Political Campaigning in Referendums: Framing the Referendum Issue

De Vreese and Semetko document the importance of the media and information ﬂows in direct democracy elections. The stud...

Author: Holli A Semetko

148 downloads 583 Views 908KB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

De Vreese and Semetko document the importance of the media and information ﬂows in direct democracy elections. The study creatively combines public opinion surveys, media content, and electoral analyses to examine the dynamics of citizen choice in the referendum. Direct democracy appears to be the wave of the future. This book illustrates the challenges this poses for our theories of democratic citizenship, campaign processes, and the quality of democratic choices. Professor Russell Dalton, University of California, Irvine At last! This study ends the shameful scholarly neglect of the unique characteristics of referendum campaigns and the information environment that shapes their outcome. De Vreese and Semetko use panel surveys, focus groups, content analyses, and interviews with journalists and campaign managers to assess how these important campaigns are framed and conducted and how they affect political systems in general as well as individual politicians, the media, and citizens. This highly readable pioneering study, which uses Denmark’s referendum on the euro as a prototypical test case, is essential reading for everyone interested in political communication in democratic societies. Professor Doris A. Graber, University of Illinois at Chicago Political science research has not kept pace with the relatively recent arrival of “referendum democracy” in a number of the world’s democracies, and particularly with the spate of referendums associated with the project of deepening European integration. In this book, de Vreese and Semetko have helped us to ﬁnally catch up! Professor David Farrell, Manchester University

Political Campaigning in Referendums

Referendums are an increasingly important means of enacting or preventing legislation in countries around the world. Despite the simplistic nature of the Yes or No vote, the referendum issue is often complex and multifaceted. The framing of the referendum issue in the information environment can thus be crucially important to the process of crystallizing public opinion and mobilizing support during the campaign. This book reviews the research on campaigns and elections and investigates the effects of campaigning in referendums, drawing on panel survey data, media content data, focus groups, and interviews with journalists and campaign managers. The authors argue that the media coverage not only inﬂuences public perceptions of the campaign, the referendum issue, and the party leaders but that, in a close race, it also shapes the voting and the political future of the incumbent party. The ﬁrst study to investigate the dynamics and effects of a referendum campaign on politicians, media, and citizens, this innovative volume will be of interest to students and researchers of political science, communications, and EU politics. Claes H. de Vreese is Associate Professor and Dutch Science Foundation Fellow in the Amsterdam School of Communications Research at the University of Amsterdam. He is currently a Visiting Scholar at the Center for European Studies at Harvard University. Holli A. Semetko is Vice Provost for International Affairs and Director of the Claus M. Halle Institute for Global Learning, and Professor of Political Science at Emory University; and former Professor and Chair of Audience and Public Opinion Research at the University of Amsterdam.

Routledge research in political communication

1 Political Campaigning in Referendums Framing the referendum issue Claes H. de Vreese and Holli A. Semetko

Political Campaigning in Referendums Framing the referendum issue

Claes H. de Vreese and Holli A. Semetko

First published 2004 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Simultaneously published in the USA and Canada by Routledge 270 Madison Ave, New York, NY 10016 Routledge is an imprint of the Taylor & Francis Group This edition published in the Taylor & Francis e-Library, 2004. © 2004 Claes H. de Vreese and Holli A. Semetko All rights reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data Vreese, C. H. de (Claes Holger), 1974– Political campaigning in referendums : framing the referendum issue / Claes H. de Vreese and Holli A. Semetko. p. cm. Simultaneously published in the USA and Canada. Includes bibliographical references and index. 1. Advertising, Political. 2. Political campaigns. 3. Referendum. 4. Elections. 5. Voting. I. Semetko, Holli A. II. Title. JF2112.A4V73 2004 328.2⬘3–dc22 2004004146

ISBN 0-203-33503-1 Master e-book ISBN

ISBN 0-415-34941-9 (Print Edition)

Contents

List of ﬁgures List of tables Preface

vii viii x

PART I

The importance of information in referendums

1

1 The dynamics of a referendum campaign

3

2 The context of the campaign

23

PART II

The political parties’ campaigns 3 The Yes and No camps

43 45

PART III

The campaign in the news: polls, personalities, and the multi-faceted issue of the euro

63

4 The campaign in the news

65

5 Elite framing of the issue

92

vi

Contents

PART IV

Campaign effects

105

6 Agenda-setting in a referendum

107

7 Polls, strategy news, and political cynicism

120

8 Primed by the referendum

136

9 Campaign effects of the vote

153

PART V

The consequences of referendums

169

10 Conclusions

171

Appendices Notes References Author index Subject index

185 198 202 218 222

Figures

2.1 2.2 4.1 4.2 4.3 5.1 5.2 6.1 7.1 7.2 7.3 7.4 7.5 8.1 8.2 9.1 10.1

Attitudes towards the European Union: support for membership Opinion polls in the ﬁnal month of the campaign Visibility of the referendum in television news Share of news topics on television in August and September 2000 Share of actors from the Yes and No camps Television news framing of the referendum: conﬂict news framing Television news framing of the referendum: consequence news framing Effects of exposure to news coverage about the referendum on the salience of the issue Predicting political cynicism Predicting negative campaign evaluation Predicting change in political cynicism Predicting change in negative campaign evaluation Predictors of negative change in perception of role played by polls in the referendum campaign Explaining government evaluation, low political involvement citizens Explaining government evaluation, high political involvement citizens Inﬂuences on the likelihood of voting No Public support for the euro one year after the referendum

35 37 72 74 89 99 100 115 125 126 127 128 134 145 145 164 182

Tables

1.1 1.2 2.1

2.2 2.3 2.4 2.5 2.6 3.1 3.2 3.3 3.4 4.1 4.2 4.3 4.4 4.5 4.6 5.1 5.2 5.3 5.4 6.1 6.2

National referendums on European integration issues, 1972–2003 Typology of information and campaign effects Overview political parties: percentage of votes in national (1998) and European (1999) elections and Yes or No recommendation in referendum Vote intention general elections and vote in referendum Party and voter alignment: a classiﬁcation of party success Editorial positions of major television news programs and national daily newspapers in the referendum Percentage of electorate voting Yes and No and turnout in referendums on European issues, 1972–2000 Attitudes towards the euro preceding the referendum Overview of political parties’ and movements’ activities in the referendum campaign Advertising strategies by political party Party issue agenda I Party issue agenda II Television news agenda during referendum campaign Visibility referendum in daily newspapers Key newspaper headlines during the campaign Visibility and quality of reporting of polls Political actors in the news Visibility and evaluation of political actors in the ﬁnal month of the campaign The referendum issue on television news The referendum issue in the press Strategic news framing of the referendum The framing of the referendum issue by the media and the Yes and No camps Overview of agenda-setting studies Media agendas and public agendas during the referendum campaign

5 15

26 27 28 33 34 36 52 54 57 59 73 75 76–79 81 83 86–87 97 98 101 103 109 112

Tables 6.3 6.4 7.1 8.1 8.2 8.3 8.4 8.5 9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.9 9.10 10.1 10.2

Rank public and media agendas The multi-faceted issue of the euro: media agenda and public agenda referendum issue Aggregate level change on perceptions of the campaign Change in overall and issue-speciﬁc government performance rating during the campaign Change in evaluation of political leaders during the campaign Visibility and evaluation of selected actors in the campaign Evaluation of political actors Change stability in voting intention in general election Time of vote decision Vote by left–right political ideology Vote by government performance rating Vote by economic expectations Attitude towards European integration: seven items on EU skepticism scale Vote by EU skepticism Percentage No voters by gender, age, and education Vote intention (wave 1) and actual voting behavior (wave 2) Reasons for voting Yes Reasons for voting No Perceived consequences of the No result Predicting support for the euro one year after the referendum

ix 113 116 125 140 142

143 148–149 151 154 155 156 156 157 158 159 161 167 167 181 183

Preface

During the period in which we were writing this book, eleven countries in Europe held national referendums on issues of European integration. In 2001 and 2002, the Irish voted on the Treaty of Nice, in 2003, nine of the ten new member states of the European Union voted on the issue of membership, and in September 2003 the Swedes voted No to introducing the euro. And in the coming years, electorates in many “old” and “new” EU member states will be asked to vote in referendums that determine the fate of the ﬁrst European Constitution. Referendum campaigns and their outcomes are often unpredictable. Political elites often leave their voters in a state of uncertainty about whether or not to endorse the proposal put before the public. The campaign is therefore a central and often decisive process for deﬁning the issue and crystallizing public opinion. Yet we have only limited knowledge about the dynamics and role that a referendum campaign plays. We aim for this book to be of use to several audiences: undergraduate and graduate students in comparative politics courses, politicians and campaign managers, students and scholars of political communication, policy makers and journalists. The book is designed to address the role of information in the dynamic process of a referendum campaign. Each chapter reviews relevant literature and draws on comparative examples, before moving to our country case study. We use the Danish 2000 national referendum on the introduction of the euro to investigate questions about media contents, uses, and effects in a referendum campaign. This referendum was announced in March 2000. When voting took place on September 28, 2000, a small majority of the electorate (53 percent) voted against Denmark’s participation in the third phase of the European Monetary Union (EMU). During this referendum we had the opportunity to collect panel survey data, to conduct a content analysis of media coverage, and to carry out focus group interviews and interviews with journalists and campaign managers. This would not have been possible without substantial research grants awarded in an exceptionally short turn-around time by the Danish Research Academy, the Nordic Film Foundation, the Danish Broadcasting Associ-

Preface xi ation, and the Amsterdam School of Communications Research (ASCoR) at the University of Amsterdam. We want to express our sincere gratitude to the many individuals in these institutions who made it possible for us to fund this project. We are especially grateful to Peter Neijens, Professor and Director of ASCoR who himself is an expert on survey research and referendums, and who has provided encouragement on the project throughout its various phases. We would also like to thank Doris Graber who has been aware of this project from the very beginning and has offered valuable insight and helpful criticism at various points along the way. She continues to be an inspiration and a mentor. The survey was conducted by GfK Denmark. Erik Christiansen at GfK could not have been more helpful and we are thankful for his friendship and support. The content analysis was competently and meticulously conducted by four native Danes in Amsterdam: Kristina Rohde Larsen, Joachim Aaen, Silja Marit Zetterqvist, and Rikke Friis. Today these individuals are ex-Amsterdam students and dispersed across the globe, currently in Denmark, Switzerland, and India. At the Danish Broadcasting Association (DR), Lotte Lindegaard and Erik Nordahl Svendsen from the Audience Research Department and Lisbeth Knudsen and Soeren Knudsen from the News Department were extremely helpful in obtaining access to the newsroom. We are also especially grateful for the assistance of Kaare R. Schou and Mikael Dyrby at the competing television network (TV2). We would also like to express our gratitude to all those who agreed to be interviewed for this project. This includes the panel survey respondents and focus group participants who remain anonymous, as well as the campaign managers and the journalists who gave generously of their time to answer an academic’s questions during an unusually busy time in the newsrooms. We presented different parts of this work to colleagues at several conferences including those of the American Political Science Association, the European Consortium for Political Research, the International Communication Association, and the World Association for Public Opinion Research. Many colleagues provided valuable feedback at these meetings and in personal conversations with the authors, including Kees Aarts, Susan Banducci, Sam Barnes, Shaun Bowler, Kees Brants, Wouter van der Brug, Russ Dalton, Wolfgang Donsbach, Cees van der Eijk, Bob Entman, Esteban Escobar-Lopez, Doris Graber, Ron Inglehart, Jeff Karp, Jan Kleinnijenhuis, Robert Luskin, Juan Diez Nicolas, Thomas Patterson, Jacques Thomassen, David Swanson, and David Weaver. We would also like to thank the anonymous reviewers of the book manuscript for their helpful suggestions. We are of course entirely responsible for any shortcomings or omissions. Heidi Bagtazo and Grace McInnes at Routledge have been very helpful in the process of bringing this manuscript to publication, and Andreas Schuck has done meticulous work on the index and appendices.

