No title

P1: FNN Archives of Sexual Behavior [asb]

PL186-228369

October 21, 2000

17:7

Style file version Nov. 19th, 1999

Archives of Sexual Behavior, Vol. 30, No. 1, 2001

Sadomasochistically Oriented Behavior: Diversity in Practice and Meaning1 Laurence Alison, Ph.D.,2,4 Pekka Santtila, Ph.D.,3 N. Kenneth Sandnabba, Ph.D.,3 and Niklas Nordling, M.Psych3

One hundred and eighty-four subjects (22 women and 162 men) who were members of two sadomasochistically oriented clubs answered a semistructured questionnaire containing items relating to a variety of sexual behaviors. Using a multivariate statistical analysis that geometrically represents the co-occurrence of individual actions as a visual array (Guttman (1954). In Lazarfeld, P. E. (ed.), Mathematical Thinking in the Social Sciences, Free Press, Glencoe, IL.) four qualitatively different sexual scripts emerged: hypermasculinity; administration and receiving of pain; physical restriction; and psychological humiliation. Although similar themes have been suggested before, this study demonstrated their empirical base. Humiliation was significantly associated more with females and with heterosexual orientation in men, while hypermasculinity was associated with males and with homosexual orientation in men. KEY WORDS: sadomasochism; homosexuality; heterosexuality; hypermasculinity; humiliation.

INTRODUCTION The current conceptualization of sadomasochism relies on a label of convenience for a set of related sexual activities of particular subcultures (Haeberle, 1978; Katchadourian and Lunde 1972/1975). Facets include physical restriction and bondage (Baumeister, 1988) and humiliation (Baumeister, 1988; Moser and Levitt, 1987; Weinberg, 1987). Additionally, Weinberg et al. (1984), Lee (1979), and Kamel (1983) refer to a subset of behaviors commonly associated with the gay 1 Research supported by grants from the Rector of Abo ˚ Akademi University and the SETA Foundation. 2 Department

of Psychology, University of Liverpool, Liverpool, United Kingdom. ˚ Akademi University, Finland. of Psychology, Abo 4 To whom correspondence should be addressed at Department of Psychology, Eleanor Rathbone Building, University of Liverpool, Liverpool L69 7ZA, United Kingdom. 3 Department

1 C 2001 Plenum Publishing Corporation 0004-0002/01/0200-0001$19.50/0 °

P1: FNN Archives of Sexual Behavior [asb]

2

PL186-228369

October 21, 2000

17:7

Style file version Nov. 19th, 1999

Alison, Santtila, Sandnabba, and Nordling

male ‘leather’ scene that, to observers, appears to be sadomasochistic in origin. These behaviors include enemas, catheters, anal fisting, and scatological practices, and are sometimes described by the subjects as displays of ‘masculinity and toughness’ (Weinberg et al., 1984, p. 387). The evidence for these facets of pain, restriction, and humiliation and Weinberg et al.’s identification of hypermasculinity (Weinberg et al., 1984) as being related sets of actions remains speculative because there have, to date, been no correlative studies examining the notionally related facets of these groups of sexual behaviors. That is, are the behaviors associated with each of these suggested facets more empirically interrelated with one another than with behaviors of the other facets? Or, more specifically, are there subsets of, for example, physically restrictive acts (handcuffs, straitjackets, use of chains) or pain creating actions (skinbranding, caning, spanking) that are empirically related? This would provide empirical support for the existence of distinct themes within sadomasochistic sexual behavior. The present study sought not to identify these sets of actions as defining of sadomasochism but rather to map out the empirical relationships between the individual acts performed by individuals engaged in the sadomasochistic subculture. The concern was to establish the extent to which individual actions bear an empirical relationship to one another. It is therefore important to bear in mind that while these sets of actions are commonly considered to be sadomasochistic and therefore take on the connotation of ‘describing’ sadomasochists, they are actually a set of descriptions about subsets of interactions that occur between individuals engaged in a variety of behaviors. For example, in Weinberg et al.’s exploration (Weinberg et al., 1984) of the ‘leather’ scene subjects engaged in dressing up in leather, using chains and analy fisting did not consider themselves sadomasochists. Similarly, previous studies have not taken account of the potential preferences that may exist for individuals in engaging in one facet (i.e., the administration or receiving of pain) over another facet (i.e., humiliation). One possibility is that individuals would only engage in a limited set of behaviors and not in others suggesting that sadomasochism is in fact a label for a number of independent phenomena. Another possibility is that individuals emphasize a particular set of behaviors but also engage in other behaviors to a more limited extent. This would suggest that it makes sense to conceptualize sadomasochism as a distinct phenomenon. Also, there have been no investigations exploring whether a preference for one facet over another is related to the sex and the sexual orientation of the subjects, and whether there is a preference for the ‘sadistic’ or the ‘masochistic’ partner to engage in one facet more than any other. Finally, the reasons for an individual developing a preference for a particular set of sadomasochistic sexual behaviors has not been thoroughly clarified. In fact, some recent studies (Nordling et al., 2000; Santtila et al., accepted) have shown evidence for associations between childhood experiences and aspects of sadomasochistic behavior whereas the importance of an ongoing process of adult

P1: FNN Archives of Sexual Behavior [asb]

PL186-228369

October 21, 2000

17:7

Style file version Nov. 19th, 1999

Sadomasochistically Oriented Behavior: Diversity in Practice and Meaning

3

socialization has been stressed by more sociologically oriented researchers. Weinberg et al. (1984) criticized traditional models for ignoring the sadomasochistic subculture. According to them, this subculture provides opportunities for persons to define and elaborate their sexual activities. Also, the finding that the sexual activity that occurs in a sadomasochistic scene is often, but not always, scripted and therefore collaborative (Weinberg, 1987), suggests that individual sexual repertoires are socially constructed. An attempt was made in the present study to explore whether the extent to which the subjects were integrated into the sadomasochistic scene (defined as the frequency of sadomasochistic sessions and involvement with pornographic material of a similar nature) would predict the flexibility of their sexual repertoires within each of the suggested sadomasochistic subthemes. In summary, we were interested in establishing what evidence existed for identifying the facets of sadomasochism previously referred to by others (physical restriction, administration and receiving of pain and psychological and physical humiliation as well as hypermasculinity), the potential relationship that these domains may have with other factors in the individuals’ lives as well as the extent to which the degree of involvement in the sadomasochistic subculture would predict the flexibility of the subjects’ sexual repertoires. It was expected that female subjects would be more likely to engage in behaviors related to humiliation (Messman and Long, 1996) whereas gay male subjects would report more behaviors related to hypermasculinity (Weinberg et al., 1984).

METHOD Subjects One hundred and eighty-four subjects (22 women and 162 men) who were members of two sadomasochistically oriented clubs participated in this study (two subjects were discarded due to lack of responses on questions related to specific sexual behaviors). Ninety-five were recruited from Kinky Club, a club for mainly heterosexual people with a variety of sexual preferences. Ninety-one subjects were recruited from the MSC-Finland association with mainly gay male members. Of the subjects, 43.0% reported being mainly heterosexual, 5.4% bisexual, and 51.6% mainly homosexual. Of all subjects, 27.0% identified themselves as mainly sadistic, 22.7% as both sadistic and masochistic, and 50.2% as mainly masochistic in their sadomasochistic behavior. Subjects were highly educated (over a third had a university degree). They also had a higher income level than the population in general (Statistical Yearbook of Finland, 1993) with half having a monthly income of more than 2000 US dollars. This study group, as with others previously explored (Baumeister, 1988; Spengler, 1977; Weinberg, 1987) is suggestive of a group of subjects who are not psychologically disturbed or dysfunctional but are rather better educated and in a

P1: FNN Archives of Sexual Behavior [asb]

4

PL186-228369

October 21, 2000

17:7

Style file version Nov. 19th, 1999

Alison, Santtila, Sandnabba, and Nordling

generally higher earning bracket than the general population. (For more details on the demographic and sexual characteristics of the males in the study group, see Sandnabba et al., 1999). Materials The Questionnaire Information selected for this study was taken from a larger survey conducted by Sandnabba et al. (1999) in which the authors were investigating the social adaptation and variety of sexual behaviors of sadomasochists. The original questionnaire was designed as a descriptive analysis of the range of behaviors that sadomasochists indulged in and as a descriptive analysis of the study group. For the purposes of the present study the focus was upon the set of activities engaged in and the empirical relationships between them. Concentration was therefore upon elements of behavior within sadomasochistic encounters. This included items pertaining to the various practices of the subjects on a dichotomous basis, that is, had they indulged in a particular behavior at any point during the last 12 months. This list included a number of activities previously listed in studies on the behavior of individuals in the sadomasochistic scene. Additionally, questions (other than the demographic information already outlined) were asked regarding the number of sadomasochistic sessions during the preceding year and the extent to which the subjects had acquired sadomasochistic pornographic magazines during the same time period for the purpose of exploring any relationships between these variables and the sadomasochistic behaviors of the subjects. Statistical Analysis: SSA Based on the joint co-occurrence of behaviors, the 184 questionnaires were analyzed using Smallest Space Analysis (SSA). SSA is a nonmetric multidimensional scaling procedure that represents the associations between variables as the inverse of distances in a statistically derived geometric space—the greater the similarity between two variables the greater their proximity in the corresponding space (Guttman, 1968). SSA is based upon the assumption that the underlying structure of complex behavioral systems is most readily appreciated if the relationship between each and every other variable is examined (Canter and Heritage, 1990). The SSA-I program computes association coefficients between all variables and rank orders them, creating a triangular matrix consisting of association coefficients for each variable against every other variable. It is these rank ordered coefficients, rather than their absolute values that are used to form the spatial representation of variables. By using these rank ordered coefficients, SSA is able to represent the variables in the smallest possible dimensionality.

P1: FNN Archives of Sexual Behavior [asb]

PL186-228369

October 21, 2000

17:7

Style file version Nov. 19th, 1999

Sadomasochistically Oriented Behavior: Diversity in Practice and Meaning

5

To find an optimal representation of the variables, SSA-I performs iterations that compare the rank order assigned to the original associations with the rank order of the distances between points in the plot (which the program calculates from a derived association matrix). With each iteration, adjustments are made to the geometric representation in an attempt to minimize the difference between the plot and the original association matrix. The closer the two sets of rank orders are, the better the fit. These iterations continue until the difference between these two matrices is at a minimum. The degree of fit between the geometric representation and the original input similarity data is indicated by Guttman’s coefficient of alienation (Guttman, 1968). The coefficient of alienation ranges from 0 (indicating a perfect fit) to 1. A coefficient of .20–.25 is considered a reasonably good degree of fit (Shye et al., 1994). The resulting configuration of points in the SSA is based solely upon the relationships among variables. Therefore, these points can be examined directly without assuming underlying dimensions as in factor analysis (Canter and Heritage, 1990). Classifying variables through an examination of the regional structure in an SSA is part of an approach to research known as facet theory (Canter, 1985). The elements of ‘facets’ in this case refer to the overall classification of behaviors into ‘types’ (i.e., hypermasculinity, administration of pain, humiliation, and physical restriction). The spatial contiguity of the variables representing these facet elements provides a test of the major underlying differences among these variables as revealed through their co-occurrence in sadomasochistic behaviors, and is therefore a test as to whether the facet structure is empirically supported (Canter et al., 1998). The postulation of facet elements goes beyond simply saying that behaviors belong in particular ‘groups.’ The principle of contiguity states that, “. . . variables which are more similar in their facet structure will also be more related empirically” (Foa, 1965, p. 264). Therefore, variables that share the same facet elements would be more highly correlated and thus should appear closer together in the multidimensional space. This idea can be extended as a general, regional hypothesis. Basically, items that have facet elements in common will be found in the same region of space. Likewise, variables that have very low intercorrelations will appear in different regions of the plot indicating dissimilarity. For further elaboration of the technique, the range of applied fields within which it has been used, and its strengths and weaknesses see Canter (1983), Elizur and Sagie (1999), Guttman (1954), and Shye et al. (1994). RESULTS A two-dimensional solution was employed to represent the association matrix, where the coefficient of alienation was .27. This is a reasonably high coefficient indicating that the fit between the representation and the actual ranked associations

P1: FNN Archives of Sexual Behavior [asb]

PL186-228369

October 21, 2000

6

17:7

Style file version Nov. 19th, 1999

Alison, Santtila, Sandnabba, and Nordling

was not particularly good. However, as Shye (1994, p. 125) states in the past it has been customary to attempt SSA solutions of increased dimensionality whenever the coefficient was considered high . . . This procedure has, however, been found lacking both on technical and on theoretical grounds [since] the coefficient of alienation is sensitive to the number of items processed.

Although our results suggest that the geometric output is not a perfect representation of the variables, the two-dimensional arrangement of variables did approximate the conceptual system of sadomasochism as hypothesized in previous work (Baumeister, 1988; Messman and Long, 1996; Moser and Levitt, 1987; Weinberg et al., 1984). Moreover, employment of a three-dimensional representation did not significantly lower the coefficient of alienation, nor, more importantly did it change the first (two-dimensional configuration) projection of the points. Thus, whilst higher dimensional models may reveal other, more subtle nuances of this set of sexual activities, the two-dimensional model does reflect an interpretable, conceptually and empirically supported model. As Shye (1994, p. 126) states that, “If a regional contiguity pattern in a two-dimensional solution conforms well to a content facet formulated in advance and if evidence for this conformity is accumulated, the purpose of theory construction has been achieved.” Figure 1 therefore shows the projection of the two-dimensional space. Table I displays the behaviors (and frequency of occurrence) separated into the hypothesized polar regions along with their associated Cronbach’s alpha coefficients. These ranged from .71 (humiliation) to .78 (administration of pain) and while not indicative of perfect scales of behavior are reasonably high given, in some of the regions, the relatively small number of items (i.e., only 5 items in humiliation). It is important, however, not to consider these as discrete regions, but rather as a polarizing spectrum of gradual change around the midpoint intersection of all the regions. Thus, cusp variables such as wrestling and bondage in the physical restriction region are associated relatively more with their adjacent hypermasculine region than with mummification, which lies near the cusp of its Table I. The Frequency of the Sexual Behaviors in the Different Regions of the SSA Space with Associated Cronbach’s α Coefficients Hypermasculinity (N = 8, α = .75) Rimming Dildo Cockbinding Watersports Enema Fistfucking Scatologia Catheter

% 77.5 70.2 68.3 50.6 42.7 32.9 18.2 10.4

Administration of pain (N = 7, α = .78)

%

Clothspins Spanking Caning Weights Hot wax Electricity Skinbranding

67.6 64.0 50.7 41.5 34.8 16.4 15.8

Humiliation (N = 5, α = .72)

%

Flagellation Verbal humiliation Gag Faceslapping Knives

81.8 70.1 53.0 37.2 10.9

Physical restriction (N = 9, α = .75)

%

Bondage Handcuffs Chains Wrestle Slings Ice Straitjacket Hypoxyphilia Mummifying

88.4 73.2 70.8 45.1 39.0 31.7 17.0 16.5 13.4

P1: FNN Archives of Sexual Behavior [asb]

PL186-228369

October 21, 2000

17:7

Style file version Nov. 19th, 1999

Sadomasochistically Oriented Behavior: Diversity in Practice and Meaning

7

Fig. 1. A two-dimensional Smallest Space Analysis of sadomasochistic behaviors in a sample of 184 subjects.

contiguous set of related variables in the humiliation region. This visual map of the correlation matrix, therefore, reveals evidence of a circular ordering of the variables from hypermasculine, through to the administration of pain, to humiliation, physical restriction and back, full circle to hypermasculine. This also suggests that sadomasochism can be conceptualized as a set of interrelated behaviors where individuals give different emphases to particular themes rather than it being a label of convenience for a number of independent phenomena. To establish whether there were relationships between the subjects’ sex, sexual orientation, and sadomasochistic preference and any of the themes, summary variables reflecting the number of behaviors present in the separate regions of the plot were formed. Thereafter, any differences between the male and female subjects and between the predominantly and exclusively heterosexual male and predominantly and exclusively gay male subjects (bisexual subjects were discarded

P1: FNN Archives of Sexual Behavior [asb]

PL186-228369

October 21, 2000

8

17:7

Style file version Nov. 19th, 1999

Alison, Santtila, Sandnabba, and Nordling Table II. Number of Sexual Behaviors in the Different Regions of the SSA Space in Groups of Gay Male and Heterosexual Male Subjects and in Groups of Sadistic and Masochistic Subjects Heterosexual (N = 64)

Gay male (N = 89)

Sadistic (N = 42)

Masochistic (N = 80)

Region

M

SD

M

SD

M

SD

M

SD

Hypermasculinity Administration of pain Humiliation Physical restriction

3.17 3.09 2.88 3.91

2.17 2.15 1.32 2.24

4.03 2.74 2.28 3.97

1.95 1.99 1.46 2.11

3.40 2.55 2.40 3.81

2.07 1.99 1.50 2.27

4.05 3.16 2.64 4.02

2.04 2.11 1.49 2.32

from this analysis) and also between the predominantly and exclusively sadistic male and predominantly and exclusively masochistic male (flexible subjects were discarded from this analysis) were explored using the Student’s t-test. (Only male subjects were included in the two latter analyses due to the low number of female subjects in the study group). Female subjects engaged in significantly more of the behaviors in the humiliation region (M = 3.14, SD = 1.70) compared to male subjects (M = 2.56, SD = 1.46), t(182) = 1.70, p < .05, whereas the male subjects engaged in more of the behaviors in the hypermasculinity region (M = 3.75, SD = 2.14) compared to female subjects (M = 2.50, SD = 1.90), t(182) = 2.62, p < .01. The means and standard deviations for the other analyses are presented in Table II. There were significant differences between the hetero- and gay male subjects in terms of their involvement in the hypermasculinity and humiliation regions. As expected, the gay male subjects were more likely to engage in a larger number of the behaviors of the hypermasculinity region (t(151) = 2.57, p < .01) compared to the heterosexual male subjects whereas the latter were more likely to engage in a larger number of humiliation behaviors (t(151) = 2.59, p < .01). In terms of the sadomasochistic preference, there were nonsignificant tendencies for the masochistic male subjects to have engaged in a larger number of behaviors in both the hypermasculinity (t(120) = 1.65, p < .10) and the administration of pain (t(120) = 1.56, p < .12) regions compared to the sadistic male subjects. Next, Pearson’s correlation coefficients were computed between the variables indicating the subjects’ involvement in the sadomasochistic subculture (number of sadomasochistic sessions, use of sadomasochistic pornography) and the number of sexual behaviors within the different regions that the subjects had engaged in during the preceding year (see Table III). As expected, positive and significant associations could be observed. However, the absolute values of the coefficients were not particularly impressive. DISCUSSION Our results are encouraging in supporting previous references to various facets of sadomasochistically oriented behavior. At a broad level, these appear to take

P1: FNN Archives of Sexual Behavior [asb]

PL186-228369

October 21, 2000

17:7

Style file version Nov. 19th, 1999

Sadomasochistically Oriented Behavior: Diversity in Practice and Meaning

9

Table III. Correlations Between Measures of Involvement in Sadomasochistic Subculture and the Number of Behaviors Engaged in the Different Regions of the SSA Space Number of behaviors engaged in the different regions Hypermasculinity Administration of pain Humiliation Physical restriction ∗p

Number of sadomasochistic sessions

Use of sadomasochistic pornographic material

.24∗∗∗ .23∗∗ .18∗ .10

.19∗∗ .24∗∗∗ .21∗∗ .19∗∗

< .05; ∗∗ p < .01; ∗∗∗ p < .001.

on different functions for men and women on the one hand and gay men and heterosexual men on the other. The female subjects significantly preferred themes associated with humiliation whereas the male subjects were more likely to engage in themes related to hypermasculinity. In other words, at opposing poles lie the facets of hypermasculinity and humiliation and in adjacent regions lie the physical aspects of restriction and the administration of pain. Elements of humiliation are significantly preferred by the women and by the heterosexual males in the study group and were revealed in the correspondence between faceslapping, flagellation, the use of a gag, the use of knives and razors, and verbal humiliation. Thus, actions significantly preferred especially by the gay men involved rimming; watersports; cockbinding; fistfucking; scatologia; and the use of dildos, enemas, and catheters. We would suggest that adjacent facets (restriction and pain) to these two regions can take on rather different meanings for each group but are important components of the full picture of sadomasochistic behavior. Our results confirmed that the following acts were associated with one another: subjecting individuals to (sadistic component) or receiving (masochistic component) electric shocks, skinbranding, caning, spanking, the use of hot wax, clothespins, and the use of weights. It is perhaps significant that the more extreme and intense of these experiences (skinbranding and electrocution) are located nearer the hypermasculine region, whereas spanking and caning appear next to the humiliation region. This may be suggestive of subtle variations in the context within which pain is administered and received and the meaning that it has for the different subjects and, potentially, for different sexes and for different sexual orientations. For example, as noted in Lee (1979), Kamel (1983), and Weinberg et al.’s examinations (Weinberg et al., 1984) of the gay ‘leather’ scene an emphasis is placed either upon the masculinity of the activities involved and/or the extent to which the bruises and welts received represent a ‘sign of love and/or surrender’ (Weinberg et al., 1984, p. 387). In contrast, the caning and spanking aspects of pain administration and reception may be more indicative of the associated facet of humiliation and the adoption of a set of activities associated with the ‘illusion of violence’ (Weinberg and Kamel, 1983). As Gebhard (1969) points out, individuals involved in the heterosexual sadomasochistic scene rarely highlight pain as a central facet. As Califia (1979, p. 21) points out, “The basic dynamic of S and M

P1: FNN Archives of Sexual Behavior [asb]

10

PL186-228369

October 21, 2000

17:7

Style file version Nov. 19th, 1999

Alison, Santtila, Sandnabba, and Nordling

is a power dichotomy, not pain.” Our results suggest that the administration and receiving of pain may take on rather different meanings depending upon the context within which it is received and, of course, depending upon the intensity of the sensation delivered/experienced. In other words, pain may, for some individuals, be rather more literal (hypermasculine) than be symbolic (humiliation). Secondly, reference has been made to the physically restrictive acts associated with sadomasochistic practices. Bondage; mummification; hypoxyphilia; the use of straitjackets, chains, handcuffs, slings and ice, as well as wrestling with one’s partner were all located in this region. Significantly, however, bondage, hypoxyphilia, and wrestling are adjacent to the hypermasculinity region whereas mummification, straitjackets, chains, and handcuffs are adjacent to the humiliation region. Once again, this emphasizes the potentially different psychological symbolism of different aspects of restriction for different parties. The advantage of the multivariate scaling analysis presented here is that it represents the relationships between each and every variable and thus allows for refinements and more subtle developments of hypotheses regarding the nature of the domain under investigation. Thus, these findings are not definitive but rather generative of more refined hypotheses regarding associations of particular actions. Subsequent studies could follow up the extent to which different individuals draw upon facets of pain and restriction in order to heighten the intensity of their experiences. This contextualizing of these components clearly has significant psychological meaning for the acting out of the subjects of this particular group but of course, whether this is a generalizable finding to other sadomasochistic groups remains open to question. The particular structure here may be indicative of differences between the styles of the clubs examined rather than between the sexuality of the subjects. Subsequent work would need to clarify this issue and explore a range of homo- and heterosexual-oriented clubs to explore whether the differences established here are an artifact of the club rather than an association with an individual’s sexual preferences. The subjects’ involvement in the sadomasochistic subculture through sexual contacts and pornography was positively associated with greater variability in their sexual behavior. Although the design of the study does not warrant any causal conclusions, the results imply that sadomasochistic behavior is best understood as a product of adult socialization processes where real or imagined sexual contact leads the subjects to adopt new behaviors and sexual scripts. This finding accords well with social constructionistic explanations of sexual behavior (Weinberg, 1987; Weinberg et al., 1984). In this study group, results suggest that for the gay male group the administration and reception of pain was a more intense and real perception and that the symbolic representation of pain was more important for the women and the heterosexual men. Similarly, the restrictive acts in this particular heterosexual group appear to have a more symbolically humiliating association than the more physically intense aspects of restriction located nearer to the hypermasculinity

P1: FNN Archives of Sexual Behavior [asb]

PL186-228369

October 21, 2000

17:7

Style file version Nov. 19th, 1999

Sadomasochistically Oriented Behavior: Diversity in Practice and Meaning

11

region preferred by the gay male subjects. The subtleties of the meanings of these actions, however, will remain speculative until further research is able to explore how individuals employ these facets within their own particular context. This may mean that the outcome of our study provides the basis for developing interviewing protocols or surveys that will help capture the meaning that restriction, pain, humiliation, and masculinization takes on for different subsets of the sadomasochistic population.

REFERENCES Baumeister, R. F. (1988). Masochism as escape from self. J. Sex Res. 25: 28–59. Califia, P. (1979, December 17). A secret side of lesbian sexuality. Advocate: 19–23. Canter, D. V. (1983). The potential of facet theory for applied social psychology. Quality and Quantity, 17: 35–67. Canter, D. V. (ed.). (1985). Facet Theory: Approaches to Social Research, Springer, New York. Canter, D. V., and Heritage, R. (1990). A multivariate model of sexual offense behavior: Developments in ‘offender profiling.’ J. Forensic Psychiat. 1: 185–212. Canter, D. V., Hughes, D., and Kirby, S. (1998). Pedophilia: Pathology, criminality or both? The development of a multivariate model of offense behavior in child sexual abuse. J. Forensic Psychiat. 9: 532–555. Elizur, D., and Sagie, A. (1999). Facets of personal values: A structural analysis of life and work values. Appl. Psychol: Int. Rev. 48: 73–87. Foa, U. G. (1965). New Developments in facet design and analysis. Psychol. Rev. 72: 262–274. Gebhard, P. (1969). Fetishism and sadomasochism. In Masserman, J. H. (ed.), Dynamics of Deviant Sexuality, Grune and Stratton, New York, pp. 71–80. Guttman, L. (1954). A new approach to factor analysis: The circle. In Lazarfeld, P. E. (ed.), Mathematical Thinking in the Social Sciences, Free Press, Glencoe, IL. Guttman, L. (1968). A general nonmetric technique for finding the smallest co-ordinate space for a configuration of points. Psychometrika 33: 469–506. Haeberle, E. (1978). The Sex Atlas, Seabury Press, New York. Kamel, G. W. L. (1983). The leather career: On becoming a sadomasochist. In Weinberg, T. and Kamel, G. W. L. (eds.), S and M: Studies in Sadomasochism, Prometheus, Buffalo, NY, pp. 73–79. Katchadourian, H., and Lunde, D. (1975). Fundamentals of Human Sexuality, 2nd Ed., Holt, Rinehart and Winston, New York. Lee, J. A. (1979). The social organization of sexual risk. Alternative Lifestyles 2: 69–100. Messman, T. L., and Long, P. S. (1996). Child sexual abuse and its relationship to revictimization in adult women: A review. Clin. Psychol. Rev. 16: 397–420. Moser, C., and Levitt, E. (1987). An exploratory-descriptive study of a sadomasochistically oriented sample. J. Sex Res. 23: 322–337. Nordling, N., Sandnabba, N. K., and Santtila, P. (2000). The prevalence and effects of self-reported childhood sexual abuse among sadomasochistically oriented males and females. J. Child Sex. Abuse 9, in press. Sandnabba, N. K., Santtila, P., and Nordling, N. (1999). Sexual behavior and social adaptation among sadomasochistically-oriented males. J. Sex Res. 36: 273–282. Santtila, P., Sandnabba, N. K., and Nordling, N. (in press). Retrospective perceptions of family interaction in childhood as correlates of current sexual adaptation among sadomasochistic males. J. Psychol. Hum. Sex., in press. Shye, S., Elizur, D., and Hoffman, M. (1994). Introduction to Facet Theory: Content Design and Intrinsic Data Analysis in Behavioral Research, Sage Publications, London. Spengler, A. (1977). Manifest sadomasochism of males: Results of an empirical study. Arch. Sex. Behav. 6: 441–456.

P1: FNN Archives of Sexual Behavior [asb]

12

PL186-228369

October 21, 2000

17:7

Style file version Nov. 19th, 1999

Alison, Santtila, Sandnabba, and Nordling

Weinberg, M., Williams, C. J., and Moser, C. (1984). The social constituents of sadomasochism. Soc. Probl. 31: 379–389. Weinberg, T. S. (1987). Sadomasochism in the United States: A review of recent sociological literature. J. Sex Res. 23: 50–69. Weinberg, T. S., and Kamel, G. W. L. (1983). S & M: An introduction to the study of sadomasochism. In Weinberg, T., and Kamel, G. W. L. (eds.), S and M: Studies in Sadomasochism, Prometheus, Buffalo, NY, pp. 17–24.

