level that are affiliated with the labor movement and are not formed around a particular employer. For example, the Coalition of Labor Union Women (CLUW), founded in 1974, strives to promote gender equality from within the labor movement. Other groups affiliated with the AFL-CIO include the Asian Pacific American Labor Alliance (APALA), the Coalition of Black Trade Unionists, Frontlash, and Pride at Work (Gray, 1993, as cited in Crain, 1995; Till-Retz, 1986). Crain argues that separate departments, such as departments dedicated to issues that concern women, known as “Women’s Departments,” while serving an important function in pressuring labor unions

156

ORLY LOBEL

from the inside, have also served to reinforce the idea that women’s needs are separate and marginal, and are left to women’s own responsibility (see also, Kates, 1989, as cited in Crain & Matheny, 1999). CLUW has defined its goals to include not only gender issues at the workplace (affirmative action, child care, pregnancy benefits, pay equity), but also more general political activity (abortion rights, child care policy, domestic violence) (Balser, 1987; Frank & Holcomb, 1990). In Israel, there is a women’s department “Naamat” affiliated with the major national union, the “Histadrut,” and there are often difficult conflicts between them. 25. Marshall (1995) argues that effective workplace desegregation is achieved when coalitions of women and people of color work together toward a common goal. On the national level, black firefighters of different local departments have formed the Black Firefighters Association. See also NAACP v. Seibels (1994), regarding suits alleging discrimination against blacks and women in local government employment in Birmingham. 26. Another example is the prohibition on organized labor struggles to engage in a secondary boycott, whereas identity groups, if not in a union context, can organize wide boycotts. See, for example, NAACP v. Claiborne Hardware Co. (1982), in which a wide economic boycott against white business for demands of racial justice in the community was upheld by the court. 27. Regarding intersection, it is helpful to consider Janet Halley’s critique of intersectionality, suggesting that rather than thinking of intersections of identity as instantiated in persons, we should think of intersectionality as found in discourses (Halley, 1998, pp. 136–137). Later, I suggest a move toward a “post identity” model, and it becomes more beneficial to think about the organizations themselves as intersecting institutions rather than people as intersecting “identities.” In addition, in order to facilitate a complex understanding of “inevitable intersectionality” it is helpful to shift identity categorical discourse from “in/out” or “on/off” categories to categories that encompass all human beings (“women” to “gender,” “homosexuality” to “sexuality”). This shift is only one instrumental step towards the elimination of rigid categories in a post-identity discourse. 28. In a later paper, arguing for an elimination of exclusive representation of unions, allowing multiple representation by identity groups, Crain explains that this call for separatism is not rooted in a fear of coalition. Rather, separatism can serve as a basis for “Rainbow coalitions,” in which differences are not suppressed but voiced, and the coalition is based not on “principles of unit” but on multiple perspectives and experiences (Crain & Matheny, 1999). 29. Even with their dramatic weakening, there is a prevalent public mistrust in the United States of unions (Heckscher, 1988). 30. The divide between class and identity in contemporary politics parallels the divide between redistribution and recognition. On the dichotomy distribution/recognition, see Fraser (1997). 31. In a different context, Kelman and Lester (1997) discuss the success of students with learning disabilities (LD) in rights claiming. Kelman and Lester argue that the claims of students with LD are currently more effective when they are inscribed in antidiscrimination language than as moral claims regarding the merits of the education system in general. Antidiscrimination claims have the quality of appearing as though they are prior to social distribution struggles. Thus, LD is an example in which antidiscrimination rights claiming has effectively united “two ordinarily oppositional political ideological camps. The first is a conservative one which radically distinguishes the moral claims of the ‘deserving poor’ from

Between Solidarity and Individualism

157

the claims of the ‘undeserving poor,’ who must be discouraged from ‘choosing’ poverty by denying their requests for public assistance. The other is what we dub, for lack of a better term, a ‘left- multiculturalist’ ideology, one which radically distinguishes the claims of social groups thought to suffer from irrational devaluation of their capacities from those whose claims are based on either need or solidarity” (pp. 13–14). 32. On different views of morality in relation to the use of rights discourse and alternatives, see generally, Haidt, Koller and Dias (1993). 33. Olsen describes the market/family dichotomy as “a human construct that entails the same kind of self-alienation that Karl Marx described in On the Jewish Question (1843), his classic essay on the state/civil society dichotomy. Marx perceived that human beings lead a ‘double life’ – one life in the state and a separate life in civil society. He referred to the political state, in which we regard ourselves as communal beings, as the ‘species-life’ of mankind; the political ‘citizen’ is abstract and universal. Civil society, however, is the realm of the particular; each member of civil society is a ‘private individual,’ separated from his community and concerned only with his own interests and desires. The dualism between species-life and individual life involves a form of self-alienation, an artificial split of the person into an abstract citizen of the state and an egoistic individual in civil society” (p. 1564). 34. See, for example, Hansmann (1990), arguing that heterogeneity of interests in worker controlled corporations could lead to inefficiency. On the negative effects of fragmentation of organized workers, see Rogers (1990). It is also interesting to notice that similar arguments concerning intragroup battles which advantage the “real” enemy have been raised by NAACP leaders when confronted with opposition of African-Americans opposed to the strategy they chose (Bell, 1976). 35. Describing this “dual labor market theory” in which employers deliberately promote racial and gender divisions within the working class (p. 1575). 36. For a critique of the feminist movement as claiming a universal experience of all women and suppressing the voices of women of color, see Harris (1990) and Spelman (1988). On the problems of litigating rights while facing disputes between members of the gay community concerning marriage, see Rubinstein (1997). On general tensions that occur in class actions, see Garth (1982), Grossberg (1989), Note (1982), Rhode (1982), Resnik et al. (1996), and Tobias (1991). 37. Even within separatist groups of organized labor we witness exclusion of minorities. For example, sex-segregated unions of women waitresses, which have been empowering for white women, have also practiced overt exclusion of women of color until the 1940s and afterwards discrimination from within (Crain, 1995, p. 73). 38. In fact, we could view “the law of work” as encompassing also contract law and corporate law, along with doctrines of many other fields such as constitutional law, torts, property, intellectual property, and welfare laws. The following table relates the legal fields of practice, pedagogy, and scholarship, along with the modes of activity of each, which are central to any reform of the workplace. Activity/Hierarchy Individual Action Collective Action

Capital Contract law Corporate law

Labor Employment law Labor law

Interrogating this framework, it becomes clear that “blind spots” are likely to occur within such divisions of thought. For example, a discussion regarding a model of participatory

158

ORLY LOBEL

democracy would most naturally be situated in the law of corporations, yet the fact that corporate theorists can refer to the existence of employment and labor law, such proposals tend to be marginalized in legal thought. It is important to realize that these internal divisions are neither “natural” nor necessary. For example, in Israel, employment and labor laws are considered one discipline (Lobel, 2001a). Pedagogically, I believe that workplace regulation should be taught in one encompassing course called “the law of work” which would help surpass rigid divisions that narrow our perspectives and prevent us from imagining broader ways to achieve social reform. 39. Even in antidiscrimination suits, class actions have declined and the central focus has become individual employment discrimination (Donohue & Siegelman, 1991). 40. Rather than active protest, workers today tend to express their dissatisfaction by diminishing their commitment and effort at work (Heckscher, 1988). 41. It is interesting that both Minow (1997) and Piore (1995) refer to the philosophy of Hannah Arendt to offer ideas of human interaction but read her in different ways. Stephen Carter (1991) has made a similar call for dialogue that does not involve a hierarchy of ‘authenticity’: “We who are black, rather than establishing a hierarchy of ‘correct’ or ‘more valuable’ black views... should be having a conversation, not a monologue, and certainly not this bitter argument. We should talk to each other rather than at each other. We should make our shared love for our people the center of our belief, and use that shared center as a model for the possibility of solidarity that does not seek to impose a vision of the right way to be black.” The power of dialogue as a mechanism for non-coercive coalition building must also take into account the differences of speech between sub-cultures. For an ample account of differences of speech and their effects on subordination, see White (1991). 42. See also Patchen (1998). 43. See, Iglesias (1993) considering multiple representation by interest groups in a given bargaining unit; proportionate interest representation through cumulative voting, and a network of multiple-interest unions; and McUsic and Selmi (1999) suggesting the creation of diverse voting and representation structures, shunning the majority rule, in order to increase participation of all members in a union. Suggestions regarding the method of voting and representation are important and, similar to the political arena, should be a focus for experimentation. What is important to my model is that such changes will still mandate intragroup discussions prior to actions vis-`a-vis the employer. 44. See also generally, Brown and Gaertner (2001), on the possibility of developing supra-identities that overcome the fragmentation of disparate loyalties without suppressing the commitment to group identity. 45. Section 8(d) of the NLRA (1988) requires employers to “bargain collectively” with employees’ representatives, concerning “wages, hours and other terms and conditions of employment.” Subjects that have been interpreted as falling within this section are regarded as mandatory for bargaining. Section 302 of the Labor Management Relations Act (1947) states that childcare benefits can be bargained over, but cannot be a mandatory subject of collective bargaining (Diktaban, 1991). In effect, statistics have shown that only approximately 2% of the nation’s establishments with ten or more employees sponsor day care centers for their workers (Diktaban, 1991, p. 413). 46. In unions in which the majority of members are women there have been successful initiatives by the union of childcare projects (Pitegoff, 1993).

Between Solidarity and Individualism

159

47. In order to better understand the nature of coalition building, it would be helpful to explore past motivations that have produced moments of solidarity. It is important to contrast the call for critical coalitions with formations of strange alliances in the past between social groups motivated by a common opposition to a third group. Intuitively, we can imagine two modes of solidarity: “solidarity against,” in which the main mutual motivation of the groups or individuals is their resistance to a third group, and “solidarity for,” which is based on a common ideal or “faith.” This is of course a crude division and in most cases these two motivations are very difficult to separate. Yet, it is important to notice the different nature, both in rhetoric and in substance, of coalitions that are dominated mainly by either one of these motivations. We have already discussed the uneasy relationships between the labor movement and both the civil rights movement and the women’s rights movement, and have mentioned the formation of strange alliances in reaction to these tensions. “A backlash against feminism and against racial justice reforms often unites business and labor groups while requiring, paradoxically, both a revived defense of the traditional family as a communal enclave away from competitive individualism and a revitalized individualism, attacking special governmental assistance through welfare and employment rights” (Minow, 1993, p. 36). Such alliances and backlashes, which are based on ad hoc solidarity between those who seem to be “natural opponents” in terms of substantive political ideology, are usually in opposition to the groups’ self-understanding and are created mainly through aversion to others. 48. Roberto Unger (1996) has criticized contemporary thought as being devoid from the workplace as a place that should be the target of efforts for change.

ACKNOWLEDGMENTS For their comments and suggestions, I thank On Amir, David Charny, Janet Halley, Duncan Kennedy, Martha Minow, Celestine Nyamu, Charles Sabel and Lucie White.

REFERENCES Ansley, F. L. (1992). U.S.-Mexico free trade from the bottom: A postcard from the border. Texas Journal of Women and the Law, 1, 193–248. Bales, R. A. (1997). The discord between collective bargaining and individual employment rights: Theoretical origins and a proposed reconciliation. Boston University Law Review, 77, 687–760. Balser, D. (1987). Sisterhood & solidarity: Feminism and labor in modern times. Boston, MA: South End Press. Beck, U. (1987). Beyond status and class: Will there be an individualized class society? In: V. Meja, D. Misgeld & N. Stehr (Eds), Modern German Sociology (pp. 341–355). New York: Columbia University Press.

160

ORLY LOBEL

Bell, D. A. (1976). Serving two masters: Integration ideals and client interests in school desegregation litigation. Yale Law Journal, 85, 470–516. Bernstein, D. E. (1993). Roots of the “underclass”: The decline of laissez-faire jurisprudence and the rise of racist labor legislation. American University Law Review, 43, 85–138. Briggs, S. (1994). Domestic partners and family benefits: An emerging trend. Labor Law Journal, 45, 749–782. Brown, R., & Gaertner, S. L. (2001). Blackwell handbook of social psychology: Intergroup processes (Vol. 3). Malden, MA: Blackwell. Carter, S. L. (1991). Reflections of an affirmative action baby. New York: Basic Books. Chayes, A. (1976). The role of the judge in public law litigation. Harvard Law Review, 89, 1281–1316. Cheit, R. E. (1982). Competing models of fair representation: The perfunctory processing cases. Boston College Law Review, 24, 1–44. Crain, M. (1990). Building solidarity through expansion of NLRA coverage: A blueprint for worker empowerment. Minnesota Law Review, 74, 953–1021. Crain, M. (1994). Between feminism and unionism: Working class women, sex equality, and labor speech. Georgetown Law Journal, 82, 1903–2001. Crain, M. (1995). Women, labor unions, and hostile work environment sexual harassment: The untold story. Texas Journal of Women and the Law, 4, 9–81. Crain, C., & Matheny, K. (1999). Labor’s divided ranks: Privilege and the United Front ideology. Cornell Law Review, 84, 1542–1626. Craver, C. B. (1998). Why labor unions must [and can] survive. University of Pennsylvania Journal of Labor and Employment Law, 1, 15–47. Crenshaw, K., Gotanda, N., Peller, G., & Thomas, K. (Eds) (1995). Critical race theory: The key writings that formed the movement. New York: New Press. Diktaban, C. A. (1991). Employer supported child care as a mandatory subject of collective bargaining. Hofstra Labor Law Journal, 8, 385–415. Donohue, J. J., & Siegelman, P. (1991). The changing nature of employment discrimination litigation. Stanford Law Review, 43, 983–1033. Dubault, R. A. (1995). The ADA and the NLRA: Balancing individual and collective rights. Indiana Law Journal, 70, 1271–1300. Eckles v. Consolidated Rail Corp., 94 F.3d 1041, 1051 (7th Cir. 1996). EEOC v. Mitsubishi Motor Manufacturing of America, Inc., 990 F. Supp. 1059 (C. D. Ill. 1998). Emporium Capwell Co. v. Western Addition Community Organization, 420 U.S. 50 (1975). Estlund, C. L. (1992). What do workers want? Employee interests, public interests and freedom of expression under the National Labor Relations Act. University of Pennsylvania Law Review, 140, 921–1004. Finley, L. (1986). Transcending equality theory: A way out of the maternity and the workplace debate. Columbia Law Review, 86, 1118–1182. Fischl, R. M. (1989). Self, others, and Section 7: Mutualism and protected protest activities under the National Labor Relations Act. Columbia Law Review, 89, 789–865. Foner, P. (1964). History of the labor movement of the United States: The politics and practices of the American Federation of Labor, 1900–1909 (Vol. 4). New York: International Publishers. Ford Motor Co. v. Huffman., 345 U.S. 330 (1953). Frank, M., & Holcomb, D. (1990). Pride at work: Organizing for lesbian and gay rights in unions. New York: Lesbian & Gay Labor Network. Fraser, N. (1997). Justice interruptus: Critical reflections on the ‘postsocialist’ condition. New York: Routledge.

Between Solidarity and Individualism

161

Frug, M. J. (1979). Securing job equality for women: Labor market hostility to working mothers. Boston University Law Review, 59, 55–103. Garth, B. G. (1982). Conflict and dissent in class actions: A suggested perspective. Northwestern University Law Review, 77, 492–535. Gaspain, F. E. (1998). Local union transformation: Analyzing issues of race, gender, class, and democracy. Social Justice, 25, 13–30. Geman, R. (1997). Safeguarding employee rights in a post-union world: A new conception of employee communities. Columbia Journal of Law and Social Problems, 30, 369–405. Gibson, J. D. (1987). Childbearing and childrearing: Feminists and reform. Virginia Law Review, 73, 1145–1182. Goldberg, M. J. (1985). The duty of fair representation: What the courts do in fact. Buffalo Law Review, 34, 89–171. Griggs v. Duke Power Co., 401 U.S. 424 (1971). Grossberg, L. M. (1989). Class actions and client-centered decisionmaking. Syracuse Law Review, 40, 709–788. Haidt, J., Koller, S. H., & Dias, M. (1993). Affect, culture, and morality, or is it wrong to eat your dog? Journal of Personality and Social Psychology, 65, 613–628. Halley, J. E. (1994). Sexual orientation and the politics of biology: A critique of the argument from immutability. Stanford Law Review, 46, 503–568. Halley, J. E. (1998). Gay rights and identity imitation: Issues in the ethics of representation. In: D. Kairys (Ed.), The Politics of Law: A Progressive Critique. New York: Pantheon. Hankinson, K. D. (1998). Navigating between a rock and a hard place: An employer’s obligation to reasonably accommodate the disabled in a unionized workplace. Journal of Contemporary Health Law and Policy, 15, 245–274. Hansmann, H. (1990). When does worker ownership work? ESOPs, Law firms, codetermination, and economic democracy. Yale Law Journal, 99, 1749–1816. Harris, A. P. (1990). Race and essentialism in feminist legal theory. Stanford Law Review, 42, 581–616. Heckscher, C. C. (1988). The new unionism: Employee involvement in the changing corporation. New York: Basic Books. Hill, H. (1977). Black labor and the American legal system. Washington: Bureau of National Affairs. Hill, H. (1987). Race, ethnicity and organized labor: The opposition to affirmative action. New Politics, 1(Winter), 31–82. Ho, L., Powell, C., & Volpp, L. (1996). (Dis)assembling rights of women workers along the global assembly line: Human rights and the garment industry. Harvard Civil Rights-Civil Liberties Law Review, 31, 383–414. Hyde, A. (1981). Economic labor law v. political labor relations: dilemmas for liberal legalism. Texas Law Review, 60, 1–33. Hyde, A. (1993). Employee caucus: A key institution in the emerging system of employment law. Chicago-Kent Law Review, 69, 149–193. Hyde, A., Sheed, F., & Uva, M. D. (1993). After Smyrna: Rights and powers of unions that represent less than a majority. Rutgers Law Review, 45, 637–669. Iglesias, E. M. (1993). Structures of subordination: Women of color at the intersection of Title VII and the NLRA. Not! Harvard Civil Rights-Civil Liberties Law Review, 28, 395–503. Issacharoff, S., & Harris, E. W. (1997). Is age discrimination really age discrimination? The ADEA’s unnatural solution. New York Univeristy Law Review, 72, 780–840. Kalleberg, A. (1995). Part-time work and workers in the United States: Correlates and policy issues. Washington and Lee Law Review, 52, 771–798.

162

ORLY LOBEL

Kamerman, S. B. (1988). Maternity and parenting benefits: An international overview. In: E. F. Zigler & M. Frank (Eds), The Parental Leave Crisis: Toward a National Policy (pp. 235–244). New Haven, CT: Yale University Press. Kay, H. H. (1993). Equality and difference: The case of pregnancy. In: P. Smith (Ed.), Feminist Jurisprudence (pp. 27–47). New York: Oxford University Press. Kelman, M. G. (1996). Progressive vacuums. Stanford Law Review, 48, 975–999. Kelman, M., & Lester, G. (1997). Jumping the queue: An inquiry into the legal treatment of students with learning disabilities. Cambridge, MA: Harvard University Press. Kessler-Harris, A. (1975). Where are the organized women workers? Feminist Studies, 3, 92–110. Klare, K. E. (1978). Judicial deradicalization of the Wagner Act and the origins of modern legal consciousness, 1937–1941. Minnesota Law Review, 62, 265–339. Klare, K. E. (1982). The quest for industrial democracy and the struggle against racism: Perspectives from labor law and civil rights law. Oregon Law Review, 61, 162–215. Lobel, O. (2001a). Agency and coercion in labor and employment relations: Four dimensions of power in shifting patterns of work. University of Pennsylvania Journal of Labor and Employment Law, 4, 121–193. Lobel, O. (2001b). Class and care: The roles of private intermediaries in the in-home care industry in the United States and Israel. Harvard Women’s Law Journal, 24, 89–137. Local Number 93 v. City of Cleveland, 478 U.S. 501, 106 S. Ct. 3063 (1986). Lynd, S. (1984). Communal Rights. Texas Law Review, 62, 1417–1441. Malin, M. H. (1992). The Supreme Court and the duty of fair representation. Harvard Civil rights-Civil Liberties Law Review, 27, 127–186. Marshall, S. I. (1995). Class actions as instruments of change: Reflections on Davis v. City and County of San Francisco. University of Southern Florida Law Review, 29, 911–949. Marx, K. (1843). On the Jewish question. In: D. McLellan (Ed.), Karl Marx (1977, pp. 39–62). Oxford: Oxford University Press. McUsic, M. S., & Selmi, M. (1997). Postmodern unions: Identity politics in the workplace. Iowa Law Review, 82, 1339–1373. Mills, J. K. (1992). Childcare leave: Unequal treatment in the European Economic Community. University of Chicago Legal Forum, 1992, 497–515. Mink, G. (1986). Old labor and new immigrants in American political development: Union, party and state, 1875–1920. Ithaca, NY: Cornell University Press. Minow, M. (1993). Partial justice: Law and minorities. In: A. Sarat & T. E. Kearns (Eds), The Fate of Law (pp. 15–77). Ann Arbor, MI: University of Michigan Press. Minow, M. (1997). Not only for myself: Identity, politics and law. New York: New Press. NAACP v. Claiborne Hardware Co., 458 U.S. 886 (1982). NAACP v. Seibels., 31 F.3d 1548 (11th Cir. 1994). NLRA § 7, 29 U.S. C. §157 (1988). NLRA §§ 2(2)-(3), 29 U.S. C. § 152(2)-(3) (1994). NLRA §§ 8(a)(5), 9(a), 9(c) 29 U.S. C. §§ 158(a)(5), 159(a) 159(c) (1994). NLRB v. Bell Aerospace Co., 416 U.S. 267 (1974). NLRB v. Downslope Industries., 676 F.2d 1114, 1119 (6th Cir. 1982). NLRB v. Peter Cailler Kohler Swiss Chocolates Co., 130 F.2d 503, 506 (2d Cir. 1942). Note (1982). Conflicts in class actions and protection of absent class members. Yale Law Journal, 91, 590–614. Note (1994). Labor-management cooperation after electromation: Implications for workplace diversity. Harvard Law Review, 107, 678–695.

Between Solidarity and Individualism

163

Olsen, F. E. (1983). The family and the market: A study of ideology and legal reform. Harvard Law Review, 96, 1497–1578. O’Melveny, M. K. (1993–1994). The Americans with Disabilities Act and collective bargaining agreements: Reasonable accommodations or irreconcilable conflicts? Kentucky Law Journal, 82, 219–248. Patchen, M. (1998). Diversity and unity: Relations between racial and ethnic groups. Belmont, CA: Wadsworth Publishing. Piore, M. J. (1995) Beyond individualism. Cambridge, MA: Harvard University Press. Pitegoff, P. (1993). Child care enterprise, community development, and work. Georgetown Law Journal, 81, 1897–1943. Resnik, J., Curtis, D. E., & Hensler, D. R. (1996). Individuals within the aggregate: Relationships, representation, and fees. New York University Law Review, 71, 296–401. Rhode, D. L. (1982). Class conflicts in class actions. Stanford Law Review, 34, 1183– 1262. Rogers, J. (1990). Divide and conquer: Further reflections on the distinctive character of American labor laws. Wisconsin Law Review, 1990, 1–147. Rubinstein, W. B. (1997). Divided we litigate: Addressing disputes among group members and lawyers in civil rights campaigns. Yale Law Journal, 106, 1623–1681. Russell, J. (1994). After the fifth sun: Class and race in North America. Englewood Cliffs, NJ: PrenticeHall. Spelman, E. (1988). Inessential woman: Problems of exclusion in feminist thought. Boston: Beacon Press. Steele v. Louisville & Nashville R. R. Co., 323 U.S. 192 (1944). Thompson, D. E., & DiTomaso, N. (1988). Ensuring minority success in corporate management. NY: Plenum Press. Thomsen, M. M. (1996). The role of business and government in the provision of child care assistance: A comparative analysis of the United States and Canada. University of Pennsylvania Journal of International Economic Law, 17, 1209–1254. Till-Retz, R. (1986). Unions in Europe: Increasing women’s participation. Labor Studies Journal, 10, 250. Tilly, C. (1998). Durable inequality. Berkeley, CA: University of California Press. Tobias, C. (1991). Standing to intervene. Wisconsin Law Review, 1991, 415–463. Trans World Airlines v. Hardison, 432 U.S. 63 (1977). Unger, R. M. (1996). What should legal analysis become? New York: Verso Books. Vacca v. Sipes, 386 U.S. 171, 190 (1971). Valdes, F. (1998). Beyond sexual orientation in queer legal theory: Majoritarianism, multidimensionality, and responsibility in social justice scholarship or legal scholars as cultural warriors. Denver University Law Review, 75, 1409–1464. VanderVelde, L. S. (1983). A fair process model for the union’s fair representation duty. Minnesota Law Review, 67, 1079–1164. Van Wezel Stone, K. (1992). The legacy of industrial pluralism: The tension between individual employment rights and the New Deal collective bargaining system. University of Chicago Law Review, 59, 575–644. Weiler, P. C. (1991). Hard times for unions: Challenging times for scholars. University of Chicago Law Review, 58, 1015–1032. Whatley, W. C. (1993). African-American strikebreaking from the Civil War to the New Deal. Social Science History, 17, 525–558.

164

ORLY LOBEL

White, L. E. (1991). Subordination, rhetorical survival skills, Sunday shoes: Notes on the hearing of Mrs. G. In: K. T. Bartlett & R. Kennedy (Eds), Feminist Legal Theory: Readings in Law and Gender (pp. 404–430). Boulder, CO: Westview Press. Willborn, S. L., Schwab, S. J., Burton, Jr., J. F., & Burton, J. F. (1998). Employment law. Charlottesville, VA: Lexis Law Publishing. Williams, W. W. (1984–85). Equality’s riddle: Pregnancy and the equal treatment/special treatment debate. New York University Review of Law and Social Change, 13, 325–380. Wolters, R. (1970). Negroes and the Great Depression: The problem of economic recovery. Westport, CT: Greenwood Publishing. Young, K. N. (October 12, 1995). A League of their own: Disneyland gay-lesbian group has influenced significant change at park. L. A. Times.

HIERARCHY-ENHANCING VERSUS HIERARCHY-LEVELING PERSPECTIVES IN ORGANIZATIONS: VALUING VALUE DIVERSITY IN THE NEW ECONOMY Wayne Eastman ABSTRACT This paper contends that diverse value systems that lead people to sympathize with their own groups in some cases and with other people’s groups in other cases can serve to reduce opportunistic behavior in organizations. In particular, it is useful for an organization to have people who espouse an “innovationist” perspective that supports reduction in hierarchy and economic disparities along with flexibility and change in work conditions. It is also useful to have people who espouse an opposing “traditionalist” perspective that supports hierarchy and inequality along with clear rules and stability in work conditions. Methodologically, this paper applies a political philosophy that assumes the fundamental status of what James Madison et al. in Federalist 10 (1937) called “factions” – group-based divisions between people – to explore the possibilities for democracy in organizational life. Substantively, the paper argues for valuing

Diversity in the Workforce Research in the Sociology of Work, Volume 14, 165–181 © 2004 Published by Elsevier Ltd. ISSN: 0277-2833/doi:10.1016/S0277-2833(04)14007-7

165

166

WAYNE EASTMAN

value diversity in workplaces. Specifically, I contend that value systems I will call “innovationist” and “traditionalist,” both of which lead people to support their own group interests in some cases and to sympathize with other groups in other cases, can be a basis for effective group problem-solving and decision-making in organizations, particularly in high-commitment, new economy organizations. The Madisonian dilemma that frames the paper can be characterized as follows: Group loyalties are useful, but a group that is dominant structurally or in a particular situation always has an incentive to inefficiently take advantage of a weaker group. So how is it possible to benefit from group loyalty and at the same time minimize exploitation? In the Federalist, Madison and his co-authors argued for a system of checks and balances to prevent the dominance of any one faction. Here, I will contend that value-based decision-making in which people are influenced by certain values as well as by their interests can ameliorate inefficient, opportunistic behavior (Williamson, 1985, 1996) caused by interest-based decision-making. (Following Williamson, opportunistic behavior in an organization will be defined here as occurring when a group or individual adheres to its interest in a situation in which it either disregards or lacks credible information as to the greater interest of another individual or group. Such inefficient behavior may also be termed exploitative.) I define the “innovationist” and “traditionalist” value perspectives in relation to group interests. The innovationist perspective sides with worker interests on certain key issues, such as narrowing inequalities and loosening rules and tight hierarchical controls, but sides with managerial interests on certain other key issues, such as an employee’s obligation to participate fully and to internalize the organization’s mission. For its part, the traditionalist perspective sides with managers on certain key issues, such as deference to authority and the upholding of inequality, but with workers on others, such as maintaining traditional work routines and resisting an expectation that the worker must in effect become a manager through wholeheartedly identifying with the organization. The paper claims that in order to be useful organizationally, innovationism and traditionalism need to be understood in a way that makes them capable of appealing across group lines. A group-based organizational politics with workers more or less uniformly on the one side and managers or investors equally uniformly on the other side fails to protect well against group-based opportunism and exploitation, and is not workable or desirable; such a politics is better repressed than surfaced. This aspect of my argument–in favor of having value systems in organizations cut across rather than reinforce group divisions–is indebted to work in political science and sociology on social conflict (Coser, 1964; Horowitz, 1985; Lijphart, 1980; Lipset, 1960; Lipset & Rokkan, 1967) that has noted the value of multiple, cross-cutting cleavages in a society compared to a stark division between groups.

Hierarchy-Enhancing Versus Hierarchy-Leveling Perspectives in Organizations

167

If diversity of values between innovationists and traditionalists is to be surfaced, respected, and valued in organizational life, the innovationist and traditionalist value systems both need to be understood as ways of thinking that appeal to certain worker interests but also to certain managerial and investor interests. A complex interplay of values with group interests allows for more viable, multiple forms of solidarity than a stark opposition of group interests, or an authoritarian denial of such differing interests (Dahrendorf, 1959; Durkheim, 1997). Politics is generally seen as, and often is, a domain of righteousness, indignation, and disputation. The prospects for surfacing innovationist and traditionalist politics in organizations are dim unless such a politics can also be understood as, and actually work as, a vehicle for problem-solving, cooperation, and compromise. I will suggest that both innovationist and traditionalist values, rightly understood, can foster insight into unfairly self-aggrandizing behavior by oneself and others, and help both managers and workers to develop strategies to resolve actual or potential conflicts between them. Instead of simply criticizing a worker who seems recalcitrant to him or her, a manager in a firm in which there is a sympathetic understanding of value diversity could present the situation as one in which neither side can have things all its own way: If the worker doesn’t like traditional rules and supervision, he or she needs to show commitment to an innovationist project of shaping the job. An employee could take the same mediating tack on his or her behalf: If the manager wants him or her to work long hours and come up with new ideas, it doesn’t work well for the manager to also employ nit-picking traditionalist rules. By giving workers and managers an argumentative but also mediating discourse, the innovationist and traditionalist perspectives do not overcome self-interest or guarantee favorable, non-opportunistic outcomes. But by cutting across group-based politics that are present in any case, innovationism and traditionalism can be genuinely helpful in reaching superior outcomes, even assuming, as I do, that the parties involved remain thoroughly self-interested. I will present a simple model aimed at supporting that claim. The essence of the model, which assumes as a starting point that people’s basic preferences are determined by their group interests, is that both innovationism and traditionalism help achieve efficient outcomes by allowing managers and workers to be persuaded to avoid certain opportunistic behavior that they would undertake in the absence of these value systems. The paper’s central idea is that valuing innovationist-traditionalist value diversity would be a sensible strategy to adopt for at least some organizations. As such, I am advocating what could be termed “soft democracy.” In “soft democracy,” the democratic idea of recognizing and respecting different values is applied to organizational life, but not necessarily any of the formal procedures of

168

WAYNE EASTMAN

democracy, in particular decisions made by voting. Other examples of advocacy of what can be called soft democracy, focusing on matters such as fairness or participative consensus rather than employee voting, at least of an adversarial sort, include Eastman and Santoro (2003a), Hsieh (2001), Mansbridge (1982, 1983), and Shapiro (1999). Eastman and Santoro (2003a) claim that valuing value diversity along many dimensions is desirable for a corporation and relate that contention to Donaldson and Dunfee (1999) and their integrative social contracts approach to business ethics; Eastman and Santoro’s argument is related to that of this paper, but focuses on a wide array of diverse values rather than the innovationist-traditionalist dimension with which this paper is concerned. “Hard democracy,” in which organizations select their top leadership (or other positions) through a binding procedure that brings managers and employees as well as investors into the picture, is a considerably trickier issue than soft democracy. Numerous academics and social reformers over the years (e.g. Dahl, 1985; Dewey, 1966; Ellerman, 1990; Mill, 1909; Pateman, 1970) have asserted that the full realization of democracy would be aided by having employees have a stake in deciding the leaders of their organizations. Eastman and Santoro (2003b) present a model of organizational politics, related to that employed in this paper, under which the competing sides or “parties” have an incentive to support efficient rather than opportunistic proposals in order to maximize their support from members of different groups within the organization. But that paper and other work that makes a case for hard democracy should be distinguished from the case made here for valuing value diversity as a form of soft democracy. Hard democracy involves a host of bargaining and other issues that go beyond the “soft democratic” issues of culture and value diversity with which the present paper is concerned.

THE LATENT POLITICS OF HIERARCHY Workplaces do not have a well-articulated politics of different and competing value systems that is comparable to that operating in democratic legislatures. But they do have a powerful latent politics that centers around how one deals with one’s hierarchical position. For the manager, a basic micropolitical choice is between softer and harder treatment of subordinates. For the worker, a basic choice is between resisting one’s subordinate status as unfair in at least some respects and accepting it. Because these choices are wrapped up with so many aspects of day to day life for superiors and subordinates – a professor’s grading policy, a secretary’s

Hierarchy-Enhancing Versus Hierarchy-Leveling Perspectives in Organizations

169

attitude about making the boss coffee, a parent’s decision about whether to spank a child, etc. – the politics of hierarchy thus understood are well-nigh ubiquitous. We constantly have to place ourselves somewhere on the dichotomies between hardness and softness for superiors, or acquiescence and resistance for subordinates, in our workplaces. Even if we believe that the easiness-toughness dichotomy is misleading and unduly limiting (as we should), and even if our ideology of hierarchy is inarticulate or ambivalent or wavering, as most people’s are, we are nevertheless politicians of hierarchy who negotiate the basic opposition between accommodation and toughness on almost as regular a basis as we speak prose (Kennedy, 2004). The opposition between being hard-line and being easy, resonant as it is within one’s hierarchical role, is flawed as a basis for understanding value politics in organizations as a general concept that can bridge hierarchical distinctions. Is innovationism soft or hard? Is traditionalism tough or easy? It depends on whether one takes the superior’s or the subordinate’s point of view. The hard-soft or tough-easy distinction fails to map into a useful definition of innovationism and traditionalism in an organization because it is in the end indistinguishable from a simple group-interest approach in which managerial interests are opposed to worker interests (or different group interests are falsely equated). Within the ranks of managers, the tension between soft control and hard control feels more like a question of what sort of management works best than an issue of redistribution down the hierarchical ladder, with softer control having a human relations, “Theory Y” flavor and harder control having a more Taylorite, “Theory X” quality (McGregor & Bennis, 1985). Similarly, within the ranks of employees, the tension between accommodation and resistance to hierarchy feels like a genuine split over which strategy betters the position of individuals and of the group as a whole, not a disagreement over redistribution. But the vitality and meaning of these tensions presupposes dialogue within rather than between the different groups of an organization. If superiors and subordinates are talking to one another, the contrast between easiness and toughness loses its apparent plausibility as a way of distinguishing between innovationism and traditionalism. The politics of “easiness” and “toughness” is a latent politics for good reason, in that it doesn’t work well when surfaced and related to different groups rather than just one. The point here can be related to the criticism of apparently pro-employee managerial approaches (such as McGregor’s Theory Y perspective on employees as self-motivated) as either managerialist strategies to establish control through softer means (Domhoff, 1990; Kolko, 2000; Perrow, 1986; Weinstein, 1985), or as perspectives useful to managers in questioning their own interests but not useful to employees faced with their own dilemmas and responsibilities.