xii

Preface

Some of the arguments as well as a number of the analyses presented in the chapters in this book have been included in journal articles. These include “Primed by the Euro,” Scandinavian Political Studies, 27, 1, 45–65, “News Matters: Inﬂuences on the vote in a referendum campaign,” European Journal of Political Research, 43, 5, “Cynical and engaged: strategic campaign coverage, public opinion and mobilization in a referendum,” Communication Research, 29 (6), 615–641, and “Public perception of polls and support for restrictions on the publication of polls: Denmark’s 2000 euro referendum,” International Journal of Public Opinion Research, 14, 410–433. We acknowledge these ﬁrst publications and would like to thank the publishers of these journals for allowing us to include some of the material in this book. We would also like to thank the anonymous reviewers at those journals for their constructive criticisms. Holli Semetko moved from Amsterdam to Atlanta in the summer of 2003 to serve as Vice Provost for International Affairs and Director of the Claus M. Halle Institute for Global Learning, and Professor of Political Science at Emory University. She is grateful for the research support provided by Emory University and for the input she received from many Emory scholars interested in European politics and European studies. She also appreciates the research assistance of Hajo Boomgaarden on related projects that made it possible for her to ﬁnish this project. The move took place during the ﬁnal months of completing the manuscript. Her work on the book during this time of transition was enriched by the culinary skills of her husband Sam Cherribi and the inspiring cuddles and thoughtful questions from their daughters Miriam Isabelle and Sophia Elisa. Claes H. de Vreese Holli A. Semetko January 2004

Part I

The importance of information in referendums

1

The dynamics of a referendum campaign

Referendums and citizen initiatives are an increasingly important means of enacting or preventing legislation in countries around the world. In the United States, where there is no provision for national referendums, citizen initiatives are common practice in dozens of states and the number of times citizens have been confronted with voting on speciﬁc issues in state or local contests has increased considerably over the past two decades (Bowler and Donovan 1998; Bowler et al. 1998; Butler and Ranney 1994; Scarrow 2001). In Canada, national referendums were held on issues of major constitutional importance in the 1990s (Clarke and Kornberg 1994; Clarke et al. 2000; Johnston et al. 1996). And in Europe, national referendums continue to play an especially important role in the process towards advanced European integration (Hug and Sciarini 2000; Hug 2003). In contrast to a general election campaign where political parties provide relatively clear-cut information cues for voters, in referendums the information cues from political parties are often ambiguous. For example, parties may be internally divided over the referendum issue as has been the case in Britain on both the question of membership of the European community and on joining the euro. Or to take another example, political parties from opposite sides of the ideological spectrum may form unusual coalitions in referendums. In other words, parties that normally oppose one another in general elections may suddenly be on the same side of the issue, as in the 1994 Nordic referendums on EU membership. Referendums may also give rise to new parties or movements that come into existence speciﬁcally to take a stand on the issue. Despite the apparent simplistic nature of the referendum vote (Yes or No), the referendum issue is often multi-faceted and different aspects may trigger diverse perceptions of the issue among voters. The arguments, aspects, or framing of the issue by political actors and key information sources during a referendum campaign may therefore be unusually important to a referendum outcome (Lupia 1992, 1994; Lupia and McCubbins 1998). Previous studies have documented signiﬁcant swings in public opinion over the course of a referendum campaign (LeDuc 2002). While

4

The importance of information in referendums

longer-term factors such as partisanship or ideology have been found to be important in national elections, the short-term impact of campaign strategies and tactics can make a substantial difference in determining referendum outcomes. Perceptions of the referendum question on the part of the voters, the images they hold of groups and individuals involved, and their reactions to the discourse of the campaign can be as important to the voting decision as their opinions and beliefs on the issue itself. Contrary to regular elections, no candidates or parties appear on the ballot, thus voters must decide among alternatives that may be unfamiliar. One may therefore expect a greater degree of volatility and uncertainty in vote preferences in a referendum campaign compared to a general election campaign. This book is about the dynamics of a referendum campaign. We focus on how information, often provided by the media, can help to crystallize opinions and alter perceptions of political issues and candidates under the speciﬁc conditions of a referendum campaign. Given that referendums are held on a speciﬁc issue, the learning process about the issue is critical for voters’ understanding of the options. For well-known issues, this process of learning may be less important given that voters draw on previous knowledge and experiences, so their vote may be mostly driven by prior predispositions or well-established attitudes. For less salient issues, such as foreign policy or issues pertaining to European integration, voters may be more dependent on information that they obtain during the course of a campaign.

Referendums and the integration of Europe Between 1972 and 2003, there were no less than 28 national referendums in European countries on issues of European integration. This makes the integration of Europe the most voted-on issue in the world. Early national referendums in different European countries have determined membership in the European Community, such as Denmark’s 1972 referendum on membership and Britain’s 1975 referendum on remaining in the European Community (King 1977). Ratiﬁcation of the Maastricht Treaty was the subject of referendums in France (Appleton 1992), Denmark (Siune and Svensson 1993), and Ireland (Sinnott 1995). The question of membership was the focus of referendums in 1994 in Austria, Finland, Norway, and Sweden (Jenssen et al. 1998). The ratiﬁcation of the Amsterdam Treaty was put to a referendum in Ireland and Denmark in 1998, and adoption of the euro was the focus of a referendum in Denmark in 2000. In 2001, the provisions of the Treaty of Nice were subject to ratiﬁcation by national referendum in Ireland, and were rejected by a 54 percent majority and a low 33 percent turnout. In 2002, the Irish ratiﬁed the Nice Treaty in a second national referendum after more active and engaged campaigning by the incumbent government and the Yes camp. The enlargement of the European Union (EU) makes referendums

The dynamics of a referendum campaign

5

Table 1.1 National referendums on European integration issues, 1972–2003 Year

Country

Issue

Outcome

1972 1972 1972 1975 1986 1987 1992 1992 1992 1993 1994 1994 1994 1994 1998 2000 2001 2002 2003 2003 2003 2003 2003 2003 2003 2003 2003 2003

Denmark Ireland Norway Great Britain Denmark Ireland Denmark France Switzerland Denmark Norway Austria Finland Sweden Denmark Denmark Ireland Ireland Sweden Lithuania Latvia Estonia Poland Czech Rep Slovakia Slovenia Cyprus Hungary

EC membership EC membership EC membership Continuation of EC membership Single European Act Single European Act Maastricht Treaty Maastricht Treaty EEA Treaty Edinburgh Treaty EU membership EU membership EU membership EU membership Amsterdam Treaty EMU/euro Nice Treaty Nice Treaty (amended) EMU/euro EU membership EU membership EU membership EU membership EU membership EU membership EU membership EU membership EU membership

Yes Yes No Yes Yes Yes No Yes No Yes No Yes Yes Yes Yes No No Yes No Yes Yes Yes Yes Yes Yes Yes Yes Yes

even more common as 2003 illustrated. Nine of the ten ﬁrst-round applicant countries to the EU have held referendums determining the country’s position on membership. Malta, Slovenia, Estonia, Slovakia, the Czech Republic, Latvia, Hungary, Lithuania, and Poland all voted in favor of membership in valid national referendums. Several of the countries have a minimum turnout requirement of 50 percent and all countries, although some barely, met the required turnout level. As a result, campaigns in these countries had two primary aims: ﬁrst to convince the electorate to vote for membership and second to mobilize the electorate. The single European currency, the euro, has been the topic of previous referendums. Sweden held a referendum on the common currency in September 2003 and voted No to the euro. This campaign will be remembered in particular for the assassination of the Minister of Foreign Affairs, Anna Lindh. In June 2003, the Blair-led government in the UK decided, based on the advice of the Chancellor of the Exchequer Gordon Brown, to postpone the decision on a referendum on the euro. The question will be

6

The importance of information in referendums

revisited in 2004, and it is expected that there will be a referendum in the future. During the summer of 2003 the draft version of a new EU constitution was presented by the European Constitutional Convention, chaired by Valerie Giscard d’Estaing. This constitution was discussed during an ongoing Intergovernmental Conference in the autumn of 2003 and occupied center stage of negotiations during Italy’s six month EU presidency in the second half of 2003. By December 2003 the negotiations reached a critical phase and the completion negotiations were postponed. A number of countries debated whether to hold national referendums to ratify the constitution. Parties in Britain, France, the Netherlands, Portugal, and Spain pushed in favor of holding national referendums on the EU constitution. Denmark and Ireland are required by their constitutions to hold a binding referendum. The issue of European integration is a key issue for direct democracy. Future EU political decisions are contingent upon national referendums and to the extent that the results of the referendums are not forgone conclusions, it is important to understand the dynamics of referendum campaigns. However, while previous studies of referendum campaigns have acknowledged the importance of information about the issue, they have paid little or no attention to studying in a systematic fashion the information available in the campaign and modeling its effects on public opinion and the vote (Farrell and Schmitt-Beck 2002). Our study takes the contents, uses and effects of information in national referendum campaigns as a focal point. We begin with a discussion of different perspectives on citizens’ perceptions of politics, political attitudes and behaviors and propose a general framework for considering the role of information in referendum campaigns.

Public opinion, political attitudes, and information With the publication of Converse’s (1964) seminal article “The Nature of Belief Systems in Mass Publics,” the ﬂuctuation and apparent inconsistency of public opinion on policy matters became the subject of considerable research and scholarly debate. Converse (1964) identiﬁed weak relationships between individual citizens’ responses to identical policy questions when the same persons were interviewed in a three-wave panel study spanning several years between 1956 and 1960. As there were no intervening events that would have led to the observed changes in the opinions, he concluded that the belief systems of mass publics were far away from those of political elites, which were more coherent and consistent and organized along a liberal–conservative continuum. In a later article, Converse (1970) dubbed this ﬂuctuation of opinion phenomenon as mass publics having “non-attitudes.” This perspective was challenged when reanalysis of the panel data

The dynamics of a referendum campaign

7

showed that measurement error that had not been accounted for, and that once it was taken into account, citizens’ opinions appear to be more stable and consistent than Converse had claimed (Achen 1975). Subsequent research has pointed to the importance of question wording, and suggests measurement error can be reduced by using questions that contain more information or opposing viewpoints (Zaller and Feldman 1992). In contrast to Converse (1970), Page and Shapiro (1992) argue that a typical individual’s preferences would be located somewhere in the middle of a continuum ranging from policy preferences that may be described as perfectly stable, informed, and consistent to preferences that are entirely random and uninformed. They argue that information plays an important role in the opinion formation process and note: if citizens’ preferences are dependent upon uncertain beliefs, bolstered by incomplete bits of information – then new information or arguments that bear upon beliefs about policy alternatives can change people’s policy preferences. (Page and Shapiro 1992: 15) This perspective adds up to the accuracy of collective preferences, deﬁned as long-term preferences of individuals, as measured by surveys or polls. The measurement of collective opinion is for the most part free of “random error associated with individual attitudes,” and that “even if individual opinions or survey responses are ill-informed, shallow, and ﬂuctuating, collective opinion can be real, highly stable . . . and it can be measured with considerable accuracy” (Page and Shapiro 1992: 16–17). In an analysis of more than thirty years of surveys of public opinion in the United States, three explanations are offered for why collective opinion is solid while the measured opinions of many or most individuals seem to be shaky. First, Page and Shapiro argue that random measurement errors cancel out when drawing on large numbers of respondents. Therefore surveys provide fairly accurate information about the collectivity. Second, “temporary opinion changes by different individuals occur in offsetting directions, so that they, too, cancel out and allow collective measurements to reﬂect the more enduring tendencies of opinion” (Page and Shapiro 1992: 384). Third, “processes of collective deliberation, and reliance upon trusted cue givers, enable people to arrive at reasonable policy preferences without an extensive informational base” (p. 385). Assuming that public opinion, at least at the collective or aggregate level, can be measured in a reliable and valid manner, albeit with random measurement error, the interesting empirical question concerns the factors that shape developments in public opinion. In the context of an electoral contest, the media are key players. Dalton (2002: 39) concludes that “campaigns are now largely media events.” The centrality of the media is also highlighted in Page and Shapiro’s (1992: 386) further argument in which

8

The importance of information in referendums

they conclude that short-term movements in public opinion can “largely be accounted for (and predicted) by quantitative analyses of what news, from what sources, appears in the mass media.” In other words, information and campaigns matter. The media and the importance of mediated information could be particularly obvious in the case of low-involvement issues and under the condition of mixed or conﬂicting opinions among elites.