P1: FZN/FVK Archives of Sexual Behavior [asb]

PL186-228370

October 19, 2000

17:44

Style file version Nov. 19th, 1999

Archives of Sexual Behavior, Vol. 30, No. 1, 2001

Sexual Attitudes and Number of Partners in Young British Men Emil Kupek, Ph.D.1

The relationship between sexual attitudes and number of heterosexual partners in a survey-based and nationally representative random sample of 551 British men aged 16–25 years was examined. The main predictor of the number of partners in the last 5 years was the time since the first sexual intercourse, whereas age, marital status, education, social class, smoking, and alcohol consumption contributed on a smaller but significant level. Sexual attitudes were summarized in terms of three underlying dimensions which could be described as permissiveness, attitudes toward sexual relations of same-sex partners, and importance of orgasm for sex. None of these was a significant predictor of the number of partners in the last 5 years. Both permissiveness and number of partners were associated with the age of first sexual intercourse and other background variables indicating opportunities for social contact. In conclusion, common factors of sexual attitudes and the number of sexual partners are not directly related but rather jointly predicted by a very similar set of background variables such as age, time since first sexual intercourse, social class, smoking, and alcohol consumption. Given the absence of a significant relationship between sexual attitudes and number of young men’s partners, promoting safer sex may be a more sensible strategy than trying to change these attitudes. KEY WORDS: health promotion; sexual attitudes; number of sexual partners; alcohol; covariance structure analysis.

INTRODUCTION Sexual attitudes and behavior have come under the spotlight of mainstream health research since the late eighties. This was principally due to the need to 1 Departamento de Sa´ ude P´ublica, Centro de Ciˆencias de Sa´ude, Universidade Federal de Santa Catarina,

Campus Universit´ario, Trindade, 88.040-900 Florian´opolis-SC, Brazil; e-mail: kupek@repensul. ufsc.br. 13 C 2001 Plenum Publishing Corporation 0004-0002/01/0200-0013$19.50/0 °

P1: FZN/FVK Archives of Sexual Behavior [asb]

14

PL186-228370

October 19, 2000

17:44

Style file version Nov. 19th, 1999

Kupek

identify the risk behaviors and attitudes regarding HIV transmission as a basis for successful intervention (see Kim et al., 1977, for a review). Most of the studies focused on specific risk groups defined in terms of high risk behaviors such as substance abuse, unprotected sex with large number of partners or with partners likely to be infected with HIV. Condom use, number of sexual partners, and risk for HIV were principal behavioral items, which were often accompanied with questions on attitudes toward condom use. Particular attention was given to adolescents and young people as a clear target group for behavioral change (Bagnall et al., 1990; Baker et al., 1996; Breakwell and Fife-Shaw 1992a,b; Bryan et al., 1996; Coker et al., 1994; Dunne et al., 1994; Durbin et al., 1993; Leigh et al., 1994; Lowry et al., 1994; McEwan et al., 1992; Sikand et al., 1996). In the early nineties, first results of nationally representative sex surveys based on probability samples from France and Great Britain were published (ACSF 1992; Johnson et al., 1994), whereas some other European countries used less extensive and methodologically more vulnerable types of survey (Matthews et al., 1995). In the US, the General Social Survey provided some basic information on sexual lifestyle (Anderson et al., 1992) for this period. However, these surveys did not pursue an in-depth investigation of the relationship between sexual attitudes, such as their structure, and their relationship toward number of sexual partners. For example, although many questions in the British National Survey of Sexual Attitudes and Lifestyles (BNS) dealing with attitudes, views, and opinions were related to important outcome measures (Johnson et al., 1994), this was done on the level of individual items. The only analysis which looked into global patterns of response in terms of principal components analysis excluded some important sexual attitudes (Johnson et al., 1994, pp. 387–422). Principal component scores were compared in a descriptive way for a range of demographic variables but not to any behavioral outcome. The aim of this work is to build on these findings by expanding the analysis of the structure of sexual attitudes and examining its relationship with number of sexual partners. It was known from previous analyses that a large number of reported sexual partners were associated with respondents’ age, gender, marital status, age of first sexual intercourse, and social class in a multivariate model (Johnson et al., 1994). As the effects of some of these factors on the number of partners were known to be nonlinear (e.g., age) and gender specific, stratified analysis on these variables seemed a more realistic choice than to try to accommodate all these complex effects within one model. In this analysis, young men (16–25 years) were chosen for two reasons. First, young people are known to be more likely to try to change their attitudes and sexual behavior than the older ones are as numerous examples of recent efforts in the context of AIDS prevention campaigns have shown (Freudenberg, 1989; Matthews et al., 1995; Wellings, 1992). This is the practical reason why health promotion is usually aimed at this segment of the population. Second, gender specification was shown to be important

P1: FZN/FVK Archives of Sexual Behavior [asb]

PL186-228370

October 19, 2000

17:44

Sexual Attitudes and Number of Partners in Young British Men

Style file version Nov. 19th, 1999

15

and maintained throughout previous analyses (Johnson et al., 1994). Young men are a clear target group for promotion of sexual health, which might benefit from the analysis of the relationship between sexual attitudes and number of partners.

SUBJECTS AND METHODS Data for the present study were collected as a part of BNS, which was conducted with a random sample of 18,876 men and women aged 16–59 years living in Great Britain. They were interviewed and invited to complete a questionnaire containing more sensitive questions. The primary aim of the survey was to obtain reliable data on risky sexual behavior and help in developing sexual health education programs. Attitudinal questions with a strong moral judgment were placed towards the end of the questionnaire after the questions on sexual behavior. The final sample achieved was representative of the target population in terms of gender, age, ethnicity, marital, and socioeconomic status (Johnson et al., 1994). A subsection of participants (N = 4548) was randomly selected to receive additional items on sexual attitudes. Data from male participants aged 16–25 years (N = 551) were selected for the present investigation. Overall response rate was 63.3%, which was higher than that of previous studies of sexual behavior in Britain and some other countries and comparable to large social surveys of less sensitive topics (Johnson et al., 1994). The acceptance rate was 71.5% among eligible respondents who could be interviewed. The number of heterosexual partners in the last 5 years was chosen for the behavioral outcome to be related to sexual attitudes and other relevant predictors. Of the 551 men aged 16–25 years, 106 reported having no heterosexual partners in the last 5 years. Of these 106, only 4 reported having a homosexual partner. Internal consistency of answers to different questions and external validation of some self-reports, as well as reasonably low item nonresponse (Kupek, 1998), all indicated good quality of the data. The methodology of the survey is discussed in more detail elsewhere (Johnson et al., 1994; Wadsworth et al., 1993). Covariance structure analysis (CSA) (also known as Structural Equations Model (SEM) and Linear Structural Models (LISREL)) was used to combine factor and regression analyses in one model. This requirement emerged from the need to summarize a large number of attitude items into a small number of interpretable response patterns (factor analysis), whose predictive value to the behavioral outcome can then be tested in a regression model. One type of relationship between sexual attitudes and number of partners is that of mutual influence (no causality implied with this term). CSA was used because more than a single regression equation is needed to express this type of relationship. This method can account

P1: FZN/FVK Archives of Sexual Behavior [asb]

PL186-228370

October 19, 2000

17:44

Style file version Nov. 19th, 1999

16

Kupek

for the intrinsic dependency structure between the variables, such as number of sexual partners, and the time since first sexual intercourse (both will be zero for those with no partners). Furthermore, CSA can properly account for the fact that factor analysis scores of sexual attitudes are not observed but model-based values and hence cannot be simply regressed on the dependent variable. The latter twostep approach—factor analysis followed by regression—might be a reasonable approximation for eliminating clearly insignificant predictors but underestimates the degrees of freedom used up in the factor analysis model. The capacity of CSA to handle simultaneously both factor analysis and regression allows an external validation of predictive value of common factors with respect to the outcome of interest. Because of the complexity of CSA, several analyses were done in the course of the model building. Transformations and Selection of the Variables Logarithmic transformation of the main outcome variable (number of partners in last 5 years) was used to stabilize its variance and linearize it with respect to the independent variables in the model. A small constant was added to make this transformation possible for respondents with no partners. Among the background variables, respondent’s age and time of sexual activity were continuous. Education was treated as an ordinal scale ranking respondents with degrees, higher education or A-level, O-level or CSE or other (mainly nonBritish) diplomas, and none of the above. Smoking comprised four categories: nonsmokers, ex-smokers, light smokers, and heavy smokers. Alcohol consumption was ranked as none, low, moderate, and high. Marital status originally comprised seven categories: married, cohabiting with opposite-sex partner, cohabiting with same-sex partner, divorced, separated, widowed, and single. As only two respondents lived with same-sex partners, these were combined with those cohabiting with opposite-sex partners. Those divorced, separated, or widowed were treated as being in one category on the grounds that they all have had a live-in relationship in the past. Previous analysis for men (Johnson et al., 1994, p.128) showed that with respect to the number of lifetime partners, these respondents are closer to single men than to the married men. This provided a basis on which to attempt an ad hoc ordinal scale regarding respondents’ experience of a formal (marriage) or at least socially binding live-in relationship (e.g., common friends). Along this line, respondents were ranked highest if married, those cohabiting were given the second rank, those divorced/separated/widowed were placed third, and single men formed the last category. The bias due to treating ordinal scale variables as interval scale was reported to be sufficiently minor in comparison with other possible

P1: FZN/FVK Archives of Sexual Behavior [asb]

PL186-228370

October 19, 2000

17:44

Style file version Nov. 19th, 1999

Sexual Attitudes and Number of Partners in Young British Men

17

sources of model misspecification (Bentler and Newcomb, 1992) for many practical applications.

Testing Multivariate Normality The assumption of a multivariate normal distribution is important for maximum likelihood estimates of parameters’ variance and chi-square test for overall goodness of fit in CSA to be valid. Both graphical investigations of linearity between ordered Mahalanobis distances and chi-square distribution (Krzanowski, 1988) and correlation were used to demonstrate this assumption (Filliben, 1975; Sharma, 1996).

Factor and Covariance Structure Analysis As the viability of the multivariate normal distribution assumption was confirmed with these data, a maximum likelihood estimation method was used both for factor analysis and CSA. A range of CSA goodness of fit indices as well as the distribution of normalized residuals were examined. Several models were systematically tested for meaningful modifications and the key findings presented. The fit of the factor analysis and regression parts of the model were also examined. SAS software was used (SAS Institute Inc., 1989a). It is important to remember that CSA fits the covariance matrix calculated from the raw data. Although the variables not related by a regression equation appear unconnected in graphical representations of CSA, their covariance is set to the value calculated from the raw data, not to zero as the absence of an arrow might suggest. To avoid excessive statistical details, only p values or the rejection criterion ( p > .05) were provided for some analyses. The covariance matrix used for the CSA is available on request.

RESULTS Basic descriptive statistics for sexual attitudes and the number of partners is given in Table I. The majority (82%) of respondents were single, 9% were married, and 9% cohabited with a partner of the opposite sex. Half of the respondents reported first sexual intercourse between the age of 16 and 20, and 28% said it occurred before the age of 16. Almost one-fifth (19.5%) of the 551 men aged 16–25 years reported having had no heterosexual partners in last 5 years, 20.1% reported one partner, 10.7% reported two partners, 22.5% reported three or four partners, 14.8% reported between five and nine partners, 10.4% said they had had

P1: FZN/FVK Archives of Sexual Behavior [asb]

PL186-228370

October 19, 2000

17:44

Style file version Nov. 19th, 1999

18

Kupek Table I. Percentage of Answers by Response Category (N = 551) Attitudes

1

2

3

4

5

DK/M

Sex before marriage Adultery—marriage Adultery—cohabitation Unfaithfulness—regular partners One-night stands Sex between 2 men Sex between 2 women Abortion Older people want less sex Unfaithfulness is harmless Companionship more imp. than sex Sex, no orgasm not satisfying—men Sex, no orgasm not satisfying—women Sex better with one partner Sex most important part of marriage Sex better in longer relationship

2.3 50.0 33.3 27.8 30.9 57.4 47.7 19.1 0.6 1.6 14.1 5.6 4.2 6.6 2.5 10.5

4.4 34.0 33.9 34.4 20.9 10.7 15.4 14.5 33.1 16.1 52.4 39.4 30.1 36.1 16.0 52.7

11.2 12.1 20.7 21.5 20.3 4.0 6.4 32.5 28.5 10.4 20.4 19.0 22.9 26.8 20.3 21.3

10.2 1.3 4.3 7.8 11.0 4.2 4.6 9.8 24.0 54.4 9.6 28.3 29.7 21.0 50.5 7.1

70.7 1.2 6.6 6.8 15.4 21.6 23.5 18.2 2.8 15.4 0.6 1.7 1.5 2.3 8.0 0.7

1.1 1.3 1.3 1.7 1.6 2.1 2.4 5.8 10.9 2.2 2.8 6.1 11.7 7.2 2.7 7.7

Note. 1: agree strongly; 2: agree; 3: neither agree nor disagree; 4: disagree; 5: strongly disagree; DK/M: “don’t know” and not answered; see Appendix for exact wording of the questions.

ten or more partners, and 1.9% did not respond to this item (for more details see Johnson et al., 1994).

The Structure of Sexual Attitudes: Factor Analysis Factor analysis for attitude variables was a distinctive part of the CSA model, whose results (factor scores) could then be related to the outcome. Three factors were found to be optimal for this model. Two alternative hypotheses—that there was no common factor and that more than three common factors were needed—were both rejected on the basis of the chi-square for the likelihood ratio test ( p < .01) and the drop from the third to fourth eigenvalues (3.8 to 0.8). The factor loadings were rotated to a simple solution according to the varimax criterion (Krzanowski, 1988), which yields uncorrelated factors with variables predominantly loading on one factor (Table II). The first factor had high loadings on attitudes towards unfaithfulness (i.e., in marriage, cohabitation, and with regular partner), one-night stands, the harm caused by unfaithfulness, and whether monogamous sex was valued. This factor was interpreted as permissiveness versus restrictiveness in sexual attitudes as all variables had this in common. Thus, respondents with low scores on this factor evaluated unfaithfulness and one-night stands as wrong, disagreed with the statement that unfaithfulness was harmless, and agreed that sex is better when sticking to one partner. Those with high scores on the permissiveness factor expressed opposite attitudes.

P1: FZN/FVK Archives of Sexual Behavior [asb]

PL186-228370

October 19, 2000

17:44

Style file version Nov. 19th, 1999

Sexual Attitudes and Number of Partners in Young British Men

19

Table II. Varimax Rotated Factor Pattern (Standard Error) Attitudes

Factor I

Factor II

Factor III

Sex before marriage Unfaithfulness in marriage Unfaithfulness in cohabitation Unfaithfulness—regular partners One-night stands Sex between 2 men Sex between 2 women Abortion Older people want less sex Unfaithfulness is harmless Companionship more imp. than sex Sex, no orgasm not satisfying—men Sex, no orgasm not satisfying—women Sex better with one partner Sex most important part of marriage Sex better in longer relationship

28 (7) 64 (7) 80 (11) 80 (12) 36 (10) 5 (24) 8 (23) 23 (9) 3 (7) −41 (7) 20 (7) 1 (17) 1 (15) 27 (7) −7 (8) 11 (6)

21 (9) 18 (10) 4 (16) 5 (16) 8 (14) 94 (38) 98 (40) 36 (14) 10 (7) −9 (9) −4 (8) 3 (24) 4 (22) 4 (7) 9 (12) 6 (6)

−7 (7) 10 (9) −3 (8) −1 (9) 2 (10) 12 (38) 11 (40) 0 (15) 7 (6) 7 (7) −11 (6) 91 (20) 87 (17) 1 (6) 31 (7) 7 (6)

Note. Values are multiplied by 100 and rounded to the nearest integer. The loadings with the highest contribution are in bold.

The second factor was predominantly determined by attitudes toward sexual relations of same-sex partners. The interpretation of the third factor is importance of orgasm for sex as the loadings for two questions regarding this issue for men and women had far the highest loadings. The model residuals were low and normally distributed. The permissiveness factor accounted for 37% of the variation in attitudes, whereas the homosexual acceptance and importance of orgasm factors accounted for 34% and 29% respectively. Relationship Between Permissiveness and Number of Partners: Covariance Structure Analysis Before CSA modeling, ordinary least squares regression was used to examine the predictive value of factor scores for the number of partners in last 5 years. A bivariate model was fitted for each of the identified factors. Only permissiveness score was statistically significantly related to the outcome ( p = .05), whereas the other two factors’ scores were not ( p = .21 and p = .18 for factors II and III, respectively). Only sexual attitudes, which were identified as key contributors to the permissiveness factor, were selected for the CSA. Background variables known to influence sexual behavior and attitudes (Johnson et al., 1994) were added to the model to explore their effects in a systematic way. As a result of this analysis, a model was found, which satisfied both statistical and interpretational criteria (Fig. 1 and Table III). The factor structure of

P1: FZN/FVK Archives of Sexual Behavior [asb]

PL186-228370

October 19, 2000

17:44

Style file version Nov. 19th, 1999

20

Kupek

Fig. 1. Relationship between attitudes, number of partners and background variables. (Note: Error terms not displayed.)

permissiveness was retained, but no relationship with the number of partners could be confirmed statistically. They appear unrelated to each other but simultaneously influenced by background variables such as age, time since first sexual intercourse, social class, smoking, and alcohol consumption. Older men with a longer period of sexual activity, higher education and social class, higher alcohol and cigarette consumption, and unmarried men reported more partners in the last 5 years. The only attitude associated with number of partners was that on “one-night stands.” Respondents who were more approving of “one-night stands” reported having more sexual partners. Both observed and latent attitudes were associated with background variables. The higher the age and education, the higher the tolerance of adultery. Also, older men more often agreed that unfaithfulness was harmless. Married men thought that sex was better when sticking to one partner more often than did those who were never married. This attitude was also associated with longer time since first sexual intercourse and lower alcohol consumption. Other possible associations of the background variables with the remaining attitudes in the model were tested and rejected both on the basis of likelihood ratio and univariate t-test ( p > .05). The above model achieved an excellent overall fit. The chi-square value for the likelihood ratio test against independence model (χ 2 = 27.37; df = 41; p = .95)

P1: FZN/FVK Archives of Sexual Behavior [asb]

PL186-228370

October 19, 2000

17:44

Style file version Nov. 19th, 1999

Sexual Attitudes and Number of Partners in Young British Men

21

Table III. Parameter Estimates for the Model Presented in Fig. 1 Parameter code Factor analysis a1 a2 a3 a4 a5 a6 Background b1 b2 b3 b4 b5 b6 b7 c1 c2 c3 c4 c5 d1 d2 d3 d4 d5 d6 d7 d8

Raw

Standard error

t-test value

Standardized

1.000 1.145 1.397 1.404 −1.215 0.670

— 0.157 0.180 0.256 0.207 0.148

— 7.305 7.765 5.484 −5.879 4.524

0.546 0.460 0.551 0.451 −0.566 0.326

0.149 0.247 0.041 0.337 −0.046 0.059 −0.142 −0.027 0.099 0.038 0.070 0.054 0.351 0.769 0.908 0.418 −0.269 −0.303 −0.436 0.242

0.049 0.080 0.016 0.121 0.019 0.014 0.033 0.017 0.053 0.013 0.026 0.018 0.179 0.276 0.073 0.141 0.096 0.086 0.155 0.110

3.021 3.096 2.624 2.768 −2.369 4.281 −4.342 −2.795 1.874 2.775 2.657 3.022 1.964 2.780 12.463 2.969 −2.806 −3.532 −2.824 2.194

0.156 0.151 0.142 0.136 −0.128 0.192 −0.176 −0.283 0.126 0.264 0.183 0.223 0.082 0.105 0.687 0.119 −0.119 −0.196 −0.107 0.081

Note. Error term estimates for attitude questions, number of partners, and permissiveness are not displayed as they are of secondary importance here. The value of parameter “a1” was fixed to 1.

clearly rejected the alternative hypothesis that the similarity between the databased and model-based covariances is spurious. The goodness-of-fit index and its adjusted version (Joreskog and Sorbom, 1981; SAS Institute 1989b) were .99 and .98 respectively. Upper bound for root mean square error approximation was only .002 and the probability of close fit was 1.0 (Browne and Cudeck, 1993). The distribution of asymptotically standardized residuals was symmetric and within 95% confidence limits expected on the basis of the sampling variation. Several variations of the model presented were examined and rejected on the basis of the likelihood ratio and univariate t-test, using p < .05 as the criterion: a) no correlation between measurement errors for the attitudes on unfaithfulness in marriage, cohabitation, and regular partnership (χ 2 = 71.73, df = 44), b) no predictive value of attitude on one-night stands on the number of partners (χ 2 = 32.13, df = 42),

P1: FZN/FVK Archives of Sexual Behavior [asb]

PL186-228370

October 19, 2000

22

17:44

Style file version Nov. 19th, 1999

Kupek

c) including permissiveness factor as predictor of the number of partners in the last 5 years (χ 2 = 24.40, df = 40), d) including number of partners in the last 5 years as predictor of permissiveness (χ 2 = 25.23, df = 40).

DISCUSSION The aim of this work was to examine the structure of sexual attitude and the relationship between sexual attitudes and number of heterosexual partners. Sexual attitudes can be summarized in terms of three underlying factors: permissiveness, acceptance of homosexuality, and importance of orgasm for sex. This result expanded previous analysis of eight sexual attitudes, which included both men and women (Johnson et al., 1994) by showing that additional attitudes could be summarized in terms of the importance of orgasm for sex. Permissiveness and number of partners among young men were not found to be related directly, but were jointly determined by the time since first sexual intercourse, age, and other variables indicating opportunity for social contacts. It should be noted that permissiveness is largely defined in terms of sexual morality by using the strong evaluation word “wrong” in question 39, as well as by large weight given to the items on unfaithfulness (cf. Appendix). Permissiveness was associated with the same background variables as the number of partners except for marital status and education. Smoking, alcohol consumption, earlier age of first sexual intercourse and thus longer time of sexual activity, lower social class, and younger age were all associated with higher permissiveness score. Respondents with longer times since first sexual intercourse reported more partners in the last 5 years. Time since first sexual intercourse was the strongest predictor of the number of partners and explained 47% of the variance of number of partners. Various other studies have found early start of sexual activity strongly associated with number of partners (ACSF, 1992; Coker et al., 1994; Durbin et al., 1993). Higher educational qualifications and social class were also found to be associated with higher number of partners in the last 5 years in our sample. The role of these factors has been suggested for the general population (Johnson et al., 1994). Both factors are associated with a greater opportunity for social contact (e.g., leaving parental home to receive university education elsewhere). Smoking and alcohol consumption were associated with, both, more permissive sexual attitudes and larger number of partners in last 5 years (Table II). This is likely to reflect the degree of acceptance of peer culture norms (e.g., going out to pubs or similar places where social contacts are often made). Other British studies also found social drinking and smoking to be associated with having a boyfriend/girlfriend (Health Education Authority, 1992) and more sexual partners (Bagnall et al., 1990; Breakwell and Millward, 1997; McEwan et al., 1992). Some

P1: FZN/FVK Archives of Sexual Behavior [asb]

PL186-228370

October 19, 2000

17:44

Sexual Attitudes and Number of Partners in Young British Men

Style file version Nov. 19th, 1999

23

US studies reported similar relationship between drinking and smoking, and having a number of sexual partners (Andersson and Dahlberg, 1992; Lawry et al., 1994; Leonard and Ross, 1997). Because sexual partnership may depend on opportunity for social contacts, it seems reasonable to find more permissive sexual attitudes being associated with indicators of the social context, which typically mediates these contacts. The attitude towards adultery was associated with age and education (Fig. 1). Although 54% of men aged 16–19 years thought sex outside marriage was always wrong, this figure was about 5% lower among men aged 20–25 years. Education made a more palpable difference: 60% among respondents with degrees and 92% among those with no qualifications disapproved of adultery. Liberal attitudes have been found to be associated with higher education (Johnson et al., 1994). They also seem to develop gradually among young men as the association with age suggested. Men who had had first sexual intercourse earlier, were never married, and drank more alcohol were likely to disagree with the statement that monogamy brings a more satisfying sexual life than changing partners (“Sex better with one partner”). They also had more sexual partners and higher permissiveness scores. Thus number of sexual partners and permissiveness are not directly related but both of them are associated with the same background variables that seem to be related to sexual lifestyle—monogamy versus multiple partnerships. The latter was found to be more attractive to the men who did not have a partner or had had first sexual intercourse a short time ago. Among the attitude items, only “one-night stands” showed a small but significant effect on number of partners. This may be due to two main reasons. First, the length of time since initiation of sexual activity, age, and related variables such as marital status, social class, and education would be expected to determine the number of partners to a large extent as they define some important aspects of biological and social framework of sexual behavior. Second, the sexual attitudes analyzed here were fairly general social norms regarding sexual behavior as opposed to more specific attitudes regarding personal action in the near future. Only the latter were shown to have sufficient predictive power regarding behavioral outcomes of interest to inform a successful intervention (Ajzen, 1988, 1991; Terry et al., 1993). However, the knowledge of general norms is a necessary background against which specific conative aspects of the attitudes can be selected. It has been argued that sexual behavior is generally difficult to predict because it is impulsive, interactive (i.e., also depends on partner’s behavior), and is affected by situational factors such as alcohol consumption. For example, only carefully executed studies achieved good prediction of condom use based on sexual attitudes by asking about intention to use a condom (Terry et al., 1993). An additional difficulty is that prediction depends on accuracy of self-reported behavioral outcome, which in turn depends on respondents’ memory and honesty, but the survey

P1: FZN/FVK Archives of Sexual Behavior [asb]

24

PL186-228370

October 19, 2000

17:44

Style file version Nov. 19th, 1999

Kupek

method still seems a viable method despite its limitations (ACSF, 1992; Breakwell and Fife-Shaw, 1992b; Johnson et al., 1994; Mathews et al., 1995; Siegel et al., 1998). Low to moderate correlations between the three variants of attitudes towards unfaithfulness—in marriage, cohabitation and, regular partnership—significantly contribute to the overall fit of the models examined. Although a common factors model would normally assume these correlations to be zero (Dunn et al., 1993; Joreskog and Sorbom, 1981; Krzanowski, 1988), there are good reasons for this not to hold in this particular case. First, all three items varied the context of the question whereas its core—unfaithfulness—remained the same. Hence any confounding is likely to affect the supposedly unique error variances of these items in a similar way, thus resulting in significant correlations between them. Second, the confounding is probable in this case because only one of the three latent factors found to predict the observed attitudes—permissiveness—was used in CSA modeling. The reason for exclusion of the other two common factors identified in the factor analysis prior to the CSA was their lack of predictive power regarding the number of partners in the last 5 years. Correlated measurement errors in a common factors model indicate that the manifest variables measure something else or something in addition to what has been accounted for by the latent variable (Joreskog, 1993). The two omitted factors regarding tolerance of homosexual behavior and importance of orgasm for sex are known additional sources of variation, which act as confounders in the CSA models examined. There may well be other unknown confounders that enhance this effect. Although some of the above findings might have been anticipated, their empirical confirmation is important in two ways. First, the structure of a wider range of sexual attitudes has been identified and its relationship with number of partners examined for nationally representative data. Second, this may be useful for historic and between-country comparisons. On the other hand, the exploratory nature of the study (i.e., absence of firm hypotheses) and large-scale survey perspective may impose limitations for such comparisons. Sexual attitudes and behavior in this case seem to be parallel developments determined to a large extent by a common social background. If permissiveness is seen as an example of sexual morality, the acquisition of which is supposed to regulate important aspects of sexual behavior, then its impact on partner acquisition among men aged 16–25 years seems to be insignificant in comparison with some general factors that enhance opportunities for social contact. Although there is a widespread agreement in terms of sexual attitudes, this does not correspond to the variations in number of partners. One important implication for sexual health promotion is that normative evaluation of sexual behavior is not likely to change young men’s partner acquisition, so the promotion of safer sex seems a much more meaningful campaign. The fear that such campaigns aimed at STD prevention and/or improving knowledge of contraception methods can make young people

P1: FZN/FVK Archives of Sexual Behavior [asb]

PL186-228370

October 19, 2000

17:44

Style file version Nov. 19th, 1999

Sexual Attitudes and Number of Partners in Young British Men

25

more promiscuous by changing their sexual attitudes also seems unfounded from this perspective.