170

WAYNE EASTMAN

INNOVATIONIST AND TRADITIONALIST VALUES AS BELIEF SYSTEMS THAT CUT ACROSS GROUP INTERESTS I begin with the assumption that both managers and workers tend to see things from the perspectives of their differing hierarchical interests, and that both as rational maximizers have a strong tendency to want it all in their relationship. Here, I use “wanting it all” as shorthand for engaging in suboptimal exploitation of another. In a “wanting it all” mode, managers want employees to defer to them and to follow their rules and codes, and at the same time to be zealously dedicated, creatively inspired participants in the life of the organization. For their part, maximizing workers want managers to treat them as equals and let up on rule enforcement and discipline, and at the same time not pressure them with expectations to participate actively, enthusiastically, and creatively. When the irresistible force of maximizing managers collides with the immovable object of maximizing workers, what happens is unpredictable, but often undesirable for the organization. A group-based politics in which the interests of managers are identified with one value position – “managerialism,” we might call it–and the interests of workers are identified with an opposing value position – “unionism,” we might call it–has little value in reducing the damage from undue self-aggrandizement by both groups. To the contrary, a managerialist politics is apt to make managers feel more righteous about wanting to have it all in regard to workers. For workers, the effects are reversed, with group-based unionism rationalizing employee desires to have it all. Whatever the value of managerialist and unionist ideologies for other purposes, such as collective bargaining, they do not work well in creating beliefs in people that make them sympathetic to interests other than their own. If one assumes that organizational politics are group-based in a straightforward way, with unionist values reinforcing the interests of organizational subordinates and managerialist values reinforcing the interests of organizational superiors, organizational politics is often worth suppressing rather than surfacing. In that case, unionist management becomes an undesirable permissiveness that allows employees to get away with murder, while managerialist management becomes an excessive pursuit of managerial domination. But rightly understood, these are the flaws not of organizational politics of all kinds, but of the group-based organizational politics, and more precisely of the way in which group-based politics reinforces rather than counters understandable though undesirable tendencies by groups in organizations to get away with what they can. If an opposition of values is applied to organizational life without the counterproductive assumption that one side defends the interests of workers while the other side defends the interests of managers, that opposition can be a good way to counter

Hierarchy-Enhancing Versus Hierarchy-Leveling Perspectives in Organizations

171

undue self-aggrandizement by both groups. The innovationist-traditionalist value opposition has that desirable property. Compared to traditionalism, innovationism favors an egalitarian approach toward workplace hierarchy, with fewer honorifics and less deference from workers, and less aggressive discipline from managers. This is more or less in keeping with a group-based sense of pro-worker unionism compared to pro-manager managerialism. But the next, and equally crucial, part of the distinction between innovationism and managerialism goes in a very different direction. Compared to innovationism, traditionalism as an ideology or value system is more accepting of worker passivity. Innovationism is more rigorous in its demands for active, enthusiastic, and creative participation by employees. The flip side of traditionalist insistence on employees’ duty to obey rules and accept their subordination is that, having done so, workers have a zone of free space in which they can disengage from managerial responsibility and active participation. The flip side of innovationist support for equality and openness to reducing hierarchical distance is that employees are pressured to take on more of the burden of possessing correct values and participating in managing themselves and the organization. Thus understood, which value system is more oriented toward advancing the interests of workers becomes highly debatable, as does the parallel question of which value system is more oriented toward advancing managerial interests. More importantly, both the innovationist and traditionalist value systems act as a check against the excesses of worker and managerial self-interest. Managers who are trying to have it all their way with workers are rebuked both by innovationism, with its egalitarian bent, and by traditionalism, with its greater tolerance for employee passivity and noninvolvement. Workers trying to have everything their way with management are similarly rebuked by innovationism, with its pressure on employees to be creative and to participate in organizational life, and by traditionalism, with its demands for deference and acceptance of inequality. Of course, the protection supplied by the combination of innovationist and traditionalist values against opportunistic behavior by workers or managers is not perfect. The traditionalist manager may go overboard with demands for deference, while the innovationist manager may go overboard with expectations for employee participation and creativity. The innovationist worker may go too far with demands for equality unmatched by acceptance of responsibility, and the traditionalist worker may go too far with passive disengagement. But the respective values held by each of these workers or managers militate against exploitation. Excesses in rule enforcement by a manager that result in the manager trying to manage every aspect of the worker’s life and psyche are not in keeping with a traditionalist belief that the subordinate has a separate, distinctive perspective. Mere acquiescence by a manager to a worker’s demands is contrary

172

WAYNE EASTMAN

to an innovationist belief in managers’ striving to have workers internalize correct beliefs. Further, a competition of the different value systems militates against exploitation. Innovationism questions the exploitative errors of traditionalism, while traditionalism seizes upon those of innovationism. Instead of being defined in terms of the toughness-easiness dichotomy or of fidelity to worker or managerial interests, innovationism and traditionalism in organizations are better understood as ways of expressing but also of splitting the interests of workers and managers. I assume that neither value system contravenes the group interests of managers or workers; that is, innovationist managers and workers and traditionalist managers and workers are no less (but also no more) biased in favor of their group interests than apolitical managers and workers. Traditionalist managers tilt more toward managerial interests on certain issues than innovationist managers do, but less so on other issues. Both innovationist and traditionalist managers try to control their subordinates, but the basic strategies of the two are quite different. The traditionalist strategy, in addition to assuming the value of hierarchical distance and deference, assumes, in keeping with theory X (McGregor & Bennis, 1985), that the preferences of the subordinate will incline him or her to unsatisfactory conduct. Such unsatisfactory conduct is deterred by known rules, violation of which is subject to penalties of sufficient magnitude to make it irrational for the subordinate to engage in the conduct. Unsatisfactory conduct is also deterred in the traditionalist control strategy by supplying rewards for identifiable desirable conduct, with the aim of having rewards of sufficient magnitude to make it rational for the subordinate to do the right thing. For the traditionalist manager, the credibility of his or her threats to penalize and promises to reward constitute a central, defining dilemma. If it actually pays the traditionalist manager to penalize or reward, there is no problem. But much of the time, it will not pay to do so. If it is easier under the circumstances not to penalize or not to reward one’s office subordinate, the manager will not do so under the traditionalist’s understanding of human motivation. But if the manager does not carry out his or her threats and promises, the whole edifice of penalties and rewards that makes desirable conduct by the subordinate rational is subject to crashing down. Thus, the effective traditionalist manager is one who defies the belief system’s definition of rationality by remaining committed to threats and promises even when it is not in his or her interest to do so. For all its assertiveness in applying penalties and rewards, the traditionalist control strategy is unassertive when it comes to trying to change the underlying or intrinsic preferences of the subordinate, which in traditionalism are assumed to be substantially contrary to the interests of the institution and/or society. The innovationist control strategy, on the other hand, tries to mold or reveal the

Hierarchy-Enhancing Versus Hierarchy-Leveling Perspectives in Organizations

173

preferences of the subordinate in favor of doing the right thing even if doing so is not apparently in his or her interest. Because the point of innovationist control is to achieve identification by the subordinate with the right values for the institution, the innovationist manager is at least somewhat dubious about the penalties and rewards that are central to the traditionalist vision of control. Extrinsic incentives, even when successful, presume reluctance by the subordinate to do the right thing, and in doing so counter the innovationist vision of an employee imbued with the right intrinsic values (Maslow, 1998; McGregor & Bennis, 1985). From a rational choice perspective, the innovationist manager is faced with an obvious dilemma. If the manager sincerely forgoes any kind of penalty or reward of the subordinate, he or she would seem to be powerless to change the subordinate, and hence reliant on the subordinate’s innate good will and diligence. Just as the traditionalist manager who truly adheres to traditionalism’s view of rational, extrinsic motivation would seem to be ineffective by virtue of his or her inability to carry out threats and promises, the innovationist manager who truly adheres to innovationism’s belief in intrinsic motivation would seem to be ineffective by virtue of his or her opposition to penalties and rewards. The innovationist manager does have a major penalty and reward available, however. Through his or her concern for good character, he or she is trying to establish a reliable reputation with the subordinate as a morally admirable figure whose judgments carry weight. Assuming the manager does so, his or her approval or disapproval of the subordinate is an extremely powerful reward or punishment. It is also a reward or punishment that, unlike many of the traditonalist’s, is credible, given the plausible assumption that people will tend to have a bad will toward a person who acts badly and a good will toward a person who acts well. Although the innovationist manager questions rewards and punishments, his or her authority may in fact control the subordinate more effectively than the traditionalist apparatus of rewards and penalties. Just as managers have politically-fraught strategies of control, employees have similarly fraught strategies of resistance. The innovationist worker’s strategy in a nutshell is to demand equality and challenge hierarchy, while conceding the innovationist manager’s general point (if not agreeing to it in all specific cases) on the need for the subordinate to identify with institutional and social values. The traditionalist worker’s strategy is to resist the innovationist manager’s demand for identification with broader values, while conceding the traditionalist manager’s general case for inequality and hierarchical authority (if not all specific instances of them). Both of these resistance strategies place the subordinate in a position in which he or she is collaborating with as well as resisting the manager’s control efforts. Unlike the dichotomy between toughness and accommodation, which results in completely different behavior from the superior’s and the subordinate’s

174

WAYNE EASTMAN

position, the dichotomy of innovationism and traditionalism looks basically the same whether one is on the top or the bottom. As described here, both the innovationist and traditionalist belief systems are pure or ideal types. In practice, managers and workers will rely on a combination of innovationist and traditionalist claims, rather than exclusively one or the other. But the claim here is that many people in organizations do have a bent toward a traditionalist or an innovationist approach, and that this bent is a useful one, because an optimal level of hierarchy inheres in the creative tension between innovationist and traditionalist visions (of subordinates as well as managers), not in a purportedly apolitical theory of efficient control. Assuming that people often have an interest in doing things that are not in the collective interest, the traditionalist apparatus of rewards and penalties has obvious utility. The utility of the innovationist focus on good character and good will is less obvious, but is also substantial. Persuading a person that doing the right thing is in the interest of the person himself or herself, rightly conceived, is a crucial aspect of trying to make the give-and-take of everyday life go more smoothly. A central problem for human interactions outside the fairly narrow domain in which monetary exchange works well is how to deal with inefficient incentives to get one’s way through having a bad character. Having a bad character can make a variety of threats credible that would not be otherwise; a convincing reputation for malice or envy or anger can be extremely useful for its possessor in certain interactions, even if it is highly counterproductive socially. The innovationist vision of doing the right thing because it is right, rather than because of extrinsic incentives, is a key part of making organizational processes work well. The good will championed by Immanuel Kant is at least as important to organizations as the managerial panopticon described by Jeremy Bentham. What works best, though, is neither innovationism triumphant nor traditionalism triumphant, but a constant, ongoing cooperative competition in which both sides attempt to persuade the cross-pressured and to get adherents of innovationism and traditionalism to live up to their faiths rather than simply adhere to group interests. Problem-solving in the firm works not through dispassionate rationality alone (even if we could somehow strip away bias, there is often deep ambiguity about which interest is more significant and ought to prevail) but through politics (Pfeffer, 1981) – specifically through a politics in which innovationist and traditionalist perspectives both are respected and represented.

VALUING VALUE DIVERSITY A fruitful back and forth between innovationist and traditionalist values may be possible to achieve through a “we believe in diversity” approach. Such an approach

Hierarchy-Enhancing Versus Hierarchy-Leveling Perspectives in Organizations

175

already has a strong scholarly grounding (see, e.g. Kanter, 1977 for a structuralist analysis of minority group behavior in an organization), as well as a presence in many firms through efforts to open up the workplace to members of traditionally underrepresented groups. The argument here for valuing value diversity, although not dependent on arguments for valuing other types of diversity, is a reasonable next step, and is probably more likely to succeed in a firm already committed to a valuing diversity approach. In a firm committed to valuing value diversity, one could hear a line like the following: “Our firm is a better firm when we have innovationists and traditionalists on all of our teams. To help us come up with better ideas and to make sure ideas undergo questioning and scrutiny, it’s important to have both tradition-minded and innovation-minded perspectives represented.” In such an experimentalist firm, belief systems could also be presented as a matter of psychology–some people work best as innovationists and others as traditionalists, just as some people work best as extraverts and others as introverts. One criterion on which managers could be evaluated would be working effectively with value diversity in problem-solving and in the personal styles of employees. In a “we believe in diversity” firm, a subordinate could more readily say something like the following to a superior: “If you want me to put in long hours and come up with all kinds of new ideas on my own, that’s fine. But in that case, it doesn’t work well to yell at me, look over my shoulder constantly to see I’m not day-trading or looking at porn, or get down on me for my messy clothes. I’d prefer the innovative arrangement, where I’m expected to do a lot on my own but things are pretty equal between us. But I’ll work under a traditional arrangement, too. What I don’t want from you is a traditional line on discipline and an innovative line on taking initiative.” Similarly, superiors’ messages to subordinates could more readily be couched in a way that avoids “having it all” approaches with their potential for managerial opportunism. A manager could say to a worker. “I don’t want you to punch a time clock – I prefer a flexible, innovative style. But to make that work, there’s got to be buy-in on your part. You’ve got to convince me that you’re going above and beyond the call in your work.” Another example of a value-based tack a superior could take in trying to deal with a recalcitrant employee: “If you don’t want to be expected to do a lot on your own or to volunteer a lot, I understand. Things can work okay that way. But in that case, as a traditional kind of employee you need to be respectful and diligent about doing your job as I define it. Or if your concern is that you don’t want too much inequality between us, that you don’t want to kowtow to me, and so on, I understand that, too. Things can work that way between us, too. But then as a flexibility-minded kind of employee you need to take initiative and come up with better ways to do things on your own.”

176

WAYNE EASTMAN

As should be clear from the examples, the idea in a pro-value diversity setting is that one is not committed to the righteousness of either innovationism or traditionalism, but to the value of both value systems and their interplay as a means of countering the omnipresent tendencies for both managers and workers to exploit the other. Lest this discussion be seen as a way of trying to enforce a narrow hegemony of moderate innovationism and traditionalism, I should stress that that’s not what I have in mind. There can absolutely be a place for radicals in the pro-value diversity firm. For example, as a radical innovationist employee or manager, one would have a very high level of desire for equality and aversion to coercive management, but also a very high level of willingness to try new things, take initiative, live with ambiguity, and work without clear, predictable assurances of reward. At the same time, the pro-value diversity firm would have its limits in terms of what is acceptable. “Having it all” perspectives on behalf of one group or another would be repressed in the culture of the pro-value diversity firm, because of the close connection between such approaches and group-based exploitation. For purposes of cooperative workplace relations (as opposed to say collective bargaining), one could not openly classify oneself as a unionist in the sense of one who believes that organizations systematically rip off workers, and that workers should both have far more equality relative to managers, take much less initiative, and do far less work. On the other side, one could not openly classify oneself for purposes of cooperative interactions with workers as a managerialist who believes the current balance in organizations tilts too far in favor of subordinates, and that managers currently receive too little compensation and respect and too much responsibility relative to employees. (Or rather, one could espouse any of these group-interest political positions, but without having the protection and support that would be extended in the pro-value diversity firm to those who espouse the organizationally correct group-splitting innovationist and traditionalist lines.)

A MODEL OF HOW GROUP-SPLITTING POLITICS COUNTERS EXPLOITATION The intuitive argument just made about how innovationism and traditonalism can serve as vehicles for one-on-one dialogue that counters group-based opportunism can be formalized into a simple model of decision-making when group interests conflict. (The model here is inspired by game-theoretic modeling (Gintis, 2000; Schelling, 1960) in its use of payoffs, its eschewal of neoclassical assumptions about markets and competition, and its focus on collective rationality, but does not

Hierarchy-Enhancing Versus Hierarchy-Leveling Perspectives in Organizations

177

involve the analysis of strategic interaction, as standard game-theoretic models do. I make the following simplifying assumptions: (1) There are conflicting group interests; (2) A manager’s or worker’s basic payoff from a given position X that he or she can take is equal to his or her group interest; (3) The organization’s payoff from position X is equal to the sum of manager and worker group interests; (4) A manager or worker may be influenced by values such as traditionalism or innovationism in addition to interests. We can now attach some numbers and labels to the assumptions. Suppose a manager has the power to take a position that has a managerial interest payoff of +1 but a worker interest payoff of −3. Although the position is opportunistic from the organizational perspective, it will be taken by the manager. Now assume, though, that that particular position has an ideological payoff of −2 for innovationists and +2 for traditionalists. An innovationist manager will then desist from taking the action, (since the ideological disvalue of the action exceeds its group interest value), resulting in a desirable outcome that would not be achieved absent ideology. Of course, it could be traditionalism rather than innovationism that leads to the desirable action being taken against group interest. Further, neither innovationism nor traditionalism, nor their combination, is a guarantee of an optimal outcome. Group-splitting value systems reduce exploitation, but does not eliminate it. Indeed, some of the time an action that is exploitative will be reinforced by a belief system; in fact, that will happen just as often as a belief system counters an exploitative action. But in that case, the person would perform the exploitative action anyway. A significant reason for the value of diverse values in the model is that they counter interests only when there is a conflict that brings a need for value-based mediation, not in all cases. If a manager or worker simply applied a −2 value-based payoff to a situation in which there is no conflict between the interests of the relevant groups, it would create certain problems. Suppose the group interest payoffs for a given action are +1 for managers and +1 for workers. In this case, it would be unfortunate rather than desirable for the controlling person to apply a −2 value-based payoff to the desirable action and hence avoid taking it. Although the model as presented applies to situations involving some conflict of interests, it does not presuppose zero-sum conflict; rather, the perspective here is one assuming mixed-sum interactions in which there is an optimal outcome for the collectivity. Therefore, the model is concerned with the classic game-theoretic

178

WAYNE EASTMAN

issue (Schelling, 1960) of how to achieve a cooperative, collectively desirable outcome when doing so may not accord with the individual or group-based rationality of decision-makers. Since having value-based payoffs makes it possible for people to be persuaded to take actions that result in higher organizational payoffs, it is a desirable trait from a collective perspective to have value-based preferences. But there is a collective action problem, in that being persuadable when the other side is not reduces the interest-based payoff to one’s group compared to the side that acts simply on behalf of group interests. Thus, a norm of reciprocity and fairness in the organization is important to make value diversity work. To be open to having one’s interests countered by values that give one a sympathy for a member of another group, one wants the member of the other group to be open as well. If managers are to persuade workers of managerial commitment to recognizing and valuing workers’ value-based preferences, they would do well to show that managers’ values can be known and can form a valid basis for worker claims to managers.

AMBIGUITY AS TO WHAT IS INNOVATIONIST OR TRADITIONALIST Ambiguity as to whether a given organizational idea is innovationist or traditionalist is likely to be common. Economic issues in organizations tend to be harder to separate from cultural ones than in the general political arena. Sometimes the distinction is clear, to be sure. A proposal to allow employees to wear what they like is a primarily cultural issue as to which there is an obvious innovationisttraditionalist dimension, while a proposal to widen pay differentials is a primarily economic issue as to which there is a similarly clear value component. But fairly often proposals to change organizational life blend economic and cultural, and also innovationist and traditionalist, themes. Suppose a managerial initiative aims to enhance productivity by reducing bureaucracy and hierarchy, expecting employees to act more flexibly and independently, and moving from a system of providing stable, equal rewards to all reasonably adequate employees to providing higher rewards for employees who respond especially effectively to the more free-form environment. Such a proposal is ambiguous in terms of our categories, in that it is culturally innovationist and economically traditionalist. One could of course try to eliminate the ambiguity by saying that one sphere or another prevails, or that one or another understanding of what it means to be innovative or traditional is wrong. But such efforts to repress ambiguity are likely to be less successful in the organizational sphere than in the general political sphere. An effort to pigeonhole

Hierarchy-Enhancing Versus Hierarchy-Leveling Perspectives in Organizations

179

the reformist management proposal as truly cultural and hence truly innovationist, or truly economic and hence truly traditionalist, is not likely to generate consensus. For value diversity to be workable in an organization, though, it is not necessary that proposals be clearly labeled as innovative or traditional, and in fact ambiguity may well be desirable. The national political system works its way through issues like immigration on which the overall politics are ambiguous. Aspects of such ambiguous issues have clear ideological valences (liberals are more receptive to the multi-racial aspects of immigration, conservatives to the labor cost reducing aspects, and so on). Given that, a critical scrutiny of the other side’s proposals and competition for swing support can come into play on such ambiguous issues. So, too, with the domain of ambiguous issues within an organization. A management proposal for simultaneously loosening up the workplace and allowing high-fliers to fly higher is not at all lacking in value implications. Like the immigration issue, its overall value ambiguity goes along with having clearly identifiable subparts. As such, value-based politics can operate on the management proposal just as they can on an issue like immigration. In an organization, a committed innovationist would push for the ambiguous proposal to be less inegalitarian, or actually egalitarian, and for it to be even more culturally innovationist. A committed traditionalist would push for it to be less culturally innovationist, and for it to be even more economically inegalitarian. In doing so, both of them would try to refine their particular proposals so that they appealed to others not fully committed to their value system. Innovationist advocates would have to deal with the fact that highly paid, culturally innovationist managers may well not be partial to innovationist economic egalitarianism and that economically egalitarian workers may not be receptive to being expected to act more like managers. Traditionalists have similar challenges in getting inegalitarian but also culturally innovationist managers to see the value of traditionalism on corporate culture and in getting traditionalist but also egalitarian workers to support a more inegalitarian pay distribution. The task of innovationist and traditionalist proponents in the organization is thus not only to persuade “centrists” who are in between the positions, but also members of their own side who are lukewarm in their adherence to either the economic or the cultural aspect of innovationist or traditionalist values. A final observation on the ambiguities of innovationism and traditionalism is in order: It needs to be acknowledged that the case made in the paper for valuing value diversity in the form of innovationism and traditionalism is a contextual one. A highly educated workforce may well assimilate the notion of value diversity more readily, as may a workforce already receptive to the idea of valuing diversity in general. Further, innovationist-traditionalist value diversity is more likely to be a feasible notion to introduce in a high-commitment (or,

180

WAYNE EASTMAN

alternatively, a “twenty-first-century” or “new economy”) organization (Ancona et al., 1999; Powell, 2001) with a need to have employees as well as managers engage in innovative practices and identify themselves with the job, as well as a need for more traditional procedures and practices. Just as managers in such high-commitment firms are likely to be more receptive to initiatives in areas such as work/family balance (Osterman, 1995), they may well have a greater incentive to share power and greater openness to value diversity initiatives.

CONCLUSION It can pay off in efficiency terms for people in an organization to have different values as to what kinds of policies most benefit the organization. Specifically, it can be good to have some who support innovation-oriented policies that flatten hierarchies and reduce inequalities while at the same time expecting employees to be flexible and identify with their jobs, and others who support tradition-oriented policies that uphold hierarchy and inequality but that also uphold employee prerogatives and limit demands on employees to identify with the job. While the innovationist position is more oriented toward worker interests in certain respects, such as narrowing wage disparities with management, the traditionalist side is more pro-worker in other respects, such as protecting worker interests in job classifications; the same applies in relation to managerial interests. The benefits from having diverse value systems depend on the way in which they lead people to have sympathy with other groups, not only with their own. Given the positive consequences of an informal interplay in a firm between innovationist and traditionalist value systems, firms – especially “new economy” firms with highly educated workforces–should consider experimenting with programs to encourage managers and employees to value these types of value diversity.

REFERENCES Ancona, D. G., Kochan, T., Van Maanen, J., Scully, M., & Westney, E. (Eds) (1999). Managing for the future: Organizational behavior and procedures (2nd ed.). Cincinnati, OH: South-Western College Publishers. Coser, L. (1964). Functions of social conflict. New York: Free Press. Dahl, R. A. (1985). A preface to economic democracy. Berkeley: University of California Press. Dahrendorf, R. (1959). Class and class conflict in industrial society. Palo Alto, CA: Stanford University Press. Dewey, J. (1966). Democracy and education. New York: Free Press. Domhoff, W. (1990). The power elite and the state: How policy is made in America. New York: Aldine de Gruyter.

Hierarchy-Enhancing Versus Hierarchy-Leveling Perspectives in Organizations

181

Donaldson, T., & Dunfee, T. W. (1999). Ties that bind: A social contracts approach to business ethics. Boston: Harvard Business School Press. Durkheim, E. (1997). The division of labor in society. New York: Free Press. Eastman, W., & Santoro, M. (2003a). The importance of value diversity in corporate life. Business Ethics Quarterly, 13, 433–452. Eastman, W., & Santoro, M. (2003b). Voting for your boss: An economic argument for workplace democracy. University of Pennsylvania Journal of Law and Social Change, 6, 37–54. Ellerman, D. (1990). The worker-owned firm. Boston: HarperCollins. Gintis, H. (2000). Game theory evolving. Princeton, NJ: Princeton University Press. Horowitz, D. L. (1985). Ethnic groups in conflict. Berkeley, CA: University of California Press. Hsieh, N. (2001). Justice and the ownership and governance of firms. Unpublished manuscript. Kanter, R. M. (1977). Men and women of the corporation. New York: Basic Books. Kennedy, D. (2004). Legal education and the reproduction of hierarchy. New York: New York University Press. Kolko, G. (2000). The triumph of conservatism. New York: Times Books. Lijphart, A. (1980). Democracy in plural societies. New Haven, CT: Yale University Press. Lipset, S. M. (1960). Political man. Garden City, NY: Doubleday. Lipset, S. M., & Rokkan, S. (1967). Cleavage structures, party systems, and voter alignments. In: S. M. Lipset & S. Rokkan (Eds), Party Systems and Votoer Alignment (pp. 1–64). New York: Free Press. Madison, J., Jay, J., & Hamilton, A. (1937). The federalist papers. New York: Random House. Mansbridge, J. (1982). Fear of conflict in face to face democracies. In: F. Lindenfeld & J. Rothschild-Whitt (Eds), Workplace Democracy and Social Change. Boston, MA: Porter Sargent Publishers. Mansbridge, J. (1983). Beyond adversary democracy. Chicago: University of Chicago Press. Maslow, A. (1998). Toward a psychology of being (3rd ed.). Hoboken, NJ: Wiley. McGregor, D., & Bennis, W. G. (1985). The human side of enterprise: 25th annniversay printing. New York: McGraw-Hill/Irwin. Mill, J. S. (1909). Principles of political economy. London: Longman’s, Green & Co. Osterman, P. (1995). Work/family programs and the employment relationship. Administrative Science Quarterly, 40, 681–700. Pateman, C. (1970). Participation and democratic theory. Cambridge: Cambridge University Press. Perrow, C. (1986). Complex organizations: A critical essay. New York: McGraw Hill. Pfeffer, J. (1981). Power in organizations. New York: HarperCollins. Powell, W. (2001). The capitalist firm in the twenty-first century: Emerging patterns in western enterprise. In: P. DiMaggio (Ed.), The Twenty-First-Century Firm: Changing Economic Organization in International Perspective. Princeton, NJ: Princeton University Press. Schelling, T. (1960). The strategy of conflict. New York: Oxford University Press. Shapiro, I. (1999). Democratic justice. New Haven: Yale University Press. Weinstein, J. (1985). The corporate ideal in the liberal state, 1900–1918. Boston: Beacon Press. Williamson, O. (1985). The economic institutions of capitalism. New York: Free Press. Williamson, O. (1996). The mechanisms of corporate governance. Oxford: Oxford University Press.

WOMEN, MEN, CAREER AND FAMILY IN THE U.S. YOUNG PHYSICIAN LABOR FORCE Susan W. Hinze ABSTRACT Drawing upon the 5% Public-Use Microdata Sample (PUMS) from the 1990 and 2000 Censuses (with comparisons to the 1980 Census through the work of Uhlenberg & Cooney, 1990), this paper examines the changing characteristics of the U.S. young physician labor force (aged 30–49). Currently, over 45% of medical degrees are earned by women, but gendered work-family patterns persist. Measures examined include income, hourly wages, mean work hours, part-time and overtime work, practice setting, marital status, and children. For a sub-sample of physicians married to physicians, I also examine income and work hour differentials. Close attention is paid to whether a marriage premium and/or a motherhood penalty in wages exists and persists over time. Implications of the documented workforce diversity are discussed for organizations within which physicians are employed. The relatively recent influx of women into professional occupations marks an unprecedented demographic transition. Despite marked progress, gender differences in professional career paths and patterns remain evident: occupational and intra-occupational sex segregation persist; women earn less than men even Diversity in the Workforce Research in the Sociology of Work, Volume 14, 185–217 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0277-2833/doi:10.1016/S0277-2833(04)14008-9

185

186

SUSAN W. HINZE

when controlling for occupation, industry, experience and hours worked; women are less likely than men to be in leadership positions within the workplace; and women more often than men work part-time or reduced hour schedules, leading to cumulative disadvantages in later life (Bickel, 2000; England & Farkas, 1986; Epstein et al., 1999; Harrington, 1993; Lorber, 2000; Padavic & Reskin, 2002; Rapoport et al., 2002). Family lives vary as well, with professional women less likely to marry and have children (Hewlett, 2002). Finally, the newly documented “family wage gap” divides working women without children from working women with children; in short, a wage penalty for motherhood exists (Budig & England, 2001; Crittenden, 2001; Waldfogel, 1998). While the gendered inequalities in the workforce are not new and have been fairly well-documented for professional women, perhaps what is new is the focus on organizational adaptations to the new realities of work and family life. The scholarship on women, men, work and family has shifted from a narrow view of how women’s roles as wives and mothers influence career progression (or lack thereof ) to a broader examination of how organizations within which women and men labor are structured in ways that disadvantage those with caretaking responsibilities (Harrington, 1999; Rapoport et al., 2002; Williams, 2000). Some evidence suggests that when women work like men, they reap similar benefits (e.g. Baker, 1996). Yet the reality is that women’s career paths and patterns differ markedly from men’s. In short, many women (and, increasingly, more men) cannot or will not adapt themselves to the masculine ideal worker model designed for those free of care giving responsibilities. Work patterns of professional men are shifting as well, as more professional men marry professional women. What are the implications of shifting work and family patterns for workplaces “designed for men with stay-at-home wives” (Landers et al., 1997, p.166)? Increasingly, scholars and employers recognize the mismatch between the structure of the workplace (modeled on a typical American worker as a male breadwinner earning a family wage, whose wife was a full-time homemaker) and an increasingly diverse and varied workforce in which the typical worker is apt to be a woman who is part of a dual-career household (Christensen, n.d). The radical shift in the composition of the physician workforce (49% of medical students in 2003 were women, compared with 5% in 1960 (Barzansky & Etzel, 2003; Martin et al., 1988)), combined with the fluid medical environment of shifting alliances and organizational arrangements (Hafferty & Light, 1995) positions medicine as an institution ripe for examination. The occupation of physician is in flux as physicians traverse the dual terrains of a dynamic, rapidly changing medical field and a dynamic, rapidly shifting gender and family system. This paper offers an examination of career and family patterns of young physicians over a 20 year period, from 1980 to 2000, a time period during which

Women, Men, Career and Family in the U.S.

187

the percentage of women increased from 12% of the young physician workforce (aged 30–49) to 32% of the young physician workforce. What has the influx of women into the profession wrought between 1980 and 2000? Have income differentials lessened over time? Do women work the same kinds of hours as men? Are they located in the same kinds of practices? Are women as likely to marry and have children? Have men’s work hours shifted over time with the influx of women and an increase in dual-career marriages? What happens to work and family patterns when physicians marry physicians? I have two broad objectives in this paper. First, I employ U.S. Census data to gain a better understanding of the changing characteristics of the U.S. young physician labor force. Second, the findings presented here are used as a springboard for a discussion of the implications of workforce diversity for organizations within which physicians are employed. After reviewing the literature on physician career and family lives, I detail the sample and methods utilized. Next, the findings are presented in line with career characteristics (specifically, income, work hours and practice setting) and family characteristics (specifically, marital and parental status). A core question frames this section: What are the implications of the work-family patterns for organizational adaptations? I conclude the data presentation with an in-depth look at physicians married to physicians (hereafter referred to as MD2 couples). Again, the focus is on how the patterns unfolding in the young physician workforce challenge organizations that employ physicians. Finally, I conclude with a more general discussion of how the changing demographics of the physician workforce, along with some persistent inequalities, draws our scholarly attention to the necessity of altering current structures within which physicians work.

THE CASE OF PHYSICIANS The larger societal shift from an economy organized around manufacturing to one organized around the service sector means service workers now account for 70% of all workers (Rosenstock & Lee, 2000). Many of those service jobs are in the healthcare industry, where women comprise 80% of the workforce. New realities of the modern workplace include longer hours, compressed workweeks, shift work, reduced job security and an increase in part-time and temporary work (Rosenstock & Lee, 2000). While these realities are more problematic for workers with less education, they still characterize many jobs in the service sector, including that of physician. Drawing on interviews with leaders in the Boston health care industry, Harrington, Bookman et al. (2001) document accumulating

188

SUSAN W. HINZE

pressures for all levels of workers. In particular, they note that work-family issues remain “relatively invisible” in the health care workforce especially given the domination of health care professions by men whose “identity centered heavily on work” (Harrington et al., 2001, p. 6). But increasing numbers of women in the health care workforce, along with increases in the number of dual-career men, mean that medical organizations must begin taking families into account. The following review of the literature will explore, in general, the status of women relative to men in the medical profession. Women physicians are an interesting case, because they have more economic resources than most women workers, but also strong pressures to invest heavily in work (Grant et al., 1990). Have women achieved equity? Following Rapoport et al. (2002), I will use gender equity to refer to the fair allocation of opportunities and constraints for women and men. As they note, the emphasis is on fairness rather than equality “to honor the reality that different life goals and priorities, as well as differing capabilities, shape individuals’ wants and needs at work” (Rapoport et al., 2002, pp. 10–11). While a close look at the income gaps between women and men does not allow for a full consideration of fair allocation of opportunities and constraints, it does provide insight into one particularly persistent form of gender inequality.

THE GENDERED PAY GAP Women physicians continue to earn significantly less than men. According to the American Medical Association (AMA), in 2000 women’s annual unadjusted earnings were $120,000 compared to $195,000 for men, a differential of 62.5% (Kane & Loeblich, 2003). A range of factors help account for this differential, including: medical specialty, practice setting, hours worked, and board certification. In short, women tend to be located in less lucrative specialties and practice settings, work fewer hours and weeks per year, and are less likely to be board certified (Baker, 1986; Bird, 1996; Bobula, 1980; Hinze, 2000; Lorber, 2000; Riska, 2001). Other variables often used for predicting income in studies of gender differences in income include residency length, medical school quality, foreign medical graduate, and productivity measures (patients seen per hour) (Kehrer, 1976; Langwell, 1982; Oshfeldt & Culler, 1986). The focus of most of the empirical studies of gender differences in physician income is on reducing the income gap through the use of control variables. Because the residual effect is often identified as discrimination, accounting for the income gap using human capital variables that reduce the residual reinforces

Women, Men, Career and Family in the U.S.

189

the interpretation that women are not victims of discrimination; rather, their own choices result in reduced incomes. However, “explaining” the gap with numerous controls has proven problematic. Data sets are often limited due to inadequate measures or lack of generalizability, and the findings are inconsistent. Some evidence suggests that if women and men physicians bring the same education, training, and experience to the table, they will be paid the same regardless of gender (e.g. Baker, 1996).1 Other data contradict this conclusion; several studies suggest that even when controlling for a range of human capital and structural variables, the pay gap persists (e.g. Bashaw & Heywood, 2001; Bird, 1996; Ness et al., 2000; Wallace & Weeks, 2002). In other work (Hinze, 2000), I have shown how the inclusion of family context variables, generally ignored in traditional human capital or structurally oriented models, “explain” much of the income gap. Specifically, spouse income, and spouse work hours are important predictors of the wage gap, as are more subjective indicators like sacrifice of work for family. Even so, the elimination of a direct effect of gender on income does not mean that gender discrimination is absent because the indirect effect of gender on income is considerable (Hinze, 2000). In short, discrimination may operate indirectly through a range of factors, such as specialty choice, practice setting, experience and training. To provide an example, if a woman avoids a surgical specialty because she perceives an old boy network hostile to women, and if she chooses a practice setting because masculine work norms are absent and policies are more family friendly, then are her “choices” responsible for her lesser income? An alternative reading is that the structure of medical specialties and practice settings discourage many women from entry. To push the example, if a woman turns down a fellowship (in effect, additional training) because she is “constrained” by her family situation (her spouse cannot relocate for his job), then is she voluntarily choosing a lower income pathway? Chances are that men are not similarly constrained; consider that some studies show approximately half of male physicians are married to stay-at-home spouses (Hinze et al., 1999; Sobecks et al., 1999). In short, while empirically rigorous examinations of the effect of gender on income are necessary for furthering our understanding of women’s equity in the profession, they are not sufficient. It may be tempting to conclude that gender discrimination is absent if the control variables eliminate the direct effect of gender, but such a conclusion ignores the indirect effects and contributes to a narrow view of equity. Gendered assumptions and norms characterize organizations in the ways that “force” choices. Heyman (2000) argues that gender inequalities result from the inability of current societal structures to accommodate the needs of working families. Women often make different “choices” in the workplace because of socially constructed barriers, like inadequate daycare options, domestic demands,

190

SUSAN W. HINZE

and gendered assumptions in the workplace (e.g. good workers can work all the time, anytime). As Williams (2000) notes, the focus on women’s work-family conflict at an individual level (e.g. characteristics associated with higher levels of conflict, or coping strategies) can draw attention away from inflexible organizations that discriminate against those with care giving responsibilities. In other words, as Williams (2000) forcefully states: work-family conflict is gender discrimination.

WORK HOURS According to the AMA, women in 2001 worked 49 hours per week compared to men’s 57 hours per week (Kane & Loeblich, 2003). Male dominated professions are often characterized by overwork. As women increase in numbers, will overwork demands decline? In a study of work norms and professional labor markets, Landers et al. (1997) argue that despite the increasing interest in shorter hour options with women’s entrance into the professions, an economic model of work norms suggests that such accommodations may remain infrequent. In short, “long work hours” become a screen for valuable, yet hard to observe characteristics such as commitment or ambition. This leads to a “rat race” equilibrium in which firms are unable to offer short hours to the growing number of professionals who might want them because the private costs of innovation exceed the social benefits. How prevalent are reduced hour schedules among men and women physicians over time? In particular, how common are reduced hour schedules for young physicians (≤ 35 hours per week) and has the entrance of women reduced overtime schedules (≥ 55 hours per week)?2 What factors are associated with reduced work hours? In general, marriage and parenthood have been found to be predictive of reduced work hours for women, but not for men (Grant et al., 1990; Weisman & Teitelbaum, 1987). In particular, physicians married to physicians “choose” to work fewer hours than physicians not married to physicians (Brotherton & LeBailly, 1993). Evidence suggests that when physicians reduce their work hours voluntarily, the perceived rewards outweigh the perceived costs (Lundgren et al., 2000). McMurray et al. (2002) find that physicians with part-time schedules experience increased feelings of control at work, higher levels of satisfaction and lower stress than those with full-time or over-time schedules. However, despite shorter workweeks, women are more likely than men to report serious problems resulting from overwork (Collins et al., 1997). And some evidence suggests that women physicians with reduced hour schedules report considerable or extreme concern over their career sacrifice (Gareis et al., 2003).

Women, Men, Career and Family in the U.S.

191

PRACTICE SETTING In the 1980s, a dramatic change occurred in the mode of medical production as the economic and administrative control of medicine shifted to forces outside the profession (Riska, 2001). With a rise in the corporate model of medicine, we have witnessed a major shift in the employment patterns of physicians, from self-employed in solo and group practices, to salaried employees in large group practices (Collins et al., 1997; Riska, 2001). In 2001, the AMA reports 61.5% of physicians as self-employed (Kane & Loeblich, 2003). Available evidence reveals that women are more likely to be employees than men, and are more likely to be located in patient-focused and clinically oriented medicine over research-focused and academically oriented medicine (Lorber, 1984, 2000; Martin et al., 1988; Riska, 2001).3 Self-employed physicians (those who have an ownership interest in their practice) earn approximately 33% more than the median income of employed physicians, in part because they have more patient visits and work hours in a typical week of practice, (Kane & Loeblich, 2003). This paper uses a large, nationally representative 5% Census of physicians to examine work characteristics and patterns for young women and men physicians with an eye toward change over time in income, work hours and practice characteristics. In particular, I will examine how the income gap changes between 1980 and 2000. I will explore changes over two decades in average work hours for women and men, and will pay close attention to part-time and over-time work schedules. In addition, I will attempt to answer the following questions: What are the gender differences over time in practice location? Are there within practice setting income differences between women and men? Where are the gaps widest and where are they narrowest and how have those gaps changed over time?