Information and cues The degree of consensus or division between elites is a key point in Zaller’s (1992) work. This is especially relevant to referendum campaigns in which the debate is not only between two camps, but often within each camp. Given the news value of conﬂict, despite what may appear to be elite consensus in terms of the sheer numbers of political spokespersons sharing the same view on an issue, one or two dissenting voices may become magniﬁed in campaign communication. The division of opinion between elites may become the predominant focus of campaign information simply because of this. The degree of consensus or division between elites can result in very different campaigns and can also affect the ways in which citizens vote on referendum issues (Butzer and Marquis 2002). Lupia (1992) addresses the question of how information affects electoral outcomes in the environment of a direct legislation election. His game theoretic model assumes that information cues are perfectly credible, whereas in reality, the multiple information cues that exist in referendums are often far from being perfectly credible. In the context of direct legislation elections (the initiative and the referendum), he later found that even those voters with low levels of “encyclopedic information” about the proposal may vote in such a way that emulates highly informed voters, by making use of information shortcuts (Lupia 1994). In the case of the California insurance reform elections, for example, these shortcuts simply involved learning about the industry’s ofﬁcial position on the issue. Referendums are held in contexts of information complexity and uncertainty. The dynamics of the campaign may vary, depending on the issue of the referendum. LeDuc (2002) identiﬁes three types of referendum campaigns: 1

2 3

opinion formation campaigns in which voters are not expected to have fully formed opinions on the issue and where opinions form as the campaign progresses; opinion reversal campaigns in which reasonably well-known issues take new directions over the course of the campaign; uphill struggle campaigns in which opinion is much ﬁrmer and less subject to rapid change or sudden reversal.

The dynamics of a referendum campaign

9

Voters have strong cues based on partisanship, ideology, and are receptive to arguments by credible and familiar political leaders. In all three types of referendum campaigns, if elites are not divided along the customary left–right dimension, this can bolster the ambiguity. Moreover, in referendums campaigns, if citizen attitudes appear to be rather unstable, new information matters. This type of information typically comes through the mass media (Bennett and Entman 2001; Dalton 2002).

Campaign effects Given the key characteristics of referendums – volatile electorates, uncertainty in elite cues, and issue complexity – campaigns may matter more than in other electoral situations. The effects of a campaign are conditional upon the overall volatility present in the political system and the extent to which voting in the referendum is aligned with party politics. While voters are believed to be rational, it is often not obvious what exactly their motives are. In general elections vote choice may be inﬂuenced by habit, class tradition, religion, ideology, speciﬁc issues, and/or political personalities. All these inﬂuential factors are mediated by information sources such as television, the press, and the Internet, as well as through interpersonal discussions. Any or all of these information sources may help to reinforce already existing views and perceptions or alter and change these. In contrast to a general election, where a considerable, though shrinking, size of the electorate is guided by ideology, partisanship, and tradition (Dalton 2000), a referendum is an entirely different political enterprise. In a referendum campaign voters may not even know the standpoint of their own parties and candidates. Once the voter is aware of the policy recommendation, the question remains whether or not this is in line with the views of the voter. While the volatility of the electorate has increased over the past decades so that more voters today shift party preference from one election to another (Klingeman and Fuchs 1995), a referendum is an opportunity to vote against the recommendation of one’s own party without punishing them in a direct sense, since the outcome of a referendum does not determine, at least in the short run, whether or not politicians are entitled to (remain in) power. This phenomenon has been labeled “the punishment trap” (Schneider and Weitsman 1996). Given the ambiguity, the media therefore play a key role in referendum campaigns. Despite this, previous studies of referendums tell us little about the role of the campaign. The campaign is discussed either in terms of bivariate relationship, such as between media exposure and vote intention (Borg and Esaiasson 1998) or in broad and general, descriptive terms (Clarke et al. 2000; Siune and Svensson 1993). In these studies, the campaign and information variables have not been integrated into a model of campaign effects on the vote.

10

The importance of information in referendums

In this book we distinguish two levels of effects: the individual and the institutional. Campaigns may increase voters’ awareness of an election and contribute to the interest and knowledge about politics (Graber 2001). The communication of a campaign may have speciﬁc effects on cognitive learning, political mobilization, voters’ issue agendas, and ultimately their vote choice. Given what we know about the effects of the media, we can assume that referendum campaigns may also contribute to either engagement or apathy about politics, and that the evaluation of candidates and elections may improve or worsen during a campaign. At the institutional level we discuss the effects that a referendum has on the political system, including the political parties and candidates, and on the media system.

Campaign effects on political institutions Political parties and movements that are mobilized to contest in a referendum campaign are affected institutionally, both from a short-term and in a long-term perspective. Parties are confronted with the need to raise money for the campaign and to raise the party proﬁle. Over the past decades, political campaigning has become increasingly professionalized and political parties are challenged to respond to the strategies of competitors as well as increasingly professional and critical journalists. In the case of national referendums on European issues, this mobilization and campaign activity is organized along traditional national party political lines. After more than 20 years of political contests for the European Parliament, a genuine European party system is still to emerge (Hix 1999). Consequently, campaigning in referendums on European issues, even issues that may be put to a popular vote in more countries, remains largely a national political enterprise. Despite the absence of a European-level party system, a national referendum on an issue of European integration affects the national political system and the established political parties considerably. In several referendums on European issues the traditional left–right political continuum has little meaning because center parties tend to be pro-European integration and (semi-)extremist parties, both at the left and the right end of the continuum, tend to be anti-European integration. The reluctance of the ideologically strongly proﬁled parties towards advanced integration stems from reasons such as loss of national sovereignty, undermining of national welfare state facilities as well as nationalistic considerations about immigration and integration. In planning a national referendum campaign, political parties are faced internally with strategic choices about who will be visible in the campaign. Externally, political parties are faced with the potential strategic alliances within the Yes or No camps. These decisions may not be fully under the control of the political parties or the candidates. Political parties often appoint key campaigners, but in many cases these responsibilities become

The dynamics of a referendum campaign

11

the Prime Minister’s or another incumbent’s, who may also come with negative (or positive) connotations. In a country in which there is a turnout threshold that must be reached in order for the referendum result to be valid, some parties may aim strategically to mobilize while others aim simply to demobilize voters. A continuing source of uncertainty is the lack of knowledge about what opposing parties are doing. In national politics, parties are often described as “owning” certain issues, and although this may change over time there is predictability in liberals arguing for tax reduction, privatization, and liberalization, while social democrats often campaign more on welfare state values. Issue ownership in referendums, however, is by no means clear. In a referendum campaign, parties are challenged to formulate clear campaign messages and mobilize what is often an internally divided party to stay “on message” during the campaign. A primary strategic goal is to reduce the ambiguity in cues given by a party to its voters, because voters are susceptible to consider other political options when elite cues are unclear or ambiguous (Zaller 1992). Whereas general election campaigns happen more or less on a regular basis in relatively predictable cycles, national referendums are largely unscheduled and the campaigns can run longer. LeDuc (2002) demonstrated that in referendum campaigns attitudes can be shaped by unforeseen developments or events during the campaign period. The longer the campaign, the greater the chance for unforeseen events. In a longer-term perspective, political parties are also challenged to position themselves strategically in relation to the framing of the issues related to the general thrust of a referendum. A referendum, especially in the case of a No outcome, is not the end of the road for the debate about policy on this and related issues. The parties on the winning side of the referendum also have a strategic opportunity to control the framing of the issue in the future. Political parties and candidates are therefore forced to consider the domestic political implications of a referendum. A referendum places a topic prominently on the political agenda and attitudes towards political leaders may be shaped more by the referendum issue than other issues during and in the aftermath of a referendum campaign. This implies that a successful campaign can bolster the popularity of political leaders while a campaign perceived to be less successful has the potential to damage the domestic political reputation of a party or politician. This situation also applies to citizens who may (or may not) experience reinforcement of partisan predispositions so that the strength and status of some parties is enhanced during a campaign by stronger bonds with supporters while the market positions of other parties are likely to weaken. A consequence of a referendum campaign could ultimately be in some form of punishment for incumbents were they to lose a referendum. Or, if they win, they may claim this is a reﬂection of their own competence or administrative skills.

12

The importance of information in referendums

A national referendum campaign may also have implications for more general levels of trust in political institutions and leaders. In the context of multi-level governance, we may experience spillover effects from domestic levels of governance to the European level and vice versa. For example, political cynicism or negativity towards European integration may spillover as general cynicism towards domestic politics. This is an unexplored dynamic in the literature. Previous research, for example, suggests that national issues and satisfaction with national democracy drive and predict the satisfaction with democracy at the EU level. A national referendum could help to push things in the reverse direction so that perceptions of European issues affect ideas about national systems. Such a benchmarking with the European level may lead the national system to look favorable, if it is seen to operate better and more efﬁciently than the European level, but it may also affect the perception of the national system in a negative way, if the European level of governance is seen as a more positive example. In conclusion, a national referendum brings a number of possible effects for political parties and leaders. A number of the implications emerge as short-term considerations in the form of strategic choices for campaign leaders, on issue spinning, and who is to join as coalition partners in the campaign. However, a number of issues also have longer-term implications that go beyond the immediacy of the campaign. A referendum may alter the domestic political power equilibrium and it may lead to changes in the approval ratings of parties and political leaders. In addition, referendums on European issues may of course also have implications for the future trajectory of European integration.

Campaign effects on the media system A national referendum campaign leads media organizations to prepare in ways that are in many respects similar to their preparation for other electoral contests, but there are also a number of speciﬁc characteristics to media planning in a referendum. In terms of logistics, campaign coverage is the result of internal cooperation in a news organization, most often between members of the central headquarters-based newsroom and the political unit. In Britain, for example, this means that central newsrooms of the press and broadcasting cooperate with their Parliamentary Westminster Unit (Blumler and Gurevitch 2001). In a referendum campaign on a European issue, the logistic challenges are magniﬁed by an additional component in the form of the Brussels bureau (de Vreese 2003). Prior to a campaign, news organizations typically prepare using background research that results in canned items for television and feature articles in the press. These preparations are made in anticipation of certain issues that are likely to appear in the campaign and when they do, background information for a story as well as potential interviewees and key

The dynamics of a referendum campaign

13

facts are already on ﬁle to make it possible to report in a timely manner. During campaigns, news programs and newspapers make use of format alterations. Newspapers typically devote speciﬁc pages or sections of the paper to follow the campaign. In television news, a frequently used “tool” is the use of a political editor to chair campaign segments. This type of additional anchor often takes the form of an analysis of the day’s events on the campaign trail. In terms of content, election time implies considerations of balance and fairness in political reporting (Semetko 2003). A national referendum can challenge existing conceptions of balanced news reporting. While public broadcasters in Europe tend to interpret balance in the news in proportion to a party’s power in Parliament, a referendum potentially undermines this principle as other (extra-parliamentary) actors join the race. Moreover, as the vote in a referendum is neither party nor candidate based, the balance issue becomes a question of hearing from the Yes and No camps. This can have the effect of a magnifying glass being placed on smaller parties or individuals on either side of the issue. Journalists are additionally challenged at the outset of a referendum campaign not least because, like the voters, all the actors involved may not be clear on the parameters of the two camps. This makes a referendum campaign potentially more unpredictable. There are a number of key issues at stake for news organizations at election time. Not only do most European broadcasters have a legal commitment to provide fair and balanced news coverage. The news business at election time is also concerned about public perceptions of media credibility, public trust in and satisfaction with the news business. News organizations therefore invest strategically in the campaign coverage, and this may have implications for editorial decisions and for playing-up certain aspects or themes of the campaign. Finally, the partisan sympathies of media owners, or the connections between media ownership and political parties or individual politicians is another aspect that is relevant to our understanding of what may be inﬂuencing the ways in which the campaign is being presented in the news.