APPENDIX: SEXUAL ATTITUDE QUESTIONS 39. From this card, what are your opinions about following sexual relationships a) If a men and a women have sexual relations before marriage, what would your general opinion be? b) What about a married person having sexual relations with someone other than his or her partner? c) What about a person who is living with a partner, not married, having sexual relations with someone other than his or her partner? d) And a person who has a regular partner they don’t live with, having sexual relations with someone else? e) What about a person having one-night stands? What is your general opinion about f) Sexual relations between two adult men? g) And sexual relations between two adult women? h) Lastly, what is your general opinion about abortion? Response options: 1 (always wrong), 2 (mostly wrong), 3 (sometimes wrong), 4 (not wrong at all), 5 (not wrong at all), 8 (depends/don’t know). 40. Now please would you say how far you agree or disagree with each of these things a) It is natural for people to want sex less often as they get older? b) Having a sexual relationship outside a regular one doesn’t necessarily harm that relationship? c) Companionship and affection are more important than sex in marriage or relationship? d) Sex without orgasm, or climax, cannot be really satisfying for men? e) Sex without orgasm, or climax, cannot be really satisfying for women? f) A person who sticks with one partner is likely to have a more satisfying sex life than someone who has many partners? g) Sex is the most important part of any marriage or relationship. h) Sex tends to get better the longer you know someone? Response options: 1 (agree strongly), 2 (agree), 3 (neither agree nor disagree), 4 (disagree), 5 (strongly disagree), 8 (don’t know).

P1: FZN/FVK Archives of Sexual Behavior [asb]

PL186-228370

October 19, 2000

17:44

26

Style file version Nov. 19th, 1999

Kupek

ACKNOWLEDGMENTS The author is indebted to Jane Wadsworth whose support and insightful comments were indispensable for this work, as well as to other principal investigators of the National Survey of Sexual Attitudes and Lifestyles, Anne Johnson, Julia Field, and Kaye Wellings for making the data available for the analysis. REFERENCES ACSF Principal Investigators and their associates (1992). Analysis of sexual behavior in France (ACSF)—A comparison of two modes of investigation: Telephone survey and face-to-face survey. AIDS. 6: 315–323. Ajzen, I. (1988). Attitudes, Personality and Behaviour, Open University Press, Milton Keynes. Ajzen, I. (1991). The theory of planned behaviour. Organiz. Behav. Human Decision Processes. 50: 179–211. Anderson, J. E., and Dahlberg, L. L. (1992). High-risk sexual behavior in the general population. Results from a national survey, 1988–1990. Sex. Transm. Dis. 19(6): 320–325. Bagnall, G., Plant, M., and Warwick, W. (1990). Alcohol, drugs and AIDS-related risks: Results from a prospective study. AIDS Care. 2(4): 309–317. Baker, S. A., Morrison, D. M., Carter, W. B., and Verdon, M. S. (1996). Using the theory of reasoned action (TRA) to understand the decision to use condoms in an STD clinic population. Health. Educ. Q. 23(4): 528–542. Bentler, P. M., and Newcomb, M. D. (1992). Linear structural modeling with nonnormal continuous variables. Application: Relations among social support, drug use, and health in young adults. In Dwayer, J. H., Feinlieb, M., Lippert, P., and Hoffmeister, H. (eds.), Statistical Models for Longitudinal Studies of Health, Oxford University Press, Oxford, pp. 132–160. Browne, M. W., and Cudeck, R. (1993). Alternative ways of assessing model fit. In Bollen, K. A., and Long, J. S. (eds.), Testing Structural Equation Models, Sage Publications, Newbury Park, CA, pp. 136–162. Breakwell, G. M., and Fife-Shaw, C. (1992a). Sexual activities and preferences in a United Kingdom sample of 16 to 20-year-olds. Arch. Sex. Behav. 21(3): 271–293. Breakwell, G. M., and Fife-Shaw, C. R. (1992b). Estimating sexual behaviour parameters in the light of AIDS: A review of recent UK studies of young people. AIDS Care. 4(2): 187–201. Breakwell, G. M., and Millward, L. J. (1997). Sexual self-concept and sexual risk-taking. J. Adolesc. 20(1): 29–34. Bryan, A. D., Aiken, L. S., and West, S. G. (1996). Increasing condom use: Evaluation of a theorybased intervention to prevent sexually transmitted diseases in young women. Health. Psychol. 15(5): 371–382. Coker, A. L., Richter, D. L., Valois, R. F., McKeown, R. E., Garrison, C. Z., and Vincent, M. L. (1994). Correlates and consequences of early initiation of sexual intercourse. J. Sch. Health. 64(9): 372– 377. Dunn, G., Everitt, B., and Pickles, A. (1993). Modelling Covariances and Latent Variables Using EQS, Chapman and Hall, London. Dunne, M. P., Donald, M., Lucke, J., Nilsson, R., Ballard, R., and Raphael, B. (1994). Age-related increase in sexual behaviours and decrease in regular condom use among adolescents in Australia. Int. J. STD. AIDS. 5(1): 41–47. Durbin, M., DiClemente, R. J., Siegel, D., Krasnovsky, F., Lazarus, N., and Camacho, T. (1993). Factors associated with multiple sex partners among junior high school students. J. Adolesc. Health. 14(3): 202–207. Filliben, J. J. (1975). The probability plot correlation coefficient test for normality. Technometrics. 17(1): 111–117. Freudenberg, N. (1989). Preventing AIDS: A Guide to Effective Education for the Prevention of HIV Infection, American Public Health Association, Washington.

P1: FZN/FVK Archives of Sexual Behavior [asb]

PL186-228370

October 19, 2000

17:44

Sexual Attitudes and Number of Partners in Young British Men

Style file version Nov. 19th, 1999

27

Health Education Authority (1992). Tomorrow’s Young Adults, Health Education Authority, London. Joreskog, K. G., and Sorbom, D. (1981). LISREL VI, Uppsala University Press, Uppsala. Joreskog, K. G. (1993). Testing structural equation models. In Bollen, K. A., and Long, J. S. (eds.), Testing Structural Equation Models, Sage Publications, Newbury Park, CA, pp. 294–316. Johnson, A. M., Wadsworth, J., Wellings, K., and Field, J. (1994). Sexual Attitudes and Lifestyles, Blackwell Scientific Publications, Oxford. Kim, N., Stanton, B., Li, X., Dickersin, K., and Galbraith, J. (1977) Effectiveness of the 40 adolescent AIDS-risk reduction interventions: A quantitative review. J. Adolesc. Health. 20(3): 204–215. Krzanowski, W. J. (1988). Principles of Multivariate Analysis: A User’s Perspective, Clarendon Press, Oxford. Kupek, E. (1998). Determinants of item non-response in a large national sex survey. Arch. Sex. Behav. 27(6): 581–594. Leigh, B. C., Morrison, D. M., Trocki, K., and Temple, M. T. (1994). Sexual behavior of American adolescents: Results from a U.S. national survey. J. Adolesc. Health. 15(2): 117–125. Leonard, L., and Ross, M. W. (1997). The last sexual encounter: The contextualization of sexual risk behaviour. Int. J. STD. AIDS. 8(10): 643–645. Lowry, R., Holtzman, D., Truman, B. I., Kann, L., Collins, J. L., and Kolbe, L. J. (1994). Substance use and HIV-related sexual behaviors among US high school students: Are they related? Am. J. Public Health. 84(7): 1116–1120. Matthews, M., Wellings, K., and Kupek, E. (1995). AIDS/HIV KAB Surveys in the European Community (General Population). London School of Hygiene and Tropical Medicine, London, UK. McEwan, R. T., McCallum, A., Bhopal, R. S., and Madhok, R. (1992). Sex and the risk of HIV infection: The role of alcohol. Br. J. Addict. 87(4): 577–584. SAS Institute Inc. (1989a). Proc CALIS. In SAS/STAT R User’s Guide, Version 6, Vol. 2, SAS Institute Inc., Cary, NC, pp. 245–366. SAS Institute Inc. (1989b). Introduction to structural equations with latent variables. In SAS/STAT User’s Guide, Version 6, SAS Institute Inc, Cary, NC, pp. 137–187. Sharma, S. (1996). Applied Multivariate Techniques, John Wiley and Sons, New York. Siegel, D. M., Aten, M. J., and Roghmann, K. J. (1998). Self-reported honesty among middle and high school students responding to a sexual behavior questionnaire. J. Adolesc. Health. 23(1): 20–28. Sikand, A., Fisher, M., and Friedman, S. B. (1996). AIDS knowledge, concerns, and behavioral changes among inner-city high school students. J. Adolesc. Health. 18(5): 325–328. Terry, D. J., Gallois, C., and McCamish, M. (eds.) (1993). The Theory of Reasoned Action: Its Application to Aids-Preventive Behaviour, Pergamon Press, Oxford. Wellings, K. (1992). Assessing AIDS Prevention in the General Population, Report for the EC Concerted Action ‘Assessment of AIDS/HIV preventive strategies,’ Academic Department of Public Health, St. Mary’s Hospital Medical School, London. Wadsworth, J., Field, J., Johnson, A. M., Bradshaw, S., and Wellings, K. (1993). Methodology of the National Survey of Sexual Attitudes and Lifestyles. J. R. Statis. Soc. A. 156(3): 407–421.

P1: FDO Archives of Sexual Behavior [asb]

PL186-228371

October 19, 2000

19:18

Style file version Nov. 19th, 1999

Archives of Sexual Behavior, Vol. 30, No. 1, 2001

Personality, Individual Differences, and Preferences for the Sexual Media1 Anthony F. Bogaert, Ph.D.2

The extent to which personality and individual differences predict preferences for and choices of various forms of sexual media was examined. Personality (e.g., intelligence, aggression) and individual difference factors (e.g., prior sexual experience) were assessed in 160 undergraduate men. These men also indicated their preferences for and choices of various forms of sexual media (e.g., “erotic,” female insatiability, violent). As expected, individual differences were predictive, with, for example, men lower in intelligence and higher in aggressive/antisocial tendencies having a higher preference for violent sexual stimuli than men higher in intelligence and lower in aggressive/antisocial tendencies had. In addition, as much as 50% of the variation in the preference for violent sexual materials was accounted for by an additive combination of individual differences and self-report arousal to these materials. Finally, the results indicated that, when given a choice to view different media materials, the men chose a broad range of media materials, although the “female insatiability” films were more popular than the other sexual films (e.g., “erotic” or violent). Results are discussed in relation to recent research and psychological theories that view adults as active in choosing their own social environments. KEY WORDS: pornography; personality; erotica; insatiability; violence.

INTRODUCTION Increasingly, developmental, personality, and social psychologists (e.g., Buss, 1984; Plomin and Bergeman, 1991; Plomin et al., 1977; Scarr and McCartney, 1 This research was completed in partial fulfilment of the author’s Ph.D. requirements at the University

of Western Ontario. Portions of this manuscript were completed while the author was a postdoctoral fellow with the Social Sciences and Humanities Research Council of Canada at the Clarke Institute of Psychiatry in Toronto, Canada. 2 Community Health Sciences, Brock University, St. Catharines, Ontario, Canada L2S 3A1; e-mail: [email protected]. 29 C 2001 Plenum Publishing Corporation 0004-0002/01/0200-0029$19.50/0 °

P1: FDO Archives of Sexual Behavior [asb]

PL186-228371

October 19, 2000

19:18

30

Style file version Nov. 19th, 1999

Bogaert

1983; Snyder and Ickes, 1985) are theorizing that individuals are active in choosing the situations or environments, including media environments, to which they are exposed. Although a vast array of environments is available, people will often be selective, choosing only those environments that are best suited to them and/or are congruent with their dispositions or personalities. This person/choice of environment association has recently been noted for sexuality as well (Money, 1986; Mosher, 1988). Mosher (1988), for example, argues that people will be selective in their sexual media viewing habits, directing their choices to material that is congruent with preexisting sexual scripts, fantasies, and dispositions. Thus, some people will never seek out sexual materials; others may have some inclination and seek out these materials only occasionally; and still others will be very inclined and will seek out these materials on a frequent basis. Also, some people will be more inclined than others to seek out or be interested in specific types of sexual media, including violent portrayals, or female sexual insatiability themes (i.e., sexually eager, nondiscriminating women), the latter of which have been reported to be prevalent in the sexual media (e.g., Brosius et al., 1993). In this study, these ideas are pursued further. Although there has been research on the relation between individual differences (e.g., personality dispositions) and an overall proclivity for the sexual media, there has been a relatively modest amount of work investigating individual difference correlates of interests in specific types of sexual media. This is the first purpose of the present study. In addition, previous sexual media preference research has typically investigated single or relatively few individual differences such as personality (Barnes et al., 1984a), fantasy (Malamuth and McIlwraith, 1988; McIlwraith and Josephson, 1985), positive sexual affect (Fisher et al., 1988), or prior exposure to the sexual media (e.g., Zillmann and Bryant, 1986). This work, though important in its own right, may well be limited because it fails to take a multivariate approach. A multivariate approach (e.g., principal components’ analysis, multiple regression analysis) assesses common content/variation among relevant individual difference predictors and the potential for additive, mediating, and/or interactive processes. Thus, a multivariate approach can provide a more complete, multidimensional account of sexual media preferences. This is the second purpose of the present study.

Individual Differences Theory and research suggests that a range of personality and individual difference variables may play a role in decisions regarding the viewing of sexual materials. The following discussion introduces these variables. First, antisocial tendencies may be relevant to the present investigation. Such constructs as aggressiveness and attraction to sexual aggression, which refer to a man’s belief system that sexual aggression is arousing for both the aggressor and the victim (Malamuth, 1989a,b), may relate to a preference for violent sexual

P1: FDO Archives of Sexual Behavior [asb]

PL186-228371

Preferences for the Sexual Media

October 19, 2000

19:18

Style file version Nov. 19th, 1999

31

materials. Similarly, hypermasculine sex roles (e.g., Mosher and Sirkin, 1984), which consist of three violence-related components (i.e., calloused sexual attitudes, a conception of violence as manly, and a view of danger as exciting), and Eysenck’s third major personality type or psychoticism (e.g., Eysenck and Eysenck, 1976), also interpreted as antisocial personality (e.g., Claridge, 1983), may relate to preferences for violent sexual materials. All four of these constructs predict an interest in or experience with actual sexual violence. For example, aggressiveness has been related to prior and future sexual violence (e.g., Greendlinger and Byrne, 1987), whereas attraction to sexual aggression has been related to the endorsement of rape myths and elevated arousal to rape depictions in a laboratory (Malamuth, 1989a,b). Similarly, hypermasculine sex roles relate to less distaste for realistic portrayals of rape (Mosher and Anderson, 1986), and psychoticism relates to prior exposure to rape depictions (Barnes et al., 1984b). One explanation for why these antisocial tendencies (e.g., attraction to sexual aggression) may relate to an interest in these types of violent sex films is that these films serve as arousing and potentially pleasurable sexual stimuli. For example, some prior research suggests that people may choose to view sexual material because of its stimulation or arousal properties (Brown and Bryant, 1989; Winick, 1971). In addition, more aggressive/antisocial people, relative to less aggressive/antisocial people, may prefer violent sexual content because it is a more familiar and/or a more “cognitively accessible” theme (e.g., Huesmann, 1988). Aggressive/antisocial people (relative to those lower in these characteristics) may also experience these materials with less emotional disturbance (e.g., viewing them as less distasteful, as found by Mosher and Anderson, 1986), and, given that those higher in antisocial tendencies have more experience with violence, these aggressive materials may reinforce or validate their existing strategies of interpersonal relations. Finally, aggressive/antisocial people may be more likely than nonaggressive/antisocial people to choose antisocial or unconventional sexual materials because they may be more likely to be assertive about satisfying unconventional interests and curiosities. Second, dominance motives and related constructs, such as Machiavellianism, which is the tendency to use manipulation, power, and opportunism to achieve personal goals (Christie and Geis, 1970), may also relate to interest in violent sexual themes, in part because domineering and controlling behavior may be important motives in the appreciation of aggressive sexual acts against women and children (e.g., Burgess, 1984; Groth, 1979; Malamuth, 1986). Thus, those high in dominance motives may wish to view power or dominance motives being satisfied because these themes are familiar and/or cognitively accessible, viewed as acceptable and less distasteful, and reinforce existing (or at least desired) strategies of interpersonal relations. Some existing research supports an association between such constructs and exposure to violent films. For example, Tamborini et al. (1987) assessed Machiavellianism, along with masculinity and sensation seeking, and found that all three constructs were positively correlated with preference for graphic

P1: FDO Archives of Sexual Behavior [asb]

32

PL186-228371

October 19, 2000

19:18

Style file version Nov. 19th, 1999

Bogaert

interpersonal violence in mainstream (nonsexual) films and, in combination with prior viewing habits, yielded a Multiple R of .51 in predicting preference for media violence. Such dimensions may also relate to preference for themes of female sexual insatiability, which also have been argued to satisfy dominance, control, and power motives (e.g., Cowen et al., 1988; Matacin and Burger, 1987). Dominance or related motives may be satisfied with these themes because females are portrayed as eager to succumb to, service, and satisfy the sexual interests of males. In addition to aggression/antisocial tendencies and dominance motives, those constructs related to sexual variety/novelty and experience are relevant to the present investigation. In particular, erotophilia/erotophobia or positive-to-negative sexual affect (Fisher et al., 1988), sensation seeking, the need for novel stimulation and arousal (Zuckerman, 1979), previous sexual experience, and previous pornography exposure may relate to general interest in the sexual media but also to more novel forms of it. One cognitive explanation for why these associations could occur is that high prior experience with a given behavior (e.g., prior exposure to the sexual media) provides an individual with highly “accessible” internal cues to the acceptability, normativeness, and rewards of that behavior. In addition, scoring higher on these types of variables should make viewing sexual materials a more positive emotional (e.g., erotophilia) or a more exciting and arousing (e.g., sensation seeking) experience. Some prior research does support the idea that these individual difference variables relate to sexual media preferences. For example, Becker and Byrne (1985) found that erotophiles, relative to erotophobes, are more likely to watch sexually explicit films and to have fewer errors of recall about these materials, and Zuckerman and Litle (1986) found that sensation seeking predicted prior attendance to sexually explicit films, including more novel sexual media forms. In addition, prior exposure to sexual materials relates to an interest in the general sexual media and the portrayal of more novel sexual activities (Kelley and Musialowski, 1986; Reifler et al., 1971; Tamborini et al., 1987; Zillmann and Bryant, 1986). Interestingly, however, some research suggests that men with a high level of sexual experience may view some forms of sexual material (e.g., aggressive) more realistically and hence more critically, perhaps because of having appropriate sexual comparisons (Malamuth and Check, 1983). Given these findings, men high in sexual experience may have a high interest in sexual materials generally, but may have a low interest in more extreme antisocial content within these media (e.g., sexual violence). Finally, intelligence may play a role in sexual media viewing habits. To the present author’s knowledge, no research has addressed directly whether this variable predicts media content preferences. However, given the association between lower intelligence and antisocial behavior (e.g., Giancola and Zeichner, 1994; Huesmann et al., 1987), individuals lower in IQ may prefer more aggressive media portrayals, including aggressive sexual materials, in part because they are more

P1: FDO Archives of Sexual Behavior [asb]

PL186-228371

October 19, 2000

19:18

Preferences for the Sexual Media

Style file version Nov. 19th, 1999

33

familiar and hence more cognitively accessible themes, and because they reinforce preexisting or recognized strategies of coping. They may also be more likely than individuals higher in intelligence to prefer aggressive material, including sexually aggressive material, perhaps because the portrayal of antisocial acts may be viewed less critically (e.g., Burnett, 1991; Jacobvitz et al., 1991). Summary and Objectives In the present investigation, 11 personality and individual difference variables (aggression, attraction to sexual aggression, dominance, erotophilia, hypermasculine sex roles, intelligence, machiavellianism, psychoticism, previous sexual media exposure, sensation seeking, and sexual experience) were used to predict preferences for content in the sexual media. However, given that many of these variables likely share variance, they were first subjected to principal components’ analysis to extract common factors. Then, multivariate analyses (e.g., multiple regression) were used to investigate the extent to which the resulting individual difference factors, alone and in combination, predict preference for and choice of different sexual media materials. In addition, the present study attempted to provide some insight into the consumption rates of various content (e.g., violent, female sexual insatiability themes) in sexually explicit materials. Current content analyses suggest that overt aggression (e.g., rape) is not prevalent whereas other themes such as female sexual insatiability are more common (e.g., Brosius et al., 1993). The present study tested whether similar consumption patterns occur within a laboratory setting. Consumption information (from field and lab studies) provides an ecological context to evaluate the impact of specific types of sexual media materials on society. Thus, if female sexual insatiability themes increase negative attitudes toward women (e.g., Zillmann and Bryant, 1989) and sexual violence can increase aggression (e.g., Donnerstein and Berkowitz, 1981), then consumption studies are necessary to establish whether these content elements are popular enough to have widespread societal effects. METHOD Participants University of Western Ontario undergraduates (160 males) with a mean age of 19.2 years were tested in groups of 5–15. These undergraduates received 2 credits toward their introductory psychology requirements for participating in this research. Men were tested because they are the primary consumers of explicit sexual materials. Some investigators report that men consume at least three times the amount of explicit sexual materials as compared to women (Shepher and Reisman,

P1: FDO Archives of Sexual Behavior [asb]

PL186-228371

October 19, 2000

19:18

34

Style file version Nov. 19th, 1999

Bogaert

1985). In addition, if sexual materials contribute to men’s antisocial behaviors toward women, as some have argued (e.g., Zillmann and Bryant, 1989), then men should be the primary participants of sexual media studies that investigate or have implications for antisocial effects. Personality Measures To measure aggression and dominance, the 16-item true/false Aggression (e.g., “When I’m irritated I let it be known”) and Dominance (e.g., “I try to control others, rather than permit them to control me”) scales from the Personality Research Form were employed (PRF; Jackson, 1984a). In the present study, internal consistency levels (i.e., alpha coefficients) for these scales were .69–.77, respectively. To measure attraction to sexual aggression, six 11-point items comprising the short form of the Attraction to Sexual Aggression scale (Malamuth, 1989a) were employed. This scale contains items such as, “If you could be assured that no one would know and that you could in no way be punished, how likely are you to engage in rape?” The alpha coefficient for this scale in the present study was .80. To measure sexual affect or erotophilia/erotophobia, the twenty-one-item 7-point scale, The Sexual Opinion Survey (e.g., “Masturbation can be an exciting experience”), was used (Fisher et al., 1988). In the present study, the alpha coefficient for this scale was .76. To measure Machiavellianism, the twenty-item 7-point Mach IV scale (e.g. “Never tell anyone the real reason you did something unless it is useful to do so”) was used (Christie and Geis, 1970). In the present study, the alpha coefficient for this scale was .76. To assess hypermasculinity, the 30-item forced choice scale, The Hypermasculinity Inventory (e.g., “All is fair in love and war”) was used (Mosher and Sirkin, 1984). The internal consistency value for this scale in the present study was .82. The 24-item true/false Psychoticism (“Do you sometimes like to tease small animals?”) scale was also employed (EPQ; Eysenck and Eysenck, 1975). Originally reported alpha coefficients for this scale were relatively high, although recent studies have observed lower levels (e.g., .49; Malamuth, 1986). In the present study, the alpha coefficient for this scale was .52. To assess sensation seeking, the 40-item forced choice Sensation Seeking Scale (e.g., “I like wild uninhibited parties”) was used (Zuckerman, 1979). In the present study, the alpha coefficient for this scale was .76. Finally, to measure intelligence, four subscales, Spatial (50 items), Digit Symbol (35 items), Similarities (34 items), and Comprehension (28 items), comprising the short-form of the Multidimensional Aptitude Battery (MAB; Hill and Jackson, 1984; Jackson, 1984b) were included. The MAB is a timed, multiple choice, group-administered IQ test that is similar to the well-known WAIS-R. In the present study, internal consistency levels for these four scales were .86 and above. Although usually allotted 7 min, participants were restricted to 5 min per subscale because of time constraints. This time-saving

P1: FDO Archives of Sexual Behavior [asb]

PL186-228371

October 19, 2000

19:18

Preferences for the Sexual Media

Style file version Nov. 19th, 1999

35

procedure should have little bearing on the rank ordering of individuals in intelligence (see Jackson, 1984b, pp. 42, 43).3 Sexual Experience Prior (12 items) and current (12 items) sexual experience scales were also constructed (e.g., “Have you ever engaged in oral stimulation of a female’s genitals?” and “In the past two months, have you engaged in oral stimulation of a female’s genitals?,” respectively). Internal consistency values for these scales were .88 and .92, respectively. In addition, number of sexual partners in the participant’s lifetime was assessed and converted to a standard score and then added to the standardized sum of the prior sexual experience scale to form an aggregate measure of past sexual experience. Finally, prior exposure to sexual materials (14 items) in the past year was assessed (e.g., “How often have you seen films or videotapes showing explicit portrayals of sexual intercourse?”). In the present study, the alpha coefficient for this scale was .87. Additional Measure: Social Desirability The 16-item true/false Social Desirability scale (e.g., “I am always prepared to do what is expected of me”) was also used (PRF; Jackson, 1984a). In the present study, the internal consistency value for this scale was .64. This scale assesses the inclination to respond to questionnaires in a socially favorable manner. Given that some of the media material presented may be construed as socially undesirable, this measure was included as a possible control. In addition, this scale was included because, as Jackson (1984a) has pointed out, social desirability deserves some consideration as a valid personality measure of high socialization and social conformity. As such, this construct should relate negatively to preference for unconventional and/or antisocial sexual materials (e.g., sexually violent portrayals). Media Stimuli A review of available content analyses research of sexual materials indicates that a number of themes (e.g., female as sexually insatiable, violence) occur with varying degrees of frequency (e.g., Brosius et al., 1993). To establish an array of sexual media materials for participants’ choices that reflects this diversity, a pool of “promotional” descriptions of fictional films was produced. The primary emphasis 3 Self-report

delinquency and altruism were also assessed in this study, but they had very limited predictability. In addition, prior sexual aggression was assessed in this study, but it had little variation. Thus, these measures are not reported here (but see Bogaert, 1993).

P1: FDO Archives of Sexual Behavior [asb]

PL186-228371

October 19, 2000

19:18

Style file version Nov. 19th, 1999

36

Bogaert

here was on ecological validity, where different sexual promotional descriptions were patterned after existing promotional materials in the sexual video market (e.g., Videolog, 1991). Two descriptions for each of the five sexual content themes were included: sexually violent; insatiable or sexually eager, nondiscriminating female; erotic or affectionate relationships; child sexuality; and sexual novelty (e.g., bestiality). In addition, two nonsexual control themes were included: violent nonsexual and nonviolent nonsexual media. The promotional descriptions used in the present study are shown in Table I.4 Dependent Measures and Manipulation Checks Participants rated their likelihood of viewing each of the 14 films under four different circumstances. These four viewing circumstances assessed different levels of effort and anonymity/confidentiality involved with viewing, important considerations for potentially socially undesirable materials. In addition, a composite or aggregate of four measures (as opposed to a single measure) increases reliability of measurement (e.g., Epstein, 1979). These likelihood-to-view measures were as follows: (1) renting each on video; (2) seeing each in a movie theatre; (3) ordering each through mail order; and (4) watching each if it happened to appear on late-night television. These ratings were on 7-point scales, where “1” indicated no likelihood of viewing and “7” indicated a high likelihood of viewing. In addition to these likelihood-to-view measures, from detailed descriptions of each film, participants could select two films to view in a supposed future testing session. After making the film selections and these likelihood-to-view ratings, participants used 7-point scales to assess the promotional descriptions for the content manipulations, including explicitness, sexual arousal, violence, affection, female’s eagerness to please men sexually (i.e., female sexual insatiability), novelty, and child sexualization.5 Procedure Volunteer bias (e.g., self-selection) has been reported in sexual research (e.g., Saunders et al., 1985). To reduce the possibility of such bias, participants were not recruited for an overtly sexual study. The present investigation was advertised merely as a “personality” study and as an opportunity to pretest films for future 4 Initially,

the descriptions were rated on various dimensions (e.g., violence, novelty) by a pretest sample of 39 male and female undergraduates. The descriptions differed in predictable ways, with, for example, the sexually violent films being the most violent of the sexual films (see Bogaert, 1993). The bestiality film was taken from Zillmann and Bryant (1986). 5 One procedural note: The students first read a list of all the film descriptions so that they could make their pretest choice. After this, they made their self-report ratings, following a presentation of each film description.