WORK-FAMILY LINKAGES In sociology, there exists a relatively new trend toward exploring how the intersection of family and work affects productivity and career trajectories, (Han & Moen, 1999; Moen, 1998, 2003). Increasingly, scholars are working at understanding the work-family linkages in physician lives (e.g, see Gareis et al., 2003; Grant et al., 1990; Hinze, 2000; Johnson et al., 1992; Lorber, 1982, 1984; Lundgren et al., 2000; Sobecks et al., 1999; Weisman & Teitelbaum, 1987). However, examinations of the work-family interface have been somewhat limited for the following reasons: small sample sizes (e.g. Gareis et al., 2003; Johnson et al., 1991; Lorber, 1982; Lundgren et al., 2000), a single gender

192

SUSAN W. HINZE

focus (e.g. Tesch et al., 1992), a single specialty (Brotherton & LeBailly, 1993; Weisman & Teitelbaum, 1987) or a single type of practice setting (e.g. Carr et al.’s (1998) work on academic physicians). However, as Catherine Henry makes clear in the recent inaugural address of the American Medical Women’s Association (2000), advancing equity for women in the medical profession requires an understanding of the challenges that “Dr. Moms” face as they juggle work and family lives. In this paper, I use Census data on physicians to explore two concepts that link the work and family lives of physicians: the marriage premium and the motherhood penalty. First, we know that, historically, married men earn more than single men, a premium that holds across types of jobs but does not accrue to most women (e.g. see Daniel, 1995)4 or accrues on a much smaller scale (e.g. see Waldfogel, 1997). This premium is estimated to be between 10 and 40% and may account for about 1/3 of the estimated gender based wage discrimination (Korenman & Neumark, 1991). The hourly wage premiums are large and persist with human capital controls. In their study of white men from the National Longitudinal Study of Young Men, Korenman and Neumark (1991) find little or no support for the notion that men are selected into marriage based upon wages, wage growth, or other wage-enhancing characteristics. Rather, they find the premium persists after “productivity” measures are included and that married workers tend to get higher performance ratings which increases promotion chances. Other evidence suggests that never married women earn more than women who are married, divorced or widowed (Goldin & Polachek, 1987; Korenman & Neumark, 1992). Does the marriage premium exist for physicians? Does it hold for women and men? Has it changed over time? Past research suggests that single women more often follow male career patterns than do married women (Mandelbaum, 1978) and Kehrer (1976) finds that married women physicians have higher family incomes and more opportunity to cut back on career, which reduces the possibility of a marriage premium. Second, evidence suggests that a motherhood penalty exists in wages, one not totally accounted for by a range of variables included in regression analyses, including experience (Budig & England, 2001; Crittenden, 2001). Furthermore, while some research suggests the gender gap in earnings has declined in recent years, the “family gap,” particularly for women, has increased (Waldfogel, 1998). Using the National Longitudinal Study of Young Women (NLSYW), Korenman and Neumark (1992) find that experience and tenure in jobs attenuate but do not eliminate the relationship between income and children for women. Using panel data from the NLSYW (pooled data with 30,000 woman-year observations), Waldfogel (1997) finds that the family gap persists when controlling for human capital, unobserved heterogeneity and past part-time status. Furthermore, the

Women, Men, Career and Family in the U.S.

193

penalty increases per child. She concludes that the negative effect on wages is due to either employer perceptions (i.e. discrimination against mothers) or unmeasured employee adjustments (e.g. job changes). Perhaps most thorough to date is the Budig and England (2001) study using 1982–1993 National Longitudinal Survey of Youth data. Results show a wage penalty of 7% per child. About 2/3’s of the penalty remains after controlling for elaborate measures of work experience (among other factors). The authors contend that the unexplained portion of the penalty is due to either discrimination by employers against mothers, or the unmeasured effect of motherhood on productivity. Do we see evidence of a family gap or wage penalty for mothers among physicians? Data gathered from a survey of medical faculty at Harvard indicate that women with children work fewer hours, report lower satisfaction with work, and slower career progress than women without children (Manson, 2000). Does the size of the motherhood penalty vary between all married physicians and physicians married to physicians? The latter category of workers, physicians married to physicians, is on the rise and a closer look at these couples could shed light on work-family linkages. Moen (2003), among others, argues for a scholarly focus on coupled careers. In other words, how do the career paths of husbands influence wives’ careers and vice versa? In this paper, work-family linkages will be explored separately for all physicians and for physicians married to physicians. For all physicians, I use Census data to examine: (1) whether a marriage premium exists for women and men physicians; (2) whether this premium has changed between 1980 and 2000; (3) whether a motherhood penalty on wages exists for women physicians with children compared to those without children; and (4) whether this penalty (if documented) has changed between 1980 and 2000. Because the person records for physicians allow for an examination of physicians married to physicians, I examine whether these unique couples experience either a marriage premium or wage penalties for children. Again, the focus is on how the premiums or penalties (if documented) change over time.

SAMPLE & METHODS This study draws upon the 5% Pubic-Use Microdata Sample (PUMS) from the 1990 and 2000 Censuses, with comparisons to the 1980 Census through the work of Uhlenberg and Cooney (1990). While the AMA has long collected data on physicians, they have not collected data on physicians’ family structure or context. The 1987 and 1991 Survey of Young Physicians (YPS), sponsored by the Robert Wood Johnson Foundation, was designed, in part, to correct the oversight. The

194

SUSAN W. HINZE

advantages of the YPS are that it covers a wide range of topics, including medical specialty, practice setting, hours worked, income, marital and parental status, and other professional and demographic characteristics. However, a few drawbacks exist. First, the YPS only covers physicians in their first five years of practice. Second, comparative data exists only for 1987 and 1991, a relatively short time period during which to compare changes in work-family patterns. Third, the YPS is somewhat limited in scope because while it includes marital and parental status, it does not have information on physician spouses. For these reasons, the Census Bureau data still offers necessary and important information on the work and family lives of young physicians. Using Uhlenberg and Cooney’s (1990) work on the 1980 Census, I can compare data from 1980 to 2000 in order to capture larger changes in the profession as the percentage of women increased dramatically. In addition, by appending spouse files, I can conduct an examination of the rapidly growing category of physicians married to physicians. However, there are three major disadvantages of the Census data. First, the Census does not contain any information on physician specialty and contains only crude information on practice setting. On both counts, the YPS fares better. Second, the Census imposes topcoding on reported earnings, artificially reducing the earnings of the top half of the distribution that are disproportionately male. Third, not all physicians in the 5% Census are in practice. The youngest cohorts, in particular, may still have a large number of physicians in training in medical residency programs or fellowships. Following the selection method of Uhlenberg and Cooney (1990) for comparison’s sake, I obtained person records based upon two criteria: age and occupation. All 30–49 year olds who reported their occupation as physician were selected for this study.5 In addition, I obtained person records of the spouses of physicians married to physicians and appended them to the physician records. Hence, I have information on spouses for all physicians currently married to physicians.6

MEASURES Income The 1990 and 2000 Census ask for wage/salary income in 1989 and 1999, which, following Uhlenberg and Cooney (1990) for the sake of comparability with 1980 data, I combined with self-employment income for 1989 and 1999. I chose to report median income over mean income so that the 1990 and 2000 data would be comparable to Uhlenberg and Cooney’s 1980 data. One problem

Women, Men, Career and Family in the U.S.

195

with using income statistics from Census data is the topcoding of income which, as noted above, artificially reduces the earnings of the top half of the distribution that are disproportionately male. Topcodes for 1980 were $75,000 for income from wages and salary and $75,000 from self-employment; for 1990 the respective topcodes were $140,000 and $90,000 and for 2000, they were $175,000 and $126,000.

Hourly Wages Hourly wages are calculated by dividing yearly income by reported hours worked per week multiplied by weeks worked per year. Hence, this measure takes into account gender differences in hours worked and weeks worked.

Mean Work Hours The Census asks respondents to estimate how many hours they worked per week. Part-Time and Overtime Work Because some researchers (e.g. see Jacobs & Gerson, 1998) argue that mean work hours obscure gendered patterns of part-time and over-time work, I calculated the percent working ≤ 35 hours per week and the percent working ≥ 55 hours per week.

Practice Setting The Census bureau catalogs workers into the following 8 classes: employee of a private, for-profit company, employee of a private not-for-profit organization, local government employee, state government employee, federal government employee, self-employed in own not incorporated business, self employed in own incorporated business and working without pay. I eliminated the working without pay category and combined the three government employee categories into one. While the categories are fairly crude and do not totally tap the diversity of the current medical scene, they do provide a rough overview of the kinds of practices in which physicians are located, especially the important distinction between physicians who are owners vs. employees. Because Uhlenberg and Cooney (1990) did not utilize this variable in their analysis of 1980 Census data, comparisons are limited to 1990 and 2000 data.

196

SUSAN W. HINZE

Marital Status The Census asks respondents to mark one category: now married, currently widowed, currently divorced, currently separated or never married. These categories preclude knowledge about whether respondents are ever divorced or widowed, and whether or not they are remarried.

Children The Census data I obtained is limited to person records. The only variable recording whether or not children are present in the household is presence of children by age category for women with the following codes: (1) with own children less than 6 years of age only; (2) with own children between 6 and 17 years of age only; (3) with own children under 6 years and 6–17 years; (4) no own children under 18 years. While the variable tells us whether children are present for women, it does not give us any detail about whether or not children are present for men. Furthermore, the variable does not provide information on number of children in the household for women or men, a serious problem given that some studies find an inverse relationship between number of children and income (Budig & England, 2001; Waldfogel, 1997). These are serious weaknesses of an analysis of physician person records, and one easily solved by appending household records. All of the above variables were utilized for analysis of all physicians and for MD2 couples. However, the following variables were used specifically for within couples analysis of physicians.

Income Differential I calculated an income differential for each of the 1,549 MD2 couples by subtracting her reported income from his reported income (using the combined wages & salary variable and self-employment income). I then created a variable with three categories: (1) husband earns more than wife; (2) wife earns more than husband; (3) husband and wife report equal earnings.

Mean Work Hours Differential I calculated a work hours differential for each MD2 couple by subtracting her reported weekly work hours from his and created a variable with three categories:

Women, Men, Career and Family in the U.S.

197

(1) husband works more hours per week than wife; (2) wife works more hours per week than husband; (3) husband and wife report working the same number of hours.

Practice Setting Differential I recoded practice setting to obtain a variable with 6 categories: (1) both husband and wife employed in private, for profit organization; (2) both husband and wife employed in private, not for profit organization; (3) both husband and wife reported being government employees; (4) both husband and wife report being self-employed in their own, not incorporated practice; (5) both husband and wife report being self-employed in their own, incorporated practice; and (6) husband and wife located in different practice setting categories. This variable does not tell us whether husbands and wives were located in the same practice, only whether they were located in the same general practice category as defined by the Census. In other words, a husband and wife could both report being an employee of a for-profit organization, but may be employed by two different for-profit organizations.

ANALYSIS Given the limitations of the data, I rely on purely descriptive data for this analysis. I chose not to include a multivariate analysis of income for two reasons. First, the focus of this paper is not to explain the income gap, but rather to paint a broader brushstroke picture of career and family patterns. As argued above, explaining income is necessary but not sufficient for a full understanding of gender equity. Constructing analyses that assume sameness in a range of human capital and structural characteristics for women and men neglects the extent to which women and men have different work and family experiences. Second, the Census does not include specialty location, which is critical for predicting income, and is imprecise in other variables, especially practice setting and number of children. Also, as I have argued elsewhere, spouse information is important for a thorough understanding of income differentials (Hinze, 2000). In short, the value of the U.S. 5% Census data lies in its ability to shed light on the broader changes over time for a large, nationally representative sample of young physicians. After presenting details about the changing composition of the young physician workforce between 1980 and 2000, I examine gender differences in median income, hourly wages, mean work hours, part-time work, over-time work,

198

SUSAN W. HINZE

practice setting, marital status, motherhood status (for women physicians) and the change over time in the percentage of physicians married to another physician. Next, I present median income, part-time work, over-time work, the percentage never married and the percentage with no own children by age cohorts of women physicians. Following Goldin (1997), I reflect on potential age, period and cohort effects. Third, I use aggregate data to examine work and family characteristics of MD2 couples. Finally, because much research testifies to the importance of coupled careers (e.g. see Moen, 2003), I report couples level data for income, hours worked and practice setting. Importantly, for MD2 couples, I also examine income differentials and hours worked differentials by presence of children.

FINDINGS The Physician Workforce In 1980, the 5% PUMS file contained 9,979 physicians aged 30–49 (Uhlenberg & Cooney, 1990), compared to 16,924 in 1990 and 20,417 in 2000.7 The increase in the young physician workforce is consistent with reports from the AMA, taking into account increases in the size and number of medical school classes. In the first row of Table 1 (Gender Composition), note that in 1980, women comprised 12% of physicians aged 30–49, compared to 22% in 1990 and 32% in 2000. The number of men physicians increased 49% between 1980 and 1990, but leveled off to a 6% increase by 2000. For women, there is a 226% increase between 1980 and 1990, and a 71% increase between 1990 and 2000. Overall, we see a 58% increase in the number of men physicians between 1980 and 2000, and a 459% increase in the number of women physicians. Also of interest is the number of physicians married to physicians. In the last row of Table 1, note that in 1980, there were approximately 550 MD2 couples (Uhlenberg & Cooney, 1990), compared to 976 in 1990, and 1,549 in 2000. Although the number of couples is three times as large over the two decades, the percentage change is not as dramatic. In 1980, 11% of physicians aged 30–49 were married to other physicians, compared to 11.6% in 1990 and 15% in 2000. How have physicians fared in terms of income, hours worked, practice setting and family characteristics between 1980 and 2000? Following Table 1, and recognizing that income topcodes changed for each year (as detailed above), we see the gender gap in median income closing. In 1980, women’s reported income (for 35–39 year olds) is 50% of what men earn. By 1990, women earn

1980 (N = 9,979) Men (8,820)

Women (1,159)

Gender Comp. Median income Age 35–39 All 30–49

88%

12%

$60,000 NAb

$30,000 NA

Hourly wages Mean work hrs % Working ≤ 35 hrs wk % Working ≥ 55 hrs wk

NA 55 6 53

NA 41 23 25

Practice setting Employee of private, for profit org.

NA

NA

Employee of private, not-for-profit org. Government employee Self-employed in own, not incorporated practice Self-employed in own incorporated practice

1990 (N = 16,900) %a

50

75

Men (13,116)

Women (3,784)

78%

22%

$102,210 $100,000

$58,000 50,000

$38 55 5 55 $98,950 27.1% $75,000 13.4% $58,000 13.2% $123,119 21.8% $191,928 24.3%

2000 (N = 20,417) %a

Men (13,941)

Women (6,476)

%a

68%

32%

57 50

$125,000 $130,000

$77,000 $73,000

62 56

$25 45 20 33

66 82

$50 56 5 55

$37 47 22 34

74 84

$45,933 33.9% $43,400 19.7% $42,000 17.8% $64,900 19.8% $90,000 8.6%

46

$125,750 37% $114,200 19% $ 87,000 12% $235,000 13% $305,000 20%

$72,000 42% $75,900 23% $57,000 16% $75,000 10% $106,000 10%

57

58 72 53 47

Women, Men, Career and Family in the U.S.

Table 1. Selected Work and Family Characteristics of U.S. Physicians, Aged 30–49.

66 66 32 35

199

200

Table 1. (Continued ) 1980 (N = 9,979) Men (8,820)

Women (1,159)

% Now married

1990 (N = 16,900) %a

2000 (N = 20,417)

Men (13,116)

Women (3,784)

%a

74 $53,000 9 $50,000 17 $43,000 34

48

Men (13,941)

Women (6,476)

83 $140,000 5 $121,000 11 $83,000 NA

74 $76,000 8 $81,400 17 $70,000 33

% Never married

8

18

% No children (aged 40–49 only)

NA

14

86 $110,000 5 $99,000 9 $60,000 NA

Married women Young children No children

NA

NA

NA

$50,000 $46,000

NA

$64,000 $70,000

6 (550)

51 (550)

7 (976)

26 (976)

11 (1,549)

24 (1,549)

% Divorced

% Married to MD (n)

51 72

54 67 84

SUSAN W. HINZE

Note: All 1980 data are derived from the findings reported by Uhlenberg and Cooney (1990); hence, comparisons for 1980 are limited. a For median income, women’s income as percentage of men’s income. For work hours, women’s hours as percentage of men’s hours. b Not Available.

%a

Women, Men, Career and Family in the U.S.

201

57% of men’s salaries, and 62% by 2000. Of course, median incomes are not adjusted for number of hours worked, number of weeks worked or specialty and practice setting. Economist Crittenden (2001) argues that comparing hourly earnings of women and men only compares women who work like men with men – a method that excludes most women. And Bashaw and Heywood (2001) are critical of hourly wage as a measure for physicians who are not typically paid by the hour. However, adjusting for hours worked per week and weeks worked per year, we find that men earned $38 per hour in 1990 and $50 per hour in 2000. Women earned $25 per hour in 1990 and $37 per hour in 2000. The good news is that, when adjusting for hours worked per week and weeks worked per year, women’s income as a percentage of men’s increased considerably, from 66% in 1990 to 74% in 2000. The smaller wage gap for women relative to men when comparing hourly earnings (women earn 74% of what men do in 2000) vs. median income (women earn 56% of what men do) reflects the reduction of women’s work hours relative to men’s. Note that in 1980, men worked approximately 55 hours per week, compared to 55 hours in 1990 and 56 hours in 2000. In contrast, women worked 41 hours per week in 1980, 45 hours per week in 1990 and 47 hours per week in 2000. So the gender gap in hours worked has narrowed considerably between 1980 and 2000, not because men are working less but because women are working more. Of late, much attention has been given to the pattern of professional women working part-time (e.g. Epstein et al., 1999). In 1980, 23% of women reported working ≤ 35 hours per week, compared to 6% of men. Those numbers have changed only slightly by 2000, with 22% of women and 5% of men working ≤ 35 hours per week. More striking is the change over time in women’s overtime work. In 1980, approximately 25% of women report working ≥ 55 hours per week, compared to 53% for men. By 2000, 34% of women worked ≥ 55 hours per week. In general, the part-time work status of women has remained the same over the past 2 decades, but the over-time work status of women has increased. For men, over-time work has stayed fairly consistent, from 53 to 55% working ≥ 55 hours per week from 1980 to 2000. As noted earlier, the Census “class of worker” category does not map perfectly onto the practice settings of physicians. In particular, physicians employed by hospitals, free-standing care centers, medical schools or universities are absorbed into the less precise categories listed in Table 1. However, practice setting reveals two important patterns. First, practice setting employment differs significantly by gender. Women are more likely than men in 1990 and 2000 to be employees and to work for the government; men are more likely to be self-employed. Note that more physicians work as employees of a private, for-profit organization than any other category, and the percentage employed in that category increased markedly

202

SUSAN W. HINZE

for both women and men between 1990 and 2000, from 27 to 37% for men and 34–42% for women. Second, the gap between men’s and women’s income narrowed for employees (of either for profit or not-for-profit organizations), but widened markedly for the self-employed (both in incorporated and not incorporated practices). The latter is particularly striking. In 1990, women self-employed in their own (not incorporated) practice earned 53% of what men did, but by 2000 women earned 32% of what men did. For those women self-employed in their own incorporated practice, the figures shift from earning 47% of what men did to 35% of what men did. In short, women fare better relative to men in terms of income if they are employees rather then self-employed; however, the downside of this observation is that salaries in general are higher for the self-employed. Finally, Table 1 compares marital and parental demographics across decades and between genders. In general, men are significantly more likely to be married, less likely to be divorced, and less likely to be never married than women. Between 1980 and 2000, the never married gap between women and men closed somewhat as men are more likely to be never married (from 8% in 1980 to 11% in 2000) and women are less likely to be never married (from 18% in 1980 to 16% in 2000). Income differences by marital status are somewhat striking. Comparing just the married, divorced and never married categories for men, note the “marriage premium” that accrues to married men in 1990 and 2000. Controlling for age would lessen the differential, but more detailed data by age category (not shown) reveals the same pattern within age category. For example, never married men make 55% of what married men make in 1990 and 59% of what married men make in 2000. Even if we examine men in the 40–49 year old age range, we find the premium holds, although it is not as dramatic. In 2000, 45–49 year old never married men earn 73% of what married men earn ($128,000 vs. $175,000). For 45–49 year old women in 2000, the never married earn 91% of the currently married. In summary, sizeable marriage premiums exist for married men, and smaller premiums accrue to married women. Finally, the Census data personal records only allow us to assess presence and age of children for female physicians. In order to be consistent with the data reported by Uhlenberg and Cooney, we compare the percentage of women aged 40–49 who report having no own children in 1980, 1990 and 2000. In 1980, 14% of women report having no own children, compared to 34% in 1990 and 33% in 2000. In short, women are less likely to have children between 1980 and 2000. An examination of median income for women with and without children in 2000 is revealing. In 2000, married women who report having children ≤ age 6 earn less ($64,000) than married women who report having no own children ($70,000). So there is a slight motherhood penalty, with married women physicians who have

Women, Men, Career and Family in the U.S.

203

young children making 91% of what married women physicians with no children make. Notice that the motherhood penalty does not occur in 1990. Since our measures here are imprecise (we do not have the number of children present in the home), it is likely that the motherhood penalty is underestimated since there is a strong, inverse relationship reported between number of children and earnings (Waldfogel, 1997). In summary, the data reveal a radical shift in physician demographics with a marked increase in the percentage of women who are young physicians, and a less marked increase in the percentage of physicians who are part of an MD2 couple. The gender gap in income is slowly closing, although the gender gap in hours worked is decreasing (largely because women are working more hours). Part-time work patterns for women hold constant between 1980 and 2000, but over-time for women has increased. Between 1980 and 2000, the percent of all physicians working as employees has increased, but gender differences remain, with women more likely to be employees of private organizations and the government. Over time, the income gaps within practice category have narrowed for employees but widened for the self-employed. Finally, the data presented reveal little change over time in never married rates for women, but a slight increase in never married rates for men. Married men receive sizable marriage premiums, while smaller premiums accrue for women. Women are less likely to have their own children in 2000 than in 1980, but the figures may reflect delayed childbearing more than increased childlessness. Finally, for married women, we see a slight motherhood penalty in 2000 (one not present in 1990). What are the implications of these findings for organizations within which physicians work? Persistent income differentials suggest the need for policies that insure equal pay for equal work. Women with similar work schedules and practice settings should earn as much as men. Organizations should examine their own pay scales to insure that married women and/or women with children are not penalized for their marital or parental statuses. However, the data are more convincing of the need for policies that move beyond helping women conform to masculine ideals (Brundtland, 2000). At a minimum, adequate paid leave policies can reduce the cumulative disadvantage of unpaid leaves for women, and may reduce the necessity of part-time schedules. According to Phillips et al. (2000), temporary replacement of leave takers is central to an effective leave policy; otherwise women worried about the effect of their absence on colleagues’ work and took suboptimal leaves. As well, institutional and academic culture can impede leave taking. Given the persistent pattern of part-time work for women, employers should design policies that ensure employees have flexibility in choosing reduced hour schedules. But workers who chose flexible hours or part-time schedules should

204

SUSAN W. HINZE

not be penalized by a reduction of opportunities that could lead to promotion (Rapoport et al., 2002). Landers et al. (1997) suggest the passage of maximum hours laws to help reduce the “rat race” equilibrium induced by organizations with intense work demands. In short, organizations within which physicians work may need to rethink traditional assumptions about ideal workers. Opening up leadership positions to people who understand and perhaps reflect the new physician demographics (i.e. more women and/or physicians with family obligations) might insure policies designed to promote those currently disadvantaged (Landers et al., 1997; More & Greer, 2000). In particular, professional associations (e.g. AMA, American Medical Women’s Association) can play a role in reshaping work norms so that employees who do not fit the “ideal” are not penalized (Henry, 2000). Overall, organizations should provide environments that nurture and foster equity, but do not do so at the expense of effectiveness (Bickel, 2000; Rapoport et al., 2002). While employee status is often associated with the loss of power and autonomy for medical professionals, there is reason to think that the “proletarianization” of medicine might hold advantages for women (Riska, 2001). In particular, organizations with heavily bureaucratic practices can promote open recruitment and standardized salaries. Indeed, the data presented here support Riska’s contention, given that the narrowest income gaps between women and men occur for heavily bureaucratized government employees. However, not all the news from the managed care front is positive. According to Linzer et al. (2002), HMO physicians are generally less satisfied with their jobs and more likely to intend to leave their practices than physicians in other practice settings.

Cohort Comparisons Goldin (1997), among others, demonstrates the value of tracing the experience of cohorts of women over time. One valuable feature of the Census data employed in this analysis is the ability to compare four age cohorts of women (following Uhlenberg and Cooney’s 1980 age breakdown) over a two decade time period. Limitations in the data prevent careful distinctions between age, period and cohort effects, but Table 2 is organized to provide some insights. First, an examination of income suggests all three may be operating, although the data are not sufficiently detailed to clearly separate the effects of age, period and cohort. Note that over time, the income gap between men and women grows smaller for all age women, suggesting a period effect. For example, the 30–34 year old women in 1990 earned 73% of what men earned, but in 2000 the 30–34 year old women

Year

Median Income (Female Income as % of Male Income) 1980

Age cohorts 30–34 NA 35–39 40–44

$30,000 (50) NA

45–49

NA

a Uhlenberg

% Part-Time Work (≤ 35 hrs. per wk)

% Over-Time Work (≥ 55 hrs. per wk)

% Never Married

% With No Own Children Present in the Home

1990

2000

1980

1990

2000

1980

1990

2000

1980

1990

2000

1980

1990

2000

$37,000 (73) $58,000 (57) $65,000 (53) $68,000 (54)

$44,000 (89) $77,000 (62) $90,000 (60) $94,000 (57)

20

16

16

31

39

45

24

26

25

NA

43

55

25

22

23

28

31

32

17

15

16

NA

23

32

24

21

25

20

26

29

14a

10

14

14

25

29

24

20

23

21

24

29

14a

11

11

14

42

37

Women, Men, Career and Family in the U.S.

Table 2. Age Cohorts for U.S. Women Physicians, Aged 30–49.

and Cooney (1990) report 14% of women aged 40–49 as never married.

205

206

SUSAN W. HINZE

earned 89% of what men earned. Women aged 40–44 earned 53% of what men earned in 1990 compared to 60% in 2000. However, age and cohort effects are also likely. Within each time period, women in older cohorts earn more than their younger counterparts, but for older women, their income as a percentage of men’s drops. For women in 1990, the youngest cohort earned 73% of what men did, compared to 54% for the oldest cohort. For women in 2000, the youngest cohort earned 89% of what men earned, compared to 57% for the oldest cohort. So although women’s incomes increase as they age, men’s incomes increase more. Finally, note that women aged 35–39 in 1980 earn 50% of what men earn, but their earnings as a percentage of men’s is 54% ten years later as 45–49 year olds.8 Over time, and with age, the income gap has closed for this cohort of women. Because the sex ratio in the profession has changed so dramatically over time, the differences by age seem more indicative of cohort differences (and changes in the climate during which women entered medicine) than age differences per se.9 A different pattern emerges for women aged 35–39 in 1990. They earned 57% of what men earned, and the income gap stayed the same 10 years later as women aged 45–49 earned 57% of what men earned. Women aged 30–34 in 1990 earned 73% of what men earned, but in 2000, the women aged 40–44 report earning 60% of what men earn. While this is certainly an improvement over the 53% earned by the same age cohort ten years earlier, it seems a surprising widening of the income gap for this cohort in a ten year period. Second, the patterns in part-time work reveal that 30–34 year old women in 1980 were more likely than the 30–34 year old women in 1990 and 2000 to work part-time and less likely to work over-time. Over time, all cohorts of women increase their part-time status and decrease their full-time status. Perhaps the most striking change over time is the percentage of young women working over-time, from 31% in 1980 to 45% in 2000. Finally, the marital and parental status variables by year and age cohort reveal a couple of interesting patterns. While the percentage never married does not change much between 1980 and 2000 for the youngest cohort, the percentage without children does shift. For 30–34 year old women, 43% report having no children in 1990 compared to 55% of 30–34 year old women in 2000. It seems that over time, more young women are delaying childbirth. Examining only older cohorts, note that in 1980, approximately 14%10 of women in the 45–49 year old cohort report having no own children, compared to 42% in 1990 and 37% in 2000. Clearly, more women in the oldest cohort forgo childbearing. In summary, younger women are closer to achieving parity in income, but older cohorts of women are further from parity. Organizations should tend to lifecourse issues to understand the decline in women’s income as a percentage of men’s

Women, Men, Career and Family in the U.S.

207

for older cohorts. Also, the trend in younger women delaying childbirth deserves attention. Women in the early stages of career building (i.e, still in training or newly employed) may feel forced to postpone childbearing until later ages when they have more flexibility in their careers. However, building flexibility into training programs (e.g. McMurray et al., 2002) may insure fairness and promote effectiveness over the long term.

The Case of MD2 s While the first half of this report details available data on the general physician population, the second half takes an in-depth look at physicians married to physicians. Such an examination is important for empirical and theoretical reasons. On the empirical front, it appears that MD2 relationships are on the rise. In a recent survey of over 1,200 physicians (graduated from two Ohio medical schools), researchers report 22% of male physicians and 44% of female physicians were married to other physicians (Sobecks et al., 1999) and many predict a continuing increasing trend in such marriages (e.g. Fletcher & Fletcher, 1993). Furthermore, dual-doctor families are theoretically important for their contribution to our understanding of dual-career families generally. Overall, approximately 11% (N = 550) of young physicians in 1980 report being married to another physician aged 30–49, compared to 15% (N = 1,549) in 200011 (see top row of Table 4). The difference is even more dramatic if we limit the comparison to married physicians only (not available for 1980). Of married physicians, 14% are married to another physician in 1990 compared to 19% in 2000. From Table 1 (last line), note the gender difference in percentage of physicians married to another physician. In 1980, only 6% of young male physicians report being married to a young physician, compared to 51% of young female physicians. By 2000, 11% of young male physicians report being married to a young physician, compared to 24% of young female physicians. So for men, the incidence of dual doctor marriages is on the rise, but for women it has declined. Overall, MD2 marriages are here to stay and will probably continue to increase. The work and family patterns for these couples, reported below, warrant examination by organizations adjusting to the increase in MD2 employees. In Table 3, I examine gender differences in the career and family characteristics of physicians married to physicians; in Table 4, I look more closely at couples level data for MD2 physicians. From Table 3, note that the patterns evident for all physicians (as reported in Table 1) are similar for physicians married to physicians. In general, women physicians married to physicians earn less money, work fewer

208

SUSAN W. HINZE

Table 3. Selected Work and Family Characteristics of MD2 Couples, Aged 30–49. 1990 (N = 1,952) Men (976) Women (976) % of Married physicians in MD2 couple Median income Hourly wages Mean work hours % Working ≤ 35 hrs wk % Working ≥ 55 hrs wk

9%

36%

$98,000 $32 55 4 55

$60,000 $27 44 25 28

Practice setting median income (% in type prac.) Employee of private, for $91,700 $55,000 profit org. (27%) (31%) Employee of private, $82,500 $50,000 not-for-profit org. (18%) (20%) Government employee $63,000 $48,000 (15%) (16%) Self-employed in own, not $123,500 $70,000 incorporated practice (20%) (22%) Self-employed in own $190,102 $98,000 incorporated practice (20%) (11%) % MD2 with children∗ 80 80 Median income With children No children

$78,000 $53,000

$43,000 $40,000

2000 (N =3,098) %a

Men (1,549) Women (1,549) %a 13%

32%

61 84 80

$130,000 $50 57 4 57

$75,000 $38 44 30 28

58 76 77

60

$130,000 (36%) $120,000 (22%) $99,500 (13%) $242,000 (11%) $306,000 (19%) 86

$74,500 (41%) $75,000 (23%) $60,000 (14%) $75,000 (11%) $102,000 (10%) 86

56

$135,000 $100,000

$75,000 $77,000

61 76 57 51

55 75

63 60 31 33

56 77

a The

only children variable in the 5% PUMS person records is “presence of own children in home.” We used this as a proxy for children for men, realizing there are several limitations to this method.

hours, are more likely to work part-time and less likely to work over-time than men physicians. While the patterns here replicate the general patterns for all physicians, there are a few notable comparisons. First, changes over time in median income and hours worked are in the opposite direction for MD2 women compared to all women physicians. Between 1990 and 2000, women’s median income as a percentage of men’s increased for all women physicians (Table 1) from 50 to 56%. But for MD2 women, between 1990 and 2000, women’s median income as a percentage of men’s decreased, from 61 to 58%. Hours worked shows a similar change. For all women physicians, the mean number of hours worked as a percentage of men’s increased between 1990 and 2000 (from 82 to 84%) with hours worked rising for women from 45 per week to 47 per week. But for MD2

Women, Men, Career and Family in the U.S.

209

Table 4. Couples Level Data for Physicians Married to Physicians, Aged 30–49. 1980 (N = 550)

1990 (N = 976)

2000 (N = 1,549)

Couples % of all physicians % of married only Median income differential % Husband earns more % Wife earns more Mean hours worked differential % Husband works more hours per week % Wife works more hours per week % Husbands and wives in same practice setting

11 NA NA 73 16 NA 63 15 NA

12 14 NA NA NA 12 63 17 30

15 19 $40,000 66 23 13 62 17 52

Median income With children income diff. No children income diff.

NA

NA NA

$45,000 $8,500

Mean hours worked With children hours diff. No children hours diff.

NA

14 4

15 3

%

MD2

women, the mean number of hours worked as a percentage of men’s decreased between 1990 and 2000 (from 80 to 77%). The primary reason for the decline is that men’s work hours increased from 55 to 57 hours per week over the course of the decade. Second, the practice setting data for MD2 physicians shows similar patterns to the all physicians data in Table 1. Again, more women and men are employees rather than self-employed workers, and the gender gaps in earnings by practice setting are similar. For example, in 2000, women physician employees in a for-profit organization earned 57% of what men earned; for MD2 women the comparable figure is 56%. As with all physicians, MD2 women earn considerably less than men when self-employed (31% of men’s income for self-employed in own, not incorporate practice, and 33% of men’s income for self-employed in own incorporated practice). Turning to family variables for the aggregate MD2 data in Table 3, note that men with children earned considerably more ($135,000) than women with children ($75,000). Interestingly, men MD2 s earned more with children in the home than without children in the home in 2000 ($135,000 vs. $100,000). In short, men receive a fatherhood premium. The picture for MD2 women is very different. Women MD2 s without children earned more than women with children ($77,000 vs. $75,000).

210

SUSAN W. HINZE

Finally, couples level data12 for MD2 s offer insight to how these couples share work and family characteristics. In Table 4, for each MD2 couple, I calculated a median income differential.13 Uhlenberg and Cooney (1990) report that in 1980, 73% of physician husbands reported earning more than their physician wives, while 16% of wives reported earning more than their physician husbands. By the year 2000, the picture is different. Only 66% of physician husbands (a 10% decline) report earning more than their physician wives, while 23% of physician wives (a 30% increase) report earning more than their physician husbands. Turning to work hours, note the mean difference between husbands and wives in 1990 is 12 hours vs. 13 in 2000. However, here the change over time is minimal. In 1980, Uhlenberg and Cooney (1990) report 63% of husbands report working more hours than their wives, while 15% of wives report working more hours than their husbands. By 2000, 62% of husbands report working more hours than their wives, while 17% of wives report working more hours than their husbands. In sum, while incomes are equalizing somewhat in physician couples, husbands still put in substantially more time per week on the job. In 1990, Uhlenberg and Cooney wrote of their 1980 physicians: “Clearly, the tradition of the husband’s career taking priority over the wife’s is maintained in a large majority of two-physician marriages today” (p. 376). Twenty years later, the statement still holds. For all MD2 couples, I calculated a joint practice setting variable. Approximately 30% of MD2 couples are employed in the same practice type in 1990, compared to 52% in 2000. (The data do not allow us to examine whether they are the same practices, only that they fall into the same practice setting type as defined by the Census.) Perhaps the most interesting findings in Table 4 are located in the bottom two rows. Using the presence and age of children variable from person records for women, I calculated gender income differentials for those MD2 couples with and without children in the home.14 This measure is rather imprecise, because we do not know number of children, only whether there are or are not wife’s children in the home. However, the findings are striking and suggest strong support for a motherhood penalty for women physicians. For physicians married to physicians with children, the income differential between women and men in 2000 is $45,000. For physicians married to physicians without children, the income differential between women and men is $8,500. The hours worked differential is, not surprisingly, just as dramatic (and accounts for some of the income differential). For physicians married to physicians with children, the hours worked differential between women and men in 2000 is 15 hours. For physicians married to physicians without children, the hours worked differential between women and men is three hours. In short, children matter in medical marriages.

Women, Men, Career and Family in the U.S.

211

In general, the patterns reported for MD2 parallel the patterns reported for all physicians, except that women MD2 s earned less as a percentage of men’s income between 1980–2000 than did all women physicians. As well, MD2 men increased their work hours over the same period relative to all male physicians. The aggregate MD2 data reveal a fatherhood premium and a (slight) motherhood penalty. So even in dual-career medical couples, men benefit economically from fatherhood while women are economically disadvantaged. The couples level data reveal that between 1980–2000, earnings between husbands and wives are equalizing, although men still earn more in most MD2 relationships. A major change between 1990 and 2000 is the percentage of husbands and wives employed in the same practice setting categories. Finally, the income differentials and hours worked differentials are somewhat dramatic for couples with and without children. There are several implications of the MD2 findings for the physician labor force. First, couples level data reveal the importance of tending to coupled careers. Perhaps organizations that employ physicians could provide on-site assistance for those in medical marriages who are juggling the dual demands of medical careers and families. In particular, policies and programs focused on men as well as women will make clear that the issues are not “limited” to women, and may help challenge the more traditional patterns that unfold over the lifecourse for physicians in medical marriages. Second, closer attention to the cultural expectations and values of the workplace is in order. Risman (1998) finds that women sacrifice work for family in egalitarian relationships not because men make more or women have different personality structures, but because cultural expectations and values within the workplace and within families operate in gender stereotypic ways that, in turn, affect income. “Doing gender” involves, to quote West and Zimmerman (1987), a complex of socially guided perceptual, interactional, and micropolitical activities that cast particular pursuits as expressions of masculine and feminine ‘natures’ (p.126).

Although individuals “do gender,” the focus for West and Zimmerman is on interactional and institutional arenas because gender is conceived of as an emergent feature of social situations: both as an outcome and a rationale for various social arrangements and as a means of legitimating one of the most fundamental divisions of society (p. 126).

The culture of medical work environments may pressure women and men physicians to “do gender.” Hence, beyond programs that help individuals and couples cope are programs that challenge the gendered assumptions of the organizations within which physicians work. Kochan (2002) suggests work redesigns that engage people in a collaborative process of adapting systems, processes and

212

SUSAN W. HINZE

schedules in ways that improve both work and organization performance, and personal and family life.

CONCLUSION In her work on the changing composition of the corporate workforce, sociologist Rosabeth Moss Kanter (1977) demonstrates the near impossibility of token women achieving equity. In her formulation, any population that represents less than 1/3 of a group’s members is a “tilted” ratio with little power to affect decisions and policies within an organization. In the young physicians workforce (age 30–49), women represent approximately 32% of physicians – not quite the 40% Kanter views as approximate parity – but enough to force attention to the mismatch between the current structure of medicine as a “greedy institution” in which an ideal worker devotes all his time and energy to the job (Coser, 1974), and an increasingly diverse physician workforce in which a typical medical worker is apt to be a woman who is part of a dual-physician marriage. An exploration of the gendered work and family patterns of young physicians for 1980–2000 provides information valuable for workplaces concerned about diversity. Fourteen years ago, Uhlenberg and Cooney (1990) concluded that with more doctors (namely, women), working fewer hours, the profession might need to adjust by training more physicians. They also suggest that gender equity could come about in two ways. First, women’s careers could become more similar to men’s; or, second, men physicians could reduce their career involvements and become more involved at home. Both “paths” suggest changes at the level of individuals. However, much about the patterns examined in this paper testify to the need for changes at the level of the organization. The role of the physician, especially in male-dominated specialties, is gendered (see Hinze (1999) for more on the gendered nature of the medical specialty hierarchy), and the structure of medical institutions valorizes masculine traits and is not conducive to the work of nurturing (Lenhart, 1998). Williams (2000) argues that one of the first steps toward gender equity is the dismantling of masculine norms in the workplace: Otherwise women will be further disadvantaged when they are treated the same as men in the face of norms that favor men because they are designed around men’s bodies or life patterns. Sometimes the dismantling of masculine norms can be accomplished simply by changing a single institution, such as the ideal-worker norm. In other situations, neutralizing the effect of masculine norms requires affirmative action, not as reparation for past wrongs but as counter balance to women’s current disadvantage (p. 207).