Campaign effects on citizens At the individual level, two types of campaign effects may be distinguished: effects of direct and personal campaign experiences and effects of indirect and mediated campaign experiences. Direct and personal campaign experiences include attending rallies and the extent to which voters engage in discussions with family, friends, or colleagues about politics and the issue of the referendum. Previous studies of referendum campaigns suggest that 1

direct/personal campaign engagement and involvement may affect turnout positively but does not give any particular direction to the vote,

14

The importance of information in referendums

2

that No voters rely more on information from personal communication channels than Yes voters, that ﬁrm voters (with a clear Yes or No preference) tend to engage more directly in the campaign by, for example, canvassing and attending a political meeting on the EU issue. (Jenssen et al. 1998: 94–99)

3

Indirect and mediated experiences with politics and campaigning are the most common since only a minority of the electorate is directly involved in a campaign. The vast majority of an electorate only hears about politics and election campaigns through the media and the content forms the backdrop against which political leaders, institutions, and issues are evaluated. Given the centrality of media in campaigns, common sense suggests that the media are bound to have effects on the electorate. In terms of mediated campaign experiences, previous research suggests that exposure to news coverage may increase the level of awareness and knowledge about politics and election themes (McLeod et al. 2002). However, national election studies have often either neglected the role of information effects (see Bartels 1993) or provided either weak or mixed empirical support for media exposure and attention effects on turnout and vote choice (Zaller 2002). As discussed above, in a referendum campaign, the media have a potentially greater role to play than in a general election campaign. The media are essential not only in getting the messages of the political actors across, but also as independent actors setting their own agenda and putting their own spins and deﬁnitions on the referendum issue. The vast majority of citizens get their information about politics via the media (Bennett and Entman 2001; Dalton 2002; Graber 2001). Television and newspapers are – even with the increasing informational role of the Internet – consistently identiﬁed by a vast majority of Europeans as their most important sources of information (Eurobarometer 51–59). In Canada, Norway, Sweden, Finland, and Denmark, respondents named television as the most important source of information during previous referendum campaigns (Clarke et al. 2000; Jenssen et al. 1998). Therefore, the media are potentially very important for providing cues used by citizens in subsequent considerations. Individual-level effects of media can be observed on the processing of information, on perceptions of the political system, on political affections and attitudes as well as on behavioral intentions or real behavior. Table 1.2 provides a typology of information and campaign effects. In this study, we focus on effects of the campaign on issue interpretation, on evaluation of political leaders and the government, on political cynicism, and on political participation in the form of vote choice. Each of these effects refers to concepts that have been studied previously and most frequently in the context of national parliamentary or legislative elections, but not in the

The dynamics of a referendum campaign

15

Table 1.2 Typology of information and campaign effects Cognitive and Perceptions of Political opinion changes political system participation Issue salience and framing (“agenda-setting” and “framing”) ✗

–

–

Evaluations of political leaders (“priming”)

✗

–

–

Political cynicism (“spiral of cynicism”)

–

✗

–

Turnout (“mobilization”)

–

–

✗

Vote choice

–

–

✗

speciﬁc situation of a referendum. We brieﬂy articulate the theoretical perspectives that inform our typology of information and campaign effects. Agenda-setting and framing in the context of a referendum One of the most inﬂuential ideas about campaign effects deals with the media’s ability to inﬂuence the public salience of issues, referred to as the agenda-setting function of the media (see McCombs 2004 for a review of this research). Agenda-setting is a heterogeneous ﬁeld of research originating in Lippmann’s (1922) notion of how the media shape our pictures of the world. Trenaman and McQuail concluded that “the evidence strongly suggests that people think about what they are told” (1961: 178). McCombs and Shaw (1972) coined the term agendasetting in their seminal study of the 1968 U.S. presidential election campaign in which a strong similarity was found between the most prominent issues in the news and undecided voters’ ranking of the most important political issues. Previous studies of agenda-setting have often been conducted in the context of general election campaigns (see, for example, Weaver et al. 1981), but the dynamics of agenda-setting in the context of a referendum has been neglected. A national referendum campaign is distinctively different from a general election in that the broader theme of the campaign has been deﬁned a priori. General elections may be waged over multiple issues such as the economy, health care, or national security, but in a referendum campaign the overall issue is deﬁned well in advance of the campaign. The agenda-setting dynamics may therefore be different if, and when, sub-issues emerge within the broader issue. The news media play a key role in deﬁning and emphasizing or neglecting certain aspects or subissues as they are put forward by key campaign actors.

16

The importance of information in referendums

Issues are at the core of agenda-setting, but what an issue is continues to be an open question (Lang and Lang 1981). An important empirical question is whether an “issue” on the media agenda has the same meaning as that “issue” on the public agenda. “The economy,” for example, may mean “taxes” to one and “fuel prices” to another. A referendum campaign provides an opportunity to explore a particular issue in more depth and investigate the multiple dimensions and sub-issues on the media and the public agenda. Our study discusses what a referendum issue means to the news media, the political parties and to the public. A spiral of cynicism? The effects of a referendum campaign on political cynicism Longitudinal studies of news media coverage of U.S. elections and politics document a change in the balance of news coverage from issue-based stories to strategic coverage. The latter emphasizes who is ahead and behind and the strategies and tactics of campaigning necessary to position a candidate to obtain and remain in a lead position (Jamieson 1992; Patterson 1993). Cappella and Jamieson (1997) argue that this mode of coverage invites the attribution of cynical motives to political actors in campaigns and public policy debates. By “reporting about politicians and their policies repeatedly framed as self-interest and seldom in terms of the common good – whether such characterizations are correct or incorrect – the public’s experience of their leaders is biased toward attributions that induce mistrust” (Cappella and Jamieson 1997: 142). Previous research on the effects of strategic news on political cynicism has been conducted in the United States and not in the context of a referendum. The evidence is primarily based on experiments and has not investigated cynicism and its development over the course of a campaign. We investigate how strategic news affects an electorate in a referendum campaign and how it contributes to campaign evaluations, political cynicism, and voter mobilization. The effects of a referendum campaign on public evaluations of political leaders The psychological process of priming is well known from research in cognitive psychology (Fiske and Taylor 1991). The premise of priming is that new information renders something – an issue or a trait – applicable for use in subsequent evaluations. Priming is the “ability of new information to alter the standards by which the public evaluates political leaders” (Krosnick and Kinder 1990). Priming theory posits that public evaluations of political leaders are made on the basis of how leaders perform on issues that citizens have in mind when formulating the evaluation. The connection with agenda-setting is that the issues that are emphasized in the news

The dynamics of a referendum campaign

17

may also be the ones that citizens have in mind when formulating evaluations of political leaders. An important yet unanswered question is whether a referendum campaign matters to the public’s evaluations of political leaders and the incumbent government. We know that information and cues that are most recently stored and are most accessible to people can function as “decision-makers” (see, for example, Krosnick and Brannon 1993). When evaluating the performance of a government or a political leader, therefore, citizens are likely to attach greater value to the information they have received most recently and in the greatest quantity. Citizens are, in other words, primed to use information about speciﬁc issues in general evaluations. News media do not merely bring certain issues to the public attention. They also offer evaluations of issues and persons. These evaluations in the news may in turn affect public evaluations of political leaders. McCombs et al. (2000), for example, showed that evaluations of political leaders in the news affect public perceptions of these leaders, and Herr (2002) showed that evaluations of candidates in the news may even affect vote choice. Given that news, generally speaking, tends to be either neutral or negative (Kepplinger 1998, 2002), we may hypothesize that strongly negative news about an issue that is also highly visible in the news could lead to audiences using that issue to evaluate the performance of political leaders and thus bolster negative evaluations of those leaders. Favorable news about an issue, on the other hand, that is highly visible, may prime audiences to evaluate more favorably the politicians connected with that issue in the news, as was found in a study designed to measure priming effects on public opinion about European heads of state attending a major EU Summit; a summit that was not marred by serious political disagreements (Semetko et al. 2003a). We address the visibility of a referendum in the news, and we also focus on assessing the tone of evaluations of political leaders in the news. In this way, we aim to better understand the dynamic of public evaluations of political leaders as a function of exposure to speciﬁc news content. The effects of a referendum campaign on the vote Most research on electoral behavior has been conducted in the context of presidential or parliamentary, national, and state-level legislative elections. Comparatively fewer studies have been carried out in the context of (national) referendum campaigns and national, state, or local citizen initiatives, even though this form of direct democracy is becoming increasingly common (Bowler and Donovan 1998; Bowler et al. 1998; Butler and Ranney 1994). Previous studies offer inconsistent explanations for the outcome of national referendums. Theories of voting behavior generally emphasize

18

The importance of information in referendums

the importance of party support or attachment (Budge and Farlie 1983; Campbell et al. 1960), as well as contextual factors such as economic perceptions (Lewis-Beck 1997), perceptions of issues (Nie et al. 1979) and evaluations of leaders or the top candidates (Wattenburg 1992). Some of the hypotheses about what drives party or candidate choice are also common in the literature on referendum voting. We identiﬁed a number of hypotheses about inﬂuences on referendum voting in a review of the referendum and citizen initiative literature. Among the explanations for support for a referendum measure are: ideology or party support, evaluations of or feelings about political elites, government performance or government approval, economic evaluations, issue-related contextual evaluations, political cynicism, political efﬁcacy, political knowledge and political interest, and demographic and regional characteristics. In studies of national referendums on European issues, special emphasis has been put on government approval and issue-related opinions. One perspective suggests that a national referendum is in fact also a referendum on the performance of the incumbent government, and argues that government performance ratings at the time of the referendum drives the referendum vote (see, for example, Clarke et al. 2000; Franklin et al. 1994; Franklin et al. 1995). Put simply: “While a popular government might expect to see its referendum proposals approved, an unpopular government will often see its proposals turned down” (Franklin et al. 1994: 106). This perspective is related to the literature on voting behavior in European Parliamentary elections. These elections are often referred to as “second-order national elections” in which domestic political considerations drive the vote in the European elections (Reif and Schmitt 1980; van der Eijk and Franklin 1996). Another perspective in the literature suggests that attitudes towards European integration issues are quite independent from domestic politics and that these attitudes drive the referendum vote, above and beyond the incumbent government’s popularity (see, for example, Siune and Svensson 1993; Siune et al. 1994). It is argued that voters have attitudes towards different areas of joint European policies and that these attitudes are consistent (Siune et al. 1994; Svensson 2002). In our model of referendum voting, we investigate not only these two hypotheses, but also the inﬂuence of ideology, economic expectations, and political disaffection. In addition we pay particular attention to the impact of the campaign, and the campaign coverage in the news sources to which the respondents were exposed, on the vote. Given the potentially volatile electorate, events and the news and information environment during a campaign can play a crucial role, also for voting behavior.