P1: FDO Archives of Sexual Behavior [asb]

PL186-228371

October 19, 2000

19:18

Preferences for the Sexual Media

Style file version Nov. 19th, 1999

37

Table I. Promotional Descriptions by Content Erotic sex Passionate Embrace. XXX. After a month-long separation Wendy and Tom return to one another’s arms for a night of mutual pleasure. The heat and sexual intensity of their passion sets the screen on fire! Explicit sexual behavior. Night Moves. XXX. Jim and Mary have been together for years and they know how to light one another’s sexual fires. See them come to life with passion and hot sexual desire! Explicit sexual behavior. Insatiable sex Lusty Woman. XXX. This woman craves sex! On the job, at home, or anywhere, the lusty action never stops, thanks to the hottest and most willing honey ever! Explicit sexual behavior. Working Girl. XXX. Anytime, anywhere with anyone—that’s her motto! This hot and horny working girl does everything for everybody! Explicit sexual behavior. Violent sex She Wanted It All Along. XXX. Ted manhandles sexy Alice until her pain turns to sexual pleasure. See her get spanked, beaten and mistreated, all in the name of love! Explicit sexual behavior. Spare the Rod and Spoil the Fun. XXX. Bill is on his way to the top, but one thing stands in his way—sweet and sexy Suzy. Her show of resistance is pretty convincing, but that’s nothing that an iron hand and a good stiff beating can’t cure. Soon she’s hot, horny, and begging for more! Explicit sexual behavior. Child sex On My Lap. XXX. Little Sally is the tender and sweet age of seven, but she knows just what to do to turn on her man! It all starts by sitting “on his lap.” Explicit sexual behavior. The Younger the Better. XXX. She’s only eight and still in grade school, but she is well-schooled in the ways of sex! See this little dynamo do her stuff! Explicit sexual behavior. Novel sex Strange Fantasies From the Far East. XXX. You won’t see this anywhere else guys! Fresh from Japan, this stuff is new, weird, and hot! Explicit sexual behavior. Beauties and the Beasties XXX. Get ready for the strange and bizarre! Karen finds ecstasy with a well-endowed specimen of “man’s best friend” whereas Bob finds counting sheep is not nearly as much fun as doing other things to them. Explicit sexual behavior. Violent control Hockey Fight Outtakes. Twenty minutes of the year’s best brawls. Watch the blood and teeth fly! Not rated. World of Boxing. Mohammed Ali, Mike Tyson, Sugar Ray Leonard, and others do some serious head-bashing. Not rated. Control The MTV Awards. The top 10 videos of the last 5 years are presented, including releases from U2, Peter Gabriel, New Kids on the Block, Tracy Chapman, and more. Not rated. Best of “Saturday Night Live.” Eddie Murphy, John Belushi, Bill Murray, and others are highlighted from the classic late night comedy show! Not rated.

research (i.e., “Personality and Pretesting Film Study”). Once they arrived, however, participants were told that the study contained sexual content, and that, along with needing undergraduates to pretest films, the experimenter was interested in “personality and sexual behavior.” The personality questionnaires were administered in the first hour, and the sexual questionnaires, along with ratings of the film descriptions and a question assessing whether the participant was interested in the “future” pretesting session,

P1: FDO Archives of Sexual Behavior [asb]

PL186-228371

October 19, 2000

19:18

Style file version Nov. 19th, 1999

38

Bogaert

were administered in the second hour. The 14 promotional descriptions were presented to the participants in alphabetical order by title, so that they appeared to be in a “natural” order, as they might be found by browsing through a promotional catalogue or at a video store. Participants were informed that the promotions described real films to be used in future psychological research, but at the present time needed to be assessed for popularity and content, and then anonymously and confidentially pretested (e.g., to assess video quality, believability of the actors, etc.), for which they would receive additional credit. Those who chose to participate in the additional session were asked to indicate which two films they would most like to pretest. Those choosing not to pretest were asked to state why they did not want to participate, and to indicate two films they would have pretested if they had chosen to continue. To enhance the believability of the pretest session, a calender with various times, dates, and locations for pretesting was included. Participants were told that these times and dates could not be changed, so they should plan to arrive promptly to the pretesting session. As an additional procedure to enhance believability, participants were told that they needed to select two films from the array (instead of one) because the experimenter may not be able to schedule their first choice. When the participants finished the questionnaire, they were probed for their awareness of the true purpose of the study (e.g., whether they believed the future pretesting session was real). Finally, participants were given an extensive verbal and written debriefing and thanked for their participation. RESULTS Correlations among the personality and individual differences were consistent with expectations and prior research, with, for example, antisocial and aggression scales being related. Note as well that some of the scales were related to social desirability, particularly Machiavellianism, attraction to sexual aggression, psychoticism, dominance, and prior sexual experience (see Table II).6 Manipulations of Film Description Content Single factor repeated-measure ANOVAs with Scheffe post hoc contrasts, tested for film description differences on the content manipulations. As shown in Table III, the sexual films were rated as the most explicit, the erotic sex films as the most affectionate, the violent sex films and violent nonsexual films as the most violent, the female insatiable sex films (e.g., Lusty Woman) as most likely to have female characters that were the most eager to have sex with many men 6 The

correlations among the individual difference measures in this study have, in part, been reported elsewhere (see Bogaert, 1993, 1996).

—

.34∗∗∗ −.01

.20∗ — −.06 —

.01 .01

4

.19∗ −.03 —

.53∗∗∗ .18∗

5

.05 .12 .39∗∗∗ —

.21∗ .31∗∗∗

6

.17∗ −.07 .22∗∗ .14 —

.17∗ .14

7

−.15 .03 .42∗∗∗ .33∗∗∗ .16∗ —

.28∗∗ .29∗∗∗

8

.07 .13 .46∗∗∗ .18∗ .00 .32∗∗∗ —

.21∗ .07

9

.19∗ .08 .20∗ −.07 .04 .03 .17∗ —

.25∗∗ −.20∗

10

.16∗ −.02 .17∗ .01 .09 .04 .30∗∗∗ .37∗∗∗ —

.10 −.18∗

11

.03 .13 .12 .10 .34∗∗∗ .14 .28∗∗ .01 .12 —

.06 .12

12

.23∗∗ .09 −.11 −.39∗∗∗ −.13 −.35∗∗∗ .02 .24∗∗ .11 −.14 —

−.05 −.48∗∗∗

13

October 19, 2000 19:18

< .05; ∗∗ p < .01; ∗∗∗ p < .001.

—

1 Aggression 2 Attraction to sexual aggression 3 Dominance 4 Intelligence 5 Hypermasculinity 6 Machiavellianism 7 Prior sexual media exposure 8 Psychoticism 9 Sensation seeking 10 Sexual experience (prior) 11 Sexual experience (current) 12 Erotophilia 13 Social desirability

3

Table II. Correlations Among Individual Difference Variables 2

PL186-228371

∗p

1

Archives of Sexual Behavior [asb]

Predictor

P1: FDO Style file version Nov. 19th, 1999

Preferences for the Sexual Media 39

P1: FDO Archives of Sexual Behavior [asb]

PL186-228371

October 19, 2000

19:18

Style file version Nov. 19th, 1999

40

Bogaert Table III. Mean Ratings of Film Descriptions and Likelihood-to-View Measures Aggregated Across Content Film content

Content manipulations Explicit Affection Violence Female promiscuity Child sexualized Novelty Arousal Viewing likelihood Video rental Movie viewing Mail order Night TV Combined self-report

Erotic sex

Insatiable sex

Violent sex

Child sex

Novel sex

Violent nonsexual

Control

6.5c 5.0d 1.8a 3.4bc 1.2a 3.4b 5.6d

6.7c 2.8c 2.3b 6.2d 1.2ab 4.1c 5.6d

6.7c 1.8b 6.1e 3.2b 1.2ab 5.6d 4.1c

6.7c 1.7b 4.2d 3.7bc 6.5c 6.3e 3.0b

6.7c 2.1b 3.4c 4.0c 1.4b 6.0de 4.0c

1.4b 1.1a 6.4e 1.3a 1.1a 2.8a 1.0a

1.0a 2.1b 1.5a 1.5a 1.1a 2.7a 1.1a

2.9c 1.6b 1.6b 5.4c 2.9c

3.1c 1.6b 1.7b 5.6c 3.0c

2.2b 1.3ab 1.4ab 4.3b 2.3b

1.6a 1.1a 1.2a 3.1a 1.8a

2.3b 1.4ab 1.4ab 4.7b 2.5b

4.1d 2.4c 2.4c 5.6c 3.6d

4.4d 3.0d 2.3c 5.7c 3.8d

Note. The contrasts occur across film content categories, such that, for example, explicitness scores are compared for the erotic sex films versus the insatiable sex films versus the violent sex films, etc. Means with different superscripts differ at the .05 level, using Scheffe’s test for multiple contrasts. Ratings are on 7-point scales, where 1 is low and 7 is high.

(i.e., nondiscriminant sexual behavior), and the child sex films as most likely to have a “child sexualized.” The violent, child, and novel sex films were generally rated as the most novel of all the films. The erotic and insatiable sex films were equal in arousal and found to be the most arousing of all the films. The sexually violent films and the novel sex films were found to be intermediate in terms of arousal, and the child sex films were the least arousing of the sexual films, but were rated as more arousing than the nonsexual films were. Overall, the promotional descriptions seemed to tap adequately a range of different sexual and nonsexual content from which the participants could make selections and indicate preference.7 Film Preferences As mentioned, the participants were asked whether they would be interested in participating in a future film “pretesting” session. Six men reported suspicion 7 Some

readers may note that some confounding of content occurred. For example, the child sexualization films were rated as more violent than the female insatiability, the novel sex, or the erotic films (although not quite as high as the sexually violent films) were. This is understandable and was considered acceptable because child sexualization, like rape in the sexually violent films, implies coercion; thus, a greater degree of discrimination between the child sexualization and the sexually violent films in level of violence would have lacked ecological validity.

P1: FDO Archives of Sexual Behavior [asb]

PL186-228371

Preferences for the Sexual Media

October 19, 2000

19:18

Style file version Nov. 19th, 1999

41

about the pretesting session (i.e., thought it would not occur) on questions that were designed to assess whether the participants surmised the hypothesis of the study and were eliminated from further analyses. Of the remaining 154 men, 51% chose not to participate in the pretesting session. A lack of interest in the pretesting session partly reflected the fact that participants were tested between November and March, and some would have had fulfilled their introductory course requirements after completing the present study. Indeed, 78% of the “nonpretesting” participants indicated that they did not participate because they did not have the time or because they already had enough credits; 18% indicated that they did not participate because they were not interested in any of the films. As mentioned, all the men were asked to indicate which 2 of the 14 films they would most like to watch (even if they chose not to participate in the future “pretesting” session). Figure 1 presents the percentage of choices made by these participants (i.e., including those who chose not to participate in the future “pretesting” session). In order, they are as follows: female insatiability films (i.e., 26.8%); control (e.g., Best of Saturday Night Live) films (21.3%); violent control films (18.9%); novel sex films (12.2%); erotic films (9.8%); child sex films

Fig. 1. Pretest film choices. “Insatiable” refers to female sexual insatiability films (e.g., “Working Girl”); “Control” refers to nonviolent control films (e.g., “Best of Saturday Night Live”); “Violent” refers to violent nonsexual control films (e.g., “Hockey Fight Outtakes”); “Novel” refers to novel activity sex films (e.g., “Strange Fantasies from the Far East”); “Erotic” refers to affectionate, mutually consenting sex films (e.g., “Passionate Embrace”); “Child” refers to child sex films (e.g., “On My Lap”); and “Violent Sex” refers to violent sex films (e.g., “She Wanted It All Along”).

P1: FDO Archives of Sexual Behavior [asb]

42

PL186-228371

October 19, 2000

19:18

Style file version Nov. 19th, 1999

Bogaert

(5.9%); and violent sex films (5.6%). Given that there were seven film categories, one might propose that, all else being equal, each film category has approximately a 1 in 7 (or 14.3%) chance of being chosen at random (i.e., its “expected” value). Thus, to give some indication of the relative popularity of each film category, I tested whether its observed value (i.e., % of choices) significantly exceeded or was significantly lower than its expected value, using the z approximation to the binomial test. The insatiable sex films and both control film categories significantly ( p < .05) exceeded their expected values, whereas the erotic, child, and violent sex films were significantly lower than their “expected” values. In addition, I tested whether the percentage of sexual films chosen (i.e., all five sexual categories) differed from its expected or chance value (i.e., 5 of 7 film categories or 71.4%). It was significantly lower (60%) than its expected value.8 I also tested whether the percentage of choices for each film category differed from the other (and not just from their “expected” value). The insatiable sex film choices were chosen significantly ( p < .05) more frequently than all the other film category choices were; the two control film categories were not significantly different from one another but were chosen more frequently than the remainder of the film categories were; and the novel and erotic sex films did not significantly differ from one another but were chosen significantly more frequently than the sexually violent and child sex films were, which did not significantly differ from each other. I also analysed these data, based on whether the men had actually chosen to pretest a film. The “pretesting” participants showed the following pattern: insatiable sex = 31.3%; control = 16.0%; violent control = 13.3%; novel sex = 13.3%; erotic = 12.6%; sexually violent = 8.0%; child sex = 5.3%, with the insatiable sex category significantly exceeding its expected value and the sexually violent and child sex films significantly lower than their expected values. The percentage of sexual films chosen (70.7%) did not significantly differ from its expected value (71.4%). I also tested whether the percentage of choices for each film category differed from one another (and not just from their “expected” value). Again, the insatiable sex film choices were chosen significantly ( p < .05) more frequently than all the other film category choices were; the two control film categories, the erotic sex film category, and the novel sex film were not significantly different from one another but were chosen more frequently than the sexually violent and child sex film categories were, which did not significantly differ from one another. 8 It

is of note that selecting only those who showed some interest/willingness to view the films, that is, those signing up for a pretest and those saying they did not participate because they did not have time or already had enough credits, showed very similar results (insatiable sex = 27.9%; control = 18.9%; violent control = 18.5%; novel sex = 12.9%; erotic sex = 9.4%; child sex = 6.7%; violent sex = 5.5%). The insatiable and the nonviolent control films significantly ( p < .05) exceeded their expected values, and the erotic, child and violent sex films were significantly ( p < .05) lower than their expected values. In addition, the percentage of sexual films chosen (i.e., all four sexual categories) was significantly lower (63%) than its expected or chance value (i.e., 5 of 7 film categories or 71.4%).

P1: FDO Archives of Sexual Behavior [asb]

PL186-228371

October 19, 2000

19:18

Style file version Nov. 19th, 1999

Preferences for the Sexual Media

43

Film choices for the “nonpretesting” participants showed a somewhat different pattern, with the control films, the violent control films, and the insatiable sex films (27.2%, 24.3%, 23.5%, respectively) significantly exceeding their expected values, the novel sex films having an observed value (11%) not differing from its expected value, and the child sex films (6.6%), the erotic sex films (4.4%), and the sexually violent films (2.9%) having observed values less than their expected values. The percentage of sexual films chosen (48.1%) was significantly lower than its expected value (71.4%). I also tested whether the percentage of choices for each film category differed from the other (and not just from their “expected” value). The two control film categories and the insatiable sex film category did not differ but were significantly chosen more frequently than the other film categories were; the novel sex films did not differ from the child sex films, but the novel sex film category was chosen more frequently than the erotic and the sexual violence categories were; the child, the erotic, and sexual violence film categories did not significantly differ from one another. Within each film category (i.e., erotic, insatiable, control, etc.), the four likelihood-to-view measures (i.e., on video, in a theatre, etc.) were moderately related, with the mean correlation between these self-report likelihood ratings having a value of .41 and a mean alpha of .84 (see also Bogaert, 1993). These likelihood ratings for each film category (i.e., four likelihood ratings for each of the two films) were summed. Each participant, then, had seven aggregate self-report likelihood-to-view measures, one for each of the seven categories of films. On these aggregate self-report likelihood-to-view measures, the control films (i.e., the violent control and nonviolent control) were the most popular; the insatiable and erotic sex films, the second most popular; the novel sex and violent sex films, the third most popular; and finally, the child sex films, the least popular (see Table III). However, note that the extreme antisocial sexual films (e.g., violent sex and child sex films), even though they were less popular than the other films, had a modestly high propensity to be viewed under certain conditions, such as “if it happened to appear on late-night television.”

Principal Components’ Analysis of Individual Differences To reduce the number of individual difference predictors and extract more common factors, a principal components’ analysis with varimax rotation was performed. Five components with eigenvalues greater than 1 emerged. The loadings are shown in Table IV and are defined by (1) antisocial tendencies/low desirability, (2) dominance/aggression motives, (3) sexual experience, (4) interest in sexual variation, and (5) intelligence. Participants’ factor scores on each of the five components served as individual difference measures.

P1: FDO Archives of Sexual Behavior [asb]

PL186-228371

October 19, 2000

19:18

Style file version Nov. 19th, 1999

44

Bogaert Table IV. Factor Structure of Individual Differences with Varimax Rotation Factor I II III IV (Antisocial/low (Dominance/ (Sexual (Sexual V desirability) aggression motives) experience) variation) (Intelligence)

Aggression Attraction to sexual aggression Dominance Intelligence Hypermasculinity Machiavellianism Pornography experience Psychoticism Sensation seeking Sex experience (past) Sex experience (current) Erotophilia Social desirability Eigenvalue Variance (%)

.345 .655

.742 .155

.147 −.319

−.069 .138

.046 −.011

−.232 .025 .526 .628 .139

.761 .019 .496 .144 .337

.075 −.068 .398 −.006 −.042

.222 .029 .023 .145 .701

−.005 .915 .033 .242 −.260

.745 .388 −.120 −.107

−.018 .062 .315 .025

.267 .609 .571 .763

−.065 .104 −.134 .166

.004 .242 −.037 −.164

.139 −.715

−.097 .257

.148 .178

.818 −.205

.250 .178

3.0 23.3

2.0 15.4

1.2 9.4

1.1 8.6

1.1 8.4

Individual Differences and Viewing Preferences As mentioned, the four likelihood-to-view measures (i.e., renting each on video; watching each in a movie theatre; ordering each through mail order; and watching each if it happened to appear on late-night television) for each film category were summed. These aggregate likelihood-to-view measures served as the primary preference criteria. In addition, whether the participant chose to view a sexual film in the putative “future” testing session was used to form a second preference criterion. Recall that all the men were asked to indicate which 2 of the 14 films they would most like to watch (even if they chose not to participate in the future “pretesting” session). For this measure, a score of 2 indicated that the participant had chosen two sexual films to pretest, “1” indicated that he had chosen one sexual film to pretest, and “0” indicated that he had not chosen any sexual film to pretest. This choice variable tested whether the individual differences could discriminate between the actual choice of a sexual film (e.g., erotic, insatiable, violent) and the choice of a nonsexual film (i.e., two control film categories). To maximize the number of cases, both the pretesting and nonpretesting participants (i.e., those not interested in the additional testing session) were included. Multiple regression analyses were conducted by “force entering” each individual difference component (Cohen and Cohen, 1983). Summary statistics for these analyses are shown in Table V. Note that the beta weights in these analyses

.127∗∗∗

.357

∗p

< .05; ∗∗ p < 01; ∗∗∗ p < .001.

.494∗∗∗ .244

−.242

−.110

.058∗∗ .470∗∗∗ .221

.348

−.018

.108

.050∗∗

unique variance contribution of predictor.

.019

.192∗∗∗ .224

.001

−.031

.000

.038∗

.195

.052∗∗

.282

Betaa

b

.012

.121∗∗∗

.000

.012

.079∗∗∗

rs2

.462∗∗∗ .214

−.068

.219

−.019

.258

.306

Betaa

b

.005

.048∗∗

.000

.066∗∗∗

.148∗∗∗

rs2

Child sex

.484∗∗∗ .235

−.080

.316

.091

.315

.148

Betaa

b

.007

.100∗∗∗

.008

.099∗∗∗

.022

rs2

Sexual film

19:18

a Standardized regression coefficient. b Squared semipartial correlation indicating

.535∗∗∗ .286

−.138

b

.096∗∗∗

rs2

Novel sex

October 19, 2000

.416∗∗∗ .194

.021

.438

.000

.011

.228

Betaa .309

b

Film type

.021

rs2

Violent sex

PL186-228371

−.143

.012

.003

.108

.143

.012

.109

Betaa

Factor I: antisocial/low desirability Factor II: dominance/ aggression Factor III: sexual experience Factor IV: variation Factor V: intelligence Multiple R R2

b

Betaa

Predictor

rs2

Insatiable sex

Archives of Sexual Behavior [asb]

Erotic sex

Table V. Summary Statistics for Multiple Regression Analyses on Sexual Film Preferences

P1: FDO Style file version Nov. 19th, 1999

Preferences for the Sexual Media 45

P1: FDO Archives of Sexual Behavior [asb]

46

PL186-228371

October 19, 2000

19:18

Style file version Nov. 19th, 1999

Bogaert

are similar to zero-order correlations because the principal components with varimax rotation are uncorrelated.9 In general, the individual difference components were predictive. For example, for the sexually insatiable films, a Multiple R of .53 occurred, and two components, sexual variation and dominance/aggression motives, contributed uniquely to prediction (both ps < .01). In addition, note that for the sexually violent films, a Multiple R of .49 occurred, and 4 of the 5 individual difference components, that is, antisocial tendencies/low desirability, sexual variation, intelligence (negatively), and dominance/aggression motives, contributed uniquely to prediction (all ps < .05). Note as well that the analysis for the “sexual” choice measure yielded a Multiple R of .48 and had two components, sexual variation and dominance/aggression motives, that contributed uniquely to prediction (both ps < .01). However, a notable exception to these significant results was the sexual experience factor, which was unrelated to any of the film categories. (In addition, the nonsexual control films were unrelated to any of the individual difference factors and are not mentioned further.)10 Multiple regression analyses were also performed with the participant’s selfreport arousal rating for each particular sexual film category (see manipulation checks). These analyses were performed for two reasons. First, they indicated whether the participants’ self-report arousal to each film category contributes to prediction. As mentioned, some research suggests that people choose to view sexual material because of its stimulation or arousal properties (Brown and Bryant, 1989; Winick, 1971). Second, these analyses indicated whether the individual difference components could contribute to prediction over and above self-report arousal for that sexual film category. For example, arousal may mediate the relationship between some individual difference components and viewing preference for certain films; if this is the case, the previously observed relationships should be reduced or eliminated when arousal is controlled. As shown in Table VI, the individual difference components generally contributed to the prediction over and above self-report arousal. Note as well that the inclusion of the self-report arousal measures typically increased prediction beyond that achieved by an additive combination of the components. For example, for the sexually violent films, a Multiple R of .71 (R 2 = .50) occurred and three of the previously significant components, antisocial tendencies/low desirability, intelligence (negatively), and sexual variation, still contributed uniquely. 9 Correlational

analysis among the five factors verified that the five factors were, indeed, uncorrelated (i.e., all correlations were 0.00). 10 I also tested whether a number of interactions between the individual difference factors predicted interest in the sexual media. In particular, there is evidence that the interaction between factors similar to sexual variation and aggression/dominance predict exposure to Penthouse magazine (Malamuth and McIlwraith, 1988). In addition, I thought a number of these individual difference factors (e.g., antisocial/low desirability or aggression/dominance) may be particularly predictive of antisocial sexual content when the participant was lower in intelligence (e.g., an Intelligence × Aggression/Dominance interaction) . However, the addition of these interaction terms did not significantly add to prediction in any of the regression equations (all ps > .05).

.260

.003 .041∗∗ .018

.051 .223 −.134

unique contribution of predictor.

.705∗∗∗ .496

−.193

.129

.085

−.070

.034∗∗ .638∗∗∗ .407

.212

.045

.078

.016∗

.007

.012

.459 .276

Betaa

b

.005

.041∗∗

.002

.006

.186∗∗∗ .076∗∗∗

rs2

Novel sex

.735∗∗∗ .541

−.110

.085

.048

.156

.612 .186

Betaa

b

.012

.007

.002

.023∗

.326∗∗∗ .033∗∗

rs2

Child sex

.516∗∗∗ .266

−.068

.236

.122

.277

.203 .140

Betaa

b

.004

.046∗∗∗

.014

.073∗∗∗

.031∗ .020

rs2

Sexual film

19:18

a Standardized regression coefficient. b Squared semipartial correlation indicating ∗ p < .05; ∗∗ p < 01; ∗∗∗ p < .001.

.661∗∗∗ .429

.017

.056∗∗∗

.003

.111

.027∗

b

.252∗∗∗ .032∗∗

rs2

Film type

October 19, 2000

−.129

.167

Betaa .501 .184

b

.150∗∗∗ .026∗

rs2

Violent sex

PL186-228371

.511∗∗∗ .261

.053

.012

.110

.432 .160

.089∗∗∗ .016

.327 .126

Betaa

Self-report arousal Factor I: antisocial/low desirability Factor II: aggression/ dominance Factor III: sexual experience Factor IV: variation Factor V: intelligence Multiple R R2

b

Betaa

Predictor

rs2

Insatiable sex

Archives of Sexual Behavior [asb]

Erotic sex

Table VI. Summary Statistics for Multiple Regression Analyses on Sexual Film Preferences with Self-Report Arousal

P1: FDO Style file version Nov. 19th, 1999

Preferences for the Sexual Media 47

P1: FDO Archives of Sexual Behavior [asb]

PL186-228371

October 19, 2000

48

19:18

Style file version Nov. 19th, 1999

Bogaert

Although the addition of self-report arousal generally added to prediction and many of the variables remained significant, it is also clear that many of the significant factors were reduced in magnitude, suggesting that self-report arousal may be partly mediating the relation between these individual difference factors and preference for different sexual content. To investigate this possibility further, I used structural equation modeling (i.e., EQS, Bentler, 1989) to test three different models of how the individual differences relate to the preference for insatiable and sexually violent films. (For these analyses, I focussed on these two types of films because the insatiable films are apparently popular and the sexually violent films have been well-studied for antisocial effects.) The models were (1) individual difference variables (including self-report arousal) having fully independent effects on viewing preference; (2) individual difference variables being partly mediated through self-report arousal, and (3) individual difference variables being fully mediated through self-report arousal. The “part mediation” models best accounted for the data, with the “Goodness-of-Fit” indices (i.e., NFI and CFI values) reaching upper bound values of 1.00 relative to .85 or lower for the other models.