Indeed, closer attention to gendered assumptions, cultural expectations and values in the workforce is clearly warranted (Rapoport et al., 2002; Risman,

Women, Men, Career and Family in the U.S.

213

1998; West & Zimmerman, 1987). Furthermore, discrimination as a concept should be broadened to include anything about how jobs are structured or what is rewarded that has a disparate impact on women and, in particular, mothers (Budig & England, 2001). Persistent income differentials and wage penalties for motherhood do not bode well for the future of the profession. In addition, recent evidence suggests that women and minority physicians are more likely to have second thoughts about a career in medicine (Hadley et al., 1992); as well, rates of burnout are higher for women physicians than men (Linzer et al., 2002). Given increased numbers of women in the profession, workforce initiatives that help retain women and promote equity will result in the best, most efficient physician workforce.

NOTES 1. Bashaw and Heyword (2001) argue that Baker’s results are distorted by the construction of an hourly wage for physicians who are not typically paid by the hour. 2. It is important to note that organizations benefit from recruiting and retaining women willing to work reduced hour schedules because they have more control over part-time workers. (See Jacobs & Gerson, 2000, for an extended argument on the benefits to employers of a bifurcated workforce in which workers are either over- or underworked.) 3. Lorber (2000) argues that increasing regulation and government control of medicine since 1945 have resulted in decreasing authority for physicians, making the occupation less desirable to white men. In short, as the prestige of physicians declines, the occupation becomes feminized. Some would argue that managed care positions are less prestigious and hence more likely to be occupied by women and people of color. 4. Daniel (1995) finds a marriage premium for white and black men, and for black women. He argues that white men tend not to augment the productivity of their wives, but black men do. 5. Uhlenberg and Cooney (1990) report selecting respondents based upon physician as occupation and at least 3 years of graduate education. The 1990 and 2000 education options for the 5% Pubic-Use Microdata Sample allow selection based upon an education variable with 16 categories, with 13 and above indicating a Bachelor’s, Master’s Professional degree or Doctorate. Most of our physicians reported receiving a professional degree (86%), but an additional 9% reported receiving a Doctorate. The remaining 4% reported receiving a Master’s (1.6%) or Bachelor’s (2.4%). 6. Appending 5% PUMS Household Records for all physicians proved impossible with my current resources. 7. The 5% PUMS full physician data set (with all ages of physicians, not just 30–49) contained approximately 30,000 physicians in 1990 and 36,000 in 2000, which translates into 600,000 in 1990 and 726,000 in 2000. The AMA reported 615,421 physicians in 1990 and 813,869 in 2000. Since there are fewer physicians in the Census data than reported by the AMA data file, the Census clearly undercounts physicians which may skew the results.

214

SUSAN W. HINZE

8. Of course, the 45–49 year olds in 1990 are not the exact women studied in 1980 as older women may have finished medical school and entered this age cohort. 9. I am indebted to Elsevier reviewer who offered the suggestion to consider age, period and cohort effects. 10. Uhlenberg and Cooney (1990) gave a 14% figure for all women aged 40–49, so 14% for age 45–49 is an estimate. 11. The percentage of physicians married to physicians is undoubtedly an undercount given our age limitations. We only have MD2 couples that were BOTH in the 30–49 age range. If physicians aged 30–49 were married to other physicians younger than 30 or older than 49, they would not have been selected into our sample. 12. Thanks to the Elsevier reviewers who suggested a closer examination of couples level data. 13. Income differentials for 1990 were problematic and are not reported here. 14. This technique would underestimate children in the home since presence and age of children for women is used as a proxy for men. If men had their own children (reported as stepchildren by wives), they do not show up on this calculation.

ACKNOWLEDGMENTS A version of this article was first presented, with Richard Carpiano, at the SASE annual meeting, London School of Economics: July 7–10, 2000. Thank-you to Dawn Aliberti and Richard Carpiano for assistance with the data analysis, and to the Elsevier reviewers and editors for their many helpful comments.

REFERENCES Baker, L. (1996). Differences in earnings between male and female physicians. New England Journal of Medicine, 334, 960–964. Barzansky, B., & Etzel, S. I. (2003). Educational programs in U.S. medical schools, 2002–2003. JAMA: The Journal of the American Medical Association, 290, 1190–1196. Bashaw, D. J., & Heywood, J. S. (2001). The gender earnings gap for U.S. physicians: Has equality been achieved? Labour, 15, 371–391. Bickel, J. (2000). Women in academic medicine. Journal of the American Medical Women’s Association, 55, 10–19. Bird, C. E. (1996). An analysis of gender differences in income among dentists, physicians, and veterinarians in 1987. Research in the Sociology of Health Care, 13A, 31–61. Bobula, J. (1980). Work patterns, practice characteristics, and incomes of male and female physicians. Journal of Medical Education, 55, 826–833. Brotherton, S. E., & LeBailly, S. A. (1993). The effect of family on the work lives of married physicians: What if the spouse is a physician, too? Journal of the American Medical Women’s Association, 48, 175–181. Brundtland, G. H. (2000). The woman physician in the year 2000. Journal of the American Medical Women’s Association, 55, 3–19.

Women, Men, Career and Family in the U.S.

215

Budig, J. M., & England, P. (2001). The wage penalty for motherhood. American Sociological Review, 66, 204–225. Carr, P. L. et al. (1998). Relation of family responsibilities and gender to the productivity and career satisfaction of medical faculty. Annals of Internal Medicine, 129, 532–538. Christensen, K. (n.d.). Workplace, workforce and working families. Retrieved November, 2003 from the Alfred P. Sloan Foundation Web site: http://www.sloan.org/programs/stndrd dualcareer.shtml. Collins, K. S., Schoen, C. A., & Khoransanizadeh, F. (1997). Practice satisfaction and experiences of women physicians in an era of managed care. Journal of the American Medical Women’s Association, 52, 52–64. Coser, L. (1974). Greedy institutions: Patterns of undivided commitment. New York: Free Press. Crittenden, A. (2001). The price of motherhood: Why the most important job in the world is still the least valued. New York: Metropolitan Books. Daniel, K. (1995). The marriage premium. In: M. Tommasi & K. Ierulli (Eds), The New Economics of Human Behavior (pp. 113–125). New York: Cambridge University Press. England, P., & Farkas, G. (1986). Households, employment and gender: A social, economic and demographic view. Hawthorne, NY: Aldine de Gruyter. Epstein, C. F., Seron, C., Oglensky, B., & Saut´e, R. (1999). The part-time paradox: Time norms, professional life, family and gender. New York: Routledge. Fletcher, R., & Fletcher, S. (1993). Here come the couples. Annals of Internal Medicine, 119(7), 628–630. Gareis, K. C., Barnett, R. C., & Brennan, R. T. (2003). Individual and crossover effects of work schedule fit: A within-couple analysis. Journal of Marriage and the Family, 65, 1041–1054. Goldin, C. (1997). Career and family: College women look to the past. In: F. D. Blau & R. G. Ehrenberg (Eds), Gender and Family Issues in the Workplace (pp. 20–65). New York: Russell Sage Foundation. Goldin, C., & Polachek, S. (1987). Residual differences by sex: Perspectives on the gender gap in earnings. American Economic Review, 77, 143–151. Grant, L., Simpson, L. A., Rong, X. L., & Peters-Golden, H. (1990). Gender, parenthood, and work hours of physicians. Journal of Marriage and the Family, 52, 39–49. Hadley, J., Cantor, J. C., Willke, R. J., Feder, J., & Cohen, A. B. (1992). Young physicians most and least likely to have second thoughts about a career in medicine. Academic Medicine, 67, 180–190. Hafferty, F. W., & Light, D. W. (1995). Professional dynamics and the changing nature of medical work. Journal of Health and Social Behavior, 35, Extra Issue: Forty Years of Medical Sociology: The State of the Art and Directions for the Future, 132–153. Han, S., & Moen, P. (1999). Work and family over time: A life course approach. In: M. Pitt-Catsouphes & B. K. Gogins (Eds), The Annals of the American Academy of Political and Social Science, 562, 98–110. Harrington, M. (1993). Women lawyers: Rewriting the rules. New York: Penguin Books. Harrington, M. (1999). Care and equality: Inventing a new family politics. New York: Knopf. Harrington, M., Bookman, A., Bailyn, L., & Kochan, T. (2001). Workforce issues in the greater Boston health care industry: Implications for work and family (Working Paper #WPC 0001). Cambridge, MA: MIT Workplace Center. Henry, C. (2000). AMWA Inaugural Address: Finding the balance. Journal of the American Medical Women’s Association, 55, 43–45. Hewlett, S. A. (2002). Creating a life: Professional women and the quest for children. New York: Talk Miramax Books.

216

SUSAN W. HINZE

Heyman, J. (2000). The widening gap: Why America’s working families are in jeopardy – and what can be done about it. New York: Basic Books. Hinze, S. W. (1999). Gender and the body of medicine or at least some body parts: (Re)constructing the prestige hierarchy of medical specialties. The Sociological Quarterly, 40, 217–239. Hinze, S. W. (2000). Inside medical marriages: The effect of gender on income. Work and Occupations, 27, 464–499. Hinze, S., Chirayath, H. T., Schaefer M. L., Chren, M. M., & Landefeld, C. S. (1999, August). The effect of gender and marital (a)symmetry on the work and family lives of young physicians. Paper presented at the 94th annual meeting of the American Sociological Association, Chicago, IL. Jacobs, J., & Gerson, K. (1998, November). Toward a family friendly, gender-equitable work week. Paper presented at Work and Family: Today’s Realities, Tomorrow’s Visions, Boston, MA. Jacobs, J., & Gerson, K. (2000). Do Americans feel overworked? Comparing ideal and actual working time. In: T. Parcels & D. Cornfield (Eds), Work and Family: Research Informing Policy (pp. 71–95). Thousand Oaks, CA: Sage. Johnson, C. A., Johnson, B. E., & Liese, B. S. (1991). Dual-doctor marriages: The British experience. Journal of the American Medical Women’s Association, 46, 155–159, 163. Johnson, B. E., Johnson, C. A., & Liese, B. S. (1992). Dual-doctor marriages: Career development. Family Medicine, 24, 205–208. Kane, C. K., & Loeblich, H. (2003). Physician income: The decade in review. In: Physician Socioeconomic Statistics (pp. 5–11). Chicago: AMA, Center for Health Policy Research. Kanter, R. M. (1977). Some effects of proportions on group life: Skewed sex ratios and responses to token women. American Journal of Sociology, 82, 965–990. Kehrer, B. H. (1976). Factors affecting the incomes of men and women physicians: An exploratory analysis. Journal of Human Resources, 11, 526–545. Kochan, T. A. (2002). An employment policy agenda for working families (Working Paper #WPC0003). Cambridge, MA: MIT Institute for Work & Employment Research and the MIT Workplace Center. Korenman, S., & Neumark, D. (1991). Does marriage really make men more productive? The Journal of Human Resources, 26, 282–307. Korenman, S., & Neumark, D. (1992). Marriage, motherhood and wages. The Journal of Human Resources, 27, 233–256. Landers, R. M., Rebitzer, J. B., & Taylor, L. J. (1997). Work norms and professional labor markets. In: F. D. Blau & R. G. Ehrenberg (Eds), Gender and Family Issues in the Workplace (pp. 168–205). New York: Russell Sage Foundation. Langwell, K. M. (1982). Factors affecting the incomes of men and women physicians: Further explorations. The Journal of Human Resources, 17, 261–275. Lenhart, S. (1998, November). Work and family issues facing female physicians. Paper presented at Work and Family: Today’s Realities, Tomorrow’s Visions, Boston, MA. Linzer, M. et al. (2002). Sex differences in physician burnout in the United States and Netherlands. Journal of the American Medical Women’s Association, 57, 191–193. Lorber, J. (1982). How physician spouses influence each other’s careers. Journal of the American Medical Women’s Association, 42, 149–152. Lorber, J. (1984). Women physicians: Careers, status and power. New York: Tavistock Publications. Lorber, J. (2000). What impact have women physicians had on women’s health? Journal of the American Medical Women’s Association, 55, 13–15. Lundgren, L., Ferguson, F., & Fitzgerald, T. (2000). Practice satisfaction of women physicians working reduced hours schedules. Journal of the American Medical Women’s Association, 55, 40.

Women, Men, Career and Family in the U.S.

217

Mandelbaum, D. R. (1978). Women in medicine. SIGNS: Journal of Women in Culture and Society, 4, 136–145. Manson, J. (2000). Progress note: Sex differences in career progress and satisfaction in an academic medical center. Journal of the American Medical Women’s Association, 57, 194. Martin, S., Arnold, R., & Parker, R. M. (1988). Gender and medical socialization. Journal of Health and Social Behavior, 29, 333–343. McMurray, J. E. et al. (2002). Women in medicine: A four-nation comparison. Journal of the American Medical Women’s Association, 57, 185–190. Moen, P. (1998). Women’s two roles: A contemporary dilemma. New York: Auburn House. Moen, P. (2003). It’s about time: Couples and careers. Ithaca, NY: Cornell University Press. More, E. S., & Greer, M. J. (2000). American women physicians in 2000: A history in progress. Journal of the American Medical Women’s Association, 55, 6–9. Ness, R. B. et al. (2000). Salary equity among male and female internists in Pennsylvania. Annals of Internal Medicine, 133, 104–110. Oshfeldt, R. L., & Culler, S. (1986). Differences in income between male and female physicians. Journal of Health Economics, 5, 335–346. Padavic, I., & Reskin, B. (2002.) Women and men at work (2nd ed.). Thousand Oaks, CA: Sage. Phillips, S. P., Richardson, B., & Lent, B. (2000). Medical faculty’s views and experiences of parental leave: A collaborative study by the gender issues committee, Council of Ontario Faculties of Medicine. Journal of the American Medical Women’s Association, 55, 23–26. Rapoport, R., Bailyn, L., Fletcher, J. K., & Pruitt, B. H. (2002). Beyond work-family balance: Advancing gender equity and workplace performance. San Francisco: Jossey-Bass. Riska, E. (2001). Medical careers and feminist agendas: American, Scandinavian and Russian women physicians. New York: Aldine De Gruyter. Risman, B. J. (1998). Gender vertigo: American families in transition. New Haven, CT: Yale University Press. Rosenstock, L., & Jackson Lee, L. (2000). Caution: Women at work. Journal of the American Medical Women’s Association, 55, 67–68. Sobecks, N. et al. (1999). When doctors marry doctors: A survey exploring the professional and family lives of young physicians. Annals of Internal Medicine, 130, 312–319. Tesch, B. J., Osborne, J., Simpson, D. E., Murray, S. F., & Spiro, J. (1992). Women physicians in dual-physician relationships compared with those in other dual-career relationships. Academic Medicine, 67, 542–544. Uhlenberg, P., & Cooney, T. (1990). Male and female physicians: Family and career comparisons. Social Science and Medicine, 30, 373–378. Waldfogel, J. (1997). The effect of children on women’s wages. American Sociological Review, 62, 209–217. Waldfogel, J. (1998). Understanding the ‘family gap’ in pay for women with children. The Journal of Economic Perspectives, 12, 137–156. Wallace, A., & Weeks, W. B. (2002). Differences in income between male and female primary care physicians. Journal of the American Medical Women’s Association, 57, 180–184. Weisman, C. S., & Teitelbaum, M. A. (1987). The work-family role system and physician productivity. Journal of Health and Social Behavior, 28, 247–257. West, C., & Zimmerman, D. (1987). Doing gender. Gender and Society, 1, 125–151. Williams, J. (2000). Unbending gender: Why family and work conflict and what to do about it. New York: Oxford University Press.

MOTHERHOOD AND CAREER COMMITMENT TO THE LEGAL PROFESSION Jean E. Wallace ABSTRACT The literature suggests that women are less committed to and less successful in their careers than men because of family responsibilities. I examine whether mothers practicing law are less committed to their legal careers than other women. Mothers acknowledge that they violate certain time and career norms associated with practicing law and work in different settings, which may be interpreted by others as indicators of their lack of career commitment. The survey results reveal that, despite these violations, mothers report greater career commitment than other women in law. I conclude by examining possible explanations for these findings. For decades sociologists have examined the gendered segregation of occupations and how men and women select into different occupations (Reskin, 1993). More recently, researchers have also examined how gender segregation is evident within a single occupation. For example, in the case of professional occupations, such as law, medicine and dentistry, it has been empirically demonstrated that men and women tend to work in different specialties and employment settings (Bird, 1996; Hull & Nelson, 2000; Kay & Hagan, 1998; Weisman et al., 1986). Much of this research frames the question of why men and women end up in different jobs as Diversity in the Workforce Research in the Sociology of Work, Volume 14, 219–246 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0277-2833/doi:10.1016/S0277-2833(04)14009-0

219

220

JEAN E. WALLACE

hinging on two competing theoretical frameworks that reflect gender choices or constraints. Basically, the choice or supply-side explanation suggests that gender differences in careers are due to the choices of individual women. That is, women tend to choose different occupations, specialties and work settings than men because women hold different work and family values and interests that lead them to select jobs and fields that best facilitate work-family balance (Reskin, 1993). The constraint or demand-side explanation focuses on the choices and behaviors of employers rather than employees. These explanations emphasize how institutional barriers or discrimination make it difficult for workers, particularly women, to balance work and family responsibilities that may then lead to gender inequalities at work. Such practices, which may appear gender neutral, differentially affect men and women in the ways in which employers recruit, assign jobs, mentor, promote and retain workers (Roos & Reskin, 1984). Regardless of which explanation is considered, emphasis is usually placed on explaining the choices or constraints that women face as responsible for the observed gendered outcomes. As well, both approaches seem to assume that women are generally less committed to their careers than men and both attempt to explain various career outcomes (e.g. earnings, mobility) whereby women’s careers tend to be less successful than men’s. Both explanations also suggest that the primary reason women are less committed to and/or less successful in their careers than men is because of family responsibilities (Bielby & Bielby, 1988). That is, women make certain career decisions and sacrifices as they attempt to juggle work and family, or women face certain barriers because employers assume they will place priority on their family responsibilities that invariably interfere with their work. Accordingly, to understand this situation more completely, two different approaches seem appropriate. First, instead of comparing men’s and women’s work attitudes and experiences, we should compare those of women who are mothers to those of women who are not mothers, since the key explanatory factor appears to be the presence of family responsibilities for working women. Second, rather than assuming that mothers are less committed to their careers than other women, we should examine this empirically. As Bielby and Bielby noted so many years ago in regards to work effort: “The assertion that women seek less demanding jobs and exert less effort on the job is simply a widely held but untested assumption” (1988, p. 1033). Their concerns apply equally well to the assumption regarding mothers’ commitment to their careers. In this study, I examine how mothers’ work experiences differ from that of other women in the legal profession and the extent to which these differences explain mothers’ and other women’s commitment to their careers in law. Career

Motherhood and Career Commitment

221

commitment is defined as the degree to which lawyers intend to continue practicing law. In examining motherhood and career commitment to the legal profession, I address four key questions: (1) According to mothers practicing law, why do others perceive them to be less committed to their legal careers? (2) Are mothers more likely to violate time and career norms and to work in family friendly workplaces than other women working in the legal profession? (3) Does violation of time and career norms and working in a family friendly workplace reduce career commitment for both mothers and non-mothers working in the legal profession? and (4) Are mothers practicing law less committed to the legal profession than other women practicing law? To answer these questions, I rely on two different sets of data collected from women in law. To answer Question 1, I used telephone interview data from women practicing law who were mothers at the time of the study. The exploratory stage of this study was designed to give mothers the opportunity to describe, in their own words, their work experiences and report how they see them as influencing others’ perceptions of their career commitment. The mothers explained how certain work-related decisions, that they as mothers make, are often interpreted by others as signaling less dedication to their legal careers. The reports of mothers’ first-hand experiences were used to identify key factors that may explain variations in women’s career commitment and to guide and direct the review of the literature related to the factors they identified. To answer the remaining questions, I used questionnaire data collected from women in the legal profession. The goal of this stage of the study was to operationalize the issues identified in the mothers’ interviews as relevant sociological concepts. This second stage of the study also functioned to collect a larger, more representative sample suitable for multivariate analyses. These data are used to empirically assess whether mothers and non-mothers differ in their time, career and work setting practices (Question 2), whether violation of time and career norms and working in certain settings reduces women’s career commitment (Question 3), and whether mothers are in fact less committed to their legal careers than other women (Question 4). The remainder of this paper is organized as follows. First, a more detailed description of the interviews conducted in the first stage of this study is presented. Second, the issues and concerns that mothers raised in the interviews are identified. The themes they voiced are presented in conjunction with a review of the relevant literature as a strategy for developing a model of women’s commitment to the legal profession. Third, a brief description of the second stage of the study follows and the variables used to operationalize the themes identified from the interview data are presented. Next, using OLS regression, the models generated from the interviews with mothers are estimated with the questionnaire data collected from

222

JEAN E. WALLACE

women in law. Lastly, this paper closes with a discussion of the results and the conclusions drawn from this study.

DATA AND METHODS: STAGE ONE The analyses that follow in this section are restricted to the interview responses of thirty-three mothers working in law. The data for this study come from a larger study of lawyers practicing law in the city of Calgary, Alberta, Canada conducted during the Fall of 1999. A snowball sampling technique was used to construct a sampling frame for this study. I started with a list of lawyers from contacts in the profession and asked each participant to provide the names and telephone numbers of other lawyers who might participate in the study. Potential respondents were contacted at their place of work and interviews were scheduled at a time most convenient for them. Most interviews were conducted during respondents’ work hours between the hours of 7:00 am and 6:00 pm. The interviews ranged from 20 to 75 minutes in duration, averaging about 35 minutes. The mothers in this study were asked a series of open-ended questions and the ones that are particularly relevant to this paper relate to their intentions to continue practicing law or change jobs, their work experiences related to having children, their work history and their current work arrangements and family situation. On average, the mothers who participated in this study have been married about 9 21 years (range = less than 1 year to 21 years). Almost half of the mothers (45%) had two children at the time of this study and the average age of the mothers’ youngest child was approximately five years old. Thirteen mothers worked full time in law firms, thirteen mothers worked part time in law firms and seven worked in corporate or government offices in a fulltime or part-time capacity.1 Their average age was 39 years (range = 30–48 years) and they have practiced law on average for 12 years (range = 4–24 years). Most of the mothers (79%) had taken at least one break from the practice of law since they were called to the bar and most of these leaves (88%) involved a maternity leave. On average, the mothers in this study worked 39 hours a week at the office (range = 16–55 hours). Mothers employed in a full-time capacity worked approximately 52 hours a week at the office (range = 45–66 hours) and mothers working part-time spent about 32 hours a week at the office on average (range = 16–53 hours). Fifty-eight percent of the mothers reported that they also usually work at home and half work at least one evening a week for an additional 4–5 hours a week. One third of the mothers regularly worked on weekends for an additional 3 or 4 hours a week. Virtually all of the mothers interviewed were

Motherhood and Career Commitment

223

married or remarried at the time of this study. Their average household earnings were between $200,000 and $300,000 (Canadian dollars).

RESULTS FROM STAGE ONE According to mothers practicing law, why do others perceive them to be less committed to their legal careers? The mothers interviewed suggest that while they are concerned about balancing work and family, this is not necessarily synonymous with low career commitment. In fact, many mothers recognize and resent the general assumption that because they have children their legal career is no longer important to them. They feel they are given less important and challenging work assignments, such that their opportunities for career advancement decline, which is often referred to in the literature as the “mommy track” experience (Schwartz, 1989). Turning to the literature, women with family responsibilities are believed to bring less effort to their work and they are assumed to be less motivated and productive, because they are either saving their energy for their home time or they have no energy left for work after meeting their home responsibilities (Becker, 1985; Kay & Hagan, 1998). Empirical tests of these assumptions have shown otherwise (Bielby & Bielby, 1988). The mothers interviewed in this study also feel these assumptions regarding their work effort and commitment are unfounded. They recognize that their employers expect total commitment and that they assume mothers and/or those working reduced hours are not very dedicated to their work. Epstein et al. (1999) refer to mothers working in law who choose to work part time as “double deviants” because their commitment becomes suspect in two different ways. Such deviants are less likely to receive the same career opportunities and advancements as those who are considered to be more productive and committed to their work. In this study, I explore career commitment as reported by the women themselves and as illustrated in the following quotes. The legal profession and the traditional large firm view you as if you weren’t a team player and you have to give 110% or it doesn’t count at all. I think it’s still very pervasive and there’s very little accommodation for alternatives . . . It’s in this particular environment where the people who succeed and who are rewarded are those who give up everything in their lives for their work. If you’re not willing to do that, you don’t get very much positive feedback or rewards. You’re still working all night and on the weekend to get the job done, but you’re still doing the minimum in their eyes and that’s not very satisfying. I’m incredibly committed to this place. I work only four days, so they assume you’re less committed. My commitment is very high. I called from the labor room three times the last time I was there to make sure they could contact my clients and reschedule my appointments.

224

JEAN E. WALLACE

The mothers interviewed indicated that their career decisions are largely influenced by the constraints they perceive in violating certain workplace norms. They feel that other lawyers believe that mothers are not willing to put work first and sacrifice everything else for the sake of their career. As a result, mothers are not considered to be committed lawyers and therefore they are not as successful professionally as other lawyers who are perceived to be more dedicated. Mothers described how they attempt to balance work and family demands and responsibilities and how this relates to others’ perceptions of their career commitment in terms of three general themes. These themes reflect: (1) violation of their employing organization’s time norms and expectations; (2) violation of their employing organization’s career norms and expectations; and (3) working in “family friendly” jobs in law. According to mothers interviewed in this study, they believe they are perceived by others to be less committed to their careers because they violate time and career norms and because they work in family friendly jobs.

Mothers’ Violation of Time Norms and Expectations One predominant theme voiced by the mothers interviewed refers to their employers’ time norms and expectations, which are illustrated in two different ways. One topic they mentioned is the amount of time that lawyers are expected to spend at the office. Perlow (1995) refers to this as “face time” as evidenced by one’s observed presence at work. The other issue refers to the time lawyers are expected to work outside of “normal” working hours, where lawyers are expected to work or attend professional functions after hours in the evenings, over lunch or on weekends. Both of these time norms and how mothers violate them are discussed and illustrated in greater detail below. Many mothers described the difficulties they face due to the excessive time pressures they feel from their colleagues, employers and/or clients. Most indicated that they would prefer to work shorter hours than they do, even those already working reduced hours. They believe that their employers, however, expect a total commitment in terms of their work hours. A number of mothers feel that practicing law and being a mother are both full-time jobs and that it is impossible to do both 100% and be happy. Their frustrations are echoed in the literature where law has long been recognized as a “greedy institution” and family has also been described as a greedy institution with regards to women’s roles (Coser, 1974). The mothers’ comments presented below illustrate the expectation that lawyers sacrifice everything for their careers and work long hours that often

Motherhood and Career Commitment

225

fall outside normal work hours. As in other professional occupations, the temporal boundaries between public “on duty” times and private “off duty” times are blurred rather than rigidly defined (Zerubavel, 1990). Consequently, there is little time that is entirely private and safe from the intrusion of work demands. You’re at work for ‘X’ number of hours and if someone, or a partner, calls you, you come in. You don’t even think about it and it doesn’t matter what you’re doing or when it is. If you’re on vacation or it’s midnight or whatever. The law takes precedence over everything and it’s the most important aspect of your life and you cancel things and you don’t say ‘no’ to anything. I feel we’re supposed to be available to everyone all of the time. Partners forget to ask what else we’re doing at work and in the rest of your life. Sometimes we need to leave the office. We do have husbands, wives, pets, hobbies and they forget that those are just as important . . . We’re expected to give it [outside life] up and we too readily give it up.

Merton introduced the concept of “socially expected durations” to refer to the “socially prescribed or collectively patterned expectations about temporal durations embedded in social structures of various kinds” (1984, p. 265). Socially expected durations may be applied to the work day or work week as well as to the amount of time an individual should spend in particular stages of his or her career or life cycle. Professionals often do not have rigidly fixed start and ending times when either they must be at work or alternatively they may refuse to see clients. Rather, the end of their day is usually marked by the completion of their work tasks instead of marked by the clock (Zerubavel, 1990). When mothers attempt to resolve the conflicts between paid work and family by working fewer hours, fixed schedules or refusing to work after hours, they may violate these socially expected durations and professional norms of relative flexibility. The time employees spend at work is often interpreted as evidence of their contribution and career dedication (Perlow, 1995; Thompson et al., 1999), where work hours may be used as an overt measure of performance that rests on the assumption that “time is money” (Yakura, 2001). In the case of lawyers, especially those in private practice, work productivity and commitment are often measured directly by the number of hours worked or billed (Epstein et al., 1995; Hagan & Kay, 1995; Seron, 1996; Wallace, 1997). In addition, a lawyer’s willingness to attend social functions with colleagues and/or clients is also often viewed as an indicator of their work involvement and commitment (Epstein et al., 1999). Kanter (1977) refers to such tasks as “diplomatic tasks” that involve after-hours dinners and weekend golf games with business partners and potential clients. Lawyers who do not put in long hours at the office or participate in professional activities outside regular office hours are generally perceived by others as less dedicated to their career.

226

JEAN E. WALLACE

Mothers’ Violation of Career Norms and Expectations The second predominant theme identified by the mothers interviewed in this study is the difficulty associated with negotiating and pursuing alternative career arrangements. These career deviations not only violate the time norms regarding work hours, they are also inconsistent with the socially expected durations (Merton, 1984) or the normal timing and progression of a career in law. The mothers reported how they violate the career norms associated with practicing law if they work in a non-traditional, or alternate, work arrangement (e.g. reduced hours, part time or job share). They also indicated that they violate legal career norms if they take a maternity leave. Many of the mothers interviewed in this study felt that even though working part time allows them to accommodate their family life, it is less than ideal. Mothers mentioned how working fewer hours and/or following a non-traditional career trajectory leads to status incongruities. Women with considerable previous experience and/or a strong client base are not considered to be as serious about their careers, nor do they receive the esteem and prestige their full-time counterparts receive. One mother commented, for example, on how when she started working part time she was no longer considered a “real” lawyer. As well, many pointed out how they fail to have the same career opportunities as others because they receive the less interesting, challenging files and quickly run into a dead-end career with little chance of advancement. The following quote illustrates how working part time does not always mean working the shortest hours, but often translates into career stigma and penalties. Women are clearly getting the short end of the stick and I don’t think they’re getting promoted and they’re not getting the remuneration because they’re not mainstream. When I looked at it, and they circulate those things, my billings for part-time work was more than some of those who worked full time, so I was a pretty good deal for the firm. In the one firm, I changed my status to part time and basically my feeling is that I became invisible, like a eunuch, in terms of status. I was not eligible for partnership and it was even more pervasive than that. It’s like you weren’t on the team and you weren’t given the same consideration as to the availability of work, and the good work, and I was treated like a secondclass associate.

Consistent with other studies, mothers in the legal profession recognize the trade offs of working reduced hours. Women working part time typically begin their careers working full time and then reduce their work hours upon the arrival of children (Chambers, 1989; Epstein et al., 1999; Hagan & Kay, 1995; Seron & Ferris, 1995). By working part time, they may have more time for their family and themselves, and feel less overloaded and exhausted, but the costs are less pay, less job security and fewer promotions (Barnett & Gareis, 2000; Budig & England,

Motherhood and Career Commitment

227

2001; Tausig & Fenwick, 2001; Wood et al., 1993). Recent studies show that working part time is one of the most significant earnings penalties mothers face compared to other women (Waldfogel, 1997). Other researchers have also documented how working reduced hours in law may result in fewer opportunities to work on high profile cases and falling off the partnership track and onto the mommy track (Epstein et al., 1995, 1999; Hagan & Kay, 1995). These various career penalties associated with working part time are associated with the assumption that women working part time are less committed to their careers. The second career violation mothers reported is taking a temporary leave from law for child care reasons, usually in the form of a maternity leave. Mothers in the legal profession in this and other studies report losing clients and credibility within the legal profession after taking a maternity leave (Leiper, 1998), such that they forfeit career advancement for control over their career time. As well, for women working in law firms, the most critical professional passage from associate to partner coincides with child bearing years (Epstein et al., 1999; Hagan & Kay, 1995). The following mothers’ comments represents many mothers’ sentiments as to how they felt their colleagues’ attitudes changed upon their return from a maternity leave: I think they think your dedication is different even though it isn’t. You have to juggle more but that doesn’t mean you’re not dedicated to your job. It was pretty blatant though. Like one guy said when I came back [from maternity leave] that ‘you go and have a couple of kids and you no longer care about your career.’ It was just like that.

In the literature on earnings, it is suggested that women who take leaves are penalized because they have less work experience than their counterparts who do not take leaves (Corcoran et al., 1984; Klerman & Leibowitz, 1999; O’Neill & Polachek, 1993; Waldfogel, 1997). Discontinuity in employment is seen to diminish workers’ skills and their value to their employer (Felmlee, 1995) and is also interpreted by employers as an indicator of less commitment to one’s work. Budig and England (2001) refer to this as the “motherhood penalty” where women who lose employment time due to child bearing and rearing are penalized by lower wages. As indicated in the quotes above, the mothers interviewed in this study suggest that taking a leave of absence is viewed as an indicator of their lesser career commitment. Working in Family-Friendly Jobs in Law The third theme raised by mothers is the notion that certain jobs in law are more “family friendly” than others. As Budig and England (2001) suggest, following the

228

JEAN E. WALLACE

theory of compensating differentials, mothers may trade off certain job rewards, such as higher wages or career advancements, for jobs that make it easier to combine work and family. Family-friendly jobs are characterized as requiring less energy, more flexible hours, fewer demands for travel, weekend or evening work, on-site day care, opportunities to take time out to check on or meet childcare needs, and most obviously the ability to work part time (Becker, 1991; Clark, 2001). It has been suggested that a primary motive for firms adopting work-family policies is the belief that they will increase employee loyalty (Roehling et al., 2001). Family-friendly jobs were described by the mothers in this study in a variety of ways in addition to the adjustments that they may make to their time and career practices described above. For example, a job is more family friendly if it does not demand long hours or if they are able to negotiate alternate work arrangements or take family leaves. In addition, the mothers described their employers’ family responsiveness in terms of work-family culture, the availability of alternate work arrangements, ability to control one’s work hours and the different employment settings in law. These different aspects of family-friendly jobs are discussed in greater detail below. A number of mothers mentioned their employers’ support of alternate work schedules, flexible hours or other ways of accommodating work and family demands. Some felt their employers were supportive of their family lives whereas others, as illustrated by the following quote, felt the culture of the organization was inconsistent with its policy. Our company says it believes in work-life balance and that’s its policy. Yet when it comes to being here after hours to get the work done, if you try to say you’ll do it the next day or later, that would not be an acceptable response.

The work-family culture of an organization may function to encourage or discourage employees from taking advantage of family benefits offered by employers. Work-family culture refers to the “shared assumptions, beliefs and values regarding the extent to which an organization supports and values the integration of employees’ work and family lives” (Thompson et al., 1999, p. 394). The work-family culture may reflect the time demands or expectations that employees prioritize work over family as well as the perceived negative consequences if employees use work-family benefits or devote time to family (Bailyn, 1997; Clark, 2001).2 The work-family policies of an organization may conflict with established organizational norms, such as a belief in the importance of “face time” (Perlow, 1995) that may lead to a “culture of overtime” (Fried, 1998). As a result, employees may be reluctant to use work-family benefits if they perceive it as costly to their careers (Blair-Loy & Wharton, 2002). The compatibility of jobs with family demands also refers to jobs that offer flexibility in work hours (Glass & Camarigg, 1992; Glass & Riley, 1998). Flexible

Motherhood and Career Commitment

229

schedules may allow employees to alter their daily starting and ending times at work or simply the ability to leave work when unexpected nonwork demands arise (Golden, 2001). Some have argued that the degree of control and flexibility over one’s work time is often more salient to workers than the sheer number of hours they work (Holtzman & Glass, 1999; Kalleberg & Epstein, 2001). For example, one mother suggested the following: If you can get a large firm to be flexible with you, it’s the best place. In-house lawyers [in a corporation] actually have more restrictions as to when they have to be in the office. They may not have to be there as much, but they have to be there between 8:30 and 5:00. Being in a large firm and in a senior position, I work when I need to and schedule my clients around me and being in-house would restrict my flexibility.

Many of the mothers interviewed in this study commented on how the different work settings in law vary in regards to their compatibility with family. They emphasized how working in private practice may demand longer hours but it also tends to offer more flexibility or control over when lawyers work the hours that they do, compared to working in corporations or government. The majority of lawyers work in the core of the profession, or private practice. A lawyer employed in a law firm in whatever capacity or who works as a solo practitioner is deemed to be in private practice. This excludes lawyers employed in government, business, judiciary or university settings. The following two mothers comment on the flexibility they experience in working in law firm settings. I looked at working in-house [for a corporation] as part time but I rejected it because it has greater expectations. It’s more structured so I just couldn’t have the flexibility I have now where I can work around the clock for a week and then take a week off and that flexibility is great. I may take a day off for my kids and I can do it when I want to. It may be somewhat important to know that this firm is as flexible as it is. No one is checking the clock as you leave. I make as much money as I want by choosing the hours I want to work and I can choose to take on work or not.

In sum, according to the mothers interviewed in this study, mothers in the legal profession are more likely to violate certain time and career norms and they are more likely to work in a family-friendly workplace compared to other lawyers. And, the mothers interviewed suggested that lawyers who violate time and career norms or who work in family-friendly workplaces are assumed to be less committed to the practice of law than other lawyers.