The dynamics of a referendum campaign

19

Summary of the book Much of the theory about information and campaign effects in general elections stems from research conducted in the United States, and there is a tendency to generalize from this to very different national and political contexts. Much of the national referendum research, on the other hand, is so country-speciﬁc or issue-speciﬁc that it can become difﬁcult to see its broader implications. In this book, we propose a more general framework for understanding the dynamics of a referendum campaign, and discuss the implications of our study for referendum and campaign research. We take the example of a national referendum on the introduction of the common European currency, the euro, as an illustrative and signiﬁcant case to discuss referendums on issues of European integration in more detail. Our study clearly shows that the contents and uses of media sources of information in the short term are important for the referendum outcome, above and beyond longer-term predispositions that would be expected to inﬂuence the vote. We also ﬁnd that public understanding of the referendum issue, citizens’ evaluations of political leaders and the government, and public perceptions of the campaign are all inﬂuenced by the use of media sources, and the volume and tone of that information. This book has ten chapters. In Chapter 2, we discuss the context of a referendum campaign and the transformations that often occur in the party and media systems. Key features of European multi-party systems and media markets are discussed. We also introduce our case study in this chapter: The Danish 2000 national referendum on the single European currency, the euro. The more speciﬁc context of the Danish case is discussed here, with an emphasis on the development of Danish attitudes towards European integration in comparison with other European citizens, and the transformation of the Danish party system and media system in comparison with other European countries that have held or intend to hold referendums on European Union issues. We also outline the research design and describe the key empirical components of the study. In Chapter 3 we review the literature on the development of political campaigning and political marketing in the United States and Europe. To discuss the problems confronted by opposing camps in a referendum, the chapter draws upon interviews with campaign managers, politicians, and journalists as well as an analysis of political advertising in Yes and No camps. The issue agendas of the two camps, as well as the individual parties, are discussed. We develop a typology of the effectiveness of the campaigns of both camps and apply it to our country case study. The case study is used as a baseline to illustrate cross-national variation in referendum campaigning styles. Chapter 4 analyzes the referendum in the news. Referendum campaigns vary in their intensity. Our case study draws upon more than 5,000

20

The importance of information in referendums

television news stories and 3,000 newspaper stories during the campaign to reveal a major increase in the amount of coverage of the referendum in the ﬁnal four weeks, a bias towards strategic and opinion poll-oriented news coverage, and neutral or negative evaluations of key campaign actors. We also analyze the way journalists covered the referendum and consider a number of constraints and challenges that shaped the news coverage. In Chapter 5 we investigate the elite framing of the referendum. We ﬁrst review the literature on elite framing and framing effects, and identify the way in which a Yes/No referendum issue becomes multi-faceted and complex. Our case study shows how the simple Yes/No issue of the euro was framed by each of the political parties and movements within the two camps. Broadly speaking, the Yes camp framed the euro issue as an economic one, whereas the No camp framed it as a political one that challenged sovereignty, the welfare state, and national identity. The individual variations and internal contradictions of these general frames are discussed in the context of how the news media responded to the different and ambiguous spins put forward by the parties. We argue that the frames used by the two camps in our case study are relevant to the position taken by parties and movements in referendums in most EU countries. We discuss the importance of studying the relationships between elite framing and the journalistic frames that emerge in the news, and how this can be expected to vary given greater or lesser political autonomy of broadcasting organizations and news outlets in the EU’s different national contexts. In the penultimate part of the book we discuss a number of campaign effects on the electorate. In Chapter 6 we investigate the public understanding of the referendum issue. The agenda-setting ability of the news media has not previously been discussed in the context of single-issue campaigns such as referendums. This chapter provides a new perspective on agenda-setting in the context of a national referendum campaign in which the broad issue is already at the top of the elites’ agendas. Our case study shows that attending to news sources throughout the campaign increased the public salience of the referendum issue, and we draw this conclusion based on individual-level measures of media exposure and issue salience, unlike most previous research that has explored the link between media and public agendas at only the aggregate-level. The issue of the referendum was multi-faceted, and its many dimensions were interpreted differently by the news media and the public. How the various aspects of the issue become more or less prominent is crucial to understanding public responses to the campaign. This chapter discusses this process and concludes with examples of how sub-issues or aspects of Yes/No issues can come to dominate and deﬁne referendums in various national contexts. Chapter 7 deals with public perceptions of the campaign. The literature on the effects of negative campaigning and strategic news coverage on

The dynamics of a referendum campaign

21

political cynicism and electoral participation is not insubstantial, but it is rarely applied to referendums. We explore how cynicism evolved over the ﬁnal weeks of the campaign. Most previous research has considered changes in political cynicism at the aggregate level over longer periods of time, and not over the short term during a referendum campaign. We make a clear link between news content and exposure to that content. Despite the strategic characteristics of the news coverage in this referendum, and the fact that citizens became more cynical over the course of the campaign, we ﬁnd that the electorate was not demobilized. This sheds new light on demobilization theories in a European context. In Chapter 8 the priming of public evaluations of political leaders during a referendum campaign is analyzed. The literature on the effects of information and campaigns on public evaluations of political leaders has not been applied to referendums. As referendums become more important, they can be expected to have consequences for domestic political leaders. We ﬁnd that the campaign inﬂuenced perceptions of the political leaders, and evaluations of the performance of the government. By comparing performance evaluations across a number of issues, we establish that this change in perception of political leaders was largely driven by the public evaluation of performance on the referendum issue. We discuss the relevance of this ﬁnding to politicians and electorates in other national contexts in Europe. In Chapter 9, we bring together our empirical chapters on campaign effects by investigating the vote. This chapter reviews the literature on referendum voting behavior to identify the key explanations for the vote. There has been an emphasis on ideology and party attachment, government performance and support, economic expectations and evaluations, and issue speciﬁc opinions on the issue of the referendum or related issues as explanations for the vote. Most studies have paid little or no attention to modeling campaign information effects. Our model of the vote takes into account variables that have been the subject of previous studies of voting behavior in referendums, as well as measures of campaign information sources. Our analysis of what inﬂuenced the vote reveals that information obtained during the ﬁnal weeks of the campaign had a signiﬁcant effect on vote choice. Finally, in Chapter 10, we discuss the broader implications of our study for campaign and referendum research in light of the theories about information in elections. The speciﬁc implications of our case study are outlined in the light of referendums and their consequences for democracies in general. We discuss the importance of the use of national referendums in the process of advanced European integration, and the various national contexts in which referendums may occur in the enlarged Europe. In doing so, we assess the popularity of the use of referendums. We focus on the combination of factors involving news values and media systems, the visibility of anti-EU parties and movements or prominent individuals,

22

The importance of information in referendums

and other key points of variation in national contexts that can inﬂuence EU referendum outcomes. Finally we propose factors that are likely to inﬂuence referendum outcomes on issues of European integration in the future under the condition of changing European economies and multicultural societies. A key conclusion is that our study substantiates the assumption that campaigns matter, but that the effects are conditional upon a number of contextual variables and characteristics of individual citizens. We conclude with a discussion of the strategic issues that emerge for the European Union and national governments when citizens say No.

2

The context of the campaign

Political parties in Europe differ in their policy positions on European integration. Most parties favor their country’s membership in the EU, though some oppose it. Some, while thinking that membership is generally a good thing, feel that steps towards uniﬁcation have gone far enough – or even too far. Others believe that further steps should be taken. Political parties also differ in terms of more traditional political orientations – attitudes to the proper role of government in society, welfare provision, and other matters which have increasingly over the past half century come to be subsumed within a single orientation towards government action, generally referred to as the left–right orientation (Lipset 1960; Lijphart 1980; Franklin et al. 1992). Previous studies have suggested that one’s orientation to or attitude towards the EU does not have much impact on party choice in European parliamentary elections (see van der Eijk and Franklin 1996). More recent research, however, suggests that the “pro–anti Europe” dimension is becoming increasingly relevant to understanding public attitudes towards EU politics, and this is discussed along with the standard “left–right” dimension for understanding citizens’ positions and perceptions of party positions in the political system (van der Eijk and Franklin 2003). These two orientations – pro vs. anti Europe and left vs. right – are evolving as orthogonal, with the pro–anti EU orientation cutting across the more traditional left–right orientation (see, for example, Hooghe et al. 2002). This chapter discusses the context of a referendum campaign and the transformations that occur in the party and media systems, which can make the campaign information environment more complex and more important for the outcome. Key features of European multi-party systems and media markets are discussed, and the more speciﬁc context of the Danish case is outlined. The Danish 2000 referendum on the single currency, the euro – ofﬁcially part of the third phase of the European Monetary Union – resulted in a No. The referendum outcome was close (53 percent No and 47 percent Yes) and the turnout level was high (88 percent). In this chapter, we identify both the unique characteristics of our

24

The importance of information in referendums

Danish case study as well as the more generally applicable characteristics that can be extended to other European contexts.

The transformation of the party system In a cross-national comparison based on the 1999 European elections, an inverted U shape of the national party system on the issue of European integration (with extreme left and extreme right parties more opposed to European uniﬁcation than moderate parties), was found in a number of EU countries including Portugal and to some extent in France, Germany, the Netherlands, and Denmark (van der Eijk and Franklin 2003). Denmark offers an example of a country in which the political party system also appears to become transformed into an inverted U shape in the case of European issues (de Vreese 2001c; van der Eijk and Franklin 2003). The Danish political parties can be plotted in a space with a left–right political ideology continuum and a “pro” or “anti” advanced European integration space so that (extreme) left and right wing parties gravitate towards anti-European integration and centrist parties towards pro-European integration. Denmark’s multiparty system includes more than 10 parties (see Appendix for a short introduction to the different parties). Governance in Denmark has traditionally been in the form of a coalition led by either the Conservatives/Liberals or the Social Democrats. Lijphart (1968) observed this as a characteristic of Danish politics more than three decades ago and it still applies. Throughout most of the 1980s, the Danish government was led by a Conservative coalition, in the 1990s the Social Democrats have led most coalitions. Following the general elections of 1998, Denmark had a minority coalition government led by the Social Democrats and the small center socialliberal party, Det Radikale Venstre. The Social Democrats had 64 and Det Radikale Venstre 7 of the 179 seats in the Danish Parliament. The minority status of the government implies that political deals are inherently compromises with parties from either the left or the right end of the political spectrum. The Danish political parties can be organized fairly well on the traditional left–right political continuum. However, with respect to issues of European integration, the two ends of the pole, “extreme” right and “extreme” left, share an anti-European integration point of view. This is a pattern increasingly found in different European countries and has applied to the Danish case over the past three decades. Analyses of the 1979, 1984, and 1989 European elections have reafﬁrmed this Danish Euro-party system structure with a voting pattern in European elections that is different from voting behavior in national elections (Worre 1989). Two political movements operate successfully on the issue of European politics in Denmark. The ﬁrst, “Folkebevaegelsen mod EU” (The People’s Movement against the EU) (PMaEU) originates from The People’s Move-

The context of the campaign

25

ment against the EEC but has adjusted its name and policy to the recent developments in European integration. The PMaEU deﬁnes its goals as working towards “liberating Denmark from the EU, a free inter-state collaboration demonstrating better action possibilities for Denmark outside the EU, and presenting arguments for alternatives to the EU” (www.folkebevaegelsen.dk). While originally launched to promote Danish non-participation in the European Community, the Movement today accepts the European institutions as reality and promotes its policy from within the system. The second “Europe-related” movement, “Junibevaegelsen” (The June Movement), was formed in the aftermath of the Danish “no-referendum” on the Treaty of Maastricht in 1992. The ofﬁcial program states that the June Movement “intends to do everything in its power to prevent the EU from introducing a common currency, developing joint armed and police forces and obtaining a large Federal budget” (Junibevaegelsen 1999). The movement aims at bringing together a cross-party list of Danish antiUnionists, including both supporters and opponents of Danish membership of the EU. In this respect the June Movement differs from the People’s Movement in that its core policies address the economic/ﬁnancial policies and the democratic deﬁcit of the EU as their prime issues rather than the issue of Danish membership of the Union. In relation to European politics, the Danish political landscape over the two decades has developed into so-called “Yes” and “No” blocks. In the 2000 referendum on the euro, the “Yes” block consisted of all political parties at the middle of the political spectrum: The Social Democrats, “Radikale Venstre,” “Centrum-Demkoraterne,” Liberals (“Venstre”), and the Conservatives. Like Labor in Britain, the Social Democratic Party is divided on European issues, but it ofﬁcially belongs to the “Yes” block. The “No” block consisted of the Socialistic People’s Party from the left wing, the two Movements, and the Danish People’s Party from the right wing as well as the center Christian People’s Party. The “No” block is a heterogeneous group of parties and movements from the entire political spectrum. The rationales for their anti-European integration policies differ greatly within this group. The Socialist Party and the Movements are opposed to enhanced economic and political integration and advocate more cooperation with traditional Nordic partners. The Danish People’s Party is opposed to more integration for various reasons, including a fear of open borders and free movement of immigrants and labor in Europe. For the 2000 referendum on the euro, the Christian People’s Party recommended voting No. Table 2.1. is an overview of the Danish political parties and movements participating in the 2000 referendum campaign. Referendums affect the structure of the strategic game of coalition forming in politics. As outlined in Chapter 1 political parties respond to referendums in often unpredictable and controversial manners. This referendum affected the political parties as early as one year prior to the

G G O O O – O – O O O O O –

35.9 3.9 8.9 4.3 7.6 – 7.4 – 24.0 2.4 – 2.5 2.7 86.0

% vote 1998 General Election 63 7 16 8 13 – 13 – 42 4 – 4 5 –

Seats in National Parliament 16.5 9.1 8.5 3.5 7.1 16.1 5.8 7.3 23.3 0.7 – 2.0 – 50.5

% vote 1999 European Parliament 3 1 1 0 1 3 1 1 5 0 – 0 – –

Seats in European Parliament

Y Y Y Y N N N N Y N N N N –

Yes or No policy in 2000 referendum

Notes a The June Movement and the People’s Movement do not run for national elections and only campaign for European Parliament elections and EUrelated referendums. b The Progress Party did not campaign in the Euro referendum campaign. c Freedom 2000 was established after the general election in 1998 as an offspring from the Progress Party.