DISCUSSION The results of this study support an individual difference and person/choice of environment approach to understanding preferences for types of sexual media. Preferences for viewing sexual films were related to a number of individual difference components and these increased prediction when they were allowed to combine additively. For example, using multiple regression analyses, the results were particularly strong for the sexually violent films, with 50% (R 2 = .50) of the variation being accounted for and three components (i.e., antisocial tendencies/ low desirability, lower intelligence, and sexual variation), along with self-report arousal, contributing uniquely to prediction. Thus, a man who has a combination of such characteristics—that is, a high self-report arousal to such films, is higher in antisocial tendencies (along with lower social desirability), is lower in IQ (e.g., has as an uncritical view of antisocial acts), and has a high interest in sexual variation—would be particularly likely to view sexually violent films. These results replicate and extend previous research investigating individual difference predictors of preferences for and choices of sexual materials, including sexually violent films (e.g., Barnes et al., 1984a; Fisher et al., 1988; Malamuth, 1989a,b; Mosher and Anderson, 1986). Future research should continue to investigate the important role that individual differences play in media preferences, including attempts to replicate the pattern of results in the present study. This is particularly so for the multiple regression analyses, which need replication to demonstrate stability across subject populations. Interestingly, and in contrast to the generally significant results, the “sexual experience” factor, which was primarily defined by current and prior sexual

P1: FDO Archives of Sexual Behavior [asb]

PL186-228371

Preferences for the Sexual Media

October 19, 2000

19:18

Style file version Nov. 19th, 1999

49

experience and sensation seeking, was unrelated to any sexual media preferences. It is of note that similar (nonsignificant) results occurred when the individual scales comprising this factor (e.g., sensation seeking and sexual experience) were used as predictors in these data (see Bogaert, 1993). Perhaps high sexual experience and sensation seeking may indicate a preference for more “active” types of social/ sexual behaviors (e.g., finding new sexual partners), whereas viewing sexual films may be reflective of more restrained, more sedentary types of social/sexual interests and activities. Regardless of the explanation, however, these results give little support to previous studies using these dimensions (e.g., Zuckerman and Litle, 1986). One additional area of future research would be to continue to study intelligence as a predictor of media effects. In the present study, intelligence was a defining variable of a principal component that predicted preference for violent sexual materials. In two recent studies (Bogaert et al., 1999), intelligence was also predictive of actual behavior after viewing different kinds of sexual material (e.g., violent vs. nonviolent), with men lower in intelligence (relative to those higher) being more likely to be sexually suggestive towards a young woman after viewing sexually violent content. Thus, not only might lower intelligence (relative to higher) predispose individuals to choosing more violent sexual media content, but it may also make them more susceptible to it once they have chosen to view it. It would be interesting to assess if such effects also occur, using a wider range of media content. It would also be of interest to assess a wider range of participants (e.g., nonuniversity students) because of the restriction of the range of IQ/intelligence scores that typically occurs with college samples. Future research should also further explore the relationship between aggressive/dominance motives and preference for sexual materials, particularly female insatiability themes. Dominance motives, as mentioned, may be partly satisfied by viewing these “insatiability” themes because women are portrayed as eager to succumb to, service, and satisfy the sexual interests of men. This view is consistent with the position taken by a number of feminist writers (e.g., Cowen et al., 1988; Matacin and Burger, 1987). One possible alternative explanation for this relationship is that men who are more dominant/aggressive (relative to those less dominant/aggressive) are merely more assertive about seeking out fairly common “male-oriented” sexual interests/strategies (e.g., having many willing and eager sexual partners or short-term mates, see Buss and Schmitt, 1993), and thus, are more likely than less assertive men to report a strong interest in such themes. Future research should evaluate the degree to which one or (perhaps) both of these explanations account for this relationship. Future research should also continue to investigate mediating mechanisms underlying the relationship between individual difference factors and preferences for sexual films. Although not the primary focus of the present study, I found some support for a “sexual arousal” mechanism. For example, although self-report arousal generally added to prediction and many of the individual differences remained

P1: FDO Archives of Sexual Behavior [asb]

50

PL186-228371

October 19, 2000

19:18

Style file version Nov. 19th, 1999

Bogaert

significant once self-report arousal was controlled for, the regression and structural equation analyses indicated that self-report arousal is probably partly mediating the relations between the individual difference factors and preferences for different sexual content. Thus, some men (e.g., those higher in antisocial tendencies or dominance/aggression motives) may select to view antisocial sexual material partly because of its stimulation or arousal properties. Other plausible mediating mechanisms underlying the relationship among and preferences for sexual films include cognitive components (e.g., thoughts, imagery, and memories of aggression-related activities) and/or affective/emotional responses (e.g., lower levels of disgust, fear, anxiety, etc.). These additional mediating mechanisms, along with their possible relations to an arousal mechanism, should be tested in future research. Finally, it is of note that this study, although having a laboratory context and not assessing a full array of sexual and nonsexual themes in the media, did provide some insight into university men’s consumption pattern of (sexual) media materials. The results indicated that, when given a choice to view different media materials, the men chose a broad range of media materials, although the “female insatiability” films were more popular than the other sexual films were (e.g., “erotic” or violent). These data may suggest some limits to the external and ecological (i.e., “real life settings/environments”) validity of experimental studies assessing the impact of more extreme sexual materials (e.g., sexual violence). For example, extending the present findings to some of this well-cited research (e.g., Donnerstein and Berkowitz, 1981; Malamuth, 1981) suggests that only a small percentage (e.g., 5–6%) of the men in these types of studies would have voluntarily placed themselves in a position to view a sexually violent film, given an array of other options from which to choose. In addition, these results do not seem to support some feminist writers (e.g., MacKinnon, 1983) who imply that men are very interested in sexual aggression against women. Moreover, some of the men may have chosen to view some of the extreme materials because it was perceived as being sanctioned (even encouraged) by the experimenter, even though under normal circumstances these men may not have chosen to view such materials. On the other hand, perhaps 5–6% is not a trivial percentage, given the laboratory constraints of the study. Thus, although some men may have chosen to view extreme materials because it was sanctioned by the laboratory context, perhaps other (and even more) men may have been reluctant to choose to view such films in front of the experimenter because such films are perceived as socially undesirable and, ultimately, would have been viewed under unnatural (and potentially uncomfortable) conditions. In addition, it should be noted that the men reported a higher propensity to view such antisocial films under other conditions, such as if it happened to appear on late-night television. Regardless of the interpretation of the consumption rate for these extreme materials, however, future research should focus on or include material that content analyses and preference studies, such as the present one, indicate to be more prevalent and popular within the sexual media (i.e., female insatiability materials).

P1: FDO Archives of Sexual Behavior [asb]

PL186-228371

October 19, 2000

19:18

Preferences for the Sexual Media

Style file version Nov. 19th, 1999

51

ACKNOWLEDGEMENTS I appreciate the helpful comments from Ray Blanchard, Carolyn Hafer and several anonymous reviewers. I also acknowledge the guidance and support provided by my Ph.D. thesis advisors, W. A. Fisher and P. A. Vernon. REFERENCES Barnes, G. E., Malamuth, N. M., and Check, J. V. P. (1984a). Personality and sexuality. Pers. Indiv. Diff. 5: 159–172. Barnes, G. E., Malamuth, N. M., and Check, J. V. P. (1984b). Psychoticism and sexual arousal to rape depictions. Pers. Indiv. Diff. 5: 273–279. Becker, M. A., and Byrne, D. (1985). Self-regulated exposure to erotica, recall errors and subjective reactions as a function of erotophobia and Type A coronary prone behavior. J. Pers. Soc. Psychol. 48: 760–767. Bentler, P. M. (1989). EQS: Structural Equation Program Manual, BMDP Statistical Software, Inc., University of California, Los Angeles. Bogaert, A. F. (1993). The Sexual Media: The Role of the Sexual Media. Unpublished doctoral dissertation, University of Western Ontario, London, Canada. Bogaert, A. F. (1996). Volunteer bias in males: Evidence for both personality and sexuality differences. Arch. Sex. Behav. 25: 125–140. Bogaert, A. F., Woodard, U., and Hafer, C. L. (1999). Intelligence as a moderator of antisocial pornography effects. J. Sex Res. 36: 283–291. Brosius, H.-B., Weaver, J. W., and Staub, J. (1993) Exploring the social and sexual “reality” of contemporary pornography. J. Sex Res. 30: 161–170. Brown, D., and Bryant, J. (1989). The manifest content of pornography. In Zillmann, D., and Bryant, J. (eds.), Pornography: Research Advances and Policy Considerations, Erlbaum, Hillsdale, NJ, pp. 3–24. Burgess, A. W. (1984). Child Pornography and Sex Rings, Lexington Books, DC Heath, Lexington, MA. Burnett, J. J. (1991). Examining media habits of the affluent elderly. J. Advert. Res. 31: 33–41. Buss, D. (1984). Toward a psychology of person–environment (PE) correlation: The role of spouse selection. J. Pers. Soc. Psychol. 47: 361–377. Buss, D. M., and Schmitt, D. P. (1993). Sexual strategies theory: An evolutionary perspective on human mating. Psychol. Rev. 100: 204–232. Christie, R., and Geis, F. C. (1970). Studies in Machiavellianism, Academic Press, New York. Claridge, G. (1983). The Eysenck Psychoticism Scale. In Butcher, J. N., and Speilberger, C. D. (eds.), Advances in Personality Assessment, Vol. 2, Erlbaum, Hillsdale, NJ, pp. 71–114. Cohen, J., and Cohen, P. (1983). Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences, 2nd Ed., Erlbaum, Hillsdale, NJ. Cowen, G., Lee, C., Levy, D., and Snyder, D. (1988). Dominance and inequality in X-rated videocassettes. Psychol. Women Quart. 12: 299–311. Donnerstein, E., and Berkowitz, L. (1981). Victim reactions in aggressive erotic films as a factor in violence against women. J. Pers. Soc. Psychol. 41: 710–724. Epstein, S. (1979). On stability of behaviour: I. On predicting most of the people much of the time. J. Pers. Soc. Psychol. 37: 1097–1126. Eysenck, H. J., and Eysenck, S. B. G. (1975). Manual of the Eysenck Personality Questionnaire, Hodder & Stoughton, London. Eysenck, H. J., and Eysenck, S. B. G. (1976). Psychoticism as a Personality Dimension, Hodder & Stoughton, London. Fisher, W. A., Byrne, D., White, L., and Kelley, K. (1988). Erotophobia–erotophilia as a dimension of personality. J. Res. Pers. 25: 123–151. Giancola, P. R., and Zeichner, A. (1994). Intellectual ability and aggressive behavior in nonclinical– nonforensic males. J. Psychopathology Beh. Assess. 16: 121–130.

P1: FDO Archives of Sexual Behavior [asb]

52

PL186-228371

October 19, 2000

19:18

Style file version Nov. 19th, 1999

Bogaert

Greendlinger, V., and Byrne, D. (1987). Coercive sexual fantasies of college men as predictors of self-reported likelihood of rape and overt sexual aggression. J. Sex Res. 23: 1–11. Groth, A. N. (1979). Men Who Rape: The Psychology of the Offender, Plenum Press, New York. Hill, T. D., and Jackson, D. N. (1984). Brief form of the Multidimensional Aptitude Battery (MAB). (Research Bulletin No. 617), University of Western Ontario, London, Canada. Huesmann, L. R. (1988). An information processing model for the development of aggression. Aggressive Beh. 14: 13–24. Huesmann, L. R., Eron, L., and Yarmel, P. W. (1987). Intellectual functioning and aggression. J. Pers. Soc. Psychol. 52: 232–240. Jackson, D. N. (1984a). Personality Research Form Manual, Research Psychologists Press, Port Huron, MI. Jackson, D. N. (1984b). Multidimensional Aptitude Battery Manual, Research Psychologists Press, Port Huron, MI. Jacobvitz, R. S., Wood, M. R., and Albin, K. (1991). Cognitive skills and young children’s comprehension of television. J. Appl. Dev. Psychol. 12: 219–235. Kelley, K., and Musialowski, D. (1986). Repeated exposure to sexually explicit stimuli: Novelty, sex, and sexual attitudes. Arch. Sex. Behav. 15: 487–498. MacKinnon, C. (1983). Feminism, marxism, method, and the state: Toward feminist jurisprudence. Signs 8: 635–658. Malamuth, N. M. (1981). Rape fantasies as a function of exposure to violent sexual stimuli. Arch. Sex. Behav. 10: 33–47. Malamuth, N. M. (1986). Predictors of naturalistic sexual aggression. J. Pers. Soc. Psychol. 50: 953– 962. Malamuth, N. M. (1989a). The Attraction to Sexual Aggression Scale: Part 1. J. Sex Res. 26: 26–49. Malamuth, N. M. (1989b). The Attraction to Sexual Aggression Scale: Part 2. J. Sex Res. 26: 324– 354. Malamuth, N. M., and Check, J. V. P. (1983). Sexual arousal to rape depictions. J. Abnorm. Psychol. 92: 55–67. Malamuth, N. M., and McIlwraith, R. D. (1988). Fantasies and exposure to sexually explicit material. Commun. Res. 15: 753–771. Matacin, M. L., and Burger, J. M. (1987). A content analysis of sexual themes in Playboy cartoons. Sex Roles 17: 179–186. McIlwraith, R. D., and Josephson, W. L. (1985). Movies, books, music, and adult fantasy life. J. Commun. 35: 167–179. Money, J. (1986). Lovemaps, Irvington Press, New York. Mosher, D. L. (1988). Pornography defined: Sexual involvement theory, narrative context, and “goodness-of fit.” J. Psychol. Hum. Sex. 1: 67–85. Mosher, D. L., and Anderson, R. D. (1986). Macho personality, sexual aggression, and reactions to guided imagery of realistic rape. J. Res. Pers. 20: 77–94. Mosher, D. L., and Sirkin, M. (1984). Measuring a macho personality constellation. J. Res. Pers. 18: 150–163. Plomin, R. C., and Bergeman, C. S. (1991). The nature of nurture: Genetic influences on “environmental” effects. Beh. Brain Sci. 14: 705–732. Plomin, R. C., Defries, J. C., and Loehlin, J. C. (1977). Genotype-environment interaction and correlation in the analysis of human behavior. Psychol. Bull. 84: 309–322. Reifler, C. B., Howard, J., Lipton, M. A., Liptzin, M. B., and Widmann, D. E. (1971). Pornography: An experimental study of effects. Am. J. Psych. 128: 575–582. Saunders, D. M., Fisher, W. A., Hewitt, E. C., and Clayton, J. P. (1985). A method for empirically assessing volunteer selection effects: Recruitment procedures and responses to erotica. J. Pers. Soc. Psychol. 49: 1703–1712. Scarr, S., and McCartney, K. (1983). How people make their own environments: A theory of genotypeenvironment effects. Child Dev. 43: 424–435. Shepher, J., and Reisman, J. (1985). Pornography: A sociobiological attempt at understanding. Ethol. Sociobiol. 6: 103–114. Snyder, M., and Ickes, W. (1985). Personality and social psychology. In Lindsey, G., and Aronson, E. (eds.), Handbook of Social Psychology, Vol II, 3rd Ed., Random House, New York, pp. 883–943.

P1: FDO Archives of Sexual Behavior [asb]

PL186-228371

Preferences for the Sexual Media

October 19, 2000

19:18

Style file version Nov. 19th, 1999

53

Tamborini, R., Stiff, J., and Zillmann, D. (1987). Preference for graphic horror featuring male versus female victimization. Hum. Commun. Res. 13: 529–552. Videolog. (1991). Trade Service Corporation, San Diego, CA. Winick, C. (1971). A study of consumers of explicitly sexual materials: Some functions served by adult movies. In Technical Reports of the Commission on Obscenity and Pornography, Vol. 4, U.S. Government Printing Office, Washington, DC, pp. 245–262 Zillmann, D., and Bryant, J. (1986). Shifting preferences in pornography consumption. Commun. Res. 13: 560–578. Zuckerman, M. (1979). Sensation Seeking: Beyond the Optimal Level of Arousal, Erlbaum, Hillsdale, NJ. Zuckerman, M., and Litle, P. (1986). Personality and curiosity about morbid and sexual events. Pers. Indiv. Diff. 7: 49–56.

P1: FZN Archives of Sexual Behavior [asb]

PL186-228372

October 19, 2000

19:29

Style file version Nov. 19th, 1999

Archives of Sexual Behavior, Vol. 30, No. 1, 2001

Sexual Functioning After Treatment for Testicular Cancer—Review and Meta-Analysis of 36 Empirical Studies Between 1975–2000 Grieteke Jonker-Pool, M.A.,1,2,7 Harry B.M. Van de Wiel, M.A., Ph.D.,1,3 Harald J. Hoekstra, M.D., Ph.D.,4 Dirk Th. Sleijfer, M.D., Ph.D.,5 Mels F. Van Driel, M.D., Ph.D.,6 Jean P. Van Basten, M.D., Ph.D.,4 and Heimen Schraffordt Koops, M.D., Ph.D4

Literature concerning sexual functioning after treatment for testicular cancer from 1975–2000 is reviewed. After a literature search in Medline and Psylit was conducted, as well as a search for cross-references made, a meta-analysis was performed. To describe sexual functioning, several aspects of the sexual response cycle were used: sexual desire, sexual arousal, erection, and orgasm; ejaculatory function, sexual activity, and sexual satisfaction were used as well. The number of patients included in the studies as well as treatment modalities were taken into account. A total of 36 relevant studies was screened (28 retrospective and 7 prospective studies), concerning 2,786 cases of testicular cancer. Meta-analysis revealed that ejaculatory dysfunction was reported most frequently and was related to surgery in the retroperitoneal area. Erectile dysfunction was related to irradiation, but was reported least frequently. Other sexual functions were not related to treatment modality. Meta-analysis revealed no deterioration of sexual functioning in the course of time, except a decrease in sexual desire and an increase in sexual satisfaction. Retrospective studies reported more sexual dysfunction than did prospective studies. Detailed analysis of separate studies, however, revealed a wide variation in reported sexual morbidity, as well as in assessment methods. Somatic consequences of disease and treatment may reduce ejaculation; however, 1 Department

of Medical Psychology, Groningen University Hospital, The Netherlands. Centre for Health Care Research, Groningen, The Netherlands. 3 Dutch Institute for Social Sexological Research, Utrecht, The Netherlands. 4 Department of Surgical Oncology, Groningen University Hospital, The Netherlands. 5 Department of Medical Oncology, Groningen University Hospital, The Netherlands. 6 Department of Urology, Groningen University Hospital, The Netherlands. 7 To whom correspondence should be addressed at Department of Medical Psychology, Groningen University Hospital, PO Box 30.001, 9700 RB Groningen, The Netherlands; e-mail: [email protected]. 2 Northern

55 C 2001 Plenum Publishing Corporation 0004-0002/01/0200-0055$19.50/0 °

P1: FZN Archives of Sexual Behavior [asb]

56

PL186-228372

October 19, 2000

19:29

Style file version Nov. 19th, 1999

Jonker-Pool, Van de Wiel, Hoekstra, Sleijfer, Van Driel, Van Basten, and Koops

other aspects of sexual functioning are not clearly related to disease- or treatmentrelated factors and may instead refer to a psychological vulnerability caused by one’s confrontation with a life-threatening, genito-urinary disease, such as testicular cancer. KEY WORDS: testicular cancer; sexuality; literature review; meta-analysis.

INTRODUCTION In the western countries testicular cancer is the most prevalent malignancy in males aged between 20 and 40 (Wingo et al., 1995). Although the disease is relatively rare, its incidence has profoundly increased in the past half century (Harding et al., 1995; Hoff Wanderas et al., 1995). In northern Europe the incidence is 7.5/100,000 inhabitants per year (Adami et al., 1994). In The Netherlands the incidence is 4.7/100,000 inhabitants per year (Visser et al., 1996). Incidence rates in non-Western countries are much lower (Kenya 0.08/100,000 and India 0.2/100,000 inhabitants per year, respectively) (Zimmerman and Kung’u, 1978; Raina et al., 1995). Generally, two types of testicular cancer are distinguished: 40% concerns seminoma tumors (ST) and 60% concerns nonseminoma testicular germ cell tumors (NSGCT). The highest incidence rate of seminomas is between 30 and 35 years of age, whereas nonseminomas are mainly diagnosed between 25 and 29 years of age (Hadu, 1979; Hoff Wanderas et al., 1995). Treatment is based on the histology of the primary tumor and on the results of staging investigations (Fung and Garnick, 1988; Peckham et al., 1983). In the case of a seminoma tumor, hemi-orchiectomy is followed by radiation therapy (RT). A stage I nonseminoma, currently, is successfully treated by hemi-orchidectomy alone, followed by frequent surveillance, the so-called “Wait & See” policy (W&S) (Gels et al., 1995). Disseminated disease is currently treated with cisplatin-based PCT and the resection of only residual retroperitoneal tumor mass (RRRTM). It should be noted that formerly a more invasive resection in the abdominal of retroperitoneal lymph nodes (RPLND) was performed (Gels et al., 1997). With the introduction of cisplatin-based polychemotherapy, followed by tumor-reductive surgery, the survival rate has increased to almost 90% (Gels et al., 1997). This increase in the survival rate has advanced the awareness of the long-term sequelae of diagnosis and treatment. As the disease affects relatively young men, in their genital area, it may not be surprising that in the past decades, several authors have investigated aspects of sexual functioning and fertility after treatment for testicular cancer. Unfortunately, however, it is very difficult to make a clear picture based on the outcomes of the existing studies. For instance, in the literature, loss of sexual desire varies between 2 and 69% and ejaculatory dysfunction between 17 and 100%. This wide range of reported dysfunctioning makes it difficult to use empirical results as a sound basis for patient education and counselling. To create a more

P1: FZN Archives of Sexual Behavior [asb]

PL186-228372

October 19, 2000

19:29

Sexuality and Testicular Cancer: Literature Review

Style file version Nov. 19th, 1999

57

comprehensive picture, it was decided to perform a meta-analysis on the existing data. METHODS Relevant literature was collected by performing a computer-based search (Medline, Psylit) of publications from 1975 to the end of 1999, with the following keywords: testicular cancer, cancer of the testicles, sexual functioning, sexuality, and sexual rehabilitation. As a second step all references given in the selected publications were examined. The following aspects of sexual functioning, based on Masters & Johnson (Masters and Johnson, 1966; Kaplan, 1979) and the DSM-IV (American Psychiatric Association [APA], 1994), were used to screen the literature: loss of sexual desire (libido); male erectile disorder; orgasmic dysfunction (in terms of delay or less intense orgasmic joy); and premature, delayed, or absent ejaculation. Some other relevant aspects of sexual functioning, frequently mentioned in the literature, were added: sexual activity and sexual satisfaction. A meta-analysis of the results was performed to obtain an insight into the ranges of outcomes of the reviewed field and into the eventual sources of heterogeneity. Means were weighted for number of patients included per study. To identify trends, outcomes of dependent (and co-) variables were simply interrelated across studies. RESULTS Entire Patient Population In the literature from 1975, a total of 29 retrospective and 7 prospective studies was found (see Appendix), which concerned, respectively, 2,437 and 338 cases that could be used for meta-analysis. Statistical analysis of the reported data, weighted by the number of patients per study, revealed the following mean scores (in rounded percents) on sexual dysfunctioning for the entire patient population (outcomes of prospective studies in parentheses): – loss of desire: 20% (11%) – male erectile disorder: 11.5% (14%) – orgasmic dysfunction: 20% (9%) – ejaculation disorder: 44% (51%) – decrease in sexual activity: 24% (13%) – sexual dissatisfaction: 19% (18%) The mean age was 35 (30.3) years and the mean follow-up period 6.9 (2.0) years. The mean response rate was 77% (88%). For more detailed results, that is, exact

P1: FZN Archives of Sexual Behavior [asb]

58

PL186-228372

October 19, 2000

19:29

Style file version Nov. 19th, 1999

Jonker-Pool, Van de Wiel, Hoekstra, Sleijfer, Van Driel, Van Basten, and Koops Table I. Overview of Meta-Analytic Outcomes of 36 Reviewed Studies Number of studies

Total patients n = 2,775 (100%)

Range

Weighted meana

SD

retrospect prospect

29 7

2,437 (88) 338 (12)

16–264 29–76

83.4b 53.5b

57.6 19.7

retrospect prospect Follow-up retrospect prospect Response retrospect prospect Desire retrospect prospect Erection retrospect prospect Orgasm retrospect prospect Ejaculation retrospect prospect Sexual Activity retrospect prospect Sexual retrospect Dissatisfaction prospect

24 5 24 7 25 5 15 4 20 6 13 2 19 5 10 1 16 2

2138 (88) 248 (73) 1988 (82) 338 (100) 2101 (86) 248 (73) 1521 (62) 147 (43) 1792 (74) 239 (77) 1251 (51) 71 (21) 1836 (75) 239 (71) 990 (41) 31 (9) 1400 (57) 93 (28)

20–77 years 17–51 years 0.25–25.5 years 0.25–3 years 43–100% 60–100% 2–69% 10–13% 0–35.2% 0–29% 0–53% 9%–? 17–100% 30–81% 13–40% 13%–? 4–38% 18%–?

34.7 years 30.3 years∗ 6.9 years 2.0 years∗ 76.6% 87.9%∗ 19.6% 11.2%∗ 11.5% 13.8% 19.8% 9.0%∗ 44.2% 51.3%∗ 24.2% 13%∗ 19.1% 18.0%∗

4.1 1.7 4.2 1.1 15.8 18.4 16.5 1.5 10.7 11.5 13 — 24.5 19.4 6.8 — 10.3 —

Age

Indicated are the number of studies in which the concerning variable was reported and the number of patients included in these studies. a Weighted mean: mean score of total of studies, weighted by sample-sizes of the studies involved. b Mean number of patients not weighted by number of patients. ∗ Using t-test to compare means of retro- and pro-spective studies; all means, except concerning erectile dysfunction, were significantly different, p < .001.

percentages, ranges, and standard deviations (SDs), see Table I; note that SD values of the reported sexual functions are broad, and that ranges vary widely between studies. Outcomes of Specific Treatment Modalities The study of outcomes of specific treatment modalities is important because it may highlight differences in sexual morbidity. In part of the studies, outcomes specified per treatment modality were reported. This concerned about half of the patients included in retrospective studies (n = 1,311 out of a total n = 2,437; Table II and Figs. 1–5). Only three out of seven prospective studies reported outcomes specified according to treatment modality; because of this low number the outcomes of these studies cannot be compared. Despite wide ranges and standard deviations (SD) in reported data concerning sexual dysfunction, meta-analysis revealed significant differences between treatments with respect to all sexual variables. The following mean scores (in rounded percentages) were found per specific

a RT:

radiotherapy, PCT: polychemotherapy, RPLND: retroperitoneal lymphnode dissection.

One-way Anova P: Scheff´e-test (group-differences with p < .05)

21.4 (13.8): 0–54.5 n = 846 0.0000 5—Rest of groups

50.9 (28.1): 0–100 n = 1108 0.0000 All groups different

15.9 (11.9): 7–44 n = 108 39.9 (14): 0–60 n = 273 28.4 (16.5): 0–64 n = 150 62.2 (21.5): 40.5–95 n = 358 80.5 (27.5): 38–100 n = 193 60 (0) n = 26

Ejaculatory Problems

27 (10.8): 0–40 n = 431 + ? 0.0000 All groups different, except 2 & 4

10.5 (0) n = 59 28.5 (11): 0–39 n = 85 34.3 (10.2): 8–40 n = 103 28.5 (3.9): 26–40 n = 178 0 (0) n=6 n=?

Reduced Activity

16.0 (9.2): 0–33 n = 583 + ? 0.0000 1—2 & 4

n=?

7.5 (1.4): 7–11 n = 75 15.8 (11.8): 7.5–33 n = 126 14.5 (6.2): 0–19 n = 137 19.9 (7.8): 15–33 n = 203 n=?

Sexual Dissatisfaction

19:29

13.52 (10.1): 0–48 n = 1013 0.0000 2—1,3,4,5

23.8 (14.3): 12–42.5 n = 101 22.5 (8.1): 14.5–38 n = 236 28.3 (6): 20–40 n = 67 21.5 (10): 9–39.5 n = 242 10.6 (16.3): 0–54 n = 174 54.5 (0) n = 26

Reduced Orgasm

October 19, 2000

22.85 (18.3): 2–69 n = 778 0.0000 All groups different, except 2 & 4

7.4 (10.1): 3.5–33 n = 75 24.8 (18.8): 0–48 n = 175 11.1 (5.8): 0–20 n = 150 11.3 (9.6): 0–31 n = 407 9.7 (8.7): 0–20 n = 180 23.5 (0) n = 26

Loss of Erection

PL186-228372

Total mean treatments number 1–6:

24.8 (16.8): 12–48.5 n = 108 14.1 (8.9): 2–22 n = 131 24.9 (3.6): 21–28.5 n = 137 13.4 (7.1): 2–21 n = 313 65.7 (1.6): 65–69 n = 89 n=?

Loss of Desire

Archives of Sexual Behavior [asb]

1. Surveillance (patients from 4 studies) n = 108 (8.2%) 2. RTa (patients from 9 studies) n = 417 (32%) 3. PCT (patients from 6 studies) n = 160 (12%) 4. PCT + surgery (patients from 9 studies) n = 404 (31%) 5. RPLND (patients from 6 studies) n = 193 (15%) 6. PCT + RPLND + RT (patients from 1 study) n = 26 (2%)

Total n = 1311 (100%)

Mean % (SD): range available n

Table II. Overview of Outcomes of Sexual Functioning Per Treatment-Modality (Means Weighted by Number of Patients Per Subgroup)

P1: FZN Style file version Nov. 19th, 1999

Sexuality and Testicular Cancer: Literature Review 59

P1: FZN Archives of Sexual Behavior [asb]

60

PL186-228372

October 19, 2000

19:29

Style file version Nov. 19th, 1999

Jonker-Pool, Van de Wiel, Hoekstra, Sleijfer, Van Driel, Van Basten, and Koops

Fig. 1. Surveillance (4 studies, n = 108).