DATA AND METHODS: STAGE TWO All lawyers practicing in the Province of Alberta were mailed a survey to their place of work in June 2000. The mailing list was obtained from the Law Society of Alberta and contains the names of all active members. Questionnaires were

230

JEAN E. WALLACE

sent to 1,599 women and 680 were returned, yielding a response rate of 43%. Data provided by the Law Society of Alberta in terms of all active members’ gender and employment situation by city allows for comparisons between the provincial data and that of the sample (available from author). A comparison of the sample data with the provincial statistics obtained from the Law Society of Alberta indicates that similar proportions of lawyers participated when compared by gender, practice setting (e.g. law firm, government office, etc.) and city (e.g. Calgary, Edmonton, etc.). The analyses in this section are restricted to the women who completed the survey, which consists of 270 (45%) mothers and 332 (55%) non-mothers.3 Most of the women in this study were married (73%) and they were about 39 years of age (range = 26–68). They practiced law for approximately 10 years on average (range = 1–44) and their mean annual earnings was $86,820 (range = $5,000–$400,000). Most had traditional, full-time work arrangements (87%) and they worked about 45 21 hours a week at the office (range = 3–50). About half of the women in this analysis worked in small or medium law firms (N = 305), 19% (N = 113) in government offices, 15% (N = 93) in large law firms and 15% (N = 90) in corporate offices. Measures: Mail-out questionnaires were constructed that measure the themes and issues raised by the mothers interviewed in the first stage of this study. Various indicators were included to tap aspects of lawyers’ career commitment, work time, career decisions and the family-friendliness of their workplace as mentioned in the interviews. Career commitment is measured by the mean score of six Likert items adapted from Wallace (2001): I would like to continue working in the legal profession; I would like to work in an occupation other than the legal profession (R); I would like to leave the legal profession (R); I plan to continue practicing law as long as possible; I intend to quit the practice of law in the near future (R); and I’m thinking about leaving the practice of law in the near future (R) (alpha = 0.92). Motherhood status is measured by whether (coded 1) or not (coded 0) respondents have any children living with them. Violation of Time Norms: The violation of time norms is tapped by three variables; the hours lawyers work at the office, the hours they work at home and the extent to which they participate in extra professional activities outside of normal work hours. The fewer hours lawyers work and the less they participate in professional activities, the more they violate the time norms associated with practicing law. Hours at Office is the average number of hours that respondents reported working at the office per week, including evenings and weekends. Hours at Home is the average number of hours respondents worked per week at home, including evenings and weekends. Extra Professional Activities is the number of times a

Motherhood and Career Commitment

231

month respondents attend professional activities (e.g. related to business or client development, conferences, meetings or receptions) before 8 am, over lunch, after 6 pm during the week, or on weekends (day or night). It is coded 1 if less than once a week; 2 is 1–2 times a week; and 3 is more than twice a week. This measure taps the extent to which lawyers participate in social, work-related activities outside of regular working hours. Violation of Career Norms: The violation of career norms and expectations is tapped by three variables; whether they work in an alternate work arrangement, whether they have taken a leave from law and the extent to which they have sacrificed their career for family reasons. Alternate Work Arrangement is coded 1 if they work in any arrangement other than full time (e.g. part time, reduced hours, job share) and coded 0 if they work full time. Leave Taken is coded 1 if they have taken a leave from law since they started practicing and 0 if they have not. Career Sacrifices reflects the summed total of the number of five possible career sacrifices, each coded 1, if made due to family responsibilities, and 0 if not: refused to take on additional work; refused to work late or extra hours; cut back on work time; not been available for clients; not attended social functions. This measure taps the extent to which lawyers have made career trade-offs by sacrificing various work obligations specifically for family reasons. Family-Friendly Workplace: Four variables tap the extent to which the respondent’s workplace is family friendly. If the work-family culture is supportive, if the workplace offers alternate working arrangements and flexible work hours, and if lawyers work in a large, law firm, the work setting is more family friendly. Supportive Work-Family Culture is measured by the mean score of three Likert items adapted from Thompson et al. (1999): Turning down work for family-related reasons will seriously hurt one’s career in this organization (R); Many lawyers are resentful when people in this organization take extended leaves to care for new or adopted children (R); and In this organization, lawyers who participate in work-family programs are viewed as less serious about their careers (R) (alpha = 0.79). Alternate Arrangements Available is coded 1 if their employer allows part time, reduced hours or other alternate arrangements for lawyers and 0 if it does not. Flexibility in Work Hours is measured by a single Likert item: It’s very hard for me to take time off to take care of personal or family matters (R). Work Setting is coded 1 if the employment setting is a law firm and 0 for all other settings (e.g. government, private corporations). Controls: Five control variables are included in the analyses. Career Plateau refers to the extent to which respondents believe they have a low likelihood of being promoted in their current organization. Those who feel their career has plateaued are less likely to be highly motivated and committed to their careers (Ettington, 1998). It is measured by the mean score of summed Likert items adapted from

232

JEAN E. WALLACE

Ettington (1998): I expect to advance to a higher level in the near future in this organization (R); I have reached a point where I do not expect to move much higher in this organization (R); and I’m likely to be promoted above my current level in this organization (R) (alpha = 0.85). In the commitment literature, job satisfaction has been found to be an important factor that explains why workers change jobs or careers (Kay, 1997; Neapolitan, 1980; Rhodes & Doering, 1983). Job Satisfaction is defined as the extent to which employees like their job. It is measured by the mean score of three summed Likert items adapted from Brayfield and Rothe (1951): Most days I’m enthusiastic about my job; I find real enjoyment in my job; and I definitely dislike my job (R) (alpha = 0.87). Law Experience, or the number of years they have practiced law, is also taken into account where it is expected the longer lawyers practice law the more committed they are to the profession. Earnings is also included for control purposes and is measured by the natural logarithm of the total annual earnings from the practice of law for the 1999 tax year, before taxes and other deductions are made. Several studies suggest that extrinsic rewards, such as pay, are important in becoming a significant obstacle to actually changing careers (Neapolitan, 1980). It has been suggested that the high salary lawyers earn acts as “golden handcuffs” (Arron, 1989) that deter lawyers from quitting the practice of law (Kessler, 1997). Marital Status is included to control for the possibility that women who are married are less committed to the profession. It is coded 1 if married or living common law and 0 if not. Analyses: First, to answer Question 2, differences in the means of the time and career norm variables, family friendly workplace variables and control variables are examined to see if mothers and non-mothers differ significantly along these variables (Table 1). To answer Question 3, ordinary least squares (OLS) regression analysis, in conjunction with interaction tests, was used to determine whether violation of time and career norms and working in a family friendly workplace similarly reduce career commitment for both mothers and non-mothers. OLS regression was used to estimate an equation for career commitment that included both mothers and non-mothers. These results are located in Table 2 and presented in the Pooled model. We can use this equation to interpret the variables that have significant main effects on career commitment for both mothers and non-mothers. Then, this equation was re-estimated by including all of the determinants, a dummy variable for motherhood status and cross-product motherhood-interaction terms for each of the determinants (not shown). The interaction terms that identify which regression coefficients differ significantly for mothers and non-mothers are also identified in Table 2 in the Pooled model. We can then use these results to identify which determinants have significantly different effects for mothers and non-mothers. Because of these significant interactions, OLS regression was then

Motherhood and Career Commitment

233

Table 1. Comparison of Mothers (N = 270) and Non-Mothers (N = 332). Variable

Mean Career commitment Time norms Weekly hours at office Weekly hours at home Extra activities per month

Mothers Mean (SD)

Non-Mothers Mean (SD)

3.57 (0.84)

3.48 (0.91)

41.79 (11.39) 4.76 (6.43) 1.78 (0.74)

48.70 (9.73)*** 3.19 (4.93)*** 1.73 (0.74)

Career norms Alternate work arrangement Leave taken Career sacrifices

0.25 (0.44) 0.73 (0.45) 3.23 (1.61)

0.03 (0.17)*** 0.20 (0.40)*** 1.57 (1.66)***

Family friendly workplace Supportive work-family culture Alternate arrangements available Flexibility in work hours Work setting (law firm)

2.97 (0.99) 0.68 (0.47) 3.41 (1.06) 0.65 (0.48)

3.01 (0.93) 0.56 (0.50)** 3.24 (1.06)* 0.67 (0.47)

3.21 (0.91) 3.07 (0.40) 12.49 (6.78) 11.29 (0.63) 0.85 (0.36)

2.82 (0.85)*** 3.06 (0.40) 7.48 (6.38)*** 11.08 (0.63)*** 0.63 (0.48)***

Controls Career plateau Job satisfaction Law experience Earnings (Ln) Marital status (married) ∗p

< 0.05 (one-tailed test). < 0.01 (one-tailed test). ∗∗∗ p < 0.001 (one-tailed test). ∗∗ p

used to estimate the main effects of the determinants on women’s career commitment separately for mothers and non-mothers (also presented in Table 2). This facilitates interpreting the similarities and differences in the effects of the specific variables on career commitment for mothers and non-mothers. To answer Question 4, as to whether mothers are less committed to their legal careers than non-mothers, the coefficient for motherhood status presented in the Pooled model of Table 2 is examined. Examination of the zero-order correlations (available from author) between the variables included in this analysis shows that only one correlation is 0.60 or higher.4 Specifically, the correlation between the hours mothers work at the office and whether or not they have an alternate work arrangement has a zero-order correlation of –0.63. In order to determine whether multicollinearity is a potential problem, following Fox (1991), I estimated variance-inflation factors (VIF) for

234

JEAN E. WALLACE

Table 2. Determinants of Career Commitment for Mothers (N = 229) and Non-Mothers (N = 298). Determinant Motherhood status (mother = 1)

Pooled b (B)a

Mothers b (B)

Non-Mothers b (B)

0.240 (0.136)b,**

Time norms Hours at office Hours at home Extra professional activities

0.005 (0.061)c −0.007 (−0.050) 0.055 (0.046)c

−0.004 (−0.051) −0.011 (−0.083) −0.023 (−0.020)

0.012 (0.128)* −0.000 (−0.004) 0.102 (0.084)

Career norms Alternate work arrangement Leave taken Career sacrifices

0.082 (0.032) −0.162 (−0.092)c,* −0.062 (−0.129)**

−0.031 (−0.016) −0.008 (−0.004) −0.038 (−0.073)

0.110 (0.021) −0.335 (−0.147)** −0.065 (−0.119)*

0.073 (0.080)* 0.009 (0.005) 0.118 (0.143)c,*** 0.133 (0.072)*

0.075 (0.088) −0.110 (−0.061) 0.166 (0.210)*** 0.065 (0.037)

0.099 (0.102)* 0.090 (0.049) 0.084 (0.099)* 0.186 (0.096)*

−0.201 (−0.206)*** 0.780 (0.353)*** 0.019 (0.151) −0.000 (−0.074) 0.033 (0.024)c 0.204 (0.078)**

−0.246 (0.267)*** 0.623 (0.297)*** −0.002 (−0.015) 0.000 (0.069) 0.148 (0.122)* 0.175 (0.074)

−0.162 (−0.152)** 0.885 (0.386)*** 0.050 (0.354)* −0.001 (−0.254) −0.067 (−0.046) 0.215 (0.115)*

0.306

0.352

0.323

Family friendly workplace Supportive work-family culture Alternate arrangements available Flexibility in work hours Work setting (law firm) Controls Career plateau Job satisfaction Law experience Law experience2 Earnings (Ln) Marital status (married) R2 a Unstandardized

(b) and standardized (B) regression coefficients. test, effect not in hypothesized direction. c Coefficient is significantly different for mothers and non-mothers, p < 0.10 (two-tailed test). ∗ p < 0.05 (one-tailed test). ∗∗ p < 0.01 (one-tailed test). ∗∗∗ p < 0.001 (one-tailed test). b Two-tailed

all of the variables included in the analysis. These results (available from author) suggest that multicollinearity among the predictors is not evident for either group of women. In addition, I respecified the model for mothers by first excluding the hours worked at the office and re-estimating the model. Then I excluded alternate work arrangement (with hours worked returned to the model) and re-estimated the model. The regression estimates of both respecified models for mothers did not differ from those reported in Table 2.

Motherhood and Career Commitment

235

RESULTS FROM STAGE TWO Are mothers more likely to violate time and career norms and work in family friendly workplaces than other women working in the legal profession? The mean scores presented in the Table 1 show that mothers are more likely to violate time norms by working significantly fewer hours at the office than other women. Mothers are more likely to work longer hours at home, however, and are involved in the same number of professional activities outside of regular work hours as non-mothers. Turning next to the career norms, mothers appear to violate all three career norms more so than other women practicing law. Specifically, mothers are more likely to have an alternate work arrangement, to have taken a leave from the practice of law and made more career sacrifices than non-mothers. Mothers are more likely than other women to work in organizations that offer alternate work arrangements and that offer more flexibility in work hours. Both groups of women report similar degrees of work-family support in their organizations and are equally likely to work in a law firm. Does violation of time and career norms and working in a family friendly workplace reduce career commitment for both mothers and non-mothers working in the legal profession? The Pooled model in Table 2 shows the regression results for the interaction tests for mothers’ and non-mothers’ career commitment. Four of the ten substantive variables have statistically significant interaction terms, namely, hours worked at the office, participation in extra professional activities, whether or not a leave has been taken and flexibility of work hours. Table 2 also shows that the control variable earnings has a statistically significant interaction. The discussion below is presented as follows: Those determinants that have a statistically significant effect on career commitment but do not have a statistically significant interaction (e.g. career sacrifices) are interpreted as having the same effect for both mothers and non-mothers. For those determinants that have a statistically significant interaction (e.g. hours at office), the way in which their effects for mothers and non-mothers differ are discussed below. The fact that mothers work significantly fewer hours than other women does not appear to reduce mothers’ career commitment. Rather, non-mothers who work fewer hours have lower commitment levels. Overall, the results in Table 2 suggest that complying with time and career norms associated with practicing law enhances career commitment, but that this is more applicable to non-mothers than mothers. Working longer hours at the office is important in understanding non-mothers’

236

JEAN E. WALLACE

career commitment. Violation of any of the three time norms, however, is not important in affecting mothers’ dedication to their career. Turning next to the career norms, it appears from Table 2, that when either mothers’ or non-mothers make career sacrifices, they report lower commitment levels. Also according to Table 2, the effect of whether a leave has been taken varies significantly for mothers and non-mothers – if non-mothers take a leave, which they are less likely to do, they report lower career commitment, but taking a leave has no impact for mothers.5 It appears that although mothers are more likely to violate career norms, it is somewhat less consequential to their career commitment than it is to other women who are less likely to violate such norms associated with practicing law. According to Table 2, three family-friendly workplace characteristics are important in enhancing career commitment for both groups of women – a supportive work-family culture, flexible work hours and working in a law firm setting. Flexibility of work hours also has a significant interaction. While this variable has a significant, positive effect on career commitment for both groups of women, the coefficient is considerably larger for mothers. That is, flexibility in work hours is more important in enhancing mothers’ career commitment than non-mothers’.6 The availability of alternate arrangements does not affect women’s commitment to their career in law. Are mothers practicing law less committed to the legal profession than other women practicing law? Table 2 shows the regression results for mothers’ and non-mothers’ career commitment.7 The findings suggest that, contrary to their colleagues’ perceptions, mothers are significantly more committed to their careers than non-mothers.8 What else affects women’s commitment to the legal profession? Several of the control variables are important in affecting career commitment among women in law. Women who are more satisfied in their jobs and who do not feel their career has plateaued are more committed to their careers and these are amongst the most important determinants for both groups of women (Table 2). According to Table 1, while both groups of women report similar levels of job satisfaction, mothers are more likely to feel that their career has plateaued compared to other women in law. As well, married women appear more committed to their careers than single women and mothers are more likely to be married than other women. Lastly, the effect of earnings is significantly different for mothers and non-mothers (Table 2). Mothers who earn more are more committed to their careers, whereas this variable has no significant effect for non-mothers and on average, mothers earn more than other women in law.

Motherhood and Career Commitment

237

DISCUSSION AND CONCLUSIONS This paper set out to examine why mothers are perceived to be less committed to their careers and what factors affect the career commitment of women working in law. First, using interview data I explored why mothers believe others perceive them to be less dedicated to their careers in law. From the mothers’ reports, it appears that they believe it is because they violate the profession’s time and career norms by working shorter hours, negotiating alternate work arrangements, taking leaves of absence and sacrificing their careers for family responsibilities. They suggested that violation of these norms is viewed by others as indicators of a lack of commitment to their career. Many of the mothers pointed out that, in their view, the assumption that mothers are less committed to their careers is erroneous and unfounded. Mothers also identified certain features of the workplace that yield a more family-friendly setting for working parents. These include the availability of alternate work arrangements, flexible hours and a more family-friendly work culture. In the second part of this paper I examined how mothers and non-mothers differ in their work experiences and what factors affect the career commitment of women practicing law. The concerns identified by mothers in the legal profession presented in the first part of this paper in conjunction with the review of the literature were used to generate a model of career commitment. This model was then estimated for mothers and non-mothers working in law. Not surprisingly, mothers are more likely to violate the time and career norms associated with practicing law than other women. Specifically, mothers tend to work shorter hours at the office and they are more likely to work in alternate work arrangements, have taken a leave from law and have sacrificed their careers for their family. Yet, the regression results of this study suggest that after taking these factors into account, mothers are significantly more committed to their legal careers than other women practicing law. These results raise the question: why is it that mothers report more commitment to their legal careers than other women? The results of this study suggest that two job characteristics are partly responsible, namely flexible work hours and earnings. Mothers report that their workplaces offer more flexible work hours and higher earnings than other women and both variables are more important in enhancing mothers’ commitment. Since the effect of motherhood on career commitment remains significant, however, after taking these and other factors into account, it suggests that there is something about being a mother that results in greater career commitment that is not measured and controlled in this analysis. Several possible explanations are presented below. One reason that mothers may report greater career commitment than other women is that the most highly committed mothers may return to their careers

238

JEAN E. WALLACE

upon having children, whereas the less committed women may not. Some have suggested that men show career commitment by having children, whereas women show their commitment by returning to work upon having children (Kay & Hagan, 1998). Professional women who have children and return to work have consciously decided to have children and continue their career, which may renew and sustain their career commitment (Hagan & Kay, 1995). In contrast, women who are not highly committed to their jobs may have quit when they entered the child-bearing, child-rearing stages of their life (Marsden et al., 1996; Wallace, 2001). Thus, a selection process may be working to “select out” the less committed women who do not want to try to juggle their work demands while having a family. Mothers may also select out of their jobs after childbirth because of a lack of parental leave, flexible hours, part-time work options, or other family-related concerns, which make it difficult to combine parenting and employment (Glass & Riley, 1998), especially if they are not highly committed. Other studies have also suggested that leaving one’s legal career temporarily, or even permanently, remains a legitimate option for women who want to spend time at home with their children (Brockman, 1992; Wallace, 1994). As a result, the mothers who remain in the workforce may represent those who are amongst the most highly dedicated to their careers. Longitudinal analyses that compare women’s career commitment before and after having children, as well as between women who decide to continue or discontinue practicing law upon the arrival of children, are needed to empirically assess this possible explanation. A second reason mothers may report being more committed is that they may have had to overcome greater obstacles to be similarly rewarded in their jobs as their colleagues. By overcoming difficult hurdles, making more sacrifices and expending more effort, their commitment is likely quite strong (Grusky, 1966; Kanter, 1972, 1977). One important obstacle at work has been identified as social approval. Many mothers, for example, described how they are not perceived by others to be “real lawyers” when they have children, work reduced hours, or take a maternity leave. Women who are not mothers may receive social approval (i.e. be viewed as “real lawyers”) in addition to other job rewards (e.g. more career opportunities since their careers appear less likely to have plateaued, as shown in Table 1) more readily than mothers. As a result of not having to fight to receive the same recognition as others, non-mothers may have only moderate commitment, compared to that exhibited by mothers who continue practicing law. Mothers, as token women in high-status jobs, may view themselves as overcoming discrimination and stereotypes by choosing to work hard, thereby demonstrating their dedication and capabilities in order to receive the same or even fewer rewards (Kay & Hagan, 1998; Maume & Houston, 2001). Kay and Hagan refer to this as “raising the bar” for women in law such that “a unique and exceptional set of standards is imposed” (p. 741) and women

Motherhood and Career Commitment

239

must demonstrate exceptional work commitment to be successful in their careers. Perhaps the bar is set still higher for mothers in the legal profession where they are expected to display even more perseverance and commitment than other women in law. More in-depth analysis of the obstacles and challenges that women face in attempting to combine motherhood and a demanding career and how this is related to their career commitment is needed to investigate this further. Thirdly, it should be noted that as women add family roles to their work roles it does not necessarily mean that an increased commitment to one (e.g. family) results in a decreased commitment to the other (e.g. work) (Marks, 1977). Sociologists often refer to a “scarcity” or “depletion” approach to multiple roles that emphasizes the over-demanding nature of multiple roles in terms of role strain or role overload. Others suggest that the “expansion” (Marks, 1977), “synergy” (Bailyn, 1993; Greenhaus & Parasuraman, 1999) or “enrichment” approaches (O’Neil & Greenberger, 1994; Thompson & Bunderson, 2001) should also be considered, in which multiple roles are argued to be energy and commitment creating rather than draining. For example, family activities may produce more energy for other roles because of the supportive and sympathetic atmosphere they generate (Thompson & Bunderson, 2001). As Bielby and Bielby suggest, “as women add work roles to their family roles [or vice versa], they generate the energy necessary to fulfill their commitments to the two sets of activities” (1988, p. 1056). The following single mother has 11-year old twins and works about 40 hours a week at the office and another 10 hours a week at home in the evenings. Her comments illustrate how spending enjoyable time at work or with her children result in her having “almost unlimited energy” to do them. I think that if you are doing something you are enjoying, whether you’re at home with kids, or working on a file, if you enjoy it, you will have almost unlimited energy to do it . . . Without sounding too self absorbed, I’m often amazed at the things I can accomplish in a day. I’m literally on the run from the moment I get up until I shut the books at midnight. Almost every day I’m really glad about the things that I’ve done. I’m really busy and on the whole I’m really happy with what I’m doing.

The arguments about whether multiple roles are enriching or depleting and how they relate to career commitment need to be examined empirically in future research. A second important question that the results of this study raise is why is it that the time and career norms are not that important in affecting mothers’ career commitment? The findings show that not only do mothers and non-mothers have different work experiences but also that different factors are responsible for their career commitment. Conformity to time and career norms appears more important in enhancing non-mothers’ career dedication, but violation of these norms is less

240

JEAN E. WALLACE

important in reducing mothers’ commitment. The results suggest that for mothers in the legal profession, the time and career norms are not synonymous with their career commitment. Perhaps mothers hold a career model that differs from the traditional career model held by men, where the latter is characterized by a linear trajectory laid out “with mathematical precision” (Leiper, 1998, p. 120) as lawyers advance on cue as they accumulate skills and knowledge. The following quotes illustrate how mothers recognize that to be considered by others to be successful in law, women must adopt the male career model. The female role models acted like males and practiced like a male does and it’s necessary to practice successfully. The women who’ve made it have decided to be like men and they run their careers like men. I think it’s impossible to have children and be married and be female and be successful.

But the following mother who works part-time explains how being successful in the traditional sense is not important to her. I feel a lot of pressure from my work place to work full time and it certainly has hurt my career. I don’t get the “big” promotions, but that’s not that important to me.

It should also be considered that mothers may not only reject the traditional male career model, but that they may also reject the traditional career goals and rewards associated with it. And rejection of these particular conceptualizations of career and success does not necessarily mean they are uncommitted to their careers; it may mean that they simply have different ideas regarding their career progress and goals. Qualitative research that explores how women and men and parents socially construct the notion of career would be helpful in this regard. While mothers may be highly committed to their careers, as long as time at work and a linear career path are viewed as indicators of one’s career dedication, mothers violating such norms will be viewed by others as less committed and therefore differentially rewarded by their employers. These assumptions of time as indicators of commitment and productivity, however, have been challenged empirically. For example, Weisman et al. (1986) found in their research on physicians that working fewer hours does not necessarily translate into less patient care. They suggest that documentation of women working fewer hours than men (and mothers working fewer than other women) does not indicate which specific professional activities are foregone by women and mothers. Mothers may use their time more effectively at the office than others who spend more time there and more detailed information on the use of professional time is required to understand this more completely. Leiper (1998) similarly notes that many women in law work a concentrated work day – they may spend fewer hours at the office than their colleagues, but they also avoid socializing with their colleagues at work and often eat lunch at their desk as they work through their lunch time.

Motherhood and Career Commitment

241

The following mother explains how she believes those who work part time do the same amount of work, just with fewer hours at the office. Part time practice doesn’t work from what I’ve seen, where my friends try to work shortened hours or fewer days. They still end up doing the same amount of work, they just do it at home instead.

The ways in which time is spent at work needs to be explored more extensively in terms of how women and men differ in this regard. This may help us to better understand professionals’ career commitment than by simply comparing the amount of time spent at work. In short, time spent at work does not necessarily equal effort, commitment or productivity, although they may be highly related for some workers. Organizations tend to be gendered in their evaluation of men and women and how their productivity and commitment are assessed and directly linked to the amount of time they spend at work (Maume & Houston, 2001). Research is needed that explores more directly the differences in the effort, commitment and productivity between mothers and other women and men and women, which has been examined only indirectly in the past (e.g. Bielby & Bielby, 1988). In closing, there are several limitations of this study that must be noted. First, this study focused on high-status professional women. These women are unique in several ways. Most of the women earn relatively high salaries that allow them to purchase household and childcare services that other women may not be able to afford. Purchasing such services may alleviate the domestic demands for women and simultaneously grant them more balance in their lives and allow them more time to focus on their careers. As well, those that hold part-time jobs have jobs that are generally very “good,” high-paying jobs, that have been voluntarily chosen, which is often not the case in lower-status, part-time jobs. Second, the measure of commitment is based on women’s self-reports – it is not an “objective” measure as some might argue work hours are and it does not reflect others’ perceptions of that person’s commitment. It is unclear how closely self-reports of one’s commitment mirror others’ perceptions of commitment. Moreover, just because mothers may believe or actually be more committed than other workers, it does not mean that their colleagues or employers see them this way. If, as indicated above, their employers still believe mothers are less committed, then they will likely continue to treat them accordingly with fewer promotions, lower earnings, etc. Multiple perceptual measures of commitment involving not only the respondent’s perceptions, but also those of their colleagues, in combination with measures of respondents’ effort and productivity would be helpful in resolving several of the issues raised earlier in this paper. Third, the cross-sectional nature of this study means that concerns regarding causal ordering may be raised. Some of the determinants

242

JEAN E. WALLACE

of commitment included in the regression analysis may instead be outcomes of commitment. It is important to note, however, that this model was derived from the factors identified by mothers that they report are assumed by others to be associated with commitment. That is, the time lawyers spend at work is interpreted by others to be an indicator of commitment, regardless of whether working long hours leads one to be more highly committed or being more highly committed leads one to work longer hours. A more accurate model of career commitment requires a longitudinal approach to disentangle the causes from the outcomes, as well as document possible changes in career commitment over life course and career stages. Fourth, it should be noted that the family-friendliness of a work setting, in terms of the arrangements that it offers or its willingness to negotiate them, may not be equally known to all employees. Employees who have actively sought such arrangements may be aware of the policies and practices that exist, whereas other employees may not. Thus the findings that mothers are more likely to report that their employers offer alternate work arrangements and more flexible work hours than non-mothers may be a function of mothers asking about and/or taking advantage of such provisions that non-mothers typically do not pursue. Lastly, it is important to determine whether the factors examined in this study are important in understanding gendered work experiences and outcomes in a broader sense. It was argued at the onset of this paper that regardless of whether the choice or constraint arguments are used to explain gendered career outcomes; both approaches assume that women are less committed to their careers than men because of family responsibilities. The critical question for future research must ask whether the factors associated with working mothers’ commitment account for the gendered differences observed in men and women’s work experiences and career outcomes.

NOTES 1. In both stages of this study, part-time employment is based on respondents’ self reports of their current work arrangements and not derived from their actual work hours. Although work arrangements, such as part-time status, may be related to the actual number of hours worked, in the discussions and analyses that follow, work status (labeled Alternate Work Arrangement in Stage Two) and work hours (labeled Hours at Office and Hours at Home in Stage Two) are considered separately (Tausig & Fenwick, 2001). 2. Thompson et al. (1999) also identify a third component of work-family culture that refers to management’s support and sensitivity to employees’ family responsibilities. Because lawyers in private practice do not have supervisors per se, similar to academics, this component was not examined in this study. 3. Note that 43 women who work in solo practice were excluded from this analysis because the variables tapping the family-friendly workplace are not applicable.

Motherhood and Career Commitment

243

4. It should be noted that age is often used to tap life stage (e.g. Roehling et al., 2001). It is highly correlated with law experience however (r = 0.76) and the variance-inflation factors suggest that this variable in combination with law experience is problematic. Law experience was retained because it is often used to tap career investments that are related to career commitment (Wallace, 1997). 5. Two other measures tapping leaves from law were also entered into each analysis. I re-estimated all of the equations (pooled, with interactions, for mothers only and for non-mothers only) with the amount of leave time taken and then with the number of leaves taken. Regardless of which measure is used, the results are the same (results available from author). There is a significant interaction effect such that if non-mothers take a leave, or spend a longer time on leave, or take a larger number of leaves, they are less committed to their careers, but there is no effect of any of the leave indicators for mothers. The estimates for the other variables included in the analyses are robust and retain their magnitude and significance of effects. 6. Initially, it seemed puzzling that this factor would be important to non-mothers. I thought perhaps the effect of a supportive work environment might depend on whether or not the woman is thinking of having children in the future but has not yet done so. That is, younger childless women are more likely to be still entertaining the idea of motherhood, whereas after a certain age, women are less likely to consider having children. A more family-friendly working environment might be more important to the younger women than those who have decided to forgo motherhood. To explore this idea further, I divided the non-mothers into two groups – those 35 years of age and younger (N = 182) and those 36 years of age and older (N = 115). This cut off was based on the average age at which the mothers in this study had their first child. I ran an interaction coefficient for this variable and found the interaction to be statistically significant. The results show that for the younger non-mothers work-family culture has a strong positive effect on commitment whereas for the older non-mothers there was no effect. 7. I also re-estimated the model by taking into account the age of children for mothers. Those with preschool-aged children and those with school-aged children were dummy coded and those without children were the reference category. Both mothers with preschool-aged children (b = 0.67, B = 0.60, t = 3.97) and school-aged children (b = 0.97, B = 0.49, t = 3.99) were more committed to their careers than non-mothers. It should also be noted that for the model for mothers only, when I took into account the age of the children, it had no significant effect on mothers’ career commitment (b = 0.05, B = 0.05, t = 0.77). 8. It should be noted, however, that the assumption of this model being additive is violated by the significant interaction terms.

ACKNOWLEDGMENTS This study was funded by a research grant from the Law School Admission Council (LSAC). The opinions contained in this paper are those of the author and do not necessarily reflect the position or policy of LSAC. The author wishes to thank Valerie Haines, Fiona Kay and Charles W. Mueller for their constructive comments on earlier drafts of this paper.

244

JEAN E. WALLACE

REFERENCES Arron, D. L. (1989). Running from the Law: Why good lawyers are getting out of the legal profession. Seattle, WA: Niche Press. Bailyn, L. (1993). Breaking the mold: Women, men and time in the new corporate world. New York: Free Press. Bailyn, L. (1997). The impact of corporate culture on work-family integration. In: S. Parasuraman & J. H. Greenhaus (Eds), Integrating Work and Family: Challenges and Choices for a Changing World (pp. 209–219). Westport, CT: Quorum Books. Barnett, R. C., & Gareis, K. C. (2000). Reduced-hours employment: The relationship between difficulty of trade-offs and quality of life. Work and Occupations, 27, 169–187. Becker, G. (1985). Human capital, effort, and the sexual division of labor. Journal of Labor Economics, 3, S33–S58. Becker, G. S. (1991). A treatise on the family. Cambridge, MA: Harvard University Press. Bielby, D. D., & Bielby, W. T. (1988). She works hard for the money: Household responsibilities and the allocation of work effort. American Journal of Sociology, 93, 1031–1059. Bird, C. E. (1996). An analysis of gender differences in income among dentists, physicians, and veterinarians in 1987. Research in the Sociology of Health Care, 13, 31–61. Blair-Loy, M., & Wharton, A. S. (2002). Employees’ use of work-family policies and the workplace social context. Social Forces, 80, 813–845. Brayfield, A. H., & Rothe, H. F. (1951). An index of job satisfaction. Journal of Applied Psychology, 35, 307–311. Brockman, J. (1992). “Resistance by the club” to the feminization of the legal profession. Canadian Journal of Law and Society, 7, 47–92. Budig, M. J., & England, P. (2001). The wage penalty for motherhood. American Sociological Review, 66, 204–225. Chambers, D. (1989). Accommodation and satisfaction: Women and men lawyers and the balance of work and family. Law and Social Inquiry, 14, 251–287. Clark, S. C. (2001). Work cultures and work/family balance. Journal of Vocational Behavior, 58, 348–365. Corcoran, M., Duncan, G., & Ponza, M. (1984). Work experiences, job segregation and wages. In: B. F. Reskin (Ed.), Sex Segregation in the Workplace: Trends, Explanations, and Remedies (pp. 171–191). Washington, DC: National Academy Press. Coser, L. A. (1974). Greedy institutions: Patterns of undivided commitment. New York: Free Press. Epstein, C. F., Saut´e, R., Oglensky, B., & Gever, M. (1995). Glass ceilings and open doors: Women’s advancement in the legal profession. Fordham Law Review, 64, 291–449. Epstein, C. F., Seron, C., Oglensky, B., & Saut´e, R. (1999). The part-time paradox: Time norms, professional life, family, and gender. New York: Routledge. Ettington, D. R. (1998). Successful career plateauing. Journal of Vocation Behavior, 52, 71–88. Felmlee, D. H. (1995). Causes and consequences of women’s employment discontinuity, 1967–1973. Work and Occupations, 22, 167–187. Fox, J. (1991). Regression diagnostics. Sage University Paper Series on Quantitative Applications in the Social Sciences, 79. Newbury Park, CA: Sage. Fried, M. (1998). Taking time: Parental leave policy and corporate culture. Philadelphia: Temple University Press.

Motherhood and Career Commitment

245

Glass, J. L., & Camarigg, V. (1992). Gender, parenthood, and job-family compatibility. American Journal of Sociology, 98, 131–151. Glass, J. L., & Riley, L. (1998). Family responsive policies and employee retention following childbirth. Social Forces, 76, 1035–1401. Golden, L. (2001). Flexible work schedules: Which workers get them? American Behavioral Scientist, 44, 1157–1178. Greenhaus, J. H., & Parasuraman, S. (1999). Research on work, family, and gender: Current status and future directions. In: G. N. Powell (Ed.), Handbook of Gender and Work (pp. 391–412). Thousand Oaks, CA: Sage. Grusky, O. (1966). Career mobility and organizational commitment. Administrative Science Quarterly, 10, 488–503. Hagan, J., & Kay, F. (1995). Gender in practice: A study of lawyers’ lives. New York: Oxford University Press. Holtzman, M., & Glass, J. (1999). Explaining changes in mothers’ job satisfaction following childbirth. Work and Occupations, 26, 365–404. Hull, K. E., & Nelson, R. L. (2000). Assimilation, choice or constraint? Testing theories of gender differences in the careers of lawyers. Social Forces, 79, 229–264. Kalleberg, A. L., & Epstein, C. F. (2001). Temporal dimensions of employment relations. American Behavioral Scientist, 44, 1064–1075. Kanter, R. M. (1972). Commitment and community: Communes and utopias in sociological perspective. Cambridge: Harvard University Press. Kanter, R. M. (1977). Men and women of the corporation. New York: Basic Books. Kay, F. M. (1997). Flight from law: A competing risks model of departures from law firms. Law and Society Review, 31, 301–335. Kay, F., & Hagan, J. (1998). Raising the bar: The gender stratification of law firm capital. American Sociological Review, 63, 728–743. Kessler, D. A. (1997). Professional asphyxiation: Why the legal profession is gasping for breath. Georgetown Journal of Legal Ethics, 10, 455–485. Klerman, J. A., & Leibowitz, A. (1999). Job continuity among new mothers. Demography, 36, 145–155. Leiper, J. M. (1998). Women lawyers and their working arrangements: Time crunch, stress and career paths. Canadian Journal of Law and Society, 13, 117–134. Marks, S. R. (1977). Multiple roles and role strain: Some notes on human energy, time, and commitment. American Sociological Review, 42, 921–936. Marsden, P. V., Kalleberg, A. L., & Cook, C. R. (1996). Gender differences and organizational commitment: Influences of work positions and family roles. In: A. L. Kalleberg, D. Knoke, P. V. Marsden & J. L. Spaeth (Eds), Organizations in America: Analyzing their Structures and Human Resource Practices (pp. 302–323). Thousand Oaks, CA: Sage. Maume, D. J., Jr., & Houston, P. (2001). Job segregation and gender differences in work-family spillover among white-collar workers. Journal of Family and Economic Issues, 22, 171–189. Merton, R. K. (1984). Social expected durations: A case study of concept formation in sociology. In: W. Powell & R. Robbins (Eds), Conflict and Consensus: A Festschrift in Honor of Lewis A. Coser (pp. 262–286). New York: Free Press. Neapolitan, J. (1980). Occupational change in mid-career: An exploratory investigation. Journal of Vocational Behavior, 16, 212–225. O’Neil, R. B., & Greenberger, E. (1994). Patterns of commitment to work and parenting: Implications for role strain. Journal of Marriage and the Family, 56, 101–118.

246

JEAN E. WALLACE

O’Neill, J., & Polachek, S. (1993). Why the gender gap in wages narrowed in the 1980s. Journal of Labor Economics, 11, 205–228. Perlow, L. A. (1995). Putting the work back into work/family. Group and Organization Management, 20, 227–239. Reskin, B. (1993). Sex segregation in the workplace. Annual Review of Sociology, 19, 214–270. Rhodes, S. R., & Doering, M. (1983). An integrated model of career change. Academy of Management Review, 8, 631–639. Roehling, P. V., Roehling, M. V., & Moen, P. (2001). The relationship between work-life policies and practices and employee loyalty: A life course perspective. Journal of Family and Economic Issues, 22, 141–170. Roos, P. A., & Reskin, R. F. (1984). Institutional factors contributing to sex segregation in the workplace. In: B. F. Reskin (Ed.), Sex Segregation in the Workplace: Trends, Explanations, Remedies (pp. 235–260). Washington, DC: National Academy Press. Schwartz, F. N. (1989). Management women and the new facts of life. Harvard Business Review, 67, 65–76. Seron, C. (1996). The business of practicing law: The work lives of solo and small-firm attorneys. Philadelphia: Temple University Press. Seron, C., & Ferris, K. (1995). Negotiating professionalism: The gendered social capital of flexible time. Work and Occupations, 22, 22–47. Tausig, M., & Fenwick, R. (2001). Unbinding time: Alternate work schedules and work-life balance. Journal of Family and Economic Issues, 22, 101–119. Thompson, C. A., Beauvais, L. L., & Lyness, K. S. (1999). When work-family benefits are not enough: The influence of work-family culture on benefit utilization, organizational attachment, and work-family conflict. Journal of Vocational Behavior, 54, 392–415. Thompson, J. A., & Bunderson, J. S. (2001). Work-nonwork conflict and the phenomenology of time: Beyond the balance metaphor. Work and Occupations, 28, 17–39. Waldfogel, J. (1997). The effect of children on women’s wages. American Sociological Review, 61, 209–217. Wallace, J. E. (1994). Exploring why lawyers leave the legal profession: A proposed model of professional commitment. Paper presented at the Annual Learned Societies Conference of the Canadian Association of Sociology and Anthropology, Calgary AB, Canada. Wallace, J. E. (1997). It’s about time: A study of hours worked and work spillover among married male and female lawyers. Journal of Vocational Behavior, 50, 227–248. Wallace, J. E. (2001). Explaining why lawyers want to leave the practice of law. In: J. Van Hoy (Ed.), Legal Professions: Work, Structure and Organization (pp. 117–145). London: Elsevier. Weisman, C. S., Teitelbaum, M. A., Nathanson, C. A., Chase, C. A., King, T. M., & Levine, D. M. (1986). Sex differences in the practice patterns of recently trained obstetrician-gynecologists. Obstetrics and Gynecology, 67, 776–781. Wood, R. G., Corcoran, M. E., & Courant, P. N. (1993). Pay differences among the highly paid: The male-female earnings gap in lawyers’ salaries. Journal of Labor Economics, 11, 417–441. Yakura, E. K. (2001). Billables: The valorization of time in consulting. American Behavioral Scientist, 44, 1076–1095. Zerubavel, E. (1990). Private-time and public-time. In: J. Hassard (Ed.), The Sociology of Time (pp. 168–177). New York: St. Martin’s Press.