Source: Danish Ministry of Domestic Affairs.

Social Democratic Party Radical Liberal Conservatives Center Democrats Socialist People’s Party June Movementa Danish People’s Party People’s Movementa Liberal Party Progress Partyb Freedom 2000c Christian People’s Party Unity List Turnout

Government (G) or Opposition (O)

Table 2.1 Overview political parties: percentage of votes in national (1998) and European (1999) elections and Yes or No recommendation in referendum

The context of the campaign

27

referendum when speculation was high about the timing of the referendum. The referendum also elicited some strategic decisions on the side of the parties. Labor Headquarters, for example, enquired with the Danish People’s Party about the timing of their annual congress so that this would not coincide with Labor’s plans or provide the People’s Party a clear opportunity for publicity immediately prior to the referendum. The more fundamental and ideological discussion on the stance of the political parties and movements, however, was partially in place when the referendum was called. Given the nature of the euro issue and the single currency being one of the current four Danish opt-outs in the EU cooperation, this issue had been discussed exhaustively in the early 1990s as part of the rejection of the Maastricht Treaty in a referendum in 1992. Many arguments were reiterated indirectly in response to the national referendum on the Amsterdam Treaty in 1998 that involved future plans for the EU. Most energy on the side of the parties in the euro referendum was therefore vested in the hectic kick-off phase of the campaign when getting visibility in terms of reacting to the call for the referendum and contributing to deﬁning the referendum issue was at stake. Given the capacity of European integration issues to override traditional voting along the left–right ideological alignment, it can be expected that voters in referendums on European integration issues may disregard the advice of their preferred political party. Below we present the results of an analysis of vote intention in national elections as reported in our survey alongside the ofﬁcial position of each individual party. Table 2.2 displays the relative success of the parties to align with their voters. Table 2.2 Vote intention general elections and vote in referendum Party

Share of vote

Ofﬁcial party policy

Voted Yes

Voted No

Social Democratic Party (n ⫽ 248) Radical Liberal (n ⫽ 50) Conservatives (n ⫽ 96) Center Democrats (n ⫽ 24) Socialist People’s Party (n ⫽ 113) June Movement Danish People’s Party (n ⫽ 131) People’s Movement Liberal Party (n ⫽ 333) Progress Party (n ⫽ 10) Christian People’s Party (n ⫽ 28) Unity List (n ⫽ 29)

23.4 4.7 9.0 2.3 10.6 NA 12.3 NA 31.3 1.0 2.6 2.6

Yes Yes Yes Yes No No No No Yes No No No

61.7 76.0 76.0 58.3 13.3 NA 15.3 NA 71.5 50.0 17.9 7.0

35.9 20.0 21.9 37.5 85.8 NA 82.4 NA 27.0 50.0 82.1 93.1

Notes n ⫽ 1.062. Blank votes and abstainers are not included in the table.

28

The importance of information in referendums

The analysis suggests that about 60 percent of the voters of the Labor party followed the ofﬁcial party advice and endorsed the introduction of the euro by voting Yes. Almost 40 percent of the voters went against the ofﬁcial party policy. The smaller center parties (Radikale Venstre, Centrum Democrats, and the Christian party) varied in the voter alignment. Radikale Venstre mobilized 80 percent of its electorate to follow the party line. Centrum Democrats ensured 62 percent of its voters for a Yes vote and the Christian Party, advocating a No, had 82 percent of its voters following the party line. The center-right parties Conservatives and Liberals obtained more than three-quarters of their voters following the party line. The left parties Socialistic People’s Party and the Unity List enjoyed 85 percent and 93 percent party alignment respectively. Compared to earlier referendums, the strong performance of the leftwing parties was to be expected. In the 1992 referendum, the Socialist party had 92 percent of its voters voting according to the party line. The Conservatives and the Liberals saw more of its voters vote against the party line in 2000 than in 1992. In 2000 22 percent and 28 percent voted No when their party recommend Yes while the numbers for 1992 were 13 percent and 18 percent. The Social Democrats did better with 60 percent of its voters following its Yes policy in 2000, while in 1992 64 percent of the voters voted No, despite the ofﬁcial Yes endorsing policy (Siune and Svensson 1993). Table 2.3 summarizes the parties’ performance in the 2000 referendum. A successful party/voter alignment is deﬁned as 75 percent or more of a party’s voters following the party line. A moderate performance is deﬁned as between 66–74 percent alignment and a poor performance is deﬁned as parties who saw less than two-thirds of its voters following the party policy. Table 2.3 shows the performance of the political parties in terms of how Table 2.3 Party and voter alignment: a classiﬁcation of party success

Strong voter alignment

Yes parties

No parties

Radical Liberal (g) Conservatives (o)

Socialist People’s Party (o) Danish People’s Party (o) Christian People’s Party (o) Unity List (o)

Medium voter alignment Liberal Party (o) Weak voter alignment

Social Democratic Party (o) Progress Party (o) Center Democrats (o)

Notes Strong voter alignment is deﬁned as 75 percent or more voters following the advice of the party. Medium voter alignment is deﬁned as between 66 and 74 percent of the voters following the advice of the party. Weak voter alignment is deﬁned as less than 66 percent of the voters following the advice of the party. (g) means government party and (o) opposition party.

The context of the campaign

29

successful they were in their efforts to align their voters with the party policy. The table also illustrates how a referendum on European integration can alter the political system. Those political parties sending mixed or ambiguous signals about the party’s standpoint on the referendum issue performed comparatively poorly in the attempt to mobilize their own voters to endorse the preference of the party elite.

The transformation of the media system In addition to the political system, the structure of the media landscape and its development over time is also an important component for understanding the context of a referendum campaign. The European media landscape has changed considerably over the past two decades (Semetko, de Vreese and Peter 2000), but the two most important sources of political information for the vast majority of European citizens continue to be television and newspapers, and Denmark is no exception (Eurobarometer 51–59). In most European countries the press operates autonomously, but still has ties to political parties. Television operates within the European public broadcasting ethos in which balance and fairness are meant to characterize the political reporting (McQuail 2000). All European countries nowadays operate in a dual system of broadcasting in which publicly funded networks compete with commercially funded stations. This situation includes Denmark, which was comparatively late in introducing commercial television. Denmark still has, comparatively speaking, a high, albeit decreasing, readership of national newspapers. Historically, four phases of political communication can be identiﬁed in Europe. The ﬁrst phase was characterized by the launch of partisan newspapers. In the second phase, partisan newspapers consolidated, grew in size but decreased in number. The second phase also includes the introduction of professional journalism (rather than partisanship-driven reporting) and the launch of radio and later television (which took place in 1926 and 1953 respectively in Denmark). The third phase is dominated by electronic mass media. Technologies such as cable, satellite, and local broadcasting led to a proliferation in the availability of political communication and news. Currently, political communication is entering a fourth phase characterized by diversity and push technologies where the consumers of news deﬁne when, where, and what type of political news they want to receive. In this phase newspaper readership is declining, the number of television outlets still increasing, and digital technologies and Internet services play an increasingly important role. Television and newspapers, however, remain the most important media.

30

The importance of information in referendums

The press The ﬁrst Danish newspaper, published in German, appeared in 1634. At that time 20 percent of the population of Copenhagen was bilingual. By the end of the seventeenth century, three newspapers were established, but the major expansion and breakthrough came in the last quarter of the eighteenth century. At that point, approximately 10 percent of the adult population were frequent readers of the 13 newspapers with a circulation of about 10,000 each that were published twice a week (Jensen 1997). The 1849 Danish constitution provided the formal framework for free and independent press and freedom of opinion and expression. Two trends dominated the development of newspapers: a geographical division between capital-based Copenhagen newspapers and regional papers on the one hand and the development of political alignment of the newspapers along the left–right scale of the political spectrum on the other hand (Jensen 1997: 188). Politiken was introduced towards the end of the century and played an important role in 1909 when it marked its twenty-ﬁfth anniversary by introducing the omnibus format. This meant a newspaper “for all people,” written above political party afﬁliations and geographical boundaries with room for more news, less announcements, less arguments, but more reports and interviews. By 1914 the conservative Berlingske Tidende introduced the same format. The tabloid format newspaper EkstraBladet marked the introduction of the popular tabloid format press in the early 1900s and was followed by the launch of B.T. in 1916. During the period between World War I and II the number of Danish papers in Copenhagen as well as in the province dropped by almost half, but the circulation increased in the same period. The surviving papers not only took over readers of the “dying” papers but also increased their own readership. The notion of a critical press emerged and was fueled by the 1920s’ economic depression. Exposure of social cleavages and workingclass problems became prominent in the press in this period, a trend which also took place in the emergence of the critical documentary wave in ﬁlm production. The competition between the newspapers led to concentration in the press market. The structure with four party papers in most of the bigger cities was abandoned, and a collective regional press emerged with essentially one newspaper per larger town. By the late 1950s, the contours of a new media landscape emerged in which television started to manifest itself. Television, however, was of no real importance until 1960. A total of 84 newspapers have died since 1945, leaving 39 newspapers of which 10 are national papers published in Copenhagen. About half of the 84 ceased to appear before 1960, and the most recent large-scale close-down happened in the early 1970s when the Social Democrats decided to give up publishing independent dailies (Petersen and Siune 1997). During the 1980s press concentration slowed

The context of the campaign 31 down. The relatively stable period of the last 20 years has been characterized by a stagnation/decline in the circulation of provincial dailies, while the national papers, especially the two tabloids EkstraBladet and B.T., during the 1970s and 1980s have increased theirs. In the 1990s, the tabloids experienced a sharp decline in their circulation while only JyllandsPosten was capable of increasing its circulation. The newspaper market today is still dominated by the same main players as a few decades ago. The important developments include a general decline in the percentage of the adult population reading daily newspapers. This peaked in 1969 with 91 percent of the population reading a newspaper daily, declining to 88 percent in 1979, 85 percent in 1984, 76 percent in 1989, and 75 percent in 1993 (Jensen 1997). Historically, an important development in the market structure of the Danish press is the transition from a partisan-based press to the rise of an autonomous and de-aligned national press. The political leanings weakened and the direct ties with the political parties were abandoned. However, party afﬁliations and political preferences are still part of popular knowledge and all national daily newspapers took an outspoken editorial stance in the euro referendum campaign which is also common practice during national election campaigns. Broadcasting Television in Denmark developed along the public service tradition in Britain. The British Broadcasting Company, the BBC, had its ﬁrst public performance of television with a premiere in London in 1926 and broadcast one daily program from 1936–1939 (until the outbreak of World War II). In Denmark, the ﬁrst public show took place on October 30, 1932. In 1947, the chair of the Radio Council initiated full development of television and from 1951–1954 a try out period was granted to the State Radio to also produce and broadcast television. On October 2, 1951 the ﬁrst broadcast took place, which was broadcast to approximately 4,500 television sets in the country. From the very beginning, Danish national broadcasting was ﬁnanced by license fees. The ﬁnancing of broadcasting was set up to be separated from the state budget and from commercial inﬂuences by allowing no commercials or advertising. DR (Denmark’s Radio and Television) was designed as an independent public institution. The public service ethos building on the notion that mass media may contribute to the political, social, and cultural education of citizens was dominant. While commercial networks like NBC, CBS, and ABC dominated the U.S. market, most European countries chose a model with a monopoly and a broadcasting institution with public service obligations. Television challenged the monopoly of the written press. Broadcast news in the early years was provided by the press, and carried the name