Fig. 2. Irradiation (9 studies, n = 417).

treatment modality (for more detailed results, i.e., exact percentages, SDs, ranges, and number of observations, see Table II). Surveillance (Concerning Four Studies, n = 108) – loss of desire: 25% – male erectile disorder: 7% – orgasmic dysfunction: 24% – ejaculation disorder: 16% – decrease in sexual activity: 11% – sexual dissatisfaction: 8%

P1: FZN Archives of Sexual Behavior [asb]

PL186-228372

October 19, 2000

19:29

Style file version Nov. 19th, 1999

Sexuality and Testicular Cancer: Literature Review

61

Fig. 3. PCT (6 studies, n = 160).

Fig. 4. PCT + surgery (9 studies, n = 404).

Compared to the entire patient population, relatively high percentages of loss of sexual desire and orgasmic disorders were reported, especially if we take into account that, medically, orchidectomy followed by surveillance is the least invasive treatment modality. As expected, relatively little (but still some) ejaculatory dysfunction was reported. Radiotherapy (RT) (Concerning Nine Studies, n = 417) – loss of desire: 14% – male erectile disorder: 25% – orgasmic dysfunction: 23% – ejaculation disorder: 40%

P1: FZN Archives of Sexual Behavior [asb]

62

PL186-228372

October 19, 2000

19:29

Style file version Nov. 19th, 1999

Jonker-Pool, Van de Wiel, Hoekstra, Sleijfer, Van Driel, Van Basten, and Koops

Fig. 5. Surgery (6 studies, n = 193).

– decrease in sexual activity: 29% – sexual dissatisfaction: 16% Results revealed that irradiated patients reported the highest rate of erectile disorders and the lowest rate of loss of desire, compared to other treatment groups. Chemotherapy (PCT) (Concerning Six Studies, n = 160) – loss of desire: 25% – male erectile disorder: 11% – orgasmic dysfunction: 28% – ejaculation disorder: 28% – decrease in sexual activity: 34% – sexual dissatisfaction: 15% Treatment with PCT resulted in a rather “average” pattern of sexual disorders. The outcome of ejaculatory dysfunctioning was low compared to patients treated by RT, whereas the outcome of decrease in sexual activity was relatively high compared to the mean score for the entire population. PCT + Surgery (Retroperitoneal Lymph Nodes, RPLND, or Residual Tumor Mass, RRRTM) (Concerning Nine Studies, n = 404) – loss of desire: 13% – male erectile disorder: 11% – orgasmic dysfunction: 22%

P1: FZN Archives of Sexual Behavior [asb]

PL186-228372

October 19, 2000

19:29

Style file version Nov. 19th, 1999

Sexuality and Testicular Cancer: Literature Review

63

– ejaculation disorder: 62% – decrease in sexual activity: 29% – sexual dissatisfaction: 20% Like PCT alone, PCT + surgery also resulted in a rather average pattern of reported sexual dysfunctioning. As expected, patients treated by surgery in the abdominal area reported a much higher rate of ejaculatory dysfunctioning compared to patients treated by PCT alone. With respect to these studies it must be noted that surgical therapeutic principles have changed since the introduction of cisplatin in 1977, and that the extent of surgery has evolved from “very radical” to “more concise.” The effect is that in older studies concerning PCT + surgery, patients reported more ejaculatory dysfunction than in the more recent studies. However, the most striking result in this group of patients can be seen in the reported loss of sexual desire: the combination of two invasive treatment modalities lead to less reported loss of sexual desire (13%) than either PCT (25%) or abdominal surgery alone (66%). Retroperitoneal Lymph Node Dissection (RPLND) (Concerning Six Studies, n = 193) – loss of desire: 66% – male erectile disorder: 10% – orgasmic dysfunction: 11% – ejaculation disorder: 81% – decrease in sexual activity: 0% – sexual dissatisfaction: not reported As the table shows, studies about sexual outcome after RPLND-alone reported a very high percentage of ejaculation disorder, probably due to the very radical surgical procedures used in the period that RPLND-alone was the common intervention. Especially, bilateral RPLND implicated the risk for retrograde ejaculation (e.g, Bracken, 1976; Nijman et al., 1987). The high outcome of reported loss of sexual desire in this subgroup of RPLND-treated patients (6%) needs to be interpreted with caution, as this outcome could be based on only two studies (one of them used very small subgroups, Gritz et al., 1989, and the other used rather vague assessment methods, Nijman et al., 1987). PCT + RPLND + RT (Concerning One Study, n = 26) Fossa et al. (1988) offered specified data concerning this combined treatment. Results are depicted here only to provide a picture “as complete as possible,” but

P1: FZN Archives of Sexual Behavior [asb]

64

PL186-228372

October 19, 2000

19:29

Style file version Nov. 19th, 1999

Jonker-Pool, Van de Wiel, Hoekstra, Sleijfer, Van Driel, Van Basten, and Koops Table III. Correlations Between Outcomes of Variables of 36 Studiesa Desire

Desire Erection Orgasm Ejaculation Activity Dissatisfaction Age Follow-up a Spearson’s

Erection

Orgasm Ejaculate Activity Dissatisfaction Age Follow-up

1

.74∗ .18 .13 .67 .72∗∗ .19 −.06

1

.77∗∗ .03 .59∗ .60∗ −.020 .10

1 .07 .54 .24 .54∗ −.15

1 .07 −.12 .33 .16

1 .48 −.32 −.02

.02 −.18

1 .11

1

Rho Significant if ∗ p < .01, ∗∗ p < .05.

will not be taken into account in the discussion. – loss of desire: not reported – male erectile disorder: 24% – orgasmic dysfunction: 55% – ejaculation disorder: 60% – decrease in sexual activity: not reported – sexual dissatisfaction: not reported The total means across the specified treatment modalities in Table II were 2–7% higher compared to the general means presented in Table I (29 retrospective studies, irrespective of treatment modality). Between treatments, there appeared to be significant differences in reported sexual functioning (Table III, lowest row); it must be noted that the number of patients included per sexual variable is not fixed (due to the fact that the studies vary in their report of sexual variables). Taking into account this variability the following image emerges: – loss of desire appeared to be relatively high in patients treated by Surveillance and PCT-alone (both groups, 25%) – loss of erection was generally low, but highest in patients treated with RT (25%) – reduced orgasm was relatively low in patients treated with RPLND-alone (11%) – ejaculatory problems were high (62–81%) in patients treated with PCT ± surgery, but were also reported by 40% of patients from studies concerning RT – reduced sexual activity was relatively low, especially in patients treated with Surveillance (but still 10.5%) – sexual dissatisfaction generally was reported as being less frequent than was sexual dysfunctioning (<20%), and as the least by patients treated with Surveillance (7.5%)

P1: FZN Archives of Sexual Behavior [asb]

PL186-228372

October 19, 2000

Sexuality and Testicular Cancer: Literature Review

19:29

Style file version Nov. 19th, 1999

65

Interrelations Interrelations between outcomes of sexual functioning across studies were moderately high (Table III), especially between desire and erection (Rho = 0.74, p = .001), and between sexual satisfaction and desire as well as, to some lesser degree, erectile dysfunctioning (resp (response rate) Rho = 0.72, p = .009 and .60, p = .039). But there was also a moderately strong relationship between sexual activity and desire, erection (resp Rho = 0.67, p = ns, and Rho = 0.59, p = .097), and orgasm (Rho = 0.54, p = ns). However, in the presence of high interrelations, erectile dysfunction was reported much less frequently as compared to other variables. Ejaculatory functioning, although the most prevalent sexual dysfunction, was hardly at all related to subjective reports of other functions (Rho’s between 0.03 and 0.13, p = ns). Interrelations, with ejaculation excluded, revealed an a-value of 0.79 across the studies. In general, reported sexual desire and sexual activity contributed most to the a-value across studies, and reported orgasmic failure least. Age of the respondents was minimally related to reported sexual functioning, except for orgasm (Rho = 0.54, p = 0.08), and to a lesser degree to ejaculation (Rho = 0.33, p = ns). Follow-up period also seemed not related to reported sexual functioning. Interestingly, interrelations with publication year all except one were positive, which means that respondents tended to report more problems in more recent studies, except where it concerned ejaculatory problems: those were reported less in more recent studies (Rho − 0.43). DISCUSSION The aim of this study was to review literature concerning sexual functioning after treatment for testicular cancer. Thirty-six studies (29 retrospective and 7 prospective) could be found, and these were analysed. This concerned over 20 studies more than those quoted by Heidenreich (Heidenreich and Hofmann, 1999), who suggested that concerning testicular cancer “there is only a limited number of research data available with regard to sexual and marital satisfaction of the patients.” Besides aspects of the sexual response cycle in terms of disorders as described in DSM-IV (APA, 1994), absence of ejaculation, reduced sexual activity, and sexual dissatisfaction were taken into account in screening and describing outcomes of the reviewed studies. In general, the reviewed studies varied widely in variables and treatments included, and also in reported sexual morbidity. Furthermore, comparison of outcomes was hindered because different research methods were used, and only a limited number of studies used standardized questionnaires, or healthy individual/patient comparison groups (see Appendix). Therefore we performed a meta-analysis in which the number of patients per study, as well as the type(s) of treatment was taken into account.

P1: FZN Archives of Sexual Behavior [asb]

66

PL186-228372

October 19, 2000

19:29

Style file version Nov. 19th, 1999

Jonker-Pool, Van de Wiel, Hoekstra, Sleijfer, Van Driel, Van Basten, and Koops

Erectile dysfunction was reported least frequently (mean 11.5% in available retrospective studies and 14% in prospective studies), and ejaculatory dysfunction most frequently (mean 44% in retrospective and 51% in prospective studies). With respect to specific treatment modalities, it appeared that erectile dysfunction, although not frequently reported, was related to irradiation therapy. Ejaculatory dysfunction was straightforwardly connected with surgical techniques concerning retroperitoneal lymph node dissection, disrupting important nerve fibres (i.e., retrograde ejaculation). In view of this high mean rate it may be promising that in more recent studies, in which more focal surgical techniques are used with cisplatin-based chemotherapy (resection of residual retroperitoneal tumor mass, RRRTM + PCT), less ejaculatory dysfunction is reported (Gels et al., 1997; Tekgul et al., 1994; Wood et al., 1992). There was a relative independence of reported sexual functioning and age of respondents. Furthermore there is no clear relationship between follow-up periods and sexual function. Although prospective studies offered insufficient data for statistical comparison, retrospective studies generally showed higher rates of sexual morbidity than did prospective studies. This suggests the presence of a response bias: retrospective questioning triggers another type of response compared to prospective follow-up research (Fichten et al., 1991). At first glance, these contradictory findings, as well as the wide range of outcomes, seem to preclude definite conclusions concerning sexual morbidity after treatment for testicular cancer. However, if we make a distinction between a physiological and a psychological perspective, a comprehensible pattern can be detected. Looking at the sexual responses that are perceived as (at least partially) physiological phenomena (erection and ejaculation) we see, as could be expected, that these dysfunctions are related to treatment modalities that directly affect the nerve systems or innervation or both (RT and RPLND/RRRTM), whereby erectile dysfunction generally is the least-reported problem. On the other hand, the medically and physiologically least-invasive treatment (Surveillance) leads to the lowest rate of these two problems. This means that the physiological domain shows a high level of consistency. But if we look at the sexual responses that are perceived as predominantly psychological and behavioral phenomena (sexual desire, orgasmic joy, sexual activity, and satisfaction) another pattern can be detected. All treatment modalities reported a decrease of sexual desire, orgasmic intensity, sexual activity, and sexual satisfaction, whereby patients treated by the medical, relatively less-invasive treatment reported higher sexual morbidity compared to more invasive treatments, for example, data of patients treated by “Wait and See” (orchidectomy-alone), or PCT-alone compared to PCT + surgery. This indicates that changes in sexual functioning after testicular cancer treatment on a psychological level are relatively independent from physiological changes, caused by factors such as the stage of disease and influences of treatment-type. Interrelations also underscored the distinction between physical and psychological effects of cancer

P1: FZN Archives of Sexual Behavior [asb]

PL186-228372

October 19, 2000

Sexuality and Testicular Cancer: Literature Review

19:29

Style file version Nov. 19th, 1999

67

treatment: reported sexual functions were strongly interrelated, except for ejaculation. Also interesting is the recurrent result that reported sexual dissatisfaction is much lower than is reported sexual dysfunction. Possibly men change their sexual satisfaction-values after having experienced a life-threatening genital cancer (Jonker-Pool et al., 1997). Aspects of sexual functioning that are predominantly mediated by cognition, perception, and emotion (subjective sexual experiencing and sexual motivation) may be affected by the psychotraumatic experience of having cancer, and not by physiological factors per se (Jonker-Pool et al., 1997; Van Basten et al., 1997; Van Basten et al., 1999). This result replicates findings in a female population who were treated for gynecological cancer (Weijmar Schultz et al., 1993). Outcomes of this study reveal that future research should, on the one hand, take into account effects of specific treatments, to address the risk of physiological sexual morbidity. On the other hand future medical studies should not focus only on biological– medical causes in phenomena that are bio-psychosocially regulated by their very nature (Bancroft, 1989). Fortunately there is a trend to a more comprehensive research attitude. When we take into account the publication year of the studies it appeared that older studies tended to stress more biological-organic sequelae of testicular cancer, whereas more recent studies also paid attention to psychosocial issues related to sexuality, such as marital relationship and subjective sexual distress (Fossa et al., 1988; Hannah et al., 1992; Schover and von Eschenbach, 1985). When the trends found in this meta-analytic study appear to be correct, this includes important suggestions for the coaching of patients. Patients can be accurately informed about the physiological-sexual risks to be expected after specific treatment, as well as about the psychological-sexual vulnerability that may evolve due to the intrinsic actual and symbolic danger where it concerns a genital tumor at a relatively young age. Adequate information and support may prevent, or reduce, unnecessary sexual and relational anxiety and suffering.

Bisset (1990) Blackmore (1988) Bloom (1993) Bracken (1976) Brenner (1985) Caffo (1999) Cassileth (1987) Couzijn (1982)

2

3b

5

6

9c

39

89 s

29 s

37 s

42 s

20

3

44

15

PCT

30 bil 6 uni ns

1

RPLND

14 bil ns

RPLND + RT

41 (?)

15

RT + PCT

61 bil

44 ns

14

19

PCT + RPLND

3 bil orch

Other/Unclear

19:29

8a

16

9

RT

October 19, 2000

7

48 s 37 ns (79%) 8s 66 ns (44%) 16 (80%) 39 s 49 ns (77%) 29 s 50 ns (51%) 67 ns (95%) 89 s (69%) 39 s (58%) 20 cpl

Surv

PL186-228372

4a

Arai (1997)

1b

n (resp %)

68

Author (Publication year)

Archives of Sexual Behavior [asb]

Sample Sizes and (Combinations of ) Treatments Involved in 29 Retrospective and 7 Prospective Studies Describing Sexual Functioning After Testicular Cancer

APPENDIX

P1: FZN Style file version Nov. 19th, 1999

Jonker-Pool, Van de Wiel, Hoekstra, Sleijfer, Van Driel, Van Basten, and Koops

Hannah (1992) Hartman (1999) Johansson (1992) Jonker-Pool (1997)

13a

14b

15

Kaasa (1991) Levison (1986) Moynihan (1987)

18

19

4s

15 s

15 s

32

58

2

42 ns

39

32 ns

3 ns

3 ns

2

32 uni

13 ns

8 ns

8 ns

4

4

1s

1s

3

2

2

46

122 ns

42 ns

6 ns

6 ns

69

(Continued )

32 RT + PCT ± RPLND

19 all treatment 49 PCT or RPLND or both

19:29

12

44

39

59 ns 41 s

7s

1s

1s

37 ns

11

October 19, 2000

17

34 cpl 17 s 17 ns (83%) 34 cpl 17 s 17 ns (?) 20 s 78 ns (78%) 39 (87%) 41 s 223 ns (85%) 46 s 103 ns (93%) 53 34 s (?) ns (88%) 102 (84%)

39 s 49 ns (?)

109 ns

PL186-228372

16a

Gritz (1989)

Douchez (1993) Gritz (1988)

Archives of Sexual Behavior [asb]

12b

11b

10a

P1: FZN Style file version Nov. 19th, 1999

Sexuality and Testicular Cancer: Literature Review 69

Rieker (1989)

Schover (1985)

Schover (1986) Stoter (1989) Stuart (1990)

Tinkler (1992) Tross (1984) abstract Aass (1993) prosp

23a,b

24a,b

25

28a,b

30

41 s 35 ns (100%)

36

137

28

48

13

47/121

38

63 bil

51

RPLND

2

8/121

11

4

RPLND + RT

4

3

RT + PCT

23

30

38/121

74

37

27

PCT + RPLND

21 PCT/ RPLND + 10 RT 20 RT + PCT/ RPLND 26/121 RPLND + RT + PCT 2/121: (?)

Other/Unclear

19:29

29a

18

34

79

20

PCT

October 19, 2000

27

84 s (51%) 48 ns (84%) 35 s 27 ns (63–78%) 155 (56–62%) 30 ns (?)

71

6

RT

PL186-228372

26

22

55 ns (?) 101 ns (100%) 9s 65 ns (80%) 40 s 61 ns (81%) 52 ns (43%)

Surv

Archives of Sexual Behavior [asb]

21

Narayan (1982) Nijman (1987) Rieker (1985)

n (resp %)

70

20

Author (Publication year)

APPENDIX (Continued )

P1: FZN Style file version Nov. 19th, 1999

Jonker-Pool, Van de Wiel, Hoekstra, Sleijfer, Van Driel, Van Basten, and Koops

Johnstone (1991) prosp Nijman (1988) prosp Tamburini (1989) prosp Trump (1985) prosp/abstract Van Basten (1999) prosp

32c

17 ns

9

(?)

(?)

5

39

8

(?)

13

4

10

(?)

6

5

15 prosp 29 ret

(?)

(?)

(?)

7

42 part 8 compl

17 prosp 43 ret ± RT

(?)

26 ret PCT + RT ± RPLND

October 19, 2000

Note. resp = response rate; Surv = surveillance; RT = radiotherapy; PCT = poly chemotherapy; RPLND = retroperitoneal lymph node dissection; n = responding number of patients; s = seminoma, ns = non-seminoma; orch = orchiectomy; bil = bilateral RPLND; uni = unilateral RPLND; part = partial RPLND; compl = complete RPLND; cpl = couples; prosp = prospective study; ret = retrospective. a Comparing to controls or other patient comparison-groups. b Comparing different treatment modalities within the study . c Using validated questionnaires.

36b,c

35c

5s 26 ns (100%) 29 (?)

16

32 prosp 24 ret

PL186-228372

34b

40 s 30 ns (60%) 54 ns (100%)

64 prosp 122 ret (?)

Archives of Sexual Behavior [asb]

33b

Fossa (1988) prosp?

31

P1: FZN 19:29 Style file version Nov. 19th, 1999

Sexuality and Testicular Cancer: Literature Review 71

P1: FZN Archives of Sexual Behavior [asb]

72

PL186-228372

October 19, 2000

19:29

Style file version Nov. 19th, 1999

Jonker-Pool, Van de Wiel, Hoekstra, Sleijfer, Van Driel, Van Basten, and Koops

ACKNOWLEDGMENT This study was supported by a grant from the Dutch Cancer Society, RUG 94-873. REFERENCES Aass, N., Grunfeld, B., Kaalhus, O., and Fossa, S. D. (1993). Pre- and post-treatment sexual life in testicular cancer patients: A descriptive investigation. Br. J. Cancer 67: 1113–1117. Adami, H. O., Bergstrom, R., Mohner, M., Zatonski, W., Storm, H., Ekbom, A., Tretli, S., Teppo, L., Ziegler, H., and Rahu, M. (1994). Testicular cancer in nine northern European countries. Int. J. Cancer 59: 33–38. American Psychiatric Association (1994). DSM-IV. Diagnostic and Statistic Manual of Mental Disorders, American Psychiatric Association, Washington, DC. Arai, Y., Kawakita, M., Okada, Y., and Yoshida, O. (1997). Sexuality and fertility in long-term survivors of testicular cancer. J. Clin. Oncol. 15: 1444–1448. Bancroft, J. (1989). The biological basis of human sexuality. In: Bancroft, J., Human sexuality and its problems, Churcill Livingstone, Edinburg, London, Melbourne, and New York, pp. 12–145. Bisset, D., Kunkeler, L., Zwanenburg, L., Paul, J., Gray, C., Swan, I. R. C., Kerr, D. J., and Kaye, S. B. (1990). Long-term sequelae of treatment for testicular germ cell tumors. Br. J. Cancer 62: 655–659. Blackmore, C. (1988). The impact of orchidectomy upon the sexuality of the man with testicular cancer. Cancer Nursing 11: 33–40. Bloom, J. R., Fobair, P., Gritz, E., Wellisch, D., Spiegel, D., Varghese, A., and Hoppe, R. (1993). Psychosocial outcomes of cancer: A comparative analysis of Hodgkin’s disease and testicular cancer. J. Clin. Oncol. 11: 979–988. Bracken, R. B. (1976). Sexual function and fecundity after treatment for testicular tumors. Urology 1: 35–38. Brenner, J., Vugrin, D., and Whitmore, W. F. Jr. (1985). Effect of treatment on fertility and sexual function in males with metastatic nonseminomatous germ cell tumors of testis. Am. J. Clin. Oncol. 8: 178–182. Caffo, O., and Amichetti, M. (1999). Evaluation of sexual life after orchidectomy followed by radiotherapy for early-stage seminoma of the testis. BJU International 83: 462–468. Cassileth, B. R., and Steinfeld, A. D. (1987). Psychological preparation of the patient and family. Cancer 60: 547–552. Couzijn, A. L., van Dam, F. S., and Hanewald, G. J. (1982). Chemotherapie van het gemetastaseerde testicarcinoom; ervaringen van patienten en hun partners. [Chemotherapy of metastasized carcinoma of the testis; experiences of patients and their partners] Ned. T. Geneesk. 126: 1854–1857. Douchez, J., Droz, J. P., Desclaux, B., Allain, Y., Fargeot, P., Caty, A., and Charrot, P. (1993). Quality of life in long-term survivors of nonseminomatous germ cell testicular tumors. J. Urol. 149: 498–501. Fichten, S., Libman, E., Amsel, R., Creti, L., Weinstein, N., Rothenberg, P., Liederman, G., and Brender, W. (1991). Evaluation of the sexual consequences of surgery: Retrospective and prospective strategies. J. Behav. Med. 14: 267–285. Fossa, S. D., Aass, N., and Kaalhus, O. (1988). Testicular cancer in young Norwegians. J. Surg. Oncol. 39: 43–63. Fung, C. Y., and Garnick, M. B. (1988). Clinical stage I carcinoma of the testis: A review. J. Clin. Oncol. 6: 734–750. Gels, M. E., Hoekstra, H. J., Sleijfer, D. Th., de Bruijn, H. W., Molenaar, W. M., Freling, N. J., Droste, J. H., and Schraffordt Koops, H. (1995). Detection of recurrence in patients with clinical stage I nonseminomatous testicual germ cell tumors and consequences for further follow-up: A single-centre 10 year experience. J. Clin. Oncol. 13: 1188–1194. Gels, M. E., Nijboer, A. P., Hoekstra, H. J., Heuvel, F., Molenaar, W. M., Plukker, J. T., Droste, J. H., and Schraffordt Koops, H. (1997). Complications of the post-chemotherapy resection of

P1: FZN Archives of Sexual Behavior [asb]

PL186-228372

October 19, 2000

Sexuality and Testicular Cancer: Literature Review

19:29

Style file version Nov. 19th, 1999

73

retroperitoneal residual tumor mass in patients with nonseminomatous testicular germ cell tumors. Br. J. Urol. 79: 263–268. Gritz, E., Wellisch, D., and Landsverk, J. A. (1988). Psychosocial sequelae in long-term survivors of testicular cancer. J. Psychosoc. Oncol. 6: 41–63. Gritz, E. R., Wellisch, D. K., Wang, H. J., Siau, J., Landsverk, J. A., and Cosgrove, M. D. (1989). Longterm effect of testicular cancer on sexual functioning in married couples. Cancer 64: 1560–1567. Hadu, I. (1979). Pathology of germ cell tumors of the testis. Sem. Oncol. 6: 14–25. Hannah, M. T., Gritz, E., Wellisch, D., Fobair, P., Hoppe, R. T., Bloom, J. R., Sun, G.-W., Varghese, A., Cosgrove, M. D., and Spiegel, D. (1992). Changes in marital and sexual functioning in longterm survivors and their spouses: Testicular cancer versus Hodgkin’s disease. Psycho-Oncology 1: 89–103. Harding, M., Hole, D., and Gillis, C. (1995). The epidemiology of nonseminomatous germ-cell tumors in the west of Scotland 1975–1989. Br. J. Cancer 72: 1559–1562. Hartmann, J. T., Albrecht, C., Schmoll, H.-J., Kuczyk, M. A., Kollmannsberger, C., and Bokemeyer, C. (1999). Long-term effects on sexual function and fertility after treatment for testicular cancer. Br. J. Cancer 80(5/6): 801–807. Heidenreich, A., and Hofmann, R. (1999). Quality-of-life issues in the treatment of testicular cancer. World J. Urol. 17: 230–238. Hoff Wanderas, E., Tretli, S., and Fossa, S. D. (1995). Trends in the incidence of testicular cancer in Norway 1955–1992. Eur. J. Cancer 31A: 2044–2048. Johansson, S., Steineck, G., Hursti, T., Fredrikson, M., Furst, C. J., and Peterson, C. (1992). Aspects of patient care. Interviews with relapse-free testicular cancer patients in Stockholm. Cancer Nursing 15: 54–60. Johnstone, B. G. M., Silberfeld, M., Chapman, J. A., Phoenix, C., Sturgeon, J. F. G., Till, J. E., and Sutcliffe, S. B. (1991). Heterogeneity in responses to cancer. Part II: Sexual responses. Can. J. Psychiat. 36: 182–185. Jonker-Pool, G., van Basten, J. P., Hoekstra, H. J., van Driel, M. F., Heuvel, F., Schraffordt Koops, H., and van de Wiel, H. B. M. (1997). Sexuality after testicular cancer treatment—Comparison of treatment modalities. Cancer 80: 454–464. Kaasa, S., Aass, N., Mastekaasa, A., Lund, E., and Fossa, S. D. (1991). Psychosocial well-being in testicular cancer patients. Eur. J. Cancer 27: 1091–1095. Kaplan, H. S. (1979). Disorders of Sexual Desire, Brunner/Mazel, New York. Levison, V. (1986). The effect on fertility, libido and sexual function of post-operative radiotherapy and chemotherapy for cancer of the testicle. Clin. Radiol. 37: 161–164. Masters, W. H., and Johnson, V. E. (1966). Human Sexual Response. Little, Brown, Boston. Moynihan, C. (1987). Testicular cancer: The psychosocial problems of patients and their relatives. Cancer Surv. 6: 477–510. Narayan, P., Lange, P. H., and Fraley, E. E. (1982). Ejaculation and fertility after extended retroperitoneal lymph node dissection for testicular cancer. J. Urol. 127: 685–688. Nijman, J. M., Schraffordt Koops, H., Kremer, J., and Heuvel, F. (1987). Gonadal function after surgery and chemotherapy in men with stage II and III nonseminomatous testicular tumors. J. Clin. Oncol. 5: 651–656. Nijman, J. M., Schraffordt Koops, H., Oldhoff, J., Kremer, J., and Heuvel, F. (1988). Sexual function after surgery and combination chemotherapy in men with disseminated nonseminomatous testicular cancer. J. Surg. Oncol. 38: 182–186. Peckham, M. J., Barret, A., Horwich, A., and Hendry, W. F. (1983). Orchiectomy alone for stage I testicular nonseminoma. A progress report on the Royal Marsden Hospital Study. Br. J. Urol. 55: 754–759. Raina, V., Shukla, N. K., Gupta, N. P., Deo, S., and Rath, G. K. (1995). Germ cell tumors in uncorrected cryptorchid testis at Institute Rotary Cancer Hospital, New Delhi. Br. J. Cancer 71: 380–382. Rieker, P. P., Edbril, S. D., and Garnick, M. B. (1985). Curative testis cancer therapy: Psychosocial sequelae. J. Clin. Oncol. 3: 1117–1126. Schover, L. R., Gonzalez-Mariscal, G., and von Eschenbach, A. C. (1986). Sexual and marital relationships after radiotherapy for seminoma. Urology 27: 117–123. Schover, L. R. and von Eschenbach, A. C. (1985). Sexual and marital relationships after treatment for nonseminomatous testicular cancer. Urology 25: 251–255.