THE MARRIAGE ADVANTAGE FOR MEN IN SCIENCE AND ENGINEERING ORGANIZATIONS Corinne Post, George F. Farris and Rene Cordero ABSTRACT Past research has found that married men have advantages in the labor market. In this paper we examine the extent to which the advantage of married men is evident in research and development organizations with different characteristics. Specifically, we examine the extent to which each of the following organizational characteristics enhance or limit the advantages that married men enjoy both in access to favorable work experiences on the job and in the evaluation of their performance: the extent of involvement in basic research, the proportion of funding from corporate sources, the extent of external contracting, the tightness of work schedules, and the use cross-functional teams. Using information from managers to measure organizational characteristics, we find that married men are advantaged in all types of R&D organizations. That married men receive advantages in the labor force, relative to unmarried men or to married and single women, has long been documented by labor economists (e.g. Becker, 1985; Gray, 1997; Kaufman & Uhlenberg, 2000; Korenman & Neumark, 1991, 1992; Lundberg & Rose, 2002) and by sociologists (e.g. Budig & England, 2001; Cole & Zuckerman, 1992; Etzkowitz et al., 2000). Given the Diversity in the Workforce Research in the Sociology of Work, Volume 14, 247–275 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0277-2833/doi:10.1016/S0277-2833(04)14010-7

247

248

CORINNE POST ET AL.

increasing competition among companies for innovation and creativity, however, one might think that companies dependent on research and development (R&D) would be forced to rid themselves of decision-making based on “irrelevant” social or demographic characteristics, such as whether someone is married or male. Further, given the growing number of women in the paid labor force, including women with young children; the increase in the number of dual-earner and even dual-career families; the fact that women in the labor force experience fewer and shorter work interruptions; and the blurring of the division between work and family, one would expect any such advantages to married men to erode, if not disappear. Explanations for the marriage advantage for men have been sought in variations in family characteristics, such as the presence of children or the employment status of their wives (Becker & Moen, 1999; Gray, 1997; Kaufman & Uhlenberg, 2000; Schneer & Reitman, 1993; Stroh & Brett, 1994). A recent analysis that considers such family characteristics, however, found that married men (with or without children and regardless of their wives’ employment status) continue to enjoy above-average advantages, both in terms of having greater access to favorable work experiences and in being given the benefit of the doubt in the evaluation of their performance (Post et al., 2003). In this paper, we go one step further to ask whether married men have advantages in all types of R&D organizations, or whether there are some organizational characteristics that are more likely to foster advantages for married men and some that may mitigate or reduce their advantages. Research and development work varies in a number of ways. Some is thought of as basic or pure research that often has long timelines and uncertain outcomes, while on the other end of the spectrum, development work is closely related to existing products and is often close to commercialization. Basic or pure research tends to be funded centrally in many companies, because it requires a long-term investment. More product-focused development work is often done at the behest of specific organizational units or external customers. In general, the time pressures of project-based or contract work are much greater than research that has less certain outcomes and often takes years to show progress. In both types of work, however, the push for innovation is increasingly requiring that different knowledge bases and functional expertise be brought together in cross-functional teams that can engage in concurrent engineering and new product development. Furthermore, R&D organizations have been changing in the last several decades, in ways similar to those of other businesses. They have reorganized themselves, reduced the number of hierarchical levels, and adopted flatter structures (DiMaggio, 2001; Heckscher, 1995; Kanter, 1997; Powell, 2001). The introduction of high performance work practices (such as increased flexibility and teamwork) has been especially noticeable in firms that rely extensively on

The Marriage Advantage for Men in Science and Engineering Organizations

249

knowledge workers (Osterman, 1995), in part because the increasing complexity of problem solving requires interdisciplinary and cross-functional knowledge sharing and managers in R&D are no longer in a position to be experts in all that their subordinates do (Barker, 1993; Smith, 1997). This shift to what some scholars call the post-bureaucratic organization (Heckscher & Donnellon, 1994) has been evident in science and engineering firms. For example, industrial research and development firms have reduced their involvement in basic research, have relied less on corporate funding and taken on, instead, more client-driven projects (Varma, 2000), have increasingly engaged in cross-functional cooperation, and have increasingly decentralized. Given the changing nature of research and development work and the variations in the structure of work in different R&D environments, there is reason to believe that the historical advantage that married men have received may vary with the type of organizational characteristics and the associated variations in work requirements and demands. In this paper, we look first at the variations in type of R&D work and how those differences might relate to the advantages often available to married men. We then discuss the content of those advantages, specifically the favorability of work experiences that may contribute to enhanced performance and more favorable performance ratings. Then, in a large sample of scientists and engineers from a number of U.S. based, industrial research and development laboratories, we test our hypotheses regarding the generalizability and extent of the advantages available to married men in R&D. Scientists and engineers represent between 5–8% of the United States workforce, depending on how they are defined (National Science Board, 2002; Organization for Economic Cooperation and Development [OECD], 2001), although they are one fifth of the workforce characterized as “knowledge workers.” The U.S. invests 6% of its GDP in knowledge production, the fourth largest investment in the industrialized world (OECD, 2001). Further, nonacademic science and engineering jobs are growing four times faster than the overall U.S. labor force (National Science Board, 2002). With the U.S. increasingly relying on knowledge-intensive work (Judy & D’Amico, 1997), it is fair to say that knowledge workers represent an important intellectual asset to the country and to their employers (Buckley, 2002; Cohen & Levinthal, 1990; von Krogh et al., 2000).

MARRIED MEN’S ADVANTAGE IN THE LABOR MARKET Considerable research in the literature on income differences in the labor market points to married men’s advantages relative to others. Economists have explained

250

CORINNE POST ET AL.

this advantage primarily by way of the specialization of labor inside the marriage unit. That is, they argue that wives, because they are more likely to specialize in child care and home chores, free their husbands from those responsibilities; in turn, husbands and fathers invest more time and energy in paid work outside the home (Becker, 1985; Gray, 1997; Kaufman & Uhlenberg, 2000; Korenman & Neumark, 1991). Economists argue that because organizations reward employees who work harder and better, men will have higher wages than women, given the longer hours men put in, and the lack of interruptions in their careers. Economists also argue that with fatherhood men start working longer hours at work because they are the primary wage-earners in their families. Employers will, in turn, provide them with higher wages, because they perceive husbands and fathers to be more reliable workers given their family’s dependence on the husband’s wages. Indeed, research finds that the hourly wages of men increases when they become fathers (Cappelli et al., 2000; Lundberg & Rose, 2002). While economists view wage inequality as reflecting individual differences in the time and efforts invested in work, discrimination proponents explain differences in wages and career attainment in terms of statistical discrimination (Budig & England, 2001): because women, on average, work fewer hours, are more likely than men to interrupt their careers for childcare, and because employers cannot easily tell which women will and which will not do so, employers have an incentive to hire men and reward them more than women. For example, in an analysis of the allocation of work practices that lead to higher performance, Post et al. (2003) found that the advantage married men enjoy in the evaluation of their performance is at least partially due to the advantage they enjoy in having more access to favorable work practices, even after taking into account differences in work experience, past performance, and educational attainment. Korenman and Neumark (1991) identified another intermediary process explaining the marriage wage premium for men: married men receive higher performance ratings than single men. Further, Friedman and Greenhaus (2000) found that married men have more authority on their job than unmarried men (after controlling for formal authority level) and that their authority advantage enhances married men’s success. Some gender scholars argue, on the one hand, that employers withhold higher wages and other organizational rewards from women and that peers withhold critical job information from them, because women do not conform to the image of the “Ideal Worker” (e.g. Conley, 1998). On the other hand, they argue that work is usually designed in such a way that it becomes difficult for individuals with family responsibilities to fit in, and since women are more likely to carry out family responsibilities than men, this hurts women more than men (Williams, 2000). Thus marriage is an advantage for men more than women, because it allows men to specialize in remunerated work and to accumulate work experience, and

The Marriage Advantage for Men in Science and Engineering Organizations

251

it also confers on men an image of stability and commitment. Further, work itself is designed with an ideal-typical, unencumbered worker in mind, and because married men are usually more likely to be unencumbered by family responsibilities than are married women, the structure of jobs give married men advantages over women. Reskin (2000) and Tsui & Gutek (1999) argue that the association of men with an “Ideal Worker” norm probably occurs non-consciously because it is based on shared cognitive perceptions and on intuitive beliefs about how men fit into the group. We automatically tend to categorize people based on demographic characteristics, because they are easily accessible cognitively. Thus, demographic characteristics, whether in physical appearance (Tsui & Gutek, 1999) or symbolically (Nippert-Eng, 1995), tend to be salient in most social interactions. Furthermore, the salience of demographic characteristics increases when the number of individuals in that category is small (Fiske & Taylor, 1991). For example, being the only woman in a group of workers makes gender salient to all of the participants. In such circumstances, members of the majority group will see more distinctions between themselves and members of the minority group than they will tend to see among themselves. In other words, they will cognitively see more coherence among majority group members than they would in the absence of a minority group member (Deaux, 1995). Demographic characteristics can also become salient in allocation decisions about who should be given valuable resources and in evaluation decisions about competence or worth (Ridgeway, 2001). This is especially likely when evaluation criteria are most ambiguous, informal, and when there are few rules to guide decision-making. Under such circumstances, cognitive schemas affect both self-perceptions and the perceptions of others (Fiske & Taylor, 1991; Fox, 1992).

ORGANIZATIONAL CHARACTERISTICS AND FAVORABILITY BIASES TOWARD MARRIED MEN In this section we explore and suggest ways in which organizational characteristics are likely to amplify or reduce inequality among demographic groups at work. Those hypotheses are summarized in Table 1. Basic Research Usually, a focus on basic research reflects an organization’s intent to advance the state of knowledge in a field, rather than to directly add to its bottom-line. Basic research labs tend therefore to be more isolated from the business side of an organization and are not directly in contact with external customers. Companies

252

CORINNE POST ET AL.

Table 1. Summary of Hypotheses for the Favorability Bias Toward Married Men. Favorability Bias Toward Married Men in Work Contexts and in Performance Evaluations Basic research Low High

− +

Corporate funding Low High

+ −

External contracting Low High

− +

Teamwork Low High

+ −

Schedules Loose Tight

− +

Notes: +: Favorability bias is stronger; 0: Favorability is not affected by the organizational characteristics; −: Favorability bias is weaker.

with heavy investments in basic research expect that the discoveries made in the labs will ultimately lead to commercial applications that will then show a profit, but the timelines tend to be extremely long and the outcomes quite uncertain. In contrast, applied research is directly geared toward developing commercial applications, for example the improvement of an existing process or product to answer customer needs, and the timelines, in contrast, tend to be much shorter and the outcomes much more predictable. While the trend since the mid- 80s and 90s has been for organizations to decrease their direct investment in basic research and increasingly cooperate with nonprofit research organizations, such as universities (Varma, 2000), recent years have witnessed an increase in basic research investments by for-profit organizations (Miller, 1997; National Science Board, 2002), e.g. in biomedical research. Companies heavily involved in basic research usually have to allow more technical control to their scientists and engineers, because managers in such organizations cannot themselves dictate what needs to be done when the outcomes are neither understood, nor certain. This is especially the case if the organization wants to attract better talent (for an example, see Kidder, 1981).

The Marriage Advantage for Men in Science and Engineering Organizations

253

When the criteria for evaluating competence is so ambiguous and difficult to measure, organizations are likely to rely more on past experience and past practice in order to gain some predictability over the returns on their investment. One way to do so is to hire and reward those people who conform to the prototypes or schema of what a competent scientist or engineer should be (Kanter, 1977; Pfeffer et al., 1976). And, because married men have been associated in the past with competent performance in research and development and because cognitive schema of the “ideal scientist” or “ideal engineer” are still strongly embedded in such organizations (Conley, 1998; Kidder, 1981; Kunda, 1991), we would expect that they would be given advantages over women or single men in organizations that place a strong emphasis on basic research. Thus, we expect married men’s advantages to be enhanced in companies heavily involved in basic research, and mitigated in companies with very little involvement in basic research. Corporate Funding The extent of funding provided to work units by corporate sources is one indicator of the degree of a firm’s organizational centralization. Traditional research cultures are typically more centralized, and also corporate-funded, while newer research cultures tend to be more decentralized (Varma, 2000). When projects are funded by corporate sources, they tend to be more insulated from the short-term influences of customer and business needs and more likely to be long-term and provide scientists and engineers with more autonomy (Smith-Doerr, 2004). More centralized organizations are also likely to be more bureaucratic, because they are not as subject to short term influences. Hence, they can develop rules that are more visible and more consistent in their application. For example, in a field study of women in engineering, McIlwee and Robinson (1992) found that in bureaucratic organizations, especially those that were government contractors, female engineers were likely to be promoted. In contrast, in a high-tech, innovative, non-bureaucratic organization, they found that women felt less confident about their skills, and had fewer chances of promotion into management, because of the very masculine interactional style that dominated the work in that company. Thus, we would expect that in firms that are more centralized (i.e. recipient of more corporate funds and, thus, also more bureaucratic) the advantage of married men in access to favorable work contexts and to favorable performance ratings will be reduced compared to their treatment in decentralized organizations (i.e. low corporate funding). External Contracting Organizations funded through external R&D contracts have to compete for customer attention and often have to meet much tighter deadlines than is true

254

CORINNE POST ET AL.

in organizations with central funding. Because their funds come from external sources, budgets directly impact a division’s bottom line, and thus are better monitored. Thus, the autonomy of scientists and engineers in deciding what projects to pursue is usually limited, because they need to respond to specific business and customer needs. In organizations that rely heavily on external contracting, scientists and engineers are also more likely to work on a diversity of projects, since the work is by definition project-based. Thus, they might have more opportunities to develop relationships with sponsors, gain leadership experiences, and be exposed to business-oriented managers. Where budgets are stricter, that is, where budgets are based on business requirements such as staying within the boundaries of an external contracting budget and developing a reputation for high quality work, competitive pressures will be high. In turn, these competitive pressures make it costly for organizations to discriminate on the basis of irrelevant characteristics. To the extent that the involvement of the organization in external contracting creates competitive pressures that force the organization to reduce costs and to increase performance, then they may not be able to afford giving advantages to married men unless it is related to true performance. Conversely, if low involvement in external contracting results in less competitive pressures, then one would expect married men’s advantages to persist. Teamwork Recent changes in science and engineering firms include an increased use of team and project work (Smith, 1997). This move is, in part, a reaction to the increased interdependence between knowledge-workers (Powell, 2001) and the assumption that teamwork will make it easier to capitalize on the knowledge bases of each employee to promote organizational creativity. One would expect that an organization that genuinely promotes teamwork will also show a willingness to democratize and decentralize decision-making processes. Although an extensive body of work documents a positive relationship between high-performance work practices (a component of which is teamwork) and firm-performance (Appelbaum et al., 2000; Konrad & Mangel, 2000), the research evidence suggests that Whites and men may react negatively when they are placed in teams with non-Whites or women. For example, in a study of the effects of work group composition, Tsui et al. (1992) found that as the diversity in the composition of the team increased, Whites and men were less likely to express commitment to the organization and more likely to be absent. Similarly, in a case study of a large multi-national software development firm, Daday and Burris (2002) found that women engineers are much more satisfied than their male counterparts with teamwork arrangements and that men are especially dissatisfied with teamwork when they feel that their autonomy is threatened or that there is a lack of clear leadership. Women found

The Marriage Advantage for Men in Science and Engineering Organizations

255

teamwork particularly useful, because it gave them a “better understanding of what is going on.” Thus, teamwork may affect Whites and men differently than non-Whites or women. Given that married men have in the past been able to count on advantages on the job, we would expect that the advantages of married men will be reduced in organizations that use teamwork and project teams extensively. Work Schedules Firms operating on tight deadlines experience more crisis situations (unforeseen problems that need to be resolved immediately), which in turn create opportunities to display leadership behaviors. Also, tight deadlines suggest that there is little autonomy in deciding what work to do, since the focus is always on projects that need to be completed immediately. In science and engineering, rewards go to those who demonstrate an ability to work under tight and changing deadlines, and to “put fires out,” rather than to those who foresee, plan for, and prevent potential future “fires” (Fletcher, 1999). In addition, biases and other cognitive shortcuts – such as those that affect the allocation of rewards and recognition – are more likely to be activated and less likely to be suppressed under stress and tight schedules, because there is little time to search for additional information that could reduce bias (Fiske & Taylor, 1991). Pressure to perform has been found to be associated with a reliance on irrelevant individual characteristics in decision-making processes (Koller & Wicklund, 1988). Therefore, assuming that tight schedules create time pressures that lead to “mindless” decision-making (i.e. decisions that rely on past practice because there is not time to think through new or different ways of doing things), we expect that tight work schedules will enhance the advantages of married men, while loose work schedules will mitigate them. This research is consistent with research on lawyers (Epstein et al., 1999), who because of working long hours and on tight schedules have the opportunity to exhibit a form of professional heroism, i.e. a display or signal of dedication that helps differentiate employees in an environment where objective measures of performance are difficult to come by. Among lawyers, the time pressures of the job also tend to favor men over women and those without family responsibilities over those who have them.

FAVORABLE WORK CONTEXTS One of the ways through which employees gain advantage is by having access to favorable work contexts that are crucial for their career advancement (Moore, 1998). In particular, three work contexts or work practices have been found to

256

CORINNE POST ET AL.

be associated with improving the climate for productivity among scientists and engineers (DiTomaso et al., 1993; Pelz & Andrews, 1976). These include: (a) Technical Control, defined as the ability to influence one’s own work assignments (Badawy, 1995; Connolly, 1983; Sindermann, 1982); (b) Having the Confidence of Others, defined as the belief that the four or five influential people have confidence in one’s abilities (Etzkowitz et al., 2000; McIlwee & Robinson, 1992); and (c) Technical Diversity, defined as the opportunity to work on a variety of different types of assignments (Katz, 1988; Hardy & Clegg, 1999; Peters, 1988). Technical Control In a science and engineering environment, those who are able to address uncertainty and aid in problem solving can position themselves in power-yielding, high-visibility, and crucial projects (Connolly, 1983), while at the same time making it likely that others will find themselves in low-impact, low-visibility projects. Because technical control is presumably used to create a climate favorable to one’s work, a scientist’s or engineer’s ability to influence decision-makers is associated with higher performance and with being thought of as more innovative. Specifically, Pelz and Andrews (1976) found that, in general, scientists and engineers with more autonomy perform better than those with little autonomy. Furthermore, influence is crucial for engineers and scientists within industry, because they often need to secure support from individuals outside their direct, formal line of authority (Badawy, 1995). Having the Confidence of Others Field studies of engineers highlight the importance of self-confidence (Dryburgh, 1999) and of self-promotion (e.g. McIlwee & Robinson, 1992) for getting good assignments and gaining credibility in research and development. Self-confidence increases with scientists’ and engineers’ socialization into their occupational roles (Katz, 1988) and also with age and educational attainment (Pelz & Andrews, 1976). In industrial science and engineering, sources of power for influencing work goals arise in part from one’s assertiveness and self-promotion (Sindermann, 1982). Technical Diversity Although having autonomy to choose one’s own projects may enhance performance, scientists or engineers who work in isolation are not necessarily likely

The Marriage Advantage for Men in Science and Engineering Organizations

257

to perform well. Instead, research has found that when scientists are exposed to a diversity of experiences and projects, they are deemed more useful to the organization (Katz, 1988; Pelz & Andrews, 1976). When they are exposed to a variety of technical assignments, to administrative duties, and to other “world views,” scientists and engineers are more useful to their organization, because they can see the “bigger picture”; they can incorporate administrative constraints into their research goals and come up with innovative products or solutions (Pelz, 1976). By doing administrative or service work, they develop weak ties (Granovetter, 1973) that are helpful in gaining influence, in stimulating innovation, and in developing opportunities for furthering their goals. Exposure to a diversity of experiences depends in part on one’s accumulated human capital. Presumably those with longer tenure in their profession or organization, those with Ph.D.’s, and those who have demonstrated achievement in their field will be more likely to be assigned to administrative or service work in addition to research.

PERFORMANCE IN RESEARCH AND DEVELOPMENT Measuring performance within research and development is an extremely difficult challenge for most R&D organizations. In most cases, timelines tend to be long, the work tends to be group or project-based, and the outcomes require the combination of contributions from a number of different people in different fields of expertise. The efforts to measure research performance by the use of “objective” measures such as publications or patents does not mitigate the problem of defining who made what contribution, because publications and patents, too, are almost always the product of research teams that draw from the contributions of many people over time. Further, research teams are sometimes organized hierarchically, with those in senior positions gaining credit for work that occurs within their laboratories or under their supervision, and hence, even “objective” measures have ambiguity associated with them. Thus, in many R&D environments, competence is evaluated primarily through reputational procedures that combine the evaluations of many people who may be in a position to observe the work of a given scientist or engineer. Further, some who work in science and engineering prefer to stay in more technical positions, while others desire promotions to managerial positions. Hence, they are often evaluated on either technical skills or on promotability into management. In this paper, we have chosen to use a measure based on the evaluations of managers who are familiar with the work of the scientists and engineers in their

258

CORINNE POST ET AL.

organizations and have limited our evaluation to how the scientists or engineer is assessed in terms of their readiness for promotion into management. While this is not a perfect measure, it does capture an important aspect of reputation, and it is especially useful for the purposes of this paper, because one would expect that if there is an advantage given to married men in performance evaluations, it is most likely to occur in the assessments of their leadership capabilities for managerial roles. Because it is difficult to determine who is really competent or not in science and engineering jobs, prototypes of the ideal scientist and ideal engineer have emerged. The prototypical scientists and engineers are usually White men, remote from the intricacies of everyday life, devoted entirely to work, without time for family or social life (Morse, 1995). These prototypes facilitate the cognitive, pre-conscious processes that are triggered in the allocation of key opportunities and rewards, and in the perception and attribution of success. Apparent conformity to the prototype marks one’s status as a member of the in-group or of the out-group (Tajfel, 1981; McIlwee & Robinson, 1992; Perlow, 1997). Having one’s status associated with that of the in-group provides an additional advantage when the status of the group is linked to the perception of competence or merit (Ridgeway, 2001). Further, members of the in-group, i.e. those who conform to the prototypes of competence, benefit from having their success attributed to personal characteristics (Heider, 1958), and from having their failures attributed to causes outside of their control (Jackson et al., 1993). Perceptions of competence are reinforced through everyday practices by giving greater access to favorable working conditions to some groups over others (Bourdieu & Passeron, 1977; Lareau, 2000). Thus, those who receive advantages in everyday work practices may also hoard opportunities for similar others (Tilly, 1998). Hoarding opportunities can occur both directly, for example, by inviting someone like oneself to participate in an important project, or through indirect mechanisms such as sharing job information or career development opportunities with those like oneself but not with others. That is, rather than discriminating against out-group members, which is obvious and illegal, in-group members may practice favoritism towards each other. Favoritism toward in-group members, when it is practiced in a context where members of that group fit a prototype of competence is a mechanism that is much more subtle and invisible than is overt discrimination. Past research has found that married men have advantages in research organizations (Cole & Zuckerman, 1992; Etzkowitz et al., 2000; Post et al., 2003; Xie & Shauman, 1998), but this research has not examined whether these advantages are more likely in some kinds of organizations than others. That is the goal of this paper.

The Marriage Advantage for Men in Science and Engineering Organizations

259

METHODS Sample and Data We collected data as part of a larger investigation of the work experience of scientists and engineers. The sample was drawn from 24 multi-billion dollar U.S.-based firms in several industries. The overall response rate for the survey was 55% for employees, although it averaged 65% in 18 of the 24 companies. A total of 3,163 full-time employees responded to the survey, and 690 managers provided ratings of the respondents (with about 80% overlap between the employees who returned surveys and the ratings provided by managers). We administered separate surveys to managers and to scientists and engineers; in all companies, most managers who were asked participated in the survey. We compared data on managers who rated employees who did not return the survey and on employees who did not receive ratings from managers: we found no major systematic biases in the respondents that would dramatically affect the results. Performance ratings were collected for research rather than administrative purposes, which others have found enhances the reliability and validity of measures (Wherry & Bartlett, 1982). The survey was conducted in 1990–1992. Sample size after listwise deletion was 2287 (reflecting the 20–25% of surveys for which managers ratings were not received). Ninety five percent hold a Bachelor’s degree, and one third a Ph.D. Twenty percent were born outside the U.S. The mean work experience is about fourteen years. All respondents work full-time.

MEASURES The Measures of Favorable Work Contexts Based on Pelz and Andrews (1976) and on subsequent research as previously discussed, we measure three key work contexts or practices associated with a productive work climate: technical control, having the confidence of others, and technical diversity. The Appendix provides the details of the items included in the three work practices. The Measure of Promotability Following Pelz and Andrews (1976), we asked managers to rate professionallevel scientists and engineers with whose work they were personally familiar in terms of their promotability into management (should an opening be available).

260

CORINNE POST ET AL.

Most were rated by more than one manager. The managers’ ratings were reported independently of respondents’ surveys to eliminate the risk of common method variance between respondents’ reports of their work context, education, and past performance. We measured the ratings using an 11-point scale ranging from 1 to 99, with 10-point increments. On average, each employee received 2.8 ratings and each manager rated 9.9 employees. The rating we use is an average of ratings from different managers (70% of the respondents had ratings from more than one manager).

The Measures of Organizational Characteristics After listwise deletion, the sample of managers from the 28 company locations in our study (some of the 24 companies were surveyed at two locations) is 473 managers. Each manager was asked to provide information on the type of work done in his or her work unit. Because not all questions were relevant to all managers and some may not have known the answers to questions about the organization as a whole, not everyone answered all of these questions. On average, for each company location, we have usable information from 61% of the managers who participated in the survey, with a range from 25 to 88% of managers who filled out the survey. We averaged responses from managers in each company to construct measures of company characteristics. Basic Research Managers evaluated on a scale of 1 (not at all) to 7 (very great) the extent to which their unit spent time working on basic research, both along established and along new lines (alpha = 0.86). For each company, we calculated the mean of the manager ratings. Because of the unequal sample size of managers in each company, and because we could not assume homogeneity of variance in the means, we used one-way analyses of variance and pairwise comparisons of the averaged manager ratings, by company, to determine which companies had overall higher or lower ratings on the extent of time spent on basic research and guided the creation of two dummy variables, “high [involvement in] basic research” and “low basic research.” Scientists and engineers in companies who had high ratings on this measure received the value of 1 for “high basic research” and those with low ratings received the value 1 for “low basic research.” Having an average involvement in basic research was the reference category. Corporate Funding Managers were asked to rate how much of their work unit’s budget came from corporate sources on a scale of 1 (none) to 7 (all). Based on their responses,

The Marriage Advantage for Men in Science and Engineering Organizations

261

two dummy variables were similarly constructed for “high” and “low” corporate funding. Having average corporate funding was the reference category. External Contracting Managers’ were also asked to what extent their work unit was involved in contract R&D (1 = not at all, 7 = very great) and about how much of their budget came from consulting contracts with sources external to the company (1 = none and 7 = all); both items were averaged into an index (alpha = 0.67). The procedure described above was also used with this index to categorize companies into two dummy variables: “high external contracting” and “low external contracting.” Having average external contracting was the reference category. Teamwork Managers assessed whether the following statements were true: “my work unit utilizes project teams extensively” and “my work unit utilizes cross-functional teams extensively.” Possible responses ranged from 1 (not at all) to 7 (very great) (alpha = 0.79). The two ensuing dummy variables are “high teamwork” and “low teamwork.” Using teamwork to an average extent was the reference category. Work Schedules Managers rated the statement “my work unit has to meet tight schedules” on a range from not at all true (1) to very true (7). Based on their responses, two dummy variables were constructed each with a value of 1 or 0. Working on average schedules was the reference category.

Control Variables We control for scientists and engineers’ human capital and past performance because they may affect a manager’s perception of their promotability and because we want to assess differences based on family structure independently of differences in education and experience. We asked the respondents to report the number of patents or patent applications they have had over the last five years. Because both patents and publications are highly skewed, we follow the practice of Pelz and Andrews (1976, p. 274) and measure both by 1 + loge (raw score + 0.5). We control for having a Ph.D. (0 = no PhD, 1 = holds a Ph.D.), which is an especially relevant status characteristic in research and development. We also control for work experience and experience squared (to account for the deterioration of returns to knowledge over time). Finally, because we know that Whites enjoy advantages in the workplace both in the evaluation of their

262

CORINNE POST ET AL.

performance (Smith et al., 2002) and in the allocation of favorable work contexts (DiTomaso et al., 2002), we control for race (0 = non-White, 1 = White.) We use a categorical approach to take into account the simultaneous effects of gender, being married, having dependents, and having an employed spouse: we controlled for 9 demographics categories with effect coded dummy variables (value of 1 if the respondent is a member of the group, −1 if the respondent is a member of the omitted group, and 0 otherwise.): (i) Single women without dependents; (ii) Single women with dependents; (iii) Married women with an employed husband without dependents; (iv) Married women with an employed husband with dependents; (v) Married men whose wife is not employed, without dependents; (vi) Married men whose wife is employed, without dependents; (vii) Married men whose wife is not employed, with dependents; (viii) Married men whose wife is employed, with dependents; and (ix) Other men and women, including non-married men with dependents, and women whose husbands are not employed (these groups individually had too few respondents to treat them as separate groups). Single men without dependents form the reference category. The sample sizes for the different categories are listed in Table 2. Finally, we include controls for respondents’ work preferences that were found by Pelz and Andrews (1976) to be highly predictive of effective performance. Style of Approach is the preference for broad thinking vs. detail work, and Internal Motivation is the extent to which the scientist or engineer prefers to have technical autonomy rather than to work according to a supervisor’s needs (regardless of the extent to which they have autonomy in their own job). These measures are also Table 2. The Gender and Family Structure Categories.

Non married women, no dependents Non married women, with dependents Dual earner women, no dependents Dual earner women, with dependents Non-married men, no dependents (reference group) Single earner men, no dependents Dual earner men, no dependents Single earner men, with dependents Dual earner men, with dependents Other men and women Non-married men with dependents n = 22 Single earner women, no dependents n = 12 Single earner women with dependents n = 2 Total

N

%

184 29 162 142 258 249 554 333 340 36

8 1 7 6 11 11 24 15 15 2

2287

100

The Marriage Advantage for Men in Science and Engineering Organizations

263

described in more detail in the Appendix. We centered all continuous independent variables around their means to avoid distortions in the intercept.

Analyses Dependent variables were the measures of workplace favorability (i.e. technical control, others’ confidence, and technical diversity) and the promotability ratings. We conducted hierarchical multiple regression analysis, with human capital (i.e. education, experience, past performance), work preferences, demographic categories, and dummy variables for organizational characteristics in a first step, and the interaction terms between organizational characteristics and being a married man in the second step. Because we were specifically interested in determining how general married men’s advantage was likely to be, we restricted the interaction terms to only this characteristic.

RESULTS: PERSISTENCE OF MARRIED MEN’S ADVANTAGE ACROSS ORGANIZATIONAL CHARACTERISTICS It appears from Tables 3A through 3E and Tables 4A through 4E, that married men consistently have better access than average to favorable work contexts and to favorable performance evaluations, and that their advantage does not differ in different organizational contexts, suggesting that the marriage advantage for men holds in a variety of work settings. Interactions between being a married man and working in the specified organizational characteristics were not statistically significant in any of the cases for which their respective main effects were significant.

DISCUSSION, LIMITATIONS, CONCLUSION Those who have researched the implications of organizations’ structural changes in the new economy suggest that in the context of flatter organizations, decentralized decision-making, and increasing reliance placed on teamwork women might be positioned to enjoy advantages because they are more likely to have the interpersonal skills associated with success in this new environment (Fondas, 1997). The implication of this assumption is that men’s advantage will be subdued as the workplace remunerates women for the skills they are more likely to possess

264

Table 3. Effects of Organizational Characteristics on Access to Favorable Work Practices. Technical Control (A) Basic research involvement (n = 2287) Human capital, work style preferences, white, U.S. born control variables Family structure & gender categories: – married men Low involvement in basic research High involvement in basic research

No main effects of basic research

 R2 Adj. R2

 R2 Adj. R2

0.07

0.26∗∗∗ 0.14∗∗ −0.13∗ 0.092∗∗∗

0.09

0.28∗∗∗ 0.15∗ −0.08 0.092∗∗∗ −0.02 −0.09 0.000 0.09

0.13∗ 0.08 −0.16 0.073∗∗∗ 0.01 0.11 0.000 0.07

No main effects of corporate funding

Technical Diversity

0.04∗∗∗ −0.01 −0.12∗∗∗ 0.071∗∗∗

0.05∗∗∗ −0.00 −0.11∗∗∗ 0.071∗∗∗ −0.02 −0.02

0.06

0.001 0.06

0.04∗∗∗ 0.05∗∗∗ −0.07∗∗∗

0.04∗∗∗ 0.06∗∗∗ −0.07∗∗∗

0.083∗∗∗

0.08

0.083∗∗∗ −0.01 −0.00 0.000 0.08

CORINNE POST ET AL.

 R2 Married men in low CF firms (decentralized) Married men in high CF firms (centralized)

0.15∗∗ 0.08∗ −0.11 0.073∗∗∗

 R2 Married men in low BR firms Married men in high BR firms

(B) Extent of corporate funding (n = 2287) Human capital, work style preferences, white, U.S. born control variables Family structure & gender categories: – married men Low levels of corporate funding High levels of corporate funding

Others’ Confidence

Technical Control (C) Extent of external contracting (EC) (n = 2287) Human capital, work style preferences, white, U.S. born control variables Family structure & gender categories: – married men Low levels of EC High levels of EC  R2 Married men in low EC firms Married men in high EC firms  R2 Adj. R2 (D) Use of teamwork (n = 2287) Human capital, work style preferences, white, U.S. born control variables Family structure & gender categories: – married men Low levels of teamwork High levels of teamwork  R2 Married men in low teamwork firms Married men in high teamwork firms  R2 Adj. R2

0.27∗∗∗ 0.16∗∗ 0.12

0.26∗∗∗ 0.16∗ 0.12

0.096∗∗∗

0.096∗∗∗ 0.01 −0.01

0.09

0.26∗∗∗ −0.02 0.24∗∗∗ 0.091∗∗∗

0.09

Others’ Confidence

No main effects of external contracting

0.001 0.09

0.25∗∗∗ 0.03 0.20∗ 0.091∗∗∗ −0.07 0.06 0.001 0.09

No main effects of teamwork

Technical Diversity

0.04∗∗∗ 0.04∗∗∗ 0.03

0.05∗∗ 0.05∗∗∗ 0.02

0.067∗∗∗

0.067∗∗∗ −0.01 0.01

0.06

0.001 0.06

0.04∗∗∗ 0.07∗∗∗ 0.06∗∗∗

0.04∗∗∗ 0.07∗∗∗ 0.07∗∗∗

0.050∗∗∗

0.050∗∗∗ 0.01 −0.02

0.04

The Marriage Advantage for Men in Science and Engineering Organizations

Table 3. (Continued )

0.001 0.04

265

266

Table 3. (Continued ) Technical Control (E) Schedule tightness Human capital, work style preferences, white, U.S. born control variables Family structure & gender categories: – married men Loose schedules Tight schedules  R2 Married men in loose schedules firms Married men in tight schedules firms

0.089∗∗∗

0.08

0.27∗∗∗ 0.07 −0.16∗ 0.089∗∗∗ 0.06 −0.00 0.000 0.08

No main effects of schedule tightness

Technical Diversity

0.05∗∗∗ −0.06∗∗∗ −0.02 0.055∗∗∗

0.05

0.05∗∗∗ −0.06∗∗∗ 0.00 0.055∗∗∗ −0.00 −0.05∗∗ 0.003∗ 0.05

CORINNE POST ET AL.

 R2 Adj. R2

0.28∗∗∗ 0.10 −0.16∗

Others’ Confidence

The Marriage Advantage for Men in Science and Engineering Organizations

267

Table 4. The Effects of Organizational Characteristics on Performance Ratings. Promotability (A) Basic research involvement (n = 2287) Human capital, work style preferences, white, U.S. born control variables Family structure & gender categories: – married men Low involvement in basic research High involvement in basic research  R2 Married men, low involvement BR firms Married men, high involvement BR firms  R2 Adj. R2 (B) Extent of corporate funding (n = 2287) Human capital, work style preferences, white, U.S. born control variables Family structure & gender categories: – married men Low levels of corporate funding High levels of corporate funding

2.63∗ 3.50∗∗∗ 2.29

2.50 3.88∗∗ 0.47

0.187∗∗∗

0.187∗∗∗ −0.68 4.05

0.18

0.001 0.18

No main effects of corporate funding

 R2 Married men in low CF firms Married men in high CF firms  R2 Adj. R2 (C) Extent of external contracting (n = 2287) Human capital, work style preferences, white, U.S. born control variables Family structure & gender categories: – married men Low levels of external contracting High levels of external contracting

No main effects of external contracting

 R2 Married men in low EC firms Married men in high EC firms  R2 Adj. R2 (D) Teamwork use (n = 2287) Human capital, work style preferences, white, U.S. born control variables Family structure & gender categories: – married men Low levels of teamwork High levels of teamwork

3.02∗∗ −5.17∗∗ −1.64

3.83∗∗∗ −3.89 0.16

268

CORINNE POST ET AL.