32

The importance of information in referendums

“The Press Radio News.” The ﬁrst television news, Tv-Aktuelt, was like cinema newsreels. The emergence of a genuine news program was not until later. The news program format was not ready by the early 1950s, and the political control/inﬂuences were discussed throughout the 1960s. When a news program emerged, questions such as whether television could be critical to politicians on issues or not were debated. During the 1970s, television news was critiqued from politicians with a conservative and liberal background for being red/leftist. In this period the evening news, TV-Avisen, summarized the main events of the day, while the radio provided hourly updates. With developments in technology, such as ENG (Electronic News Gathering) and most recently digital technologies, television news has become more immediate and visually stimulating. Today several news programs are broadcast daily, and with the arrival of the Internet, television news is now also available on-line. The monopoly on national broadcasting ended in 1987 with the launch of a second national television channel, TV2. The rationales for creating TV2 were to break the monopoly of DR, to introduce commercial television, and to offer a Danish alternative to the increasing availability of foreign channels (Petersen and Siune 1997). TV2 adhered to an American inspired news coverage grounded in “common sense appealing to the average citizen” (Hjarvard 1995). TV2 carried more local news whereas DR focused on explaining the news, providing historical contexts, and causes and consequences (Jensen 1997: 161). In 1993, Swedish owned Kinnevik launched a cable and satellite carried Danish language channel, TV3. This channel has gained in popularity and has a reach of approximately 70 percent. More than 70 percent of the Danes watch television daily (with an average viewing time of 100 minutes), and the majority prefer Danish programming (Andersen 1995). By 2000, the Danish broadcasting market counted several national networks and approximately 50 local networks. Most people also receive either German or Swedish television through the regular antenna. Cable or satellite connections are found in households covering about 70 percent of the population. These households receive about 80 channels, including all Danish channels and a number of international ones such as BBC and CNN. The major news programs are found on TV1 (DR) and TV2. While audience segmentation is becoming part of the broadcasting landscape, large heterogeneous audiences still turn to TV1 DR TV-Avisen and TV2 Nyhederne for their daily news. During the year 2000 an average of 700,000 viewers daily watch either of the two news programs (DR Audience Department, Gallup TV-Meter 2000). The transition in the European media landscape form the backdrop against which the referendum campaign must be understood. The Danish news and information market shares many similarities with other European countries, such as a strong, though declining, public service broad-

The context of the campaign

33

casting ethos that has set the standards for television news, and a considerable, though declining readership for national newspapers, and generally increasingly competitive and diverse news market.

The media and the euro referendum Despite the weakened afﬁliations between the press and political parties, all newspapers took an outspoken stance in the referendum issue and recommend readers (and voters) to vote either Yes or No. We include in this study the ﬁve most widely read daily national newspapers. This sample includes broadsheets and tabloid format newspapers. Four of these newspapers recommended a Yes and one a No. Television, by necessity and law, cannot take such an identiﬁable editorial stance and aims for fairness and balance in political reporting in compliance with public service broadcasting regulations. As Table 2.4 shows, a Yes perspective on the referendum issue ofﬁcially dominated the news environment. But in practice, both newspapers and television gave voice to both camps. The media coverage not only focused on the conﬂict between the two camps, but it also played up the differences within camps which provided mixed cues for the electorate.

Attitudes towards Europe European countries differ considerably in the aggregate-level public support for European integration. Denmark joined the European Community on January 1, 1973, following a national referendum held in 1972. The political debate leading up to the referendum was heated, and European integration has been a contentious issue among social groups and political elites ever since. The most notable reaction towards advanced European integration were the violent riots in the streets of the capital Copenhagen in the immediate aftermath of the Danish referendum Table 2.4 Editorial positions of major television news programs and national daily newspapers in the referendum Editorial position Yes

No

Neutral

Television

TV1 DR TV Avisen TV2 Nyhederne

– –

– –

✗ ✗

Newspapers

Politiken Berlingske Tidende JyllandsPosten B.T. EkstraBladet

✗ ✗ ✗ ✗ –

– – – – ✗

– – – – –

34

The importance of information in referendums

in May 1993 on the Treaty of Edinburgh. For the ﬁrst time since World War II, authorities opened ﬁre against a demonstrating public. While European integration has evoked emotional political debate, “Europe” has traditionally not been a prominent part of domestic politics. With the exception of the 1972 national elections due to the then forthcoming referendum on Denmark’s participation in the European Economic Community, “Europe” has been on the periphery of domestic political debate (see Andersen et al. 1999 for a historical discussion). The Danish Constitution requires that issues of sovereignty, which European integration ofﬁcially infringes, have to be put to a national, binding referendum. The Danish electorate has voted directly on European issues on several occasions since the 1972 referendum. In 1986 the government called a referendum on the introduction of the Single Act which implied adjustments of Danish legal practices and regulations to European standards. In 1992, in the referendum held to ratify the Treaty of the European Union as negotiated in Maastricht, the smallest possible margin rejected the treaty. The result was 51 percent against, 49 percent in favor. The 1993 referendum on the Treaty of Edinburgh, which had been negotiated with the European Council to clarify the Danish position after the rejection of the Treaty of Maastricht, was ratiﬁed with a two-thirds majority. Finally, in 1998 a small majority of 55 percent voted in favor of the Treaty of Amsterdam. As Table 2.5 shows turnout in the referendums is usually high, in the 75–90 percent range. The elections for the European Parliament in 1979, 1984, 1989, 1994, and 1999 have neither evoked the feelings or riots of the magnitude seen in relation to the referendums nor reached the same level Table 2.5 Percentage of electorate voting Yes and No and turnout in referendums on European issues, 1972–2000 Yes

No

Turnout

Issue

October 2, 1972

56.7

32.9

90.1

Denmark joins the European Community

February 27, 1986

42.0

32.7

74.8

Denmark adapts Single Act

June 2, 1992

40.5

41.7

82.3

Denmark rejects the Treaty of Maastricht

May 18, 1993

48.6

37.0

86.2

Denmark ratiﬁes Maastricht Treaty with Edinburgh amendments

May 28, 1998

55.1

44.9

74.8

Denmark ratiﬁes Amsterdam Treaty

September 28, 2000

46.8

53.2

87.5

Denmark does not enter euro/ third phase EMU

Source: Ministry of Domestic Affairs.

The context of the campaign

35

of turnout. In these elections, turnout in Denmark is much lower, usually around 50 percent. Denmark has traditionally had a proﬁle of a Europeskeptic country. Longitudinal survey data collected in the bi-annual Eurobarometer reports reveals that Danish approval of the European community up until 1985 was about 30–35 percent (see Figure 2.1) which was approximately 20–25 percent lower than the EU average (Eurobarometer 53 2000). By 1990 the overall average level of European approval had increased to 68 percent, with Denmark at around 60 percent. Throughout the 1990s support for European Union membership has ﬂuctuated in most countries. Recent data suggest that both the average EU approval rate and the Danish approval rate are about 50 percent which suggests that the historically predominant low level of public support for European Union membership in Denmark has leveled off alongside the EU average (Eurobarometer 53 2000). Prior to the 2000 referendum campaign, the level of Danish support for the common European currency was the second lowest in the EU. Only Britain had a lower percentage of the public voicing approval of the euro. In the years leading up to the referendum, the opposition to the euro dropped from about 60 percent to 50 percent (see Table 2.6) and the support increased to about 40 percent. The characteristics of Denmark make it a particularly relevant case to look at in more detail to understand the dynamics of a referendum. It has

80

Good thing Bad thing Neither/nor EU average good thing

70 60

%

50 40 30 20 10

20 00

19 99

19 97

19 95

19 93

19 91

19 89

19 87

19 85

19 83

19 81

0 Year

Figure 2.1 Attitudes towards the European Union: support for membership (data from Eurobarometer reports). Note Question wording: Is membership of the European Union a good thing, a bad thing or neither/nor? Figure excludes don’t know.

36

The importance of information in referendums

Table 2.6 Attitudes towards the euro preceding the referendum Tend to agree

Tend to disagree

1996 1996

Spring Fall

27 31

67 66

1997 1997

Spring Fall

34 36

61 59

1998 1998

Spring Fall

43 46

52 48

1999 1999

Spring Fall

44 42

48 50

2000 2000

Spring Fall

40 41

51 55

Source: Data from “Opinion on Single Currency” (European Commission) and Eurobarometer reports 52–54. Question wording: “There has to be one single currency, the euro, replacing the Danish ‘Krone.’ ” Do you tend to agree or to disagree? Don’t knows excluded from the table.

a number of features that are different from but also several features that are easily comparable to other countries which makes it a suitable benchmark case. The Danish political system turns into a multi-dimensional ﬁeld with the traditional left–right continuum being orthogonal to a pro–anti European integration continuum. Such a pattern was not commonplace in political systems across Europe, but it is emerging in several countries (Hooghe et al. 2002). This also includes a number of the new members to the EU. A number of former Eastern-bloc countries have also experienced both the extreme right and the extreme left propagating antiintegrationist European policies. Denmark is in a dual system of broadcasting with relatively high cable penetration and abundant availability of foreign news. This is very comparable to a number of other European countries and therefore makes the case particularly relevant. As elsewhere in Europe, newspaper readership is declining (Lauf 2001) and television consumption increasing. This is also comparable to the situation in most other European countries though newspaper readership is still considerable in Denmark. Public support for EU membership at the aggregate level in Denmark is almost the average of the EU. The Danish case illustrates that public opinion about European issues is not a uni-dimensional concept. Danes tend to support their country’s membership at more or less the level of the EU average, they tend to be below average on the question of common currency and they are above the EU average in the level of support for the enlargement of the EU. Denmark has a history of EU referendums which allows for benchmarking not only with previous referendums in Denmark, but also elsewhere. As concluded by the Initiative and Referendum Institute

The context of the campaign

37

Europe (IRI), “Danish referendums on the EU were responsible for bringing the initiative and referendum process and the questions of European integration altogether into the European domain” (Gross and Kaufmann 2002: 16). We focus on the Danish 2000 euro referendum, which was a close race during the entire campaign from March to September 2000 (see Figure 2.2). This provides an empirical example and backdrop to discuss and explore different theories of campaign effects. In each chapter, the ﬁndings are discussed and placed in a broader comparative perspective. In the next section we outline the scope and design of the study.

Scope and design of the study To investigate the dynamics of a referendum campaign, including the activities of the political parties, the media, and the electorate, a variety of methods were employed. Previous studies of political communication have often dealt with either media content only, which limits the inferences to public opinion, or they have discussed public opinion or voting behavior without paying much attention to the media. Our study of the referendum draws on four principal sources: 1 2

interviews with campaign leaders from the political parties and editors and political correspondents from the national news organizations, a systematic content analysis of the news media coverage of the referendum, 60 50

Yes No Don’t know

%

40 30 20 10

28 .8 –3 .9 30 .8 –5 .9 4. 9– 7. 9 6. 9– 9. 9 9. 9– 11 .9 11 .9 –1 3. 9 13 .9 –1 5. 9 15 .9 –1 7. 9 17 .9 –1 9. 9 19 .9 –2 1. 9 21 .9 –2 3. 9 23 .9 –2 6. 9

0

Figure 2.2 Opinion polls in the ﬁnal month of the campaign (The Gallup Institute for Danmark’s Radio and Berlingske Tidende).

38

The importance of information in referendums

3

a two-wave panel study with a representative sample of the Danish electorate, focus group interviews with members of the news audience.

4

Interviews with campaign leaders and news organizations Structured interviews were held with the campaign leaders of each of the ten political parties and two political movements participating in the campaign. The interviews addressed issues such as the campaign strategies of the political party, the organization of the campaign, the effort invested in the campaign, and campaign expenditure. In addition, the most salient issues for the political parties were identiﬁed in order to establish the political party agenda during the referendum campaign. This information was also compared to the issues raised by the parties in ofﬁcial party broadcasts, party promotion material, and on the homepages of the parties. Interviews were also held with key members of the national news organizations. A total of eight 60–90-minute interviews were conducted. In these interviews, held with Editors-in-Chief, political reporters, and senior political correspondents, themes such as the organization of the campaign coverage, the effort investigated in the coverage, and the journalistic and editorial approach to the campaign were investigated.