P1: FZN Archives of Sexual Behavior [asb]

74

PL186-228372

October 19, 2000

19:29

Style file version Nov. 19th, 1999

Jonker-Pool, Van de Wiel, Hoekstra, Sleijfer, Van Driel, Van Basten, and Koops

Stoter, G., Koopman, A., Vendrik, C. P., Struyvenberg, A., Heuvel, F., Willemse, P. H., Schraffordt Koops, H., van Oosterm, A. T., and Pinedo, H. M. (1989). Ten-year survival and late sequelae in testicular cancer patients treated with cisplatin, vinblastine, and bleomycin. J. Clin. Oncol. 7: 1099–1104. Stuart, N. S., Grundy, R., Woodroffe, C. M., and Cullen, M. H. (1990). Quality of life after treatment for testicular cancer—The patient’s view. Eur. J. Cancer 26: 291–294. Tamburini, M., Filiberti, A., Barbieri, A., Zanoni, F., Pizzocaro, G., Barletta, L., and Ventafridda, V. (1989). Psychological aspects of testis cancer therapy: A prospective study. J. Urol. 142: 1487– 1490. Tekgul, S., Ozem, H. A., Celebi, I., Ergen, A., and Demircin, B. (1994). Postchemotherapeutic surgery for metastatic germ cell tumors: Results of extended primary chemotherapy and limited surgery. Urology 43: 349–354. Tinkler, S. D., Howard, G. C., and Kerr, G. R. (1992). Sexual morbidity following radiotherapy for germ cell tumors of the testis. Radiother. Oncol. 25: 207–212. Tross, S. and J. C. Holland (1984). A controlled study of psychosocial sequelae in cured survivors of testicular neoplasms (Abstract). Proc. Am. Soc. Clin. Oncol. 3: 74. Trump, D. L., Romsaas, E. P., Cummings, K. C., and Malec, J. F. (1985). Assessment of psychologic and sexual dysfunction in patients following treatment of testis cancer: A prospective study (Abstract). Proc. Am. Soc. Clin. Oncol. 4: 250. Van Basten, J. P., Jonker-Pool, G., van Driel, M. F., Droste, J. H. J., van de Wiel, H. B. M., Schraffordt Koops, H., Molenaar, W. M., and Hoekstra, H. J. (1997). Sexual functioning after multimodality treatment for disseminated nonseminomatous testicular germ cell tumor. J. Urol. 158: 1411–1416. Van Basten, J. P., van Driel, M. F., Hoekstra, H. J., Sleijfer, D. T., van de Wiel, H. B. M., Droste, J. H. J., Schraffordt Koops, H., and Mensink, H. J. A. (1999). Objective and subjective effects of testicular cancer treatment on sexual function. Br. J. Urol. (Abstract) Visser, O., Coebergh, J. W. W., and Schouten, L. J. (1996). Incidence of Cancer in The Netherlands 1993, The Netherlands Cancer Registry, Utrecht. Weijmar Schultz, W. C. M., van de Wiel, H. B. M., Van de Haan, D. E. E., and van Driel, M. F. (1993). Sexuality and cancer in women. Ann. Rev. Sex Res. 3: 151–200. Wingo, P. A., Tong, T., and Bolder, S. (1995). Cancer Statistics. CA Cancer J. Clin. 45: 8–30. Wood, D. P., Herr, H. W., Heller, G., Vlamis, V., Sogani, P. C., and Motzer, R. J. (1992). Distribution of retroperitoneal metastases after chemotherapy in patients with nonseminomatous germ cell tumors. J. Urol. 148: 1812–1815. Zimmerman, R. R., and Kung’u, A. (1978). Testicular neoplasms in Kenyan Africans. Cancer 41: 2452–2455.

P1: FDO Archives of Sexual Behavior [asb]

PL186-228373

October 19, 2000

19:34

Style file version Nov. 19th, 1999

Archives of Sexual Behavior, Vol. 30, No. 1, 2001

Kallmann’s Syndrome and Transsexualism Bernd Meyenburg, M.D.,1,3 and Volkmar Sigusch, M.D., D.Sci.2

Until the present, in the world literature only one patient with Kallmann’s syndrome has been reported who became transsexual. This patient was seen almost 50 years ago. In this report, a second case is presented to encourage studies of the sexual and gender identity development in these patients. This patient’s rare endocrine disorder and secondary emotional problems have led to negative consequences because appropriate treatment of her transsexualism became impossible. KEY WORDS: Kallmann’s syndrome; hypogonadism; gender identity disorder; transsexualism.

INTRODUCTION Kallmann’s syndrome is an inherited disorder characterized by hypogonadotropic hypogonadism and anosmia (Kallmann et al., 1944). Hypogonadism is due to deficiency of gonadotropin-releasing hormone and anosmia is due to hypoplasia or aplasia of the olfactory bulbs and tracts. In Western countries, a prevalence of 1 in 10,000 to 1 in 60,000 persons is reported with an excess of male over female patients suggesting X-linked inheritance. Genetic studies have localized X-linked Kallmann’s syndrome at Xp 22.3 (Duke et al., 1995; Hardelin and Petit, 1995; Meitinger et al., 1990). Typically, male patients show eunuchoid growth after delayed or absent puberty, infantile genitals, sparse body hair growth, high-pitched voice, and hyposmia or anosmia associated with nasal saddle deformities. A large number of neurologic deficits and other somatic defects have been described (e.g., Bick et al., 1992), but

1 Department

of Child and Adolescent Psychiatry, University of Frankfurt Medical School, D-60590 Frankfurt am Main, Germany. for Sexual Science, University of Frankfurt Medical School, D-60590 Frankfurt am Main, Germany. 3 To whom correspondence should be addressed at Department of Child and Adolescent Psychiatry, University of Frankfurt Medical School, D-60590 Frankfurt am Main, Germany. 2 Institute

75 C 2001 Plenum Publishing Corporation 0004-0002/01/0200-0075$19.50/0 °

P1: FDO Archives of Sexual Behavior [asb]

PL186-228373

October 19, 2000

76

19:34

Style file version Nov. 19th, 1999

Meyenburg and Sigusch

the literature is void of describing psychosexual and gender identity development in these patients although problems in this area are likely to be present. Since the publication of Kallmann et al. in 1944 only one case of Kallmann’s syndrome with atypical gender identity development has been described in the literature. B¨urger-Prinz et al. (1966) reported on a 23-year-old male patient who presented in the early 1950s with a wish for sex reassignment surgery. The patient reported that he had felt like a girl since early childhood and preferred doll play and female activities. At the age of 15, he engaged sexually with older boys. Later on, when puberty did not occur, he began cross-dressing. Sexual activities were reduced to masturbating once or twice a month. Heterosexual interests were absent. Laboratory tests showed markedly decreased gonadotropic hormones. Testicular biopsy showed atrophic tubuli. Leydig cells were absent; bilateral anosmia was present. The patient was treated with gonadotropic hormones. His physical appearence rapidly virilized. It is not explained how administration of gonadotropic hormones could lead to virilization, given the biopsy results. It appears likely that Leydig cells were only diminished in number, but were not absent. After 2 months, the patient’s transsexual wishes subsided; heterosexual interests appeared; and masturbatory activity increased. He then engaged in a sexual relationship with an older woman. A 17-year follow-up (Fischer, 1971) showed a persistent absence of transsexual symptoms; the patient lived for 10 years in a stable sexual relationship with a woman. CASE REPORT We report a 58-year-old genetically male patient, L. B., who presented 20 years ago with a wish for sex reassignment surgery. Since the age of 39, L. B. has lived continuously as a female. (For reasons of clarity we refer to the patient throughout her whole course of life as “she.”) Medical history and examination showed Kallmann’s syndrome with some atypical features. Since the age of 38, the patient was frequently misdiagnosed (e.g., paranoid schizophrenia) and, in consequence, had extreme hardships like involuntary psychiatric hospitalization. Sex reassignment surgery has not been possible so the patient continues to live “in limbo” (L. B.). She was seen first by one of the authors (B. M.) when her 7-year-old daughter T. was referred for psychotherapy to help her deal with her father’s sex change. T. was dealing well with this problem, but severe psychiatric problems became soon apparent in L. B. (e.g., depression and personality disorder). Psychotherapy was started for 1 year until her therapist moved abroad. Concomitantly, the patient’s wife and daughter were seen on a regular basis. Until now, yearly follow-up studies were done with L. B. for 16 years. The patient is a 168-cm tall white genetic male, weighing 142 lbs, with feminine habitus. She has a high-pitched voice, which sounds hoarse due to heavy

P1: FDO Archives of Sexual Behavior [asb]

PL186-228373

October 19, 2000

Kallmann’s Syndrome and Transsexualism

19:34

Style file version Nov. 19th, 1999

77

smoking. Her skin is fine, there are fine wrinkles around eyes and mouth, a normal female hair-escutcheon. Axillary hair was shaved, pubic hair is female in pattern. There is no male muscle development. Her chest shows scars following four gynecomastia operations at the age of 18–30. There is a scar over the right hip; bone chips were removed to reconstruct nasal saddle deformity. Bilateral anosmia is present. Aureola and nipples are female in structure and pigmentation. The penis measures about 5 cm and has a normal urethral opening. A perineal aperture was not found. Testicles were tender and firm, palpable bilaterally in the inguinal canals measuring 1–2 cm in diameter and 4 ml in volume. The scrotum appeared rugated and is not bifid. The patient reported recurrent painless bloody urethral discharge occuring since the age of 27 every 4–5 weeks, lasting 24 hours, an unexplained phenomenon reported rarely also by other transsexual patients. Genetic studies showed a male karyotype (46, XY). Endocrine studies at the age of 38 showed low, castratelevels of testosterone without elevated gonadotropins and a rise of testosterone from 0.07 to 0.96 ng/ml after HCG-stimulation. There are reports of very high estradiol levels (320, 1920, 1480, and 3570 pg/ml), the source of which remains unknown. L. B. has persistantly denied estrogen intake. Endocrine studies were repeated at the age of 42. Estradiol levels were low, below 0.1 pg/ml, rising to 18.0 pg/ml after HCG-stimulation (5000 U HCG qd × 5 d). Testosterone levels remained low (<0.1 ng/ml) even after HCG-stimulation. LH-FSH-levels were low (LH 5.8 mU/ml, FSH 13.4 mU/ml), rising to normal levels after HCG-stimulation (LH 88.1 mU/ml, FSH 58.9 mU/ml). Prolactin levels were high (16.8 mg/ml). CT scan of abdomen and pelvis showed no abnormal masses in the region of the adrenal gland; no uterine tissue was identified in the pelvis. A prostate gland of normal size was seen in normal position with focal areas of low attenuation. In the region of the inguinal canals, two oval soft tissue masses were seen consistent with bilaterally undescended testicles. CT scan of sella showed the sella turcica of normal size; no mass lesions were identified within the sella or in the suprasellar cistern. Past History Past history data could only be obtained from the patient herself and her wife because relatives could not be reached. However, past history data given during prior hospital admissions and to prior therapists showed high coincidence in details. L. B. is the third child of a white Caucasian mother and an American Indian father. At the age of 1, she was given to foster parents for 1 year. Then she lived until the age of 131/2 years in an orphanage. Her two older sisters stayed with her mother after the parents divorced. Her father abused alcohol. Major diseases in the family are unknown to the patient. At the age of 14–17, the patient was sent

P1: FDO Archives of Sexual Behavior [asb]

PL186-228373

October 19, 2000

19:34

78

Style file version Nov. 19th, 1999

Meyenburg and Sigusch

to a reform school. After school, the patient served in the Air Force for 2 years and was discharged because of concealing female clothes in a foot locker. L. B. explained to the authorities that she kept these clothes for her girlfriend. Later the patient worked as a musician and ran a radio and TV repair shop. During the year prior to starting psychotherapy she worked as a doorman. L. B. reports that she has always felt confused about her gender identity. She dates wishes to be female to the age of 4 or 5, when she was caught observing girls in the orphanage where boys and girls lived separately. For punishment she was forced to live dressed as a girl with the other girls for some days. L. B. has always held this as the reason for her gender identity confusion. Later on her physical development began to contribute to the problem. Puberty was delayed. She never developed a male voice. Female breast tissue grew at the age of 17 and was surgically removed four times because L. B. tried desperately to live and pass as male. She started relationships with two women who became pregnant. The oldest child is a boy, presumedly fathered by L. B. at the age of 30; the boy lives with his mother. L. B. states she is not sure about her being the father. Her medical condition (undescended testicles) appears also to be incompatible with fertility. The second child, a girl, was born when the patient was 34. L. B. continues to live with the daughter’s mother whom she married while still living in the male role. There are doubts also about L. B. being the father of this child. The patient reports that she has had difficulties in performing sexual intercourse. During intercourse she experienced painful feelings in her penis; only a few drops of watery fluid were emitted. Sperm counts could not be made. Further examination of her undescended testicles were refused. After trying desparately to live and function as a male until her late thirties, L. B. felt increasingly insecure about her gender identity. At the age of 39, the patient was referred to a major medical center for psychological and endocrine work-up. Repeatedly, L. B. became emotionally upset and suicidal, once setting fire to herself, because she felt treated like a “freak” or “pervert.” This led to three subsequent psychiatric admissions where diagnoses of chronic paranoid schizophrenia, adjustment disorder with depression, and borderline personality disorder were made. L. B.’s endocrine problems were attributed to exogenous intake of estrogens. Sex reassignment surgery was never considered feasable by her prior therapists. Meanwhile, the patient’s wife had sought counselling by a child psychiatrist because she was then worried about her 7-year-old daughter’s development. Although the daughter was found dealing exceptionally well with this situation, it soon became apparent that her father, L. B., was in urgent need of further psychotherapy. Course of Therapy Supportive psychotherapy was started with sessions twice weekly. The review of her past history and past examinations was painful for L. B. Her most urgent

P1: FDO Archives of Sexual Behavior [asb]

PL186-228373

October 19, 2000

Kallmann’s Syndrome and Transsexualism

19:34

Style file version Nov. 19th, 1999

79

wish was now to become completely female, to have sex reassignment surgery. Because her case was highly unusual, great caution was taken not to make premature decisions. Administration of estrogens was refused by the patient. L. B. became increasingly depressed and suicidal. She spent whole sessions crying and screaming in pain, complaining that nobody was willing to help her and everybody treated her as a freak. Tricyclic antidepressant medication gave her only minor relief. Three months after starting therapy, L. B. again became seriously suicidal so that hospitalization was necessary. During this time, thorough medical studies were done with the results reported here. When confronted with the results of medical and psychiatric evaluations, that is, Kallmann’s syndrome and transsexualism (according to DSM-III classification), L. B. very angrily refused the latter diagnosis. She remains convinced until now that her endocrine problems are the reason for her gender identity disorder. Subsequently, the patient was referred to a urologist experienced in sex reassignment surgery. Here L. B. felt being treated like a “common transsexual” and became so upset that she left the office, screaming and sobbing loudly. The urologist refused to see this emotionally unstable patient again; an experience L. B. has repeated with other psychiatrists also. In her 1-year long therapy, it became evident that she had felt her gender identity disorder to be present since early childhood. Surprisingly, she reported in one more quiet session that she was unaware of the anatomical sex difference until the age of 15 when she asked a girl to show her genitals to her. When the interpretation was given that it was too painful for her to acknowledge the sex difference openly because she always felt like a female but knew very well about her genitals being male, she responded with shock, anger, and irritation. Her memories of the orphanage, where she tried to observe girls secretly, demonstrate that L. B. was driven early to find a solution for her gender identity problem. During the past 14 years, L. B.’s situation has remained essentially unchanged. Attempts to continue the therapy by a psychiatrist experienced in treating patients with gender identity disorders failed because L. B. continued to demand immediate sex reassignment surgery. With increasing age L. B. has resigned to living “inbetween,” in her view as a full-time woman with male genitals. She has never left her residential neighbourhood where she is well known. Painful teasing, frequent in the past, has stopped. At the age of 52, 9 years after terminating psychotherapy, bilateral inguinal orchidectomy was performed. Only fibroid tissue was found, and no further histologic studies were done. At this time, she was put on estrogens by her family doctor. Testosterone and estradiol levels continued to be low before surgery (estradiol <10 ng/l, testosterone <20 ng/ml). Gonadotropin levels were found to be low as well: LH <0.3 U/l, FSH <0.3 U/l, TSH 0.18 MU/l. Meanwhile (2000.2) L. B. has given up the wish for sex reassignment surgery stating it would no longer be of any use. She continues to live with her wife and daughter whose successful career is a source of major pride for her. Her mental state has improved greatly, she is no longer prone to depression and negative emotional outbursts.

P1: FDO Archives of Sexual Behavior [asb]

PL186-228373

October 19, 2000

80

19:34

Style file version Nov. 19th, 1999

Meyenburg and Sigusch

DISCUSSION Cases like this raise the question of the significance of biologic forces in the development of gender identity disorders. Sociogenetic studies emphasize today how much transsexuals have been shaped as a certain species throughout the centuries by medical science, especially by psychiatry and surgery, and how much social recognition they have gained in Western countries in the unfolding of a “neosexual revolution” (Sigusch, 1998), even including special laws in several European countries. In the case of L. B., the specific etiology of her transsexualism remains open. It is likely that co-occurrence of traumatic childhood events, for example, being separated from parents and being given away into an orphanage, and biologic factors due to her Kallmann’s syndrome led to gender identity disorder. DSM-IV diagnostic criteria for gender identity disorder demand absence of a physical intersex condition. The appropriate diagnosis would now be gender identity disorder not otherwise specified, examples include intersex conditions with accompanying gender dysphoria. ICD-10 diagnostic criteria, however, allow diagnosis of transsexualism (F 64.0) because accompanying Kallmann’s syndrome does not constitute an exclusion criterion. The presence of Kallmann’s syndrome and also a very unstable mental condition in the case of L. B. led to the result that sex reassignment measures were never seriously considered and later on even refused forcing her to continue to live “in-between,” Furthermore, L. B. considered her endocrine disorder as the only reason for her gender identity disorder. She always refused to consider herself as transsexual. As a result she found herself “stuck in the middle.” Because of this stressful situation, she became emotionally even more disturbed. It might be hypothesized that living “in-between” without sex reassignment has enabled L. B. to continue to live with her family; living as a “real” woman might have brought on even more problems for her and her family. Her quiet acceptance of her fate now at the age of 59 appears to support this different point of view that at least L. B. was highly ambivalent about going all the way to become fully female, a phenomenon also seen in therapies with other transsexual patients (Meyenburg, 1999). REFERENCES Bick, D., Franco, B., Sherins, R. J., Heye, B., Pike, L., Crawford, J., Maddalena, A., Incerti, B., Pragliola, A., Meitinger, T., and Ballabio, A. (1992). Intragenic deletion of the KALIG-1 gene in Kallmann’s syndrome. New Engl. J. Med. 326: 1752–1755. B¨urger-Prinz, H., Albrecht, H., and Giese, H. (1966). Zur Ph¨anomenologie des Transvestitismus bei M¨annern [On phenomenology of the transvestism in men], 2nd ed., Enke, Stuttgart. Duke, V. M., Winyard, P. J. D., Thorogood, P., Soothill, P., Bouloux, P. M. G., and Woolf, A. S. (1995). KAL, a gene mutated in Kallmann’s syndrome, is expressed in the first trimester of human development. Mol. Cell. Endocrinol. 110: 73–79.

P1: FDO Archives of Sexual Behavior [asb]

PL186-228373

October 19, 2000

Kallmann’s Syndrome and Transsexualism

19:34

Style file version Nov. 19th, 1999

81

Fischer, P.-A. (1971). K¨orperliche Befunde bei psychosexuellen St¨orungen. In Giese, H. (ed.), Die Sexualit¨at des Menschen [Human sexuality], 2nd ed., Enke, Stuttgart, pp. 940–955. Hardelin, J.-P., and Petit, C. (1995). A molecular approach to the pathophysiology of the X chromosomelinked Kallmann’s syndrome. Bailli`ere’s Clin. Endocrinol. Metabol. 9: 489–507. Kallmann, F. J., Schoenfeld, W. A., and Barrera, S. E. (1944). The genetic aspects of primary eunuchoidism. Am. J. Ment. Defic. 48: 203–236. Meitinger, T., Heye, B., Petit, C., Levilliers, J., Golla, A., Moraine, C., Dalla Piccola, B., Sippell, W. G., Murken, J., and Ballabio, A. (1990). Definitive localization of X-linked Kallmann syndrome (hypogonadotropic hypogonadism and anosmia) to Xp22.3: Close linkage to the hypervariable repeat sequence CRI-S232. Am. J. Hum. Genet. 47: 664–669 [Erratum: 47: 883]. Meyenburg, B. (1999). Gender identity disorder in adolescence: Outcomes of psychotherapy. Adolescence 34: 305–313. Sigusch, V. (1998). The neosexual revolution. Arch. Sex. Behav. 27: 331–359.

P1: FDO Archives of Sexual Behavior [asb]

PL186-228374-378

October 21, 2000

17:8

Style file version Nov. 19th, 1999

Archives of Sexual Behavior, Vol. 30, No. 1, 2001

BOOK REVIEWS Male, Female: The Evolution of Human Sex Differences. By David C. Geary. American Psychological Association, Washington, DC, 1998, 397 pp., $39.95 (members/affiliates), $49.95. Reviewed by Michael C. Seto, Ph.D.1

Darwinian thinking has had a tremendous impact on the scientific study of human behavior. In particular, Trivers’s discussion of parental investment (Trivers, 1972) has provided a powerful theory for integrating empirical findings and generating testable hypotheses in cognitive, social, and developmental research on human sex differences. The elegant premise of parental investment theory is that a sex difference in the minimum required investment for reproduction leads to different benefits and costs for pursuing mating opportunities versus caring for offspring. In humans, females have a greater minimum investment than men have, because of the time and energy involved in pregnancy. From this difference, a panoply of other sex differences logically follow. Geary states his ambition early: “The purpose of this book is to explain the different ways in which men and women pursue one of the most fundamental goals in life, to reproduce, and to describe how the accompanying reproductive strategies are related to a wide array of physical, behavioral, social, emotional, cognitive, and neural sex differences” (p. xi). This book is important reading for anyone interested in understanding sex differences. The writing is entertaining and lucid, and Geary demonstrates a strong grasp of the relevant biological, psychological, and anthropological literatures. In chapter 1, he cogently argues that theories based on Darwinian selection offer powerful paradigms for understanding human sex differences. In chapters 2 and 3, he presents evidence for the process of natural selection in nonhuman animals, including modern observations of physical changes in Darwin’s finches that correspond to changes in environmental conditions on the Gal´apagos Islands (see Weiner, 1995). In chapters 4 and 5, he applies the theory to humans, paying special attention to cross-cultural and historical data on parental investment, intrasexual 1 Law

and Mental Health Program, Centre for Addiction and Mental Health, Clarke Division, 250 College Street, Toronto, Ontario M5T 1R8 Canada. 83 C 2001 Plenum Publishing Corporation 0004-0002/01/0200-0083$19.50/0 °

P1: FDO Archives of Sexual Behavior [asb]

84

PL186-228374-378

October 21, 2000

17:8

Style file version Nov. 19th, 1999

Book Reviews

competition, and mate choice. He then elaborates on these topics in a detailed discussion of cognitive, social, and developmental research on sex differences, in chapters 6, 7, and 8. In the final chapter, Geary considers the implications of human sex differences for contemporary societies, focusing in particular on violence, anxiety and depression, academic achievement, and occupational interests. Chapter 6 is pivotal: Geary connects sex differences in social and sexual behavior (the foci of chapters 4 and 5) to sex differences in physique, play, cognitive abilities, and brain structures (the foci of chapters 7 and 8). The crucial links between these sets of findings are the cognitive, emotional, and motivational predispositions that characterize the evolution and development of human minds. Geary speculates that these predispositions reflect a fundamental human goal, to control resources that support survival and reproduction, such as territory, food, social dominance, and alliances. For example, he notes that both men and women value social relationships, but differ in the focus of their social activities (and therefore how they achieve control over resources). In this context, he cites the finding that four out of five women value friendly, reciprocal relationships more than the average man does, whereas three out of four men value antagonistic, competitive relationships more than the average woman does (Geary uses this reader-friendly format to express the magnitude of sex differences throughout the book). I particularly enjoyed Geary’s detailed discussion of the complexities in considering the research on cognitive and social development, including the evocative and reciprocal effects of individual predispositions and environmental factors. Geary makes a good case for the relevance of Darwinian selection to sex-typical cognitive and emotional processing. He describes research using the Profile of Nonverbal Sensitivity, a test that involves watching a 45-minute film of a woman’s nonverbal behavior, including facial expressions, body posture, and vocal intonation (speech that had been blurred so that individual words could not be heard). Across samples varying in age and nationality, females show a moderate advantage over males in the accuracy of their judgements regarding the emotional cues in the film, with two out of three females doing better than the average male does. Other evidence suggests that women are more accurate than men are in detecting disgust, fear, or sadness in both female and male faces, and in detecting anger in female faces. However, men are better at detecting angry expressions in male faces than in female faces. Along with evidence for sex differences in language abilities, social relationship skills, and childhood play, these findings speak to the selection pressures of intrasexual competition: males are more likely to compete with other males physically or through the attainment of status and material resources, whereas females are more likely to compete with other females in terms of their sexual and social relationships. A laudable feature of this book is that Geary does not bother to address the naturalistic fallacy (inferring values from facts, e.g., because sex differences in aggression or the ability to accurately detect emotions in others exist, they must

P1: FDO Archives of Sexual Behavior [asb]

PL186-228374-378

October 21, 2000

Book Reviews

17:8

Style file version Nov. 19th, 1999

85

therefore be “natural” and “good”) or the deterministic fallacy (traits that have been selected for in the past cannot be influenced by the environment) regarding Darwinian theories. Hopefully, this portends the demise of nature versus nurture debates in the public and scientific arenas. Genetic programs require environmental input to operate, so the interesting questions become: what kinds of programs are likely to exist? What are these programs designed to do? What kinds of input are important for these different programs? This book can be seen as a worthy response to the recent challenge raised by the evolutionary biologist E. O. Wilson in his book, Consilience: The Unity of Knowledge. Wilson (1998) compellingly argued that psychologists will have incomplete theories to the extent they are not cognizant of biological theories, sociologists will have incomplete theories to the extent they are not cognizant of psychological and biological theories, and so on, until the physical and social sciences are vertically integrated. Geary makes a valuable contribution to our knowledge by attempting to integrate biological, psychological, and anthropological findings on the nature and magnitude of human sex differences. Sex differences are group differences; they do not speak directly to the opportunities, limitations, rights, and liberties of individuals. Equitable treatment of men and women and the reduction of phenomena such as violence and depression are unlikely to be achieved without an accurate understanding of sex differences. In Geary’s own words, “At the broadest level, there are both scientific and social benefits that can be accrued by using sexual selection as a framework for understanding sex differences in modern society” (p. 330). REFERENCES Trivers, R. L. (1972). Parental investment and sexual selection. In Campbell, B. (ed.), Sexual Selection and the Descent of Man 1871–1971, Chicago, Aldine, pp. 136–179. Weiner, J. (1995). The Beak of the Finch, Vintage, New York. Wilson, E. O. (1998). Consilience: The Unity of Knowledge, Knopf, New York.