Table 4. (Continued ) Promotability  Married men in low teamwork firms Married men in high teamwork firms R2

0.186

 R2 Adj. R2

0.018

0.001 0.18

2.81∗ −0.97 4.40∗∗

2.51∗ −1.76 3.97∗

(E) Schedule tightness (n = 2287) Human capital, work style preferences, white, U.S. born control variables Family structure & gender categories: – married men Loose schedules Tight schedules  R2 Married men in loose schedules firms Married men in tight schedules firms  R2 Adj. R2

0.185∗∗∗

0.18

0.186 −2.30 −2.99

0.185∗∗∗ 1.50 1.09 0.000 0.18

because they are more valued in the current economy. The primary goal of this paper was to determine whether the advantages in access to favorable work contexts and evaluations of married men held under different organizational contexts. In the present analyses, we classified each of the 28 company locations in which scientists, engineers, and their managers were sampled as exhibiting high, average, or low involvement in basic research, dependence on corporate funding, engagement in external contracting, use of team and project work, and tightness of work schedules. The data clearly show that overall married men’s advantage in access to favorable work contexts and to favorable performance ratings does not differ across a variety of organizational settings, independently of any differences among scientists and engineers in education, work experience, past performance, work preferences, and after taking into account demographic factors (race and ethnicity). While not the primary focus of this analysis, we found several organizational characteristics were associated with greater or lesser access to favorable work contexts and to performance ratings. The results from these analyses show that scientists and engineers working in organizations with low involvement in basic research, low levels of corporate funding (i.e. decentralized), limited external contracting, or in either those that embrace or those that shun teamwork have more access to favorable work experiences that are associated with higher performance. Conversely, scientists and engineers working in organizations that

The Marriage Advantage for Men in Science and Engineering Organizations

269

are heavily involved in basic research, highly centralized (i.e. high corporate funding), or either in those working under very loose or very tight deadlines are likely to have less access to favorable work experiences. Scientists and engineers working in organizations that are minimally involved in basic research are more likely to be evaluated as promotable into management. This makes sense because organizations that do basic research have fewer scientists who want to leave the “bench” and go into management, whereas in organizations that do limited basic research, many of the scientists and engineers are likely to want to be promoted into management jobs (Bailyn, 1982). Scientists and engineers are also more likely to be evaluated as promotable into management in organizations that operate under tight schedules, but they are less likely to be evaluated as promotable in organizations that make minimal use of teamwork. What the present analyses tell us is that while access to favorable work contexts and performance evaluations themselves vary in different organizational settings, the advantages that married men have in both access to favorable work experiences and in performance appraisal are consistent across a wide variety of organizational settings. Thus, it appears that there is little reason to believe that the structural changes undergone by corporations in the wake of the new economy will equalize married men with others in their access to favorable work contexts or to favorable performance evaluations. While noteworthy, these results must be interpreted in light of the usual limitations of survey research. Specifically, the organizational characteristic variables are constructed using self-reported information from managers. To be sure, the measures of organizational characteristics are also quite crude; two of the five characteristics are measured with single-items. We rank organizations as centralized, decentralized, or average using the relative ratings given by managers. A more objective way of categorizing organizations would be to compare managers’ averaged ratings of company characteristics to an objective measure (for example comparing managers’ rating of their company’s degree of centralization to an objective measure of centralization). The difficulties in measuring company characteristics are by no means exclusive to this study. In the study of high performance work practices, it is the norm to rely on manager’s subjective assessments; oftentimes, managers assess the existence or non-existence of specific practices (Appelbaum et al., 2000). In this study, not only do we rely on assessments from multiple managers in each company, but our measures are also based on Likert scales rather then on yes/no questions, which allows for a more refined measure of organizational characteristics. Another limitation to the study is that we do not include a variable that assesses the extent to which scientists and engineers are interested in being promoted. While the inclusion of such a variable in the analyses might help explain why some demographic groups of scientists

270

CORINNE POST ET AL.

and engineers receive higher or lower promotability ratings from their managers, we do not expect that the moderating effect of organizational characteristics on the relationship between belonging to a group and promotability assessments would be altered by controlling for interest in promotion. The size and scope of the data under scrutiny give these findings much generalizability to other settings and populations, but especially to large companies with research and development investments. Future research could expand the validity of these findings by testing them with knowledge workers outside of research and development laboratories and in other types of science and engineering settings, e.g. in non-traditional, start-up, innovative, or smaller firms.

REFERENCES Appelbaum, E., Bailey, T., Berg, P., & Kalleberg, A. L. (2000). Manufacturing advantage: Why high-performance work systems pay off. Ithaca, NY: Cornell University Press. Badawy, M. K. (1995). Developing managerial skills in engineers and scientists: Succeeding as a technical manager. New York: Van Nostrand Reinhold. Bailyn, L. (1982). Resolving contradictions in technical careers or “What if I like being an engineer?” Technology Review, Nov/Dec: 40–47. Barker, J. R. (1993). Tightening the iron cage: Concertive control in self-managing teams. Administrative Science Quarterly, 38, 408–437. Becker, G. S. (1985). Human capital, effort, and the sexual division of labor. Journal of Labor Economics, 3, S33–S59. Becker, P. E., & Moen, P. (1999). Scaling back: Dual-earner couples’ work-family strategies. Journal of Marriage and the Family, 61, 995–1007. Bourdieu, P., & Passeron, J.-C. (1977). Reproduction: In education, society, and culture. London: Sage. Buckley, J. E. (2002). Rankings of full-time occupations by earnings, 2000 [Electronic version]. Monthly Labor Review, 125, 46–57. Budig, M. J., & England, P. (2001). The wage penalty for motherhood. American Sociological Review, 66, 204–225. Cappelli, P., Constantine, J., & Chadwick, C. (2000). It pays to value family: Work and family tradeoffs reconsidered. Industrial Relations, 39, 175–198. Cohen, W. M., & Levinthal, D. A. (1990). Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarter, 35, 128–152. Cole, J. R., & Zuckerman, H. (1992). Marriage, motherhood, and research performance in science. In: H. Zuckerman, J. R. Cole & J. T. Bruer (Eds), The Outer Circle: Women in the Scientific Community (pp. 157–170). New Haven, CT: Yale University Press. Conley, F. K. (1998). Walking out on the boys. New York: Farrar, Strauss, and Giroux. Connolly, T. (1983). Scientists, engineers, and organizations. Belmont, CA: Brooks/Cole Engineering Division. Daday, G. K., & Burris, B. H. (2002). The effects of teaming-structure on race, ethnicity, and gender differences in a high-tech corporation: A case study. Paper presented at the meeting of the American Sociological Association, Chicago, IL.

The Marriage Advantage for Men in Science and Engineering Organizations

271

Deaux, K. (1995). How basic can you be? The evolution of research on gender stereotyoes. Journal of Social Issues, 51, 11–20. DiMaggio, P. (2001). The twenty-first century firm. Princeton, NJ: Princeton University Press. DiTomaso, N., Farris, G. F., & Cordero, R. (1993). Diversity of the technical work force: Rethinking the management of scientists and engineers. Journal of Engineering and Technology Management, 10, 101–127. DiTomaso, N., Post, C., Smith, D. R., Farris, G. F., & Cordero, R. (2002). Three forms of inequality: Advantage, the absence of advantage, and disadvantage. Paper presented at the meeting of the American Sociological Association, Chicago, IL. Dryburgh, H. (1999). Work hard, play hard: Women and professionalization in engineering – adapting to the culture. Gender and Society, 13, 664–682. Epstein, C. F., Seron, C., Oglensky, B., & Saut´e, R. (1999). The part-time paradox: Time norms, professional life, family and gender. New York: Routledge. Etzkowitz, H., Kemelgor, C., & Uzzi, B. (2000). Athena unbound: The advancement of women in science and technology. Cambridge: Cambridge University Press. Fiske, S. T., & Taylor, S. E. (1991). Social cognition. New York: McGraw-Hill. Fletcher, J. K. (1999). Disappearing acts: Gender, power, and relational practices at work. Cambridge, MA: MIT Press. Fondas, N. (1997). Feminization unveiled: Management qualities in contemporary women. The Academy of Management Review, 22, 257–282. Fox, M. F. (1992). Gender, environmental milieu, and productivity. In: H. Zuckerman, J. R. Cole & J. T. Bruer (Eds), The Outer Circle: Women in the Scientific Community (pp. 188–204). New Haven: Yale University Press. Friedman, S. D., & Greenhaus, J. H. (2000). Work and family – allies or enemies: What happens when business professionals confront life choices? New York: Oxford University Press. Granovetter, M. S. (1973). The strength of weak ties. American Journal of Sociology, 78, 1360–1380. Gray, J. S. (1997). The fall in men’s return to marriage: Declining productivity effects or changing selection? Journal of Human Resources, 33, 481–504. Hardy, C., & Clegg, S. R. (1999). Some dare call it power. In: S. R. Clegg & C. Hardy (Eds), Studying Organizations: Theory and Method (pp. 368–387). Thousand Oaks, CA: Sage. Heckscher, C. (1995). White collar blues: Management loyalties in an age of corporate restructuring. New York: Basic Books. Heckscher, C., & Donnellon, A. (1994). The post-bureaucratic organization: New perspectives on organizational change. Thousand Oaks, CA: Sage. Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley. Jackson, L. A., Sullivan, L. A., & Hodge, C. N. (1993). Stereotype effects on attributions, predictions, and evaluations: No two social judgments are quite alike. Journal of Personality and Social Psychology, 65, 69–84. Judy, R. W., & D’Amico, C. (1997) Workforce 2020: Work and workers in the 21st century. Indianapolis: Hudson Institute. Kanter, R. M. (1977). Men and women of the corporation. New York: Basic Books. Kanter, R. M. (1997). Rosabeth Moss Kanter on the frontiers of management. Cambridge, MA: Harvard Business Review Press. Katz, R. (1988). Managing professionals in innovative organizations. Cambridge: Ballinger Publishing Company. Kaufman, G., & Uhlenberg, P. (2000). The influence of parenthood on the work effort of married men and women. Social Forces, 78, 931–949.

272

CORINNE POST ET AL.

Kidder, T. (1981). The soul of a new machine. Boston, MA: Brown Little. Koller, M., & Wicklund, R. A. (1988). Press and task difficulty as determinants of preoccupation with person descriptors. Journal of Experimental Social Psychology, 24, 256–274. Konrad, A. M., & Mangel, R. (2000). The impact of work-life programs on firm productivity. Strategic Management Journal, 21, 1225–1237. Korenman, S., & Neumark, D. (1991). Does marriage really make men more productive? Journal of Human Resources, 26, 282–307. Korenman, S., & Neumark, D. (1992). Marriage, motherhood, and wages. Journal of Human Resources, 27, 233–255. Kunda, G. (1991). Engineering culture: Control and commitment in a high-tech corporation. Philadelphia: Temple University Press. Lareau, A. (2000). Home advantage: social class and parental intervention in elementary education. Lantam, MD: Rowman & Littlefield. Lundberg, S., & Rose, E. (2002). The effects of sons and daughters on men’s labor supply and wages. Review of Economics and Statistics, 84, 251–268. McIlwee, J. S., & Robinson, J. G. (1992). Women in engineering: Gender, power, and the workplace culture. Albany, NY: State University of New York Press. Miller, J. A. (1997). Discovery research re-emerges in DuPont. Research-Technology Management, January-February, 24–28. Moore, G. (1998). A climbing kit for ambitious engineers. Engineering Management Journal, August, 164–166. Morse, M. (1995). Women changing science: Voices from a field in transition. New York: Plenum Press. National Science Board (2002). Science and engineering indicators. Arlington, VA: National Science Foundation. Nippert-Eng, C. E. (1995). Home and work: Negotiating boundaries of everyday life. Chicago: University of Chicago Press. Organization for Economic Cooperation and Development (2001). OECD science, technology, and industry scoreboard 2001 – Toward a knowledge-based economy. Washington, DC: OECD. Osterman, P. (1995). Work/family programs and the employment relationship. Administrative Science Quarterly, 40, 681–700. Pelz, D. C. (1976). Problem solvers vs. decision makers. In: D.C. Pelz & F. M. Andrews (Eds), Scientists in Organizations: Productive Climates for Research and Development (Rev. ed.). Ann Arbor, MI: Institute for Social Research, University of Michigan. Pelz, D. C., & Andrews, F. M. (1976). Scientists in organizations: Productive climates for research and development. Ann Arbor, MI: Institute for Social Research, University of Michigan. Perlow, L. A. (1997). Finding time: How corporations, individuals, and families can benefit from new work practices. Ithaca, NY: ILR Press. Peters, T. J. (1988). A skunkworks tale. In: R. Katz (Ed.), Managing Professionals in Innovative Organizations (pp. 433–441). Cambridge: Ballinger Publishing Company. Pfeffer, J., Salancik, G. R., & Leblebici, H. (1976). The effect of uncertainty on the use of social influence in organizational decision making. Administrative Science Quarterly, 21, 227–245. Post, C., DiTomaso, N., Smith, D. R., Farris, G. F., & Cordero, R. (2003). Gender, family & work contexts: Effects on manager rated performance among scientists & engineers. Paper presented at the meeting of the Academy of Management Association. Seattle, WA. Powell, W. (2001). The capitalist firm in the twenty-first century: Emerging patterns in western enterprise. In: P. DiMaggio (Ed.), The Twenty-First Century Firm (pp. 33–68). Princeton, NJ: Princeton University Press.

The Marriage Advantage for Men in Science and Engineering Organizations

273

Reskin, B. F. (2000). The proximate causes of discrimination: Research agenda for the twenty-first century. Contemporary Sociology, 29, 319–329. Ridgeway, C. L. (2001). The emergence of status beliefs: From structural inequality to legitimizing ideology. In: J. T. Jost & B. Major (Eds), The Psychology of Legitimacy (pp. 257–277). Cambridge: Cambridge University Press. Schneer, J. A., & Reitman, F. (1993). Effects of alternate family structures on managerial career paths. Academy of Management Journal, 36, 830–843. Sindermann, C. J. (1982). Winning the games scientists play. New York: Plenum Press. Smith, D. R., DiTomaso, N., Farris, G. F., & Cordero, R. (2002). Favoritism, bias, and error in performance ratings of scientists and engineers: The effect of power, status, and numbers. Sex Roles, 45, 337–358. Smith, V. (1997). New forms of work organization. Annual Review of Sociology, 23, 331–339. Smith-Doerr, L. (2004). Women’s Work: Gender equality vs. hierarchy in the Life sciences. Boulder, CO: Lynne Rienner. Stroh, L. K., & Brett, J. M. (1994). Dual-earner dads vs. traditional dads: Can we account for differences in salary progression? Paper presented at the meeting of the National Academy of Management, Dallas, TX. Tajfel, H. (1981). Human groups and social categories. Cambridge: Cambridge University Press. Tilly, C. (1998). Durable inequality. Berkeley, CA: University of California Press. Tsui, A. S., Egan, T. D., & O’Reilly, C. A. (1992). Being different: Relational demography and organizational attachment. Administrative Science Quarterly, 37, 549–579. Tsui, A. S., & Gutek, B. A. (1999). Demographic differences in organizations: Current research and future directions. Lanham: Lexington Books. Varma, R. (2000). Changing research cultures in the U.S. Science, Technology and Human Values, 25, 395–416. von Krogh, G., Nonaka, I., & Nishiguchi, T. (2000). Knowledge creation: A source of value. New York: St. Martin’s Press. Wherry, R. J., & Bartlett, C. J. (1982). The control of bias in ratings. Personnel Psychology, 35, 521–551. Williams, J. (2000). Unbending gender: Why family and work conflict and what to do about it. Oxford University Press. Xie, Y., & Shauman, K. A. (1998). Sex differences in research productivity: New evidence about an old puzzle. American Sociological Review, 63, 847–870.

274

CORINNE POST ET AL.

APPENDIX Measures of Favorable Work Contexts Technical Control 5-Item Scale (alpha = 74) Think of one person (other than yourself) who has most weight in the choice of your work goals. To what extent do you feel you can influence this person in recommendations or decisions concerning your technical goals? (1 = not at all; 7 = completely) In your present job to what extent do you have freedom to carry out your own ideas? (1= none/slight; 7 = utmost) In your present job to what extent do you have responsibility for initiating new activities? (1= none/slight; 7 = utmost) Does your immediate supervisor provide considerable freedom for subordinates to explore, discuss, and challenge ideas on their own? (1 = not at all; 7 = completely) To what extent do you feel you can influence the direction of the group/project team? (1 = not at all; 7 = very much) Having the Confidence of Others Consider four or five people in your organization whose opinion of you matters most to you. In general, how much confidence do you believe they have in your abilities? (1 = no confidence; 7 = complete confidence) Technical diversity Proportion of total time spent working on projects, which is spent on non-major projects. (For a typical month, what percent of your working time would you estimate you spend on your major project? On other projects?) (0 = all work time spent on major project; 1 = all work time is spent on non-major project)

Work Preferences Internal Motivation (alpha = 0.72) How much importance would you personally attach to each of these in a job (disregarding whether or not your present job provides them): (1= slight or none; 7 = utmost) To have responsibility for initiating new activities To have freedom to carry out my own ideas To work where top management is deeply committed to the goals of my work

The Marriage Advantage for Men in Science and Engineering Organizations

275

To be consulted by my direct supervisor in his/her decisions To make full use of my present knowledge and skills Style of Approach I prefer to map out broad features of important new areas, leaving detailed study to others (1= not at all; 7 = very great)

Organizational Characteristics Extent of Involvement in Basic Research (alpha = 0.86) To what extent does your work unit spend time on basic research along established lines? (1 = not at all; 7 = very great) To what extent does your work unit spend time on basic research along new lines? (1 = not at all; 7 = very great) Corporate Funding Research and development departments get their financial resources or budgets from various sources. Please think of the sources of finances for your work unit. About how much of your budget comes from corporate? (1= none; 7 = all) External Contracting (alpha = 0.67) Research and development departments get their financial resources or budgets from various sources. Please think of the sources of finances for your work unit. About how much of your budget comes from consulting contracts – external sources? (1= none; 7 = all) To what extent does your work unit spend time on contract R&D? (1 = not at all; 7 = very great) Teamwork (alpha = 0.79) To what extent is it true that your work unit utilizes project teams extensively? (1= not at all; 7 = very great) To what extent is it true that your work unit utilizes cross functional teams extensively? (1= not at all; 7 = very great) Work Schedules To what extent is it true that your work unit has to meet tight schedules? (1= not at all; 7 = very great)

UNDERSTANDING THE EFFECTS OF WORKFORCE DIVERSITY ON EMPLOYMENT OUTCOMES: A MULTIDISCIPLINARY AND COMPREHENSIVE FRAMEWORK Sheryl Skaggs and Nancy DiTomaso ABSTRACT In this paper, we develop a conceptual framework for understanding the impact of workforce diversity on labor market outcomes. We argue that to understand the impact of workforce diversity, we must consider the effects of power (the distribution of valued and scarce resources), status (the relationships among people and groups), and numbers (the compositional effects of the unit), whether in the work group, job, occupation, firm, or society. We then discuss the mechanisms that generate and reproduce these dimensions of inequality and explain how they contribute to everyday practices such as allocation decisions and evaluative processes and ultimately lead to sustained or durable inequality (e.g. labor force outcomes including attitudes, behaviors, and material and psychic rewards).

Diversity in the Workforce Research in the Sociology of Work, Volume 14, 279–306 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0277-2833/doi:10.1016/S0277-2833(04)14011-9

279

280

SHERYL SKAGGS AND NANCY DITOMASO

INTRODUCTION Over the past several decades, important research has emerged examining the effects of workforce diversity on an array of employment outcomes. Although this literature has provided some insight into the broader picture of workplace inequality, a great deal of work remains to more precisely understand the mechanisms through which workforce diversity yields differential consequences within work groups, jobs, occupations, firms and even societies. To date, the literature in sociology, psychology, and management has primarily focused on differences in individual characteristics leading to inequality (Reskin, 2003; Tilly, 1998). While this approach may demonstrate why differences exist in such areas as promotions, pay, training, turnover, job satisfaction and the like, it does little to explain how structural and group processes lead to, as well as perpetuate, unequal outcomes (Reskin, 2003). To overcome the limitations of the current literature, we propose a comprehensive framework for examining these processes by drawing on relevant theory and research from a number of fields. In particular, we focus on the literatures in sociology related to stratification and inequality (e.g. race and sex segregation), in psychology on intergroup relations, and in management on diversity and organizational demography. In sociology (and economics), the tendency has been to frame the key arguments in terms of efficiency vs. discrimination. In psychologically oriented social psychology, the emphasis has been on cognitive processes such as social identity theory, social categorization, or similarity-attraction theories. In sociologically oriented social psychology, expectation states theory has also addressed the impact of long term inequality on processes within groups. The management literature has drawn primarily from social psychology, but has also given attention to compositional issues at the level of the work group (including the top management team) and has developed a theory of organizational demography. Rather than merely providing a review of this literature, our goal is to combine key elements of these various literatures in a way that will allow for improved understanding of how workforce diversity contributes to workplace inequality. We begin this paper by first developing a comprehensive framework that pulls together theoretical conceptualizations and empirical findings from sociology, psychology, and management. We then discuss how these literatures relate to each other and attempt to link the structure of inequality to the mechanisms that contribute to its reproduction. We discuss how these mechanisms work at the level of everyday practices in work groups, firms, or other social units and then discuss how these everyday practices contribute to the reproduction of inequality over time. Finally, we consider some of the methodological challenges of doing research on workforce diversity. We believe such discussions are critical to the process of integrating theory and research in this area. Our framework is provided in Fig. 1.

Understanding the Effects of Workforce Diversity on Employment Outcomes

281

Fig. 1. The Effects of Workforce Diversity on the Generation and Reproduction of Inequality: Power, Status, and Numbers.

THE STRUCTURE OF INEQUALITY Drawing on the work of Sachdev and Bourhis (1984, 1985, 1987, 1991), who in a series of articles attempted to integrate social psychological work on social identity with sociostructural work on group relations, we argue that to explain

282

SHERYL SKAGGS AND NANCY DITOMASO

Table 1. Research Literature on Power, Status, and Numbers.

Sociology

Power (Distribution of Resources)

Status (Relationships Among Groups)

Numbers (Compositional Characteristics of Groups)

Stratification

Expectation states master status Devaluation

Statistical discrimination

Dual and segmented labor markets Politics

Psychology

Management

Social capital –

Theory of rents Competitive advantage Power and opportunity (Kanter)

Gendered/racialized labor force Cultural capital Social identity Social categorization Similarity/attraction System justification Social identity Social categorization Similarity/attraction

Compositional effects Group competition

Contact hypothesis Majority/minority relations

Diversity Group composition Organizational demography Numbers (Kanter)

how inequality gets generated and reproduced, one must understand the structural characteristics of power, status, and numbers. We define power as the distribution of, access to, and control over scarce and valuable resources; status as the relationships of deference between and among groups, and numbers as shorthand for the compositional characteristics of groups (size, proportionality, average characteristics, and so on). Thus, in order to explain inequality, we must understand the distribution of resources, relationships of deference, and the composition or representation of groups within a social unit (whether work group, job, occupation, firm or society). Workforce diversity is an issue only because of its relationship to groups. As such, power, status, and numbers address group, not individual processes. We believe that much of the literature on workforce inequality can be organized within the context of these three broad topics. Table 1 provides a summary organized by power, status, and numbers of key areas of the literature we examine below with regard to workforce diversity.

Power and Stratification Within the sociological literature, the question of how differences between men and women or whites and non-whites lead to differential wages, authority, training, promotions and countless other material or symbolic outcomes continues to serve as a nagging reminder of where research on workforce inequality has fallen short.

Understanding the Effects of Workforce Diversity on Employment Outcomes

283

The explanations for inequality, however, tend to differ by field. While most economists have typically stressed the importance of human capital investments by individual workers (Becker, 1964), sociologists (often drawing on the work of institutional economists) have developed counter arguments that emphasize the characteristics of firms, industries, and other institutional factors as sources of workforce inequality. This is especially the case as concerns about stratification have joined conceptions of economic sociology and the sociology of institutions (Granovetter, 1985). The work on stratification within sociology over the last several decades that challenged human capital explanations of inequality has been referred to as the “new structuralism” (Baron & Bielby, 1980, 1984; England, 1982). From the view of sociologists, the fact that organizational and industry characteristics explained how people were allocated to different types of jobs implied a selection process that differentiated individuals on the basis of irrelevant personal characteristics. That is, it implied processes of discrimination. In contrast, economists argued that if hiring decisions made by employers were truly discriminatory, they would prove to be quite costly for the employer (Aigner & Cain, 1977; Arrow, 1973; Becker, 1957, 1964). Thus, it was assumed that employers would derive no benefit from using discriminatory screening techniques. Hence, research in this context pitted efficiency explanations with those about discrimination, and sociological work has examined the extent to which efficiency drives decisions about who gets what types of jobs and with what level or types of rewards. Subsequent empirical research has provided little support for efficiency claims regarding workforce inequality, i.e. that people are allocated to jobs based only on what they contribute to productivity. If discrimination is the basis for workplace inequality, though, that does not explain how it comes about. Tilly’s work (1998) on “durable inequality” provides a conceptual integration of the research on workforce inequality that identifies the four main processes by which “categorical inequality” (e.g. by race, ethnicity, gender or class) are linked over time to unequal outcomes. He argues that exploitation (primarily on the part of employers) and opportunity hoarding (primarily on the part of nonelites) “cause the installation of categorical boundaries within organizations, while emulation [the spread of similar kinds of organizational arrangements] and adaptation [the incorporation of similar categories into new areas] reinforce those effects” (Tilly, 1998, p. 114). Much of the research on inequality, especially in sociology and economics, over the last several decades has called attention to the actions of both employers and coworkers in the creation and reproduction of inequality. Our discussion will concentrate especially on the processes of opportunity hoarding and exploitation as the bases for inequality.

284

SHERYL SKAGGS AND NANCY DITOMASO

For instance, in a study of personnel practices within a Washington state agency, Bridges and Nelson (1989) provided evidence suggesting the importance of employer decisions and coworker influence on wage setting practices above and beyond pure market forces. In a more recent study, Tomaskovic-Devey and Skaggs (1999), examining productivity differences for a sample of North Carolina businesses, found no evidence of variation in female or minority aggregate productivity compared to white males, meaning that the lower wages of women and minority workers were not justified on the basis of differences in productivity. Interestingly however, their findings indicate that rather than employers increasing their profits by paying women and minorities less, as one might expect, the true beneficiaries of women’s lower earnings were white and male workers. These results suggest that, contrary to economic arguments, cost and efficiency are not necessarily the key mechanisms driving employment decisions and outcomes. Instead, research demonstrates that employer or coworker stereotypes and bias contribute substantially to disparate outcomes for women and minorities when such members comprise the numerical minority in jobs, work groups, and firms (see Reskin et al., 1999 for a review). Social Closure and Opportunity Hoarding The recognition that coworkers have the ability to hoard opportunities by excluding or limiting access to jobs or training has led to renewed interest in processes of social closure (Murphy, 1988; Parkin, 1979; Weber, 1968, p. 342). One type of social closure involves members of the dominant group, generally white males, who preserve their advantage by creating and changing organizational rules in such a way that women and minorities are unqualified for or denied equal access to high status positions (Reskin, 1988; Tomaskovic-Devey, 1993). Members of privileged groups maintain their status, among other ways, by differentiating themselves from members of subordinate groups. Differentiation provides a basis for discrimination and becomes a critical strategy in hiring and promotions. Privilege is maintained by mechanisms that physically (sex and race) or symbolically (language, educational and training credentials) set members of the dominant and subordinate groups apart. The ability of privileged members to exclude those from subordinate groups depends on their power within a given workplace – generally derived from numerical representation, seniority, or authority level (Bridges, 1982; Reskin & Roos, 1990; Williams, 1987). Membership in the dominant group enables privileged individuals to gain access to vital information about intra-firm job opportunities, production changes and training programs, all of which tend to be synonymous with higher pay and authority.

Understanding the Effects of Workforce Diversity on Employment Outcomes

285

Examining such processes, Tomaskovic-Devey and Skaggs (2002), in a study of job level sex segregation, found that rather than earnings differentials primarily being about the devaluation of women’s work (i.e. the usual way we think about discrimination taking place), as had often been argued (Baron & Newman, 1990), they were more closely tied in this study to a labor process that limited or excluded access to on-the-job training for typically female jobs. Although the study did not allow for a direct test of social closure, the results do suggest that such processes operate in firms, however obscurely, and they can have real implications for employment outcomes. In fact, it is frequently the subtle practices (e.g. control of informal training) carried out by employees within the dominant group (either with the consent of employers or as a group effort to protect positions of status or authority) that help maintain job segregation. In Kanter’s classic study of men and women in the corporation, she argued that the efforts of management to control uncertainty, the difficulty of evaluating performance, the need for efficient communication, and concerns about loyalty all contributed to “homosocial reproduction” (1977a, p. 48), meaning that those in management positions tended to hire people like themselves, that is, from the same social backgrounds and with a similar physical appearance and presentation of self. Despite the entry of women and minorities into many workplaces, concerns are still raised about their ability to “fit in.” One of the implications of this is that those who differ in social or demographic characteristics have more difficulty developing useful social networks with people who can help them in their careers (DiTomaso et al., 1988; Ibarra, 1993, 1995). Social closure may also be a mechanism available to subordinate groups in competition with other subordinate groups. For example, Waldinger (1997) found that low-skilled immigrant workers have some power to provide access to jobs for their family and friends, even when it is against the employers’ wishes or interests. Waldinger (drawing from Parkin, 1979) labels this “usurpationary closure” or “exclusionary closure.” By refusing to train workers who are not part of their group, current employees can create costs to employers who refuse to hire those recommended by the employees. Further, employers who fail to hire those recommended by current employees may find that friction emerges across employee groups and that this, too, can lead to unproductive outcomes. Thus, employers who use employee networks to fill vacancies may save themselves the cost of ads and the uncertainty of hiring those about whom one has little information, but doing so is at the cost of excluding those who are not part of the networks of current employees (Waldinger, 1995). In contrast, hiring friends and family members of current workers adds a dimension of social obligation among workers that can have positive outcomes for the employer, although there may be a downside as well if employees then form cliques that may work against employer interests (Waldinger, 1997).

286

SHERYL SKAGGS AND NANCY DITOMASO

Social closure processes may also be located at a more structural level supported by key organizational actors. This is consistent with Cockburn’s (1991) research in which she found that rules established by white male employers were often used by male employees to secure their advantage. For instance, at the retail firm she studied, male managers and office workers developed a strong social world outside of work from which their female colleagues were excluded. The men frequently gathered after work and on weekends (e.g. meeting for drinks after work, attending conferences and participating in an all-male golfing fraternity) exchanging valuable industry and firm-specific information, thus, making it difficult for women to stay informed of insider news and informal rules. By hiring women to do work similar to men, but in different positions, with different titles, and at lower wages, employers were able to contribute to a structure whereby male employees can informally exclude women from information networks and thus, impede their opportunities for advancement (see also Hacker, 1983; Kanter, 1977a; Reskin & Roos, 1990). In this case, the power differences in the positions of men and women lead to cumulative advantage or disadvantage. The same processes are at work as well between whites and non-whites. Since rewards, promotions, authority and other outcomes tend to be directly linked to specific jobs and workgroups, when data permit, it is preferable to examine social closure processes at the job or workgroup level. In other words, it is at the job or work group level that those able to engage in social closure have the most to lose or gain. Nonetheless, Weeden (2002) makes a strong argument for how social closure can and does occur at the occupational level as well. In particular, Weeden (2002) found evidence of the shaping of occupational earnings structures through the use of institutionalized closure practices such as licensing, formal education, and voluntary certification. In this case, dominant group advantage is protected by the creation and enforcement of institutionalized restrictions that limit access to certain occupations (also see Abbott, 1988). Gatekeepers provide the means to carry out such practices and to preserve the advantages of similar others. Given the level of occupational sex segregation that still exists, the implications for material and symbolic outcomes are profound. In summary, then, social closure processes that contribute to hoarded opportunities can result from the actions of coworkers who are members of the dominant group, from the actions of subordinates who are in competition with other groups, from institutional factors (such as in the process of rule making), and can be facilitated by employers as well as by coworkers. Exploitation When economists claim that employers have no incentives to discriminate, they are often making the assumption that the incentive employers have for efficiency (and

Understanding the Effects of Workforce Diversity on Employment Outcomes

287

hence, for “non-discrimination”) makes exploitation both irrelevant and unlikely. When sociologists claim that employers participate in discrimination, however, they are often assuming that discrimination by employers is for the purposes of exploitation, while discrimination by coworkers in the form of opportunity hoarding is a different and unrelated process. The relationship between the two processes, however, are brought together in Sorensen’s (2000) analysis of inequality as a theory of “rents” or unearned profit taking at the expense of others. Drawing from neo-classical economics, Sorensen defines rent (2000, p. 1536) as “payments to assets that exceed the competitive price or the price sufficient to cover costs and therefore exceeding what is sufficient to bring about the employment of the asset.” Like neo-classical economics, Sorensen recognizes that the ability of property holders to derive rent from their property depends on their ability to control supply. Otherwise, those who control productive assets are able to generate only “quasi rents,” meaning rents that last only until someone else is able to offer a better alternative through competition or innovation in the form of new product development or new business models (Sorensen, 2000). The ability of the owners of productive assets to derive rents from their property, therefore, depends both on their ability to protect their property rights and their claims to the rent derived from their ownership or control over property broadly defined and also depends on their ability to keep those who directly produce the product or service from capturing a larger share of the rent for themselves by limiting the supply of labor or skill (what Sorensen labels as “composite rents” when the value of this skill depends on asset specificity within a particular firm or organization). Hence, according to Sorensen (2000), a theory of exploitation in the form of rents calls attention to the inherent conflict and the antagonist interests of those who compete over the ability to control productive assets and the distribution of rents derived from those assets. Both types of competition require collective action, often in conjunction with protection from the state. While there are many ways that those who own or control productive assets attempt to protect themselves from competition or innovation (regulations, tariffs, advertising, and such), in order to capture a larger share of the rent generated from their property also requires efforts to use the state to destroy the property claims of others, especially the property claims of workers or occupational groups who attempt to limit access to a larger labor supply or to readily available training or skill (e.g. seniority, licensing, or union contracts). Further, in order for inequality to be reproduced over the long term, the property rights and distributive claims of privileged or dominant groups need to become embedded in social institutions (Granovetter, 1985) through laws, customs, and ideologies about merit and deservedness. For this reason, even when institutional protections are available to subordinate groups, they may still be at a disadvantage, because as

288

SHERYL SKAGGS AND NANCY DITOMASO

Lin (2001) argues in his analysis of social capital, both agents and institutions act more predictably and more consistently for the benefit of groups with power and resources. Dominant groups often can use their resources to build on and protect their wealth accumulation, including by passing along the benefits of their wealth accumulation through the development of either competence or credentials or both on the part of their offspring (Sorensen, 2000, p. 1548). Challenges to exploitation and to opportunity hoarding are possible only if subordinate groups are sufficiently politicized to conceive of alternative ways of organizing society. As part of the “class struggle,” therefore, various groups of workers must engage in collective action and also must use the power of the state to protect their claims to property rights in their jobs or to maintain a monopoly of access to the training or skill required to do their jobs (e.g. from licensing or limiting access to training). Professional workers often undertake collective action to prevent others from access to doing their work and often purposely limit the access to professional training (Abbott, 1988). Less skilled workers attempt to use laws that guarantee property rights in jobs, e.g. through rights to seniority and through the enforcement of contracts. Thus, inequality is an outcome of the struggle over claims to property rights and to the distribution of rents associated with ownership or control over productive assets. Both dominant and subordinate groups must use collective action in order to capture rents over the long term, and they often must use the power of the state to institutionalize their claims or rights. Subordinate groups, especially, attempt to limit the supply of labor by union contracts, restricted access to training, and mechanisms such as licensing and state regulations. In attempting to restrict the supply of labor, they also frequently engage in social closure and opportunity hoarding, often along the lines of race/ethnicity, gender, and class categories (Tilly, 1998). Dominant groups, therefore, also engage in activities to try to destroy or undermine the ability of subordinate groups to capture a larger share of rents (Sorensen, 2000). In doing so, they often attempt to undermine mechanisms of social closure or opportunity hoarding, not for the purpose of creating fairness, but for the purpose of eliminating or destroying the monopoly rents captured by some workers. Indeed, Sorensen (2000) argues that in the last several decades, capitalists in the U.S. have been able to undermine the rights of workers, including those who have in the past been protected by unions or the welfare state and those who had gained a monopoly, with state support, over the certification and use of their skill. Thus, workers often use opportunity hoarding and social closure that contributes substantially to race/ethnic, gender, and class inequality in order to protect themselves from exploitation by capitalists, while capitalists attempt to destroy the protections of the state that redistribute rents to workers or that limits

Understanding the Effects of Workforce Diversity on Employment Outcomes

289

their access to a larger supply of labor. For capitalists, such actions both create more “efficiency” and greater exploitation. As Sorensen explains: . . . nothing guarantees that a competitive wage moves a worker above the poverty line. The rent destruction in the labor markets, except perhaps for the highest paid managers, leaves a labor market more flexible and more fluid, for fewer groups have anything to protect. The result is less structure, meaning less positional inequality, but more inequality overall. Thus, while greater homogenization overall may not have resulted from these recent trends, the destruction of rents in the labor market has created a labor market with fewer structural supports for the returns to labor. . . . Nothing guarantees that efficient labor markets create good lives. Rents are required in modern society to provide decent standards of living for the poorest part of the population. These rents are provided from the state in the form of income support and other welfare goods (p. 1553).

The analysis of exploitation as the competition over rents derived from ownership or control of productive property, however, leaves one issue unexamined (at least in Sorensen’s analysis). That is, destroying distinctions in the labor market that protect the rights or property claims of some workers at the expense of others may create a “closer match between actual individual productivity and wage level” (Sorensen, 2000, p. 1552), but it also may remove distinctions that had in the past made it difficult for workers to organize across race/ethnicity, gender, or class lines. Thus, it may be that employers acted differently depending on the level of political organization among employees. As such, they may have fostered divisions among workers on the basis of categorical distinctions (e.g. by pitting strike breakers from one categorical group against strikers from another categorical group) in order to “divide and conquer” whenever they thought workers had the social movement and organizational strength to demand concessions from employers through the power of state support. Otherwise, when worker power seemed to be waning, employers may have sought to destroy the organizational and legal protections of workers by creating fewer distinctions among workers and more leveling, i.e. more “efficiency.” In doing so, however, they also may subject themselves to more conflict and to more efforts at renewed collective action. Thus, the ability to exploit or to hoard opportunities requires more than structural position or collective action. It also requires that inequality becomes part of the fabric of everyday life. That is, as Tilly (1998) suggests, the categorical distinctions that are made in the labor market and reflected in organizational forms must become part of the taken-for-granted solutions to familiar problems that get reproduced and supported from many institutional sources. In other words, exploitation and/or opportunity hoarding must come to seem as reasonable or natural outcomes of the distribution of merit or competence. While there is always a tendency for politicization and mobilization of subordinate groups, dominant

290

SHERYL SKAGGS AND NANCY DITOMASO

groups endeavor to head off such collective action by creating explanations for inequality that reinforce the notion that the current distribution of resources is due to merit or competence (e.g. Herrnstein & Murray, 1994). As such, neither requires the use of force or chicanery. Expectation states theory (Ridgeway, 2001), in fact, predicts that the distribution of resources comes to be associated with merit or worth (i.e. comes to seem consistent with the economists’ arguments about efficiency or human capital). In other words, institutional structures that support inequality have to be supported by legitimating myths (Jackman, 1994; Sidanius & Pratto, 1999) or ideologies that come to be accepted by both dominant and subordinate groups. Those doing research on workplace inequality have been trying to understand how these processes come together, for example, by examining the connection between macro and micro processes and by trying to identify the mechanisms that link categorical or structural inequality to outcomes. This research has endeavored to incorporate research on social psychology into theories of inequality, and as such have drawn both on expectation states theory and on the growing research on cognitive psychology. Noteworthy in this regard has been recent work by Reskin (2003) and DiMaggio (1997).