Content analysis of news media coverage To investigate the news and information environment during the campaign, we conducted a content analysis of the news media coverage of the referendum campaign. A systematic and reliable analysis of media content dealing with a campaign is an appropriate method to investigate the diet of information available to an electorate. Content analysis is a frequently utilized method as a “research technique for the objective, systematic, and quantitative description of the manifest content of communication” (Berelson 1952: 18).1 The content analysis of the referendum in the media focused on television and newspapers. The main evening television news programs of the two largest networks were selected for analysis. DR TV1 broadcasts its 30 minutes main bulleting at 21.00 on weekdays and at 18.30 on weekends. TV2 has its main bulletin daily from 19.00–19.25 daily. The two programs TV-Avisen (TV1) and Nyhederne (TV2) at the time of the referendum had an average rating of 16 percent and 15 percent respectively which is an average share of 58 percent for the 19.00 TV2 newscast and 39 percent for the 21.00 TV1 news program (Media Research 2000). The main evening news bulletins of TV1 and TV2 were analyzed during the entire campaign from the day the referendum was called (March 9, 2000) until referendum day (September 28, 2000). All stories in each bulletin were coded to provide an overview of the visibility of the campaign

The context of the campaign

39

and to give an encompassing picture of the general news environment during the campaign. For stories about the referendum an additional analysis took place that involved a number of story characteristics, including the placement of the story, the type of news story, the main topic of the news story, its actors and sources. The codebook can be obtained from the authors. In the presentation of the results in the different chapters we discuss the individual variables. The second most important source of information after television for news about European Union matters in Denmark is the daily newspaper (Eurobarometer 5 2000). For the study we included the ﬁve largest and most widely read daily newspapers with the highest circulation. All front pages were screened for news about the euro and/or the referendum. These stories were coded using the same codebook as described above for television news. In addition an overview of all articles, editorials, ads, and Letters to the Editor inside the newspapers were coded for its main topic to provide a more detailed picture of the news environment during the campaign. The newspapers were analyzed for the ﬁnal month of the campaign, from August 28–September 28, 2000. The content analysis was conducted by four Danish native speakers who were MA students in the International Masters program at the University of Amsterdam. Coders were trained and supervised and questions were resolved during regular meetings and in the supervision of individual coders. The reliability test yielded satisfactory results. Reliability in content analyses aims at ensuring that a study is replicable for future researchers. In a systematic content analysis where several individuals may be involved in coding and analyzing media content, the reliability, i.e. both the reproducibility and the accuracy of coding is essential to draw conclusions from the data (Peter and Lauf 2002). The speciﬁc reliabilities of the measures are reported when the measures are discussed in the different chapters. The content analysis of the news media coverage of the campaign was designed to be linked to both the interviews with news practitioners within the news organizations and to data about the audience use of news and attitudes towards the referendum. Given the combination of a content analysis of the media coverage with a two-wave panel survey (see below) we were able to make a compelling link between being exposed to speciﬁc news outlets and its effects.

Two-wave panel study To study the dynamics of the campaign a two-wave panel study with a representative sample of the Danish electorate was carried out. A panel has the advantage over cross-sectional surveys that it allows for tracking changes over time, in this case over the course of the campaign. Repeated cross-sectional surveys allow for tracking changes at the aggregate level,

40

The importance of information in referendums

but a panel survey allows for analyzing changes at both the aggregate/ collective level and at the individual level. For this study, the ﬁrst wave was carried out between August 31 and September 5, 2000, about one month ahead of the referendum. 1,382 paper and pencil questionnaires were distributed via traditional mail and 1,065 returned leading to a response rate of 77 percent in wave one. The ﬁeld period for the second wave was in the immediate aftermath of the referendum (September 29–October 1, 2000). The wave included 1,370 respondents. A total of 1,066 questionnaires were returned which means that the response rate in wave two was 79 percent. Panel mortality was 10.5 percent between the two waves which means that a total net sample of 962 respondents were included in both waves. Response rates are often of this magnitude in Scandinavia (Andersen et al. 1999; Granberg and Holmberg 1991). The survey was administrated by Gfk Danmark that also maintains the panel. To test the suitability and representativity of the panel for this study, we included a number of measures, including voting intention, on a wave four months prior to the referendum. The distribution of the vote intention mirrored that of other publicly available polls at the time, such as those ﬁelded by Gallup. A number of measures were repeated in the two waves of the panel study. These include vote intention and actual vote, economic expectations, issue perception, evaluation of key political leaders, evaluation of overall government performance and handling of different issues, political cynicism as well as perceptions of the campaign itself. Additionally, in the ﬁrst wave respondents were probed for political efﬁcacy, interest in social and political issues, and general attitudes towards the EU and European integration. In the second wave measures were included for campaign involvement, media use, perception of the role of polls in the campaign as well as time of vote decision. Additionally gender, age, and education level were recorded as socio-demographic data for each respondent. A translated version of the questionnaire can be obtained from the authors. The wording of individual variables and the treatment in data analysis are discussed in the different chapters.

Focus group interviews To further explore the role of news coverage in the campaign and to investigate the public perception of the campaign, focus group interviews were conducted. Focus groups are appropriate for gathering rich data from distinct subgroups, for generating insights into less well understood areas, and for enhancing understanding of complex cognitive processes (Morgan 1998; Wimmer and Dominick 1997). Focus group interviews were held with members of the electorate in mid-August, about a month before Election Day.2 Participants in the focus groups were recruited by the Danish afﬁliate of AIM Nielsen Media Research and were recruited to

The context of the campaign

41

ensure participation of a broad range of the electorate in terms of age, education, and gender. Four groups were created using vote intention in the referendum as an additional selection criterion. The groups were homogeneous in terms of vote intention in the referendum leading to one “yes,” one “no,” and one “undecided” group. Additionally, a fourth group of younger persons (age 18–27) was recruited since experience shows that a not too great age range is an asset to the group dynamics. A total of 33 participants (16 women and 17 men) participated in the groups.3 Participants were invited to the Headquarters of the Danish Broadcasting Association in the capital Copenhagen. They were informed that sessions would take approximately two hours which is slightly longer than usually recommended for focus groups (Morgan 1998), but considered appropriate given the use of news stories as material throughout the session. Each group consisted of no fewer than seven and no more than 10 participants, which is a standard group size (Adams 2000). A professional moderator from the Audience and Research Department hosted all sessions and they followed an identical procedure, including a welcome, screening of three selected news stories and subsequent completion of pen and pencil questionnaire, an association game, and a debrieﬁng. All sessions were recorded by video cameras and relevant passages, excluding for example the introduction and the association games, were transcribed. The three news stories shown during the sessions were selected from the television news coverage of the campaign in the period March 2000, when the referendum was announced, till August 2000 when the focus group interviews were held. An archival search identiﬁed all stories broadcast about the euro referendum in that period. The three stories were selected because they a) were thematically representative of the coverage of the campaign so far, b) were journalistically and stylistically typical for the news coverage, and c) because the topics of the news stories were deemed to also play a role in the ﬁnal month of the campaign.

Linking survey data and media content analysis The link between data on the content of the mass media, individuals’ media use, and public opinion data tends to be rather weak (Bartels 1993; Freedman and Goldstein 1999). Often studies of the media and of public opinion are not integrated in a single research design, which implies that the codebook for the content analysis is not easily compatible with the questionnaire used in the survey. The majority of campaign studies has relied on a combination of media content analysis and cross-sectional public opinion data. While such data can provide information about the relationship between the media and the electorate’s opinions and behavior, there are inherent limitations in the type of tests that can be performed and the nature of the relationship that can be established. A content analysis and a one-point cross-sectional survey does, for example,

42

The importance of information in referendums

not allow for causality testing which raises the chicken–egg question when discussing media and campaign effects. For example, does a high correlation between political interest and reading broadsheet quality newspapers imply that politically interested individuals read more news or does reading news increase levels of political interest? Establishing causality requires a pre-test of speciﬁc attitudes and predispositions, that is establishing reliably the predispositions that respondents brought to the study, and then being able to accurately characterize the media they were actually exposed to and attentively watched/read, before repeating identical measures in a post-test. This demand, in terms of a research design, essentially calls for an experimental investigation with manipulation of realistic stimulus material and well-developed post-test measures. Alternatively, however, a panel study with repeated measures assessing the variable(s) of interest in both wave one and wave two combined with individual level data on news exposure, and a content analysis of the actual news media coverage also provides a compelling link between media content data and public opinion data. To investigate public opinion and the dynamics over the course of the referendum campaign, we used a two-wave panel survey and focus group interviews. While there is no one ideal method to explore the complex issues of campaign effects, a combination of several methods will at least contribute to avoiding the apparent danger of drawing generalizations from ﬁndings that may be biased given mere methodological constraints. Cross-sectional survey data tell us little about change over time, about shifts in opinion. Focus groups often draw on problematic sampling procedures which endanger the generalizability of the results. Experiments have strong internal validity (with the isolation of one independent variable in the experimental manipulation), but may suffer from low external validity due to stimulus shortcomings and laboratory setting. Our design therefore aimed for a combination of more than one approach (Hovland 1959).

Part II

The political parties’ campaigns

3

The Yes and No camps

In theory, a campaign functions to empower the electorate: knowledge levels should increase, voters are encouraged to vote. The process by which these civic goals are achieved has changed as news sources have become the most important intermediary between voters and parties or candidates. Campaigns no longer belong to the party domain with its local party organizations and volunteers. As press–party loyalties have declined and the outlets for electronic news have diversiﬁed, politicians are challenged to operate in and respond to a more complex communications environment. On their part, politicians operate using increasingly professional campaigning including a growing number of campaign staff, sophisticated targeting of key voters, increased expenditure on publicity, and a growing (conscious) use of campaign techniques in government. The literature thus far on the evolution of political campaigning does not address referendums speciﬁcally, but expectations may be formulated based on investigations of general elections. This chapter outlines changes in political campaigning and applies this to the speciﬁc case of a referendum campaign.

Three phases of political campaigning Research from national political contexts suggests a professionalization of politics. A transition has taken place from campaigns being short, decentralized, and dependent on physical manpower to a stage of permanent campaigning in which campaign professionals such as pollsters, marketing consultants, and spin-doctors play key roles. Observers of political campaigning by-and-large identify a three-phase typology of the evolution of campaigning (Farrell 1996; Farrell and Webb 2000; Norris 2000). The ﬁrst phase covers the period from the late 1800s to the 1950s and is characterized by a strong party organization with de facto several almost independently operating, locally organized campaigns. Volunteers engage in the campaign by distributing pamphlets, canvassing and organizing local party meetings. This ﬁrst phase is most effectively distinguished from phase two by the arrival of television. In the second phase preparations for

46

The political parties’ campaigns

campaigns are lengthier and more extensive and the campaign planning is centrally organized and controlled. The party system undergoes a professionalization transition, electoral research is introduced, and press conferences, staged media events and political style become important. This phase covers the period from the 1950s up until late 1980s/early 1990s. The third phase, which according to the literature is still emerging, is characterized by an increase in the use of new information and communication technologies, sophisticated targeting of segments of voters, increased expenditure, and growing (conscious) use of campaign techniques during periods in ofﬁce in addition to the traditional campaign cycle. These developments are implemented both internally in the party organizations as well as a means of external communication with party members and potential voters. The time horizon of campaigning is redeﬁned towards some level of permanent campaigning. Campaign professionals become increasingly important and the electorate is conceived of as existing of segments. Consultants in areas ranging from fashion to speech writing are involved in developing and implementing the campaign. Many of the changes in the emerging third phase of political campaigning have taken place behind the screens, in the organization of the campaign. The gradual evolution towards permanent campaigning encompasses a number of important dimensions of which centralization of the campaign, electoral research, new technologies and sophisticated publicity strategies are the most important. First, contemporary political campaigning is characterized by a considerable degree of centralization of the campaign management and streamlined feedback loops. The “war room” phenomenon describes the