Sexual Orientation: Toward Biological Understanding. Edited by Lee Ellis and Lynn Ebertz. Praeger, Westport, Connecticut, 1997, 276 pp., $65.00. Reviewed by Martin L. Lalumi`ere, Ph.D.2

This edited volume is a collection of chapters based on papers presented at the First International Behavioral Development Symposium: Biological Basis of Sexual 2 Law and Mental Health Program, Centre for Addiction and Mental Health, 250 College Street, Toronto,

Ontario M5T 1R8 Canada.

P1: FDO Archives of Sexual Behavior [asb]

86

PL186-228374-378

October 21, 2000

17:8

Style file version Nov. 19th, 1999

Book Reviews

Orientation and Sex-Typical Behavior, held in Minot, North Dakota in 1995 under the leadership of Lee Ellis and funded by a grant from the Eugene Garfield Foundation. This volume is mainly on sexual orientation. A companion volume on sex-typical behavior has also been published (Ellis and Ebertz, 1998) and is reviewed by Slob (2001) in this issue. The editors sought to produce a book that would be accessible to nonspecialists. The 11 chapters cover distinct yet interrelated topics and thus provide the uninitiated with a conceptual map to navigate through this fascinating territory. The chapters are also quite short and straight to the point. This said, I believe that only the chapters by Meyer-Bahlburg (prenatal estrogens), Ellis (prenatal stress), and Halpern and Crothers (cognitive abilities) would be easily understandable to nonspecialists; the other chapters require technical knowledge to be fully appreciated. Also, the heavy use of acronyms reduces the reader-friendliness of the book. The editors also sought to provide a collection of chapters describing research “on the leading edge of the scientific work on sexual orientation” (pp. xix, xx). All 11 chapters provide expert reviews of the latest scientific work on human and nonhuman endocrinology, neuroanatomy, prenatal influences, sexual differentiation, cerebral laterality, and cognitive abilities, and their role in the development of sexual orientation. Many chapters also include previously unpublished data. Surprisingly, there is no chapter on behavioral or molecular genetics, and no mention anywhere of possible immunological influences in the development of sexual orientation, three areas of great interest to scientists currently involved in the “biological” study of sexual orientation. There are also no discussions of the apparent Darwinian paradox that homosexuality poses. As the title suggests, the focus of the chapters is on “biological” processes: “About the only idea that all of the contributors to this volume share is that sexual orientation can be fully comprehended only within a framework that accommodates biological variables. This does not deny the role of learning and the social environment but simply recognizes that biological variables are also important” (p. xx). Here the editors appear to commit what Wrangham and Petersen (1996) call “Galton’s error,” that is, to separate nature and nurture and to believe that learning and social influences are outside the realm of biology. In fact, it becomes clear throughout the book that Ellis and Ebertz simply mean that it is quite likely that influences other than psychosocial ones operate in the development of sexual orientation, and that these influences have effects very early in development. All 11 chapters offer clues as to what these influences may be. Quite a few chapters describe research on nonhuman species. Grober examines sex reversal in bluehead wrasse and shows how social cues can be “biological.” These fish come in three morphs: terminal phase males, who are aggressive and court females; initial phase males, who outnumber females and engage in mass spawning without courting; and initial phase females. When the terminal male

P1: FDO Archives of Sexual Behavior [asb]

Book Reviews

PL186-228374-378

October 21, 2000

17:8

Style file version Nov. 19th, 1999

87

is removed, the largest initial phase fish (male or female) becomes the terminal male. Thus, an initial phase female can experience complete sex reversal, both in anatomy and behavior. Grober shows that it is possible to experimentally study the neuroendocrinological causes of sex and role reversal in this and related species. Kinsley et al. review the effects of different prenatal stressors on rats’ sexual behavior. The types of stressors used in experimental manipulation include handling during pregnancy, overcrowding, saline injections, uncontrollable noise or electric shock, restraint, and excess heat or light. As originally reported by Ward (1972), prenatal stress affects male copulatory behavior without noticeable alteration in reproductive physiology. The effects sometimes include demasculinization in behavior (less copulation, less ejaculation) but more consistently include feminization (increased lordosis). The rearing environment (e.g., cohabitation with adult females) can alleviate the effects of stress. The effects of prenatal stressors have been noted on different aspects of neurochemistry (change in levels of hormones and neurotransmitters) and neuroanatomy (in particular, changes in sexdimorphic brain structures). The chapter by Ellis on the effects of prenatal stress in humans offers a good complement to this chapter. Bakker and Slob focus on the role of prenatal estrogens as a masculinizing agent in male rats. They review the experimental research on the effect of reducing estrogen formation in the brain by the administration of an aromatase inhibitor. Treated rats tend to show more or less normal masculinization but do not show defeminization. As a result, the sexual motivation and behavior of rats treated with the inhibitor seem completely determined by the sex of the available sexually mature partner: A male partner produces female-like sexual behavior in treated males, and a female partner produces male-like behaviors! When castrated and treated with estrogens, however, the treated males behave as if they were homosexual. This and other evidence presented suggest that “the psychosexual differentiation of the male rat requires the neonatal action of estrogens, derived from neural aromatization of gonadal testosterone” (p. 106). Although rats have been the species of choice for these types of studies, other species have been studied as well. Perkins and Fitzgerald provide an extremely interesting review of their research with domestic sheep, along with a very good story on how Perkins started studying sexual behavior in sheep. Their review tells us that an astonishingly high 10% of domestic rams show a clear homosexual preference in choice tests. A ram is considered homosexual in choice tests if it consistently shows interest in and copulatory behavior with other rams and consistently shows no sexual interest in ewes. The sexual behavior and orientation of sheep seem influenced by a number of factors, including administration of androgens (in ewes), having a male twin (in rams), and having a very reproductively successful mother (in rams). However, the presence or absence of same-sex or cross-sex conspecifics during development, or the presence of infant sexual play, do not affect sexual maturity or sexual performance; infant males who mount

P1: FDO Archives of Sexual Behavior [asb]

88

PL186-228374-378

October 21, 2000

17:8

Style file version Nov. 19th, 1999

Book Reviews

other infant males are not more likely to prefer rams when they reach adulthood. Data from small samples suggest that homosexual sheep show a greater surge of testosterone when exposed to rams than to ewes, and that the number of estradiol receptors in the amygdala varies by sex and sexual orientation (ewes = homosexual rams > heterosexual rams). It should be noted that homosexual behavior, but not homosexual preferences, has been observed among wild populations. It remains a mystery why so many domestic sheep show a homosexual preference without any direct intervention of the investigators. What about humans? Meyer-Bahlburg examines the role of estrogens. He notes that both prenatal androgens and estrogens seem necessary for sexual differentiation of the brain and behavior in rats (the two-pathway model); one has masculinizing effects whereas the other has defeminizing effects (estrogen has feminizing effects in females at puberty). Meyer-Bahlburg addresses the question of whether the two-pathway developmental model also applies to human sexual orientation. Studies of the effect of diethylstilbestrol (DES) suggest a role for estrogens, but a simple toxic effect of DES cannot be eliminated as an alternative explanation. Studies of androgen insensitivity in males (in which female external physiology is observed) do not support the idea of a major role for estrogens. At this point, the one-pathway developmental model involving androgens has more empirical support than the two-pathway model. Dittmann presents a short review of another medical condition that could provide useful information on the role of prenatal hormones in the development of sexual orientation. Congenital adrenal hyperplasia (CAH) in girls leads to an increased production of androgens and frequent masculinization of the external genitalia. Like DES, CAH is associated with somewhat elevated rates of homosexuality. In accord with recent data, results presented by Dittmann suggest that CAH females are more likely to report homosexual fantasies, a homosexual orientation, and are more likely to appear to be masculine in many aspects of their lives than do their non-CAH sisters. This group difference is particularly evident among CAH females suffering from the salt-wasting variant of this condition. Most DES and CAH females, however, are heterosexual. The review by Swaab et al. focuses on the hypothalamus, a brain structure that is related to sexuality in other species. The authors suggest that the hypothalamus and other parts of the brain of homosexual men are not “feminized,” as is usually suggested, but belong to a “third sex.” They base this suggestion on results of morphological studies showing differences as a function of sexual orientation for structures that are not sexually dimorphic, and a lack of differences for structures that are sexually dimorphic. As their review of the empirical data suggests, however, their opinion, like the more common opinion that homosexual men have feminine-like brain structures, is premature. Not very much is known about particular sex differences in the structure of the brain and the few existing studies have used different methodologies and have produced inconsistent results. One wonders why there are not more studies on this very important topic.

P1: FDO Archives of Sexual Behavior [asb]

PL186-228374-378

October 21, 2000

Book Reviews

17:8

Style file version Nov. 19th, 1999

89

Another way to study the brain-sexual orientation connection is to examine the cerebral laterality and the cognitive abilities of heterosexual and homosexual men and women. Holtzen reanalyzed data from his previous study (Holtzen, 1994) and showed that male and female homosexuals are less right-handed than their heterosexual counterparts. Holtzen’s study is one of the most elegant studies on handedness and sexual orientation: it includes four groups (homosexual men and women and heterosexual men and women); the comparison groups are heterosexual relatives of the homosexual men and women; the handedness measure provides continuous scores; and the analysis considers potential confounding factors (age and education). The chapter by Herman-Jeglinska et al. presents new data on handedness among transsexuals. Strangely, the sexual orientation of the transsexuals and of the control subjects was not specified. Most female-to-male but only about half of male-to-female transsexuals have a homosexual sexual orientation (Blanchard et al., 1987). Herman-Jeglinska et al. report that the transsexual subjects were less likely than the control subjects to be right-handed. The group difference was larger for female-to-male transsexuals, and greater for subjects without familial nonright-handedness. These results have been confirmed in a recent meta-analysis involving over 23,000 subjects: sexual orientation is related to handedness in both men and women, in that homosexuals are more likely to be nonright-handed than heterosexuals; the effect is stronger among women (Lalumi`ere et al., 2000). Finally, Halpern and Crothers provide a succinct review of sex differences and sexual orientation differences in cognitive abilities. They propose to study cognitive abilities using an information processing approach rather than the more traditional content approach (e.g., verbal vs. spatial). Halpern and Crothers’s new data are at odds with previous data and current theories: homosexual males produced a pattern of results that was more “masculine” than the pattern of data obtained from heterosexual males! This result is very surprising because the cognitive tasks chosen by the authors have produced very large sex differences in previous studies, as they did in this study. By putting this chapter last, the editors remind us that much remains to be learned in this field of inquiry. Still, this volume leads one to believe that the study of the early determinants of sexual orientation has been very productive and that major scientific discoveries are just around the corner.

REFERENCES Blanchard, R., Clemmensen, L. H., and Steiner, B. W. (1987). Heterosexual and homosexual gender dysphoria. Arch. Sex. Behav. 16: 139–152. Ellis, L., and Ebertz, L. (eds.). (1998). Males, Females, and Behavior: Toward Biological Understanding, Praeger, Westport, CT. Holtzen, D. W. (1994). Handedness and sexual orientation. J. Clin. Exp. Neuropsychol. 16: 702–712. Lalumi`ere, M. L., Blanchard, R., and Zucker, K. J. (2000). Sexual orientation and handedness in men and women: A meta-analysis. Psychol. Bull. 126: 575–592. Slob, A. K. (2001). Review of Males, Females, and Behavior: Toward Biological Understanding. Arch. Sex. Behav. 30: 90–91.

P1: FDO Archives of Sexual Behavior [asb]

PL186-228374-378

90

October 21, 2000

17:8

Style file version Nov. 19th, 1999

Book Reviews

Ward, I. L. (1972). Prenatal stress feminizes and demasculinizes the behavior of males. Science 175: 82–84. Wrangham, R., and Peterson, D. (1996). Demonic Males, Houghton Mifflin, Boston.

Males, Females, and Behavior: Toward Biological Understanding. Edited by Lee Ellis and Linda Ebertz. Praeger, Westport, Connecticut, 1998, 322 pp., $69.50. Reviewed by A. Koos Slob, Ph.D.3

This is the second edited volume from the First International Behavioral Development Symposium: Biological Basis of Sexual Orientation and Sex-Typical Behavior held in Minot, North Dakota in 1995 organized by Ellis. The first volume (Ellis and Ebertz, 1997) is reviewed by Lalumi`ere (2001) in this issue. The focus of the current volume is on how biological factors influence gender differences in behavior, without denying the importance of learning and social environment variables. The book starts with an excellent foreword by Diamond, entitled “Sexual Development: Nature’s Substrate for Nurture’s Influence.” Diamond nicely illustrates the necessity of talking about biological influences on the adult behaviors of animals, including humans. In the 1990s, we definitely left the “sociological era” in which, for instance, it was taken for granted that a person’s sexual identity and body image was largely considered a function of socialization. The political climate has permanently changed and it is a relief that international meetings as mentioned earlier can be organized without condemnation or without being accused of nonpolitically correct science! Diamond’s foreword is highly recommended: it is well-written, quite easy to understand, and up-to-date. What I have said about Diamond’s foreword does not hold for all the chapters in the book. They are all written by very knowledgeable scientists, and they represent solid research, but one needs quite a bit of biological training and knowledge to really understand all the articles. Many studies deal with perinatal biological manipulations in laboratory animals, including the rat (Lustig, Csaba, Hull, Wilson, Vathy, Field) and monkeys (Dixson). Human studies include genetic factors in Tourette syndrome (Comings), twin studies (Hoyenga, Miller, Segal), and sex differences in intellectual giftedness (Lippa). In conclusion, the studies presented in this book are bound to be relevant to the real world as science continues to probe the complex mysteries of sex differences and similarities in behavior. Let us hope that the book will add to the ease with which we all come to manifest and understand both our gender and our sexual identities. 3 Department

of Endocrinology & Reproduction, Erasmus University Medical Center, PO Box 1738, 3000 DR Rotterdam, The Netherlands.

P1: FDO Archives of Sexual Behavior [asb]

PL186-228374-378

October 21, 2000

Book Reviews

17:8

Style file version Nov. 19th, 1999

91

REFERENCES Ellis, L., and Ebertz, L. (eds.). (1997). Sexual Orientation: Toward Biological Understanding, Praeger, Westport, CT. Lalumi`ere, M. L. (2001). Review of Sexual Orientation: Toward Biological Understanding. Arch. Sex. Behav. 30: 85–90.

Sex At Risk: Lifetime Number of Partners, Frequency of Intercourse, and the Low AIDS Risk of Vaginal Intercourse. By Stuart Brody. Transaction Publishers, New Brunswick, New Jersey, 1997, 222 pp., $32.95. Reviewed by Michael C. Seto, Ph.D.4

In this book, Brody reviews the research on correlates of partner number (chapters 1–6) and frequency of sexual intercourse (chapters 7–12), and then discusses the risk of infection with AIDS among heterosexuals (chapters 13–20). He concludes with two chapters summarizing his views and criticizing the conflation of politics and science in the study of sexual behavior, particularly pertaining to the epidemiology of AIDS. The reviews of research on number of partners and frequency of sexual intercourse are dense. Brody describes findings under loosely related subheadings (e.g., Contraceptive Choice, Psychological Variables, Substance Use) and then concludes each section with concise summaries. As acknowledged in the preface, Brody eschews declaring a theoretical perspective, in order to minimize bias in the discussion of findings and to allow readers to draw their own conclusions. However, organizing the information presented in these chapters under one or more theoretical perspectives would have helped readers to interpret the evidence. For example, one could argue that Darwinian selection theory can integrate many different findings about sexual behavior, relating them back to the ultimate causes of sexual behavior. Topics that have been studied from a Darwinian perspective include intrasex variations in casual sexual activity, sex differences in partner and relationship choice, and the role of personality in sexual behavior (Buss, 1994; Wright, 1994). The detailed reviews in the first two sections set the stage for the third section of the book, which presents Brody’s central thesis: the risk of infection with AIDS for heterosexuals engaging in vaginal intercourse, at least in the West, has been greatly exaggerated. Relatively high risk groups have been identified: intravenous (IV) drug users, homosexual and bisexual men, blood transfusion recipients, and individuals from Pattern-II countries (e.g., sub-Saharan Africa, the 4 Law

and Mental Health Program, Centre for Addiction and Mental Health, Clarke Division, 250 College Street, Toronto, Ontario M5T 1R8 Canada.

P1: FDO Archives of Sexual Behavior [asb]

92

PL186-228374-378

October 21, 2000

17:8

Style file version Nov. 19th, 1999

Book Reviews

Caribbean, etc.). The risk from vaginal intercourse in the West is typically inferred from individuals with AIDS who acknowledge heterosexual activity but deny sexual contact with these high risk groups and IV activity. Brody goes on to question this inference in a series of reasoned arguments. First, heterosexual transmission, when it does occur, is likely to involve anal intercourse. Second, some people lie or forget about their experiences with homosexual or bisexual men, receptive anal intercourse, or IV drug use. Third, Pattern-II transmission is likely to be linked to differences in hygiene and public health. However, even basic epidemiological data are difficult to obtain from Pattern-II countries, and the necessary research on the cofactors for the development of AIDS within these populations has not yet been done. Brody makes an obvious but often overlooked point while discussing the research on AIDS and risky sexual behavior: sex is rewarding and humans are motivated to engage in rewarded activities. The only certain way to achieve zero risk for infection with a sexually transmitted disease is to not have sexual intercourse at all, which is not an acceptable choice for most people. Brody presents Symons’ estimates (Symons, 1993) that, even if all cases claiming heterosexual transmission through vaginal intercourse were correct, the loss in life expectancy (a standard epidemiological measure of risk) from acquiring AIDS through vaginal intercourse is lower than the loss in life expectancy due to driving regularly or being 12 pounds overweight. As Brody notes, “A dull long life is not necessarily preferable to a shorter but fulfiling [sic] life, and there may be experiences in life that are more personally meaningful in their accumulation than birthdays” (p. 161). He also quotes Symons (1993, pp. 345, 346): Natural selection has provided us with only a few ways of experiencing intense pleasure, and surely, for many people, sexual intercourse is at or near the top of this short list; hence, anything that diminishes sexual pleasure constitutes a significant cost. Yet many writers, both popular and professional, appear to assume that sex is the one area in life where we should strive to reduce risk no matter what the cost.

Most heterosexuals are objectively at low risk for developing AIDS, suggesting they are not being irrational by avoiding condom use or having multiple sexual partners (although they might be exposing themselves to the relatively greater risk of unwanted pregnancy or infection with other sexually transmitted diseases). Overall, this book is competently written, carefully and logically presenting the evidence for Brody’s central argument about the risk of acquiring AIDS for heterosexuals living in industrialized countries. However, the first two sections could have been better organized, as they sometimes make for dry reading, and some of the parenthetical or tangential comments could have been less distractingly presented in footnotes. As a plus, there are glints of sardonic humor throughout the book: Brody’s “Note on Language” at the end of the book is manifestly funny in its defiance of modern sensitivities about terminology. Another positive aspect of the book is Brody’s careful discussion of methodological concerns in conducting sex

P1: FDO Archives of Sexual Behavior [asb]

PL186-228374-378

October 21, 2000

Book Reviews

17:8

Style file version Nov. 19th, 1999

93

research, particularly regarding the use of self-report (issues include forgetting, heuristic thinking, lying, and misunderstanding questions). For example, Brody comments on the empirical impossibility of the consistently found sex difference in partner number: in survey after survey, men report having more sexual partners than women, even though, by definition, each heterosexual interaction between two new partners involves a person of each sex (for another explanation, see Wiederman, 1998). A major strength of this book is Brody’s enthusiasm for challenging orthodox views. His succinct indictment of people who have conflated politics and science in setting AIDS policy over the past 15 years is important reading for anyone concerned about the application of scientific knowledge and the politicization of sex research and education (see Fumento, 1993). AIDS continues to be a major focus of public health policies. Some advocates of monogamy and condom use and some AIDS activists may criticize this book, suggesting that it underestimates the risk for heterosexuals and thereby hampers efforts to prevent the spread of AIDS, but the existing research evidence does not support their positions. The vectors of greatest concern appear to be anal intercourse and IV drug use, especially among individuals with already compromised immune functioning. Public education efforts advocating condom use, monogamy, and nonpenetrative activities among already low-risk heterosexuals have served only to increase sexual anxiety as we enter the third millennium.

REFERENCES Buss, D. M. (1994). The Evolution of Desire: Strategies of Human Mating, Basic Books, New York. Fumento, M. (1993). The Myth of Heterosexual AIDS, Regnery Gateway, Washington, DC. Symons, D. (1993). How risky is risky sex? J. Sex Res. 30: 188–191. Wiederman, M. W. (1998). The truth must be in here somewhere: Examining the gender discrepancy in self-reported lifetime number of sex partners. J. Sex Res. 35: 375–386. Wright, R. (1994). The Moral Animal, Pantheon, New York.

Adolescents and AIDS: A Generation in Jeopardy. Edited by Ralph J. DiClemente. Sage Publications, Newbury Park, California, 1992, 314 pp., $36.00 (hardcover), $17.95 (softcover). Reviewed by Lynn R. Surrey, Ph.D.5

The pervasive impact of HIV disease on our current culture has required health care professionals to investigate its influence from every feasible perspective. One 5 Department

of Psychology, University of Indianapolis, 1400 E. Hanna Ave., Indianapolis, Indiana 46227-3697.

P1: FDO Archives of Sexual Behavior [asb]

94

PL186-228374-378

October 21, 2000

17:8

Style file version Nov. 19th, 1999

Book Reviews

compelling area of significance has been the categorization of the disease effects at different age levels. People in the adolescent age category pose an especially difficult challenge because, as Coates points out in the foreword, many people believe that “the implementation of prevention programs . . . will cause adolescents to become more sexual or will provide a social environment that promotes sexual activity . . . ” (p. viii). Regardless, the data concerning the increasing number of teenage pregnancies and the increasing number of teenagers who have contracted sexually transmitted diseases illustrate that adolescents are sexually active. The dissemination of accurate information is critical to protect this next generation of our adult population from the spread of HIV disease. DiClemente’s preface clearly delineates the need to address HIV and its effect on the adolescent population by using a multidimensional approach. Although the disease has a biological basis, the epidemic is clearly fueled by the associated sexual behavior and choices that are the product of an entire psychosocial phenomenon. Recognizing adolescence as “a time for growth and experimentation, a period marked by establishing autonomy and confronting new challenges”(p. xii), DiClemente believes that the promotion of health and well-being for this age group requires an understanding of HIV from a classic past, present, and future perspective. His book is divided into three sections: (1) Epidemiology: The Scope of the Problem, which focuses on descriptions of adolescent behavior; (2) Prevention: Theory, Design, and Evaluation, in which the development and maintenance of HIV prevention strategies are discussed; and (3) Policy and Legal Perspectives, a section that describes health and public policy influences on adolescents and HIV. The first section of this work contains some inherent errors simply by nature of the task that the contributors are trying to accomplish. Tracking the prevalence of HIV in the adolescent population is the goal, particularly in the first chapter by Hein. Although the product is thorough in its intent and approach, the accuracy of the results are limited in terms of their current impact because the data are simply outdated. Determining characteristics of adolescent AIDS cases, seroprevalence rates, premature cause of death, and transmission category profiles are all extremely important and necessary, but in this case, all of the information presented is based on studies published or data collection conducted in the late 1980s. Nevertheless, there are some significant issues raised here that continue to be relevant and require thoughtful consideration. Most notably, it is argued that a significant impact on the transmission of HIV will be made only when behavior begins to match knowledge. This is a recurring theme voiced by many educators. Repetition of this idea is also noted by the subsequent contributors as they address the issues of adolescents in all demographic categories, and particularly in discussing specific problems for homeless and incarcerated adolescents. In many of the areas covered, the presentation of the process of monitoring outcomes and the measurement of the impact of critical effect factors is focused on this lack of connection between knowledge and behavior. Illustrations regarding knowledge, attitudes, beliefs, and behaviors, and

P1: FDO Archives of Sexual Behavior [asb]

Book Reviews

PL186-228374-378

October 21, 2000

17:8

Style file version Nov. 19th, 1999

95

the editor’s own chapter regarding the determinants of condom use among adolescents, all depict the importance of addressing this knowledge/behavior dichotomy. There are some other connected issues of importance that receive attention here, such as health care and prevention elements, cultural differences, substance use and abuse factors, and geographical influences on seroprevalence rates. The book’s second section begins with a thought-provoking chapter by Bandura regarding the social-cognitive approach to making behavioral changes. This approach is based on employing cognitive strategies and the ability of people to affect changes based on a number of substantial criteria. He presents some interesting ideas regarding the influence of a number of factors that promote change, focusing particularly on the importance of self-efficacy and its primary role in human behavior, empowerment, and effective change techniques. Other chapters in this section seem to take a broader view of prevention strategies. They focus on the use of mass media, school-based education programs, training of individual skills based on group norms, and matching training and education based on current “serostatus.” (Although there is thoughtful discussion from this broad perspective, there still seems to be a considerable gap between education strategies and making an effective connection with individual behaviors.) Although there is a conceptual significance to the importance of emphasizing prevention strategies, this section appears to miss the critical issue of most education efforts, which is, as previously discussed, addressing the ability to promote effective individual behavior change based on a person’s knowledge base. Even in Kirby’s discussion regarding schoolbased prevention programs, there is an acknowledgment that education programs in broad arenas have increased the base knowledge that individuals have regarding HIV disease, yet their individual behaviors are not influenced by this knowledge base. Schinke and Gordon show some promise in discussing interpersonal skills training for minority adolescents, but their approach seems to lack some practical consideration regarding the statistical likelihood that adolescents in lower socioeconomic groups will actually have access to computer-based instructional technology. And the final chapter here that addresses the need to focus programs based on current serostatus demonstrates the necessity to understand a host of factors which influence educational effectiveness, yet, it too ignores the critical need to help individuals translate individual knowledge into behavioral change. The final section on Policy and Legal Perspectives assimilates a number of conceptually relevant issues in consideration of the role of government agencies at local and national levels. It addresses the meshing of a number of social and political phenomena, such as the policy debate in New York City over the distribution of condoms in schools, curricula decisions about HIV prevention in educational programs at different age levels, and strategies for educating adolescents who do not attend school. D’Angelo’s interesting commentary demonstrates how the age category of adolescence has been omitted by some government agencies and the effect that this error has had on the medical care of adolescents. The vast

P1: FDO Archives of Sexual Behavior [asb]

96

PL186-228374-378

October 21, 2000

17:8

Style file version Nov. 19th, 1999

Book Reviews

discrepancy between the amount of money that is authorized for care and the amount that is actually appropriated is alarming, and this situation clearly supports the need to continue to advocate support for education and prevention and medical care programs at all levels of government. The final chapter by English presents an interesting mix of the complex legal and ethical dilemmas that may arise as a result of working with adolescents, such as consent for testing and treatment, participation in clinical trials and research, and the debate over confidentiality regarding HIV-related information. The information addressed here is done in a cursory manner, acknowledging that this realm itself is broad, yet a comprehensive review of any consequence regarding the impact of HIV disease could not be presented without it being included. DiClemente has assembled a well-qualified group of experts in an attempt to take a comprehensive look at a difficult sociocultural challenge. This work is, overall, well-documented and very well written. One major concern, as stated earlier, is that there is a significant amount of information presented that is outdated simply by the nature of the changing status of many perspectives of HIV, particularly outcome statistics and other types of data regarding prevalence, trends, and projections of the incidence of HIV in the future for this age group. It is also outdated in the use of some of its language in regard to HIV infection and disease, yet the book accomplishes the goal of presenting a balanced and comprehensive view of HIV and how it has impacted people in the age category of adolescence. Although some strategies for change are presented in a conceptual manner, the practical application of individual behavioral change based on a person’s knowledge continues to be an apparently elusive element for many educational programs. And even though the final section presents some important issues in a brief way, it balances the overall presentation of the book by completing the complex picture of how government policies and education programs might effectively address many of the issues highlighted in the previous two sections. Wells underscores the difficulty of educating about and establishing prevention techniques for HIV with this age group by acknowledging that “the scope of adolescence is intrinsically problematic for policymakers. It is a period of life that begins in childhood but ends in adulthood” (p. 234). He also states that “it is the period in which many of the behaviors are initiated that are most likely to place adolescents at risk of HIV infection, namely sexual behavior for the majority, and drug use for a smaller number” (p. 234). But he summarizes, again, the critical issue well by stating that although “It is accepted generally among health educators that educating adolescents about HIV infection and AIDS is an important step in preventing the spread of HIV. . . . Educational programs for adolescents also need to catch up with current emphases on behavior change” (p. 245).