Status Effects and Interactional Processes From the social psychology literature in sociology, workplace inequality is seen as a reflection of status beliefs that are dispersed and become accepted through everyday social interactions. Status beliefs, according to Ridgeway (2001), are derived from cultural beliefs about individuals based on certain defining characteristics such as sex or race. It is through these beliefs that some groups come to be held in higher esteem and prestige than others (Webster & Foschi, 1988). These assumptions are then transferred throughout society through interactional processes. Not only do members of the esteemed group come to accept and act in likeness to their preferred status, but cross interactions among the various group members help reinforce the status ranking overall. This occurs because subordinate groups come to accept the same status beliefs as the dominant or favored group members. Status beliefs are linked to workplace outcomes, because those with high status are also assumed to possess superior skills and competence (Ridgeway, 2001; Webster & Foschi, 1988), which then translates into greater opportunities, rewards, and access to authority. As such, employers come to use these beliefs as the basis for hiring and promotion decisions, managers and supervisors employ them in salary determinations, and advantaged employees exploit them to protect their high status. Pettigrew (1979) argues that this provides favored groups with the “benefit of the

Understanding the Effects of Workforce Diversity on Employment Outcomes

291

doubt.” In contrast, groups with lower status are excluded from such opportunities or rewards, and over time suffer the consequences of not having developed skill or competence as a result. Status beliefs, therefore, provide the basis for stereotypes that, in turn, limit or prohibit the access of subordinate groups (e.g. defined by race/ethnicity or gender) to high status occupations and jobs, and to the training that may make them candidates for such jobs. Interactional processes combined with structural factors, therefore, provide the means for ensuring that when members of the dominant group are given a distinct advantage or influence over members of the subordinate group, individuals from both groups come to see the advantage or influence as deserved and legitimate (Ridgeway, 2001). Thus, women may accept lower pay or less prestigious positions because they believe that men are more qualified or deserving (Major et al., 1984). Likewise, men may come to demand higher wages, authority, and/or power based on their status beliefs (Stuhlmacher & Walters, 1999). In this case, employers (who most likely also subscribe to conventional status beliefs) may find it difficult to challenge these expected outcomes and rewards (Ridgeway, 1997). According to Ridgeway (1997), changes in male/female representation in jobs (or race/ethnic representation) are likely to be fostered by a combination of interactional processes and structural factors (e.g. organizational practices and policies, job sorting processes, etc.). For instance, when male workers are no longer able to optimize their outcomes in certain jobs, it is not unusual for men to look for other alternatives, creating a mass departure by males from the job and an influx of females (Reskin & Roos, 1990; Tsui et al., 1992; Wharton & Baron, 1987). If, under these circumstances, the search fails to generate improved employment opportunities, male workers may attempt to resist the entry of females and are likely to engage in social closure processes in their jobs to protect their status. Male workers, for example, may successfully exclude females by fostering organizational changes in job requirements, hiring, and/or promotion practices that make it harder for women to qualify, gain access, and get rewarded on the job. As already noted, male coworkers may also refuse to help women learn the job and may keep them out of the loop when important job information is shared. In addition, even when women are able to move into male dominated jobs, status beliefs (their own as well as those of their employer) may inhibit their ability to negotiate comparable salary and benefits as those provided to men. Hence, interactional processes contribute to status beliefs permeating organizational structures and may help create, as well as support, a corresponding inequality on the job. Major (1994) argues that similar processes may exist for racial minorities, unless they become politicized and come to see their exclusion from favored positions or rewards as illegitimate. She argues that this has been one of the outcomes for African

292

SHERYL SKAGGS AND NANCY DITOMASO

Americans following the Civil Rights Movement, but that women, on average, are more likely than African Americans to accept their lower salaries as legitimate. The cognitive revolution within psychology has contributed substantially to our understanding of how status gets generated through social interaction. Especially relevant are theories regarding social identity (Tajfel, 1981) and social categorization (Turner, 1987). Studies of social identity have consistently found that the need to maintain positive identity (Brewer, 1979, 1996) contributes to an identification with people who are similar to oneself (the in group) that then leads to favorable treatment of in group members, often at the expense of members of the out group (those not like oneself). In other words, when group identification becomes salient, people tend to make comparisons that provide the basis for status differentiation. Specifically, when groups are recognized as different (i.e. placed in different categories), then the groups also tend to be ranked, with the in group thought of more favorably than the out group. The overall meaning and status of a particular group and the actions of group members can be influenced as well by the larger employment context (Ely, 1994). Research on social identity has found that these cognitive processes are not reciprocal (Brewer, 1979; Sidanius & Pratto, 1999) in the sense that out group members (meaning lower status groups) do not necessarily favor their own group in the same way as in group members do (meaning higher status groups; Tajfel & Turner, 1986). Indeed, those who come to feel that their own group is disfavored may attempt to disassociate themselves from the group (Sidanius, 1993). Hence, in settings where women have an advantage over men, group identification among women may be strong, but when women as a group see themselves as disadvantaged in terms of various outcomes and opportunities, they are likely to develop weakened social relations with other women (Ely, 1994; Major, 1994). The reactions to in group or out group status, though, may differ, depending on which group has higher status in larger contexts. Thus, even when men comprise the minority in a job, some theoretical perspectives (Ridgeway, 2001) have argued that they would still be the “in group” no matter what their proportion in a given workgroup or occupation because men constitute a “master status” in the larger society. The concept of a “master status” goes beyond numbers to the social relationships among the groups in the larger context. For that reason, one has to take into consideration who is responding to whom, and not assume that these are ever generic responses to “pure numbers” (Yoder, 1994). The concept of “master status” in the larger society (or occupation or company) transforms a theory of social psychology to a theory about competition and power (i.e. linking group competition theory to the conceptualization of in groups and out groups). Therefore, in order to formulate meaningful hypotheses about groups that have real status distinctions in the larger society (why else are we interested in

Understanding the Effects of Workforce Diversity on Employment Outcomes

293

gender or race), it does not make sense to develop generic hypotheses about social identity without linking these cognitive processes to the overall status structure of the company, the industry or the occupation (e.g. “male dominated” or “female dominated”) (Baron & Bielby, 1980; Bielby & Baron, 1986; Ely, 1994). The theory of similarity-attraction is closely associated with theories of social identity or social categorization. From the perspective of similarity-attraction, individuals in group settings are believed to favor others they perceive to be more like themselves in terms of beliefs, values and attitudes (Byrne, 1971). Because some characteristics such as sex, race/ethnicity and age are more visible than others, it is reasoned that these delimiters serve as the likely basis upon which individuals come to be attracted to others (Tsui & O’Reilly, 1989). Given the tendency to be attracted to like others, it follows that the more one has the opportunity to associate with like others in the work place, the more chance that one will respond positively. Alternatively, the more one must associate with people unlike oneself, the more likelihood of less favorable reactions. Such reactions, however, may be affected by the existence of master status roles and by the larger context, as described above. Tsui et al. (1992) found, for example, that Whites and men exhibited decreased organizational commitment, increased absenteeism, and higher turnover as the diversity of the workforce increased. Tsui and O’Reilly (1989) also found that supervisors were likely to evaluate performance more negatively and to feel less favorable toward subordinates who were dissimilar to themselves. In the same study, dissimilar subordinates also experienced more role ambiguity. Thus, the cognitive responses to other social groups may depend on the composition of the group, the characteristics associated with the group, and how those characteristics may be changing over time. Cognition, then, is not only in people’s heads, but is linked to social structure and to the everyday processes that link structure to cognition (DiMaggio, 1997). The composition of groups often affects cognition.

Composition and the Effects of Numbers Statistical discrimination provides one account of the link between cognition and group structure. In this case, it is argued that because employers often find it difficult and costly to obtain reliable information about potential employees, they may base employment decisions on average differences (perceived or actual) between groups (e.g. women vs. men or Whites vs. non-Whites). While it may appear that making decisions based on statistical averages is rational and profit-maximizing behavior, research has found that it contributes to inequality, because it operates both as a

294

SHERYL SKAGGS AND NANCY DITOMASO

sorting mechanism that restricts the types of jobs held by women and minorities and also serves as a justification for offering women and minorities lower wages (Aigner & Cain, 1977; Thurow, 1975). Statistical discrimination can influence a number of workplace decisions. For example, when recruitment and training costs are high, employers are likely to seek out employees believed to have the strongest commitment to the company, and statistical discrimination theory suggests that employers’ experiences may influence their perceptions of which employees will be successful and which will have the highest turnover rates. For instance, employers may expect women to have higher turnover rates than men due to family responsibilities. Although such stereotypical characterizations may be true on average, not all women will leave the workforce to care for children, so within group variation is likely to be more pronounced than variation across groups (Bielby & Baron, 1986). Thus, decisions on the basis of statistical discrimination can work against the interests of those women who are highly committed to the paid labor force and can negatively impact their access to opportunities and training that may enable them to earn subsequent rewards. Rather than risk making costly hiring mistakes, one of the mechanisms that employers have been found to use to avoid having to confront the possible challenges to such decisions is to segregate employees by certain distinguishing characteristics such as sex and/or race and ethnicity. As a result, there is a tendency for lower status group members to be relegated to jobs with lower pay and less onthe-job training, thus producing few opportunities for career advancement or skill acquisition. Thus, segregation is one outcome of exclusionary practices, such as social closure or opportunity hoarding, used to preserve advantage. Both employers and coworkers may influence the placement of men and women or Whites and non-Whites who could perform comparable work into distinct jobs with different duties, responsibilities, and opportunities for training and advancement. In other cases, segregation occurs through hiring practices whereby the two groups are employed in different firms with unequal opportunity structures and resources. However, segregation by job title within establishments and firms is the most prevalent mechanism generating differences in pay, promotion, and other career outcomes by gender and race/ethnicity. An extensive literature on work group composition and organizational demography (often defined in terms of easily measurable characteristics such as tenure or education among group members) has developed within management to test the effects of the cognitive, affective, and motivational processes that emerge as a result of compositional differences within groups (Cox et al., 1991; Ely, 1995; Fields & Blum, 1997; Lawrence, 1997; Milliken & Martins, 1996; Pelled et al., 1999; Pfeffer & Davis-Blake, 1987; Watson et al., 1993). This literature has tended to draw primarily from social psychology or from social contact theory to explain

Understanding the Effects of Workforce Diversity on Employment Outcomes

295

workplace outcomes. In a review article of research on work group composition, Milliken and Martins (1996) found that diversity of workgroup composition tends to have a negative effect on affect (people in diverse groups are less likely to feel positively toward each other, Tsui & O’Reilly, 1989), a positive effect on creativity (unless conflict or cultural predispositions prevents the use of diverse ideas from group members) (Cox et al., 1991; Hoffman & Maier, 1961), and affects the patterns of communication (with diverse group members less likely to talk to each other and more likely to talk to people outside the group; Ancona & Caldwell, 1988). They also found symbolic effects, such as that group members respond differently depending on whether there are others like them in the organization with authority and stature (Ely, 1994). Research from a management perspective has also looked at the impact of diversity on networks (Ibarra, 1995) and found that majority and minority group members do not have the same types of networks, nor the same effects from the networks that they have. The effects of diversity, however, depend on the psychological and sociological processes evident within the group. For example, only when stereotypes are activated do prototypical schema affect outcomes, such as selection and promotion decisions (Perry et al., 1994). In general, as might be expected, there are often more positive outcomes from diversity by functional background or education than by race/ethnicity or gender (Pelled et al., 1999), suggesting that the “master status” of individual groups comes into play in the work context as it does in the larger society. Pfeffer and Davis-Blake (1987) came to a similar conclusion in their analysis of the effects of gender composition on the wages of college administrators. They found that the larger the proportion of women in administrative positions, the lower the salaries of both men and women. They argue that as the job itself comes to be associated with women and as women are associated with lesser worth and competence, then there are negative outcomes on the salaries of everyone who holds such jobs, including men as well as women. Social contact theory has a different, but related, dynamic. For example, Kanter’s (1977a, b) work makes a point of drawing from Simmel (1908) regarding the effects of numbers (that the dynamics in a group changes when one goes from 2 people to 3, from 3 to 4, and so on in Simmel’s work up to 9). This has to do, not so much with stereotyping and exclusion, as with attentional capacity and other cognitive resources available to give some people more attention than others and with the possibility of coalition formation that can then shift the power dynamics. The intermediate link between the attentional capabilities inherent in these compositional changes and attitudinal outcomes (e.g. job satisfaction) include processes such as social identity and similarity-attraction. Kanter (1977b) includes both types of processes in her discussion of the interactional outcomes

296

SHERYL SKAGGS AND NANCY DITOMASO

which she argues are the result of the proportional differences between a minority and majority: heightened visibility leading to performance pressures (attentional capabilities), boundary heightening leading to ingroup and outgroup processes (both attention and social identity), and role entrapment leading to stereotyping, statistical discrimination and social identity. Similar to Kanter (1977a, b), Peter Blau (1977) utilizes aspects of Simmel’s work (1908) but develops a more general theory of group size and group relations. In particular, Blau (1977) emphasizes the importance not only of the actual numbers, but also of the relative proportion and status of group members compared to other groups, distinguished by characteristics such as sex, race and ethnicity. He argues that those with higher proportions and higher status in a group will be more internally cohesive among themselves and more likely to be in conflict with those of lower status and lower proportions. In other words, similar to social identity theory from psychology, Blau argues that group distinctions create strong in group/out group tendencies and pave the way for stereotypes and prejudice by majority group members toward minority group members. Blau argues that when the numbers are especially unbalanced, members of the majority will have little knowledge of or experience with minority members, because there will be few opportunities for interaction (Tolbert et al., 1999). In cases where minority and majority numerical representation is more balanced, however, interactions will likely be less constrained and hence are likely to result in more positive intragroup perceptions and behavior (Kanter, 1977b). A study by Sackett et al. (1991) demonstrates the significance of group composition in the evaluation of work performance for distinctions by gender but not by race. The researchers (Sackett et al., 1991) found that when women were less than half of the group members that they tended to be rated lower than men, but when women were more than half of the work group, they tended to be rated more positively than men. They did not find that workgroup composition affected the ratings received by Blacks compared to whites, however, because Blacks were rated less favorably than whites at almost every level of representation. Without examination of representation in the group and how it is changing, it is unlikely the true nature of employment disparity will be fully understood. For instance, similarity attraction theory leads us to expect that various outcomes will be influenced by the complex relationships between subordinates and decision makers (e.g. supervisors and employers), but this may depend on a number of factors, including whether subordinate groups are growing or declining. For members of disfavored groups, increasing their representation in the group may not itself lead to improved material or symbolic outcomes, unless the dynamics of power and status are also taken into effect. Theories of group competition address the dynamics of changing group composition.

Understanding the Effects of Workforce Diversity on Employment Outcomes

297

Group competition theory, grounded in research on racial and ethnic conflict, attempts to explain the relationship between group composition and outcomes by taking into account the competitive nature of group members in defending claims to either material or social resources. From this perspective, it is assumed that group members identify with others like themselves and are disposed to act in ways that will protect and even advance their position, often at the expense of other groups. For example, Blalock (1957) proposed that as the proportion of minority members increases, so too does the perception of threat on the part of the majority, and as threat increases, majority group members are likely to engage in greater levels of discriminatory action toward minority group members. That is, an increasing representation of minority group members provokes majority group members to respond antagonistically. However, as Tolbert and her colleagues (1999, p. 184) point out, neither Blalock (1957) nor Kanter (1977b) specifically differentiate between “the effects of numerical and social dominance [and overlook] the impact of relative group size on the attitudes and behavior of socially subordinate groups.” That is, they examine the effects of composition without considering the effects of power and status as well. That entry of a “different” group may invoke threats among current incumbents about claims to resources makes sense only if the process had been one of favoritism or hoarding of opportunity for similar others (otherwise it would just be competition and not “group” competition). Thus, entry of new or different members into a work group, occupation, or firm has the effect of interrupting the favoritism that the majority or dominant group has taken for granted. That is, the literature on group competition over scarce resources also addresses issues of social closure. Under such circumstances, for example, dominant workers may collectively restrict access to information laden informal networks or withhold training from new and different group members in order to preserve their own advantaged status. Competition theory suggests that social closure against subordinate group members is most beneficial when the threat of competition is high (Blalock, 1957; Cohen, 1998; Olzak, 1992). Accordingly, social closure will be especially likely to arise as the representation of out group members increases, and it may even invoke violence against minority group members because they are “out of place” (Sidanius & Pratto, 1999). Composition itself, however, is not the process by which these outcomes occur. Instead, one needs to introduce an intermediary process, such as social identity or social categorization to explain why majority or minority group members respond the way they do in the presence of others. Research on the social psychology of intergroup relations has a long history of addressing this distinction in what has been called “contact theory” (Brewer & Miller, 1984). Countering the frequently expressed assumption that group conflict could be eliminated if one would “just get

298

SHERYL SKAGGS AND NANCY DITOMASO

people together,” researchers beginning with Allport (1954) endeavored to specify the conditions under which getting people together would reduce or exacerbate conflict. Substantial research has now provided evidence that the structure of group relations determines the attitudes and behaviors of groups toward each other, and that as long as some groups maintain a position of dominance, higher status, and often greater numbers, they will have the ability to maintain advantage at the expense of those in less favored positions.

THE CREATION AND REPRODUCTION OF INEQUALITY Everyday Inequality Because the effects of power, status, and numbers are group based and relational (Tilly, 1998), there is a dynamic involved that requires enactment in everyday experiences as groups come into contact with each other. The mechanisms that link group characteristics to the decisions that affect the allocation of group members to certain structural positions, to the evaluative decisions that provide some groups with the benefit of the doubt and withhold credit to others, and to the processes that create segregation in the structure and composition of workgroups, jobs, occupations, firms, and the society at large are not set in motion at the outset and then never considered again. Instead, they must be reproduced through processes of institutionalization, through monitoring of gatekeepers who are often associated with privileged groups (i.e. the agents and institutions that Lin (2001) describes as working better for privileged groups than for subordinate groups), and through the ability to cumulate advantage through access to and control over financial, social, and cultural resources (Bourdieu & Passeron, 1977; Lin, 2001; Nelson & Bridges, 1999; Willis, 1977). Furthermore, these outcomes are the result of conflict or struggle over scarce resources, over the capturing of rent-producing productive assets, and over the ability to gain support from the state in protecting rights and privileges. While at the level of the society, these processes have been the subject of much social science research and have been well documented over many years, how they work in organizations to link micro processes to macro outcomes is only now receiving substantial attention (Reskin, 2003). For example, DiTomaso et al. (2003) found that U.S. born white men get greater access to favorable work experiences that then contribute to their ability to perform well, and then also get the benefit of the doubt when their performance is evaluated. Similarly, Royster (2003) found that white students were more likely to get steered to jobs than were Black

Understanding the Effects of Workforce Diversity on Employment Outcomes

299

students, who then suffer the consequences of being excluded even from blue collar jobs. Hence, the advantages that privileged groups receive in the search for a job, on the job, and in the rewards of the job contribute overall to their ability to accumulate resources that they can use to improve their social position, pass along to their children, and improve the quality of their lives and those of their communities. While the ideology of their good fortune is often that they “did it on their own,” the reality is that they often did it at the expense of those who were denied such opportunities, who were excluded from jobs or the networks that led to them, and who were discriminated against or exploited in the returns they received from their efforts. Perhaps equally important, those who are disadvantaged or disfavored also are often unable to develop the skills or capabilities that would allow them to gain access to jobs even if they did get the chance to compete for them. Over time, therefore, they may become discouraged, disgruntled, and disengaged, which then can have implications not only for their own lives but also for those around them. As Sorensen explains (2000) with regard to his theory of rents: Those possessing high credentials wish to secure an advantage to their offspring. This is facilitated by making cultural capital relevant for training. However, the very existence of credentials is also important. The superior ability of one’s offspring cannot be secured, but much can be done to secure a valuable credential for the offspring by facilitating access to institutions providing valuable credentials. In the absence of such credentials, less able offspring from high-status backgrounds might have to compete for valued employment with the more able from more humble origins (p. 1548).

Durable Inequality The long term reproduction of inequality in the labor market is associated with processes that go on at the micro level and link to macro structures. Hence, the outcomes occur both personally and professionally, both on the job and off, and through processes where social categories get distributed and reproduced unevenly across structural positions. As Tilly indicates (1998, p. 21): “. . . transactions clump into social ties, social ties concatenate into networks, and existing networks constrain solutions of organizational problems [clarifying] the creation, maintenance, and change of categorical inequality.” The distribution of material and psychic rewards that groups gain access to determine the opportunities that come their way and their ultimate well being. How people live, then, has an effect on how they behave, including their ability to develop the capabilities for doing certain jobs and the motivation for doing them well. Capability and motivation can, in turn, affect their response to the job and

300

SHERYL SKAGGS AND NANCY DITOMASO

then, again, their material and psychic rewards. Hence, there is a virtuous or vicious circle that emerges over time providing some with cumulative advantages and others with cumulative disadvantages. Social position, however, is not guaranteed. Sometimes people have bad luck, or sometimes, as Sorensen mentions, they have less capable offspring. Some make bad choices in life that no amount of social capital can protect them from, but often the structural dynamics in which they are embedded and the processes that affect how structure comes into play in their everyday experiences, contributes to outcomes that individuals may not understand and over which they may have less influence in creating than they had believed to be the case. The durability of inequality, of course, feeds back on the structural characteristics of groups, since individual advantages for those of a certain background contributes to the amount and level of power of group members, the status or esteem in which group members are held, and their ability to influence the composition of the groups in which they participate or not. While many of the processes that influence and affect these outcomes are institutionalized, they are not unchangeable. Innovation, political mobilization, and changes in sociopolitical dynamics can shift the balance of power, undermine or improve status, and can change the composition of groups over time. But, the stability of workplace inequality is often of greater weight than is acknowledged in our theories that envision individuals with certain characteristics and a given level of motivation competing openly in the market for opportunity and rewards, especially in a fair process whereby the best man or woman wins. These issues are obviously very complex and have warranted decades of research across a number of disciplines and fields. Our efforts to create a schematic representation of how these structural characteristics come together with processes of reproduction is intended to make sense out of the complexity and point in the direction of where research needs to be enhanced or developed.

MEASUREMENT CHALLENGES From a measurement standpoint, it seems, that any theoretical perspective that hypothesizes an intermediate effect between composition and attitudinal outcomes that include in group/out group dynamics would have to assume a curvilinear relationship and would probably have to analyze each group separately so as not to confound the roles of each. Thus, a curvilinear relationship would be expected from the point of view of a social identity/social categorization theory, a group contact theory, and a group competition theory. The only theory where a linear effect may make conceptual sense is similarity-attraction theory (the more people

Understanding the Effects of Workforce Diversity on Employment Outcomes

301

like me, the happier I am), although this too could require a curvilinear effect if linked to in group/out group processes. Despite the fact that too many researchers in management and organization theory treat social identity theory as if it were reciprocal (all groups will favor people of their own group and disfavor others), that is not what social identity theory necessarily predicts, nor finds. The theory is often developed from the point of view of the in group (in this case, meaning dominant or majority), while leaving more ambiguous the point of view of the out group (in this case meaning subordinate or minority). Even Tajfel and Turner (1986; see also Sidanius & Pratto, 1999) talk about the disparate responses expected from out group members than from in group members. But, again, who is the in group and who is the out group cannot be defined unless one adds the notion of “master status” to the discussion. Hence, the research has found that men are favored whether they are in the majority or minority, while women are disfavored, especially when they are in the minority, and sometimes even when they are in the majority (Williams, 1992). The in group/out group distinction makes generic or linear effects impossible. That is, whenever one introduces a notion like in group/out group in discussing these processes, by definition, one has to take into consideration: (a) that the in group members will not have the same view of things as the out group members; and (b) that the designation of the in group shifts at 50% unless other information is taken into consideration (such as the existence of a master status). But, it is evident that if men are in the majority and one is trying to understand how the “majority” responds to the entry of women into the group, men can react as the majority only until women become 50% of the group. After that, as women increase in proportion, then men become a minority. Thus, the question arises as to whether men are still the “in group” or if women take over that role. Most likely, it depends on the extent to which numbers can outweigh the influence of power and status. Based on research reviewed here, we would expect not. The three characteristics work together and cannot substitute for each other. When drawing more from social contact theory, the effects would be reciprocal, but not necessarily of the same meaning. For example, Kanter (1977b) argues that the same effects she outlined would apply to men who are in the minority as they would to women in the minority, but she specifically argues that for men, those effects may be positive (i.e. heightened visibility leading to greater privileges), while for women, they may be negative (heightened visibility leading to greater pressure). In addition, Yoder (1994) finds that these effects are confounded as well by the “gender appropriateness” of the occupation and the prestige of the job. Hence, women in science and engineering (which is male dominated) may be very satisfied, even in the minority, because they are paid well compared to other women, whereas women in secondary education may be very dissatisfied, even in

302

SHERYL SKAGGS AND NANCY DITOMASO

the majority, because they are paid poorly compared to the males in the principal positions.

SUMMARY In this paper we have endeavored to integrate the literatures in sociology, social psychology, and management regarding the impact of workforce diversity at various levels of analysis. We have argued that to understand how workforce diversity generates and reproduces inequality in the labor force, we must understand power (the distribution of scarce and valuable resources), status (the relationships of deference) and numbers (the compositional effects of groups). Each of these factors are associated with alternative mechanisms that contribute to the reproduction of inequality, and these, in turn, affect how allocation decisions are made, the assumptions inherent in evaluative processes, and the inclusion or exclusion of group members to specific contexts. What people get out of life, how they are evaluated by others, and whether they are included within the group or excluded from it, then, contributes over time to how they feel (attitudes), what they do (behavior), and the rewards that they enjoy or are denied. Inequality, hence, is affected by the diversity of the workforce through both structural and psychological processes, over time, and through everyday interaction. Future research needs to develop and elaborate the link between theory and research in its various forms and contexts.

REFERENCES Abbott, A. D. (1988). The system of professions: An essay on the division of expert labor. Chicago: University of Chicago Press. Aigner, D. J., & Cain, G. C. (1977). Statistical theories of discrimination in labor markets. Industrial and Labor Relations Review, 30, 175–187. Allport, G. (1954). The nature of prejudice. Cambridge, MA: Addison Wesley. Ancona, D. G., & Caldwell, D. F. (1988). Beyond task and maintenance: Defining external functions in groups. Group and Organization Studies, 13, 468–494. Arrow, K. (1973). The theory of discrimination. In: O. Ashenfelter & A. Rees (Eds), Discrimination in Labor Markets (pp. 3–33). Princeton, NJ: Princeton University Press. Baron, J. N., & Bielby, W. T. (1980). Bringing the firm back in: Stratification, segmentation, and the organization of work. American Sociological Review, 45, 737–755. Baron, J. N., & Bielby, W. T. (1984). The organization of work in a segmented economy. American Sociological Review, 49, 454–473. Baron, J. N., & Newman, A. E. (1990). For what it’s worth: Organizations, occupations, and the value of work done by women and nonwhites. American Sociological Review, 55, 155–175.

Understanding the Effects of Workforce Diversity on Employment Outcomes

303

Becker, G. S. (1957). The Economics of discrimination (1st ed., 2nd ed., 1971). Chicago: University of Chicago Press. Becker, G. S. (1964). Human capital (1st ed., 2nd ed., 1975). NY: National Bureau of Economic Research. Bielby, W. T., & Baron, J. N. (1986). Men and women at work: Sex segregation and statistical discrimination. American Journal of Sociology, 91, 759–799. Blalock, H. M., Jr. (1957). Percent non-white and discrimination in the south. American Sociological Review, 22, 677–682. Blau, P. M. (1977). Inequality and heterogeneity. New York: Free Press. Bourdieu, P., & Passeron, J. (1977). Reproduction in education, society, and culture. London: Sage. Brewer, M. B. (1979). In-group bias in the minimal group situation: A cognitive-motivational analysis. Psychological Bulletin, 86, 307–324. Brewer, M. B. (1996). In-group favoritism: The subtle side of intergroup discrimination. In: D. Messick & A. Tenbrunsel (Eds), Codes of Conduct: Behavioral Research and Business Ethic (pp. 160–170). NY: Sage. Brewer, M. B., & Miller, N. (1984). Beyond the contact hypothesis: Theoretical perspectives on desegregation. In: N. Miller & M. B. Brewer (Eds), Groups in Contact: The Psychology of Desegregation (pp. 281–302). Orlando, FL: Academic Press. Bridges, W. P. (1982). The sexual segregation of occupations: Theories of labor stratification in the industry. American Journal of Sociology, 88, 270–295. Bridges, W. P., & Nelson, R. L. (1989). Markets in hierarchies: Organizational and market influences on gender inequality in a state pay system. American Journal of Sociology, 95, 616–658. Byrne, D. E. (1971). The attraction paradigm. New York: Academic Press. Cockburn, C. (1991). In the way of women: Men’s resistance to sex equality in organizations. Ithaca, NY: ILR Press. Cohen, P. (1998). Black concentration effects on Black-White and gender inequality: Multilevel analysis for U.S. metropolitan areas. Social Forces, 77, 207–229. Cox, T. H., Lobel, S. A., & McLeod, P. L. (1991). Effects of ethnic group cultural differences on cooperative and competitive behavior on a group task. The Academy of Management Journal, 34, 827–847. DiMaggio, P. (1997). Culture and cognition. Annual Review of Sociology, 23, 263–287. DiTomaso, N., Post, C., Smith, D. R., Farris, G. F., & Cordero, R. (2003). Three forms of inequality: Advantage, the absence of advantage, and disadvantage. Unpublished paper, August 26. DiTomaso, N., Thompson, D. E., & Blake, D. H. (1988). Corporate perspectives on the advancement of minority managers. In: D. E. Thompson & N. DiTomaso (Eds), Ensuring Minority Success in Corporate Management (pp. 119–136). New York: Plenum Press. Ely, R. J. (1994). The effects of organizational demographics and social identity on relationships among professional women. Administrative Science Quarterly, 39, 203–238. Ely, R. J. (1995). The power in demography: Women’s social constructions of gender identity at work. The Academy of Management Journal, 38, 589–634. England, P. (1982). The failure of human capital theory to explain occupational sex segregation. The Journal of Human Resources, 17, 358–370. Fields, D. L., & Blum, T. C. (1997). Employee satisfaction in work groups with different gender composition. Journal of Organizational Behavior, 18, 181–196. Granovetter, M. (1985). Economic action and social structure: The problem of embeddedness. American Journal of Sociology, 91, 481–510.

304

SHERYL SKAGGS AND NANCY DITOMASO

Hacker, S. (1983). Mathematisation of engineering: Limits on women and on the field. In: J. Rothschild (Ed.), Machina ex Dea: Feminist Perspectives on Technology (p. 39). New York: Pergamon. Herrnstein, R. J., & Murray, C. A. (1994). The bell curve: Intelligence and class structure in American life. NY: Free Press. Hoffman, L. R., & Maier, N. R. F. (1961). Quality and acceptance of problem solutions by members of homogeneous and heterogeneous groups. Journal of Abnormal and Social Psychology, 62, 401–407. Ibarra, H. (1993). Personal networks of women and minorities in management: A conceptual framework. The Academy of Management Review, 18, 56–87. Ibarra, H. (1995). Race, opportunity, and diversity of social circles in managerial networks. The Academy of Management Journal, 38, 673–703. Jackman, M. (1994) The Velvet Glove: Paternalism and conflict in gender, class and race relations. Berkley: University of California Press. Kanter, R. M. (1977a). Men and women of the corporation. New York: Basic Books. Kanter, R. M. (1977b). Some effects of proportions of group life: Skewed sex rations and responses to token women. American Journal of Sociology, 82, 965–990. Lawrence, B. S. (1997). The black box of demography. Organization Science, 8, 1–22. Lin, N. (2001). Social capital: A theory of social structure and action. Cambridge: Cambridge University Press. Major, B. (1994). From social inequality to personal entitlement: The role of social comparisons, legitimacy appraisals, and group membership. Advances in Experimental Social Psychology, 26, 293–355. Major, B., Vanderslice, V., & McFarlin, D. B. (1984). Effects of pay expected on pay received: The confirmatory nature of expectations. Journal of Applied Social Psychology, 14, 399–412. Milliken, F. J., & Martins, L. L. (1996). Searching for common threads: Understanding the multiple effects of diversity in organizational groups. The Academy of Management Journal, 21, 402–433. Murphy, R. (1988). Social closure: The theory of monopolization and exclusion. New York: Oxford University Press. Nelson, R. L., & Bridges, W. P. (1999). Legalizing gender inequality: Courts, markets, and unequal pay for women in America. Cambridge: Cambridge University Press. Olzak, S. (1992). The dynamics of ethnic competition and conflict. Palo Alto, CA: Stanford University Press. Parkin, F. (1979). Marxism and class theory: A bourgeois critique. New York: Columbia University Press. Pelled, L. H., Eisenhardt, K. M., & Xin, K. R. (1999). Exploring the black box: An analysis of work group diversity, conflict and performance. Administrative Science Quarterly, 44, 1–28. Perry, E. L., Davis-Blake, A., & Kulik, C. T. (1994). Explaining gender-based selection decisions: A synthesis of contextual and cognitive approaches. The Academy of Management Journal, 19, 786–820. Pettigrew, T. F. (1979). The ultimate attribution error. Personality and Social Psychology Bulletin, 5, 461–476. Pfeffer, J., & Davis-Blake, A. (1987). The effect of the proportion of women on salaries: The case of college administrators. Administrative Science Quarterly, 32, 1–24. Reskin, B. F. (1988). Bringing the men back in: Sex differentiation and the devaluation of women’s work. Gender and Society, 2, 58–81. Reskin, B. F. (2003). Including mechanisms in our models of ascriptive inequality. American Sociological Review, 68, 1–21.

Understanding the Effects of Workforce Diversity on Employment Outcomes

305

Reskin, B. F., McBrier, D. B., & Kmec, J. A. (1999). The determinants and consequences of workplace sex and race composition. Annual Review of Sociology, 25, 335–361. Reskin, B. F., & Roos, P. A. (1990). Job queues, gender queues: Explaining women’s inroads into male occupations. Philadelphia, PA: Temple University. Ridgeway, C. (1997). Interaction and the conservation of gender inequality. American Sociological Review, 62, 218–235. Ridgeway, C. (2001). The Emergence of status beliefs: From structural inequality to legitimzing ideology. In: J. T. Jost & B. Major (Eds), The Psychology of Legitmacy (pp. 257–277). Cambridge: Cambridge University Press. Royster, D. A. (2003). Race and the invisible hand: How white networks exclude black men from blue-collar jobs. Berkeley: University of California Press. Sachdev, I., & Bourhis, R. Y. (1984). Minimal majorities and minorities. European Journal of Social Psychology, 14, 35–52. Sachdev, I., & Bourhis, R. Y. (1985). Social categorization and power differentials in group relations. European Journal of Social Psychology, 15, 415–434. Sachdev, I., & Bourhis, R. Y. (1987). Status differentials and intergroup behaviour. European Journal of Social Psychology, 17, 277–293. Sachdev, I., & Bourhis, R. Y. (1991). Power and status differentials in minority and majority group relations. European Journal of Social Psychology, 21, 1–24. Sackett, P. R., DuBois, C. L. Z., & Noe, A. W. (1991). Tokenism in performance evaluation: The effect of work group representation on male-female and white-black differences in performance ratings. Journal of Applied Psychology, 2, 263–267. Sidanius, J. (1993). The psychology of group conflict and the dynamics of oppression: A social dominance perspective. In: S. Iyengar & W. J. McGuide (Eds), Explorations in Political Psychology (pp. 183–219). Durham: Duke University Press. Sidanius, J., & Pratto, F. (1999). Social dominance. Cambridge: Cambridge University Press. Simmel, G. (1908). Soziologie. Berlin: Leipzig, Duncker and Humbolt. Sorensen, A. B. (2000). Toward a sounder basis for class analysis. The American Journal of Sociology, 105, 1523–1558. Stuhlmacher, A. F., & Walters, A. E. (1999). Gender differences in negotiation outcome: A meta-analysis. Personnel Psychology, 52, 653–677. Tajfel, H. (1981). Human groups and social categories: Studies in social psychology. Cambridge: Cambridge University Press. Tajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior. In: S. Worchel (Ed.), Psychology of Intergroup Relations (pp. 7–24). Chicago: Nelson. Thurow, L. (1975). Generating inequality. New York: Basic Books. Tilly, C. (1998). Durable inequality. Berkeley and Los Angeles: University of California Press. Tolbert, P. S., Graham, M. E., & Andrews, A. O. (1999). Group gender composition and work group relations. In: G. N. Powell (Ed.), Handbook of Gender and Work (pp. 179–202). Thousand Oaks, CA: Sage. Tomaskovic-Devey, D. (1993). Gender and racial inequality at work: The sources and consequences of job segregation. Ithaca, NY: ILR Press. Tomaskovic-Devey, D., & Skaggs, S. (1999). An establishment-level test of the statistical discrimination hypothesis. Work and Occupations, 26, 422–445. Tomaskovic-Devey, D., & Skaggs, S. (2002). Sex segregation, labor process organization, and gender earnings inequality. American Journal of Sociology, 108, 102–128.

306

SHERYL SKAGGS AND NANCY DITOMASO

Tsui, A. S., Egan, T. D., & O’Reilly, C. A., III (1992). Being different: Relational demography and organizational attachment. Administrative Science Quarterly, 37, 549–579. Tsui, A. S., & O’Reilly, C. A., III (1989). Beyond simple demographic effects: The importance of relational demography in superior-subordinate dyads. Academy of Management Journal, 32, 402–423. Turner, J. C. (1987). Rediscovering the social group: Self-categorization theory. Oxford, UK: Blackwell. Waldinger, R. (1995). The “other side” of embeddedness: A case-study of the interplay of economy and ethnicity. Ethnic and Racial Studies, 18, 555–580. Waldinger, R. (1997, August). Social capital or social closure? Immigrant networks in the labor market. Working Paper #26, The Lewis Center for Regional Policy Studies, UCLA. Watson, W. E., Kumar, K., & Michaelson, L. K. (1993). Cultural diversity’s impact on interaction process and performance: Comparing homogeneous and diverse task groups. The Academy of Management Journal, 36, 590–602. Weber, M. (1968)[1978]. Economy and society. In: G. Roth & C. Wittich (Eds). Berkeley: University of California Press. Webster, M., & Foschi M. (1988). Status generalization: New theory and research. Stanford: Stanford University Press. Weeden, K. A. (2002). Why do some occupations pay more than others? Social closure and earnings inequality in the United States. American Journal of Sociology, 108, 55–101. Wharton, A. S., & Baron, J. N. (1987). So happy together? The impact of gender segregation on men at work. American Sociological Review, 52, 574–587. Williams, B. (1987). Black workers in an industrial suburb. New Brunswick, NJ: Rutgers University Press. Williams, C. (1992). The glass escalator: Hidden advantages for men in the “female” professions. Social Problems, 39, 253–267. Willis, P. (1977). Learning to labor: How working class kids get working class jobs. New York: Columbia University Press. Yoder, J. D. (1994). Looking beyond numbers: The effects of gender status, job prestige, and occupational gender-typing on tokenism processes. Social Psychology Quarterly, 57, 150–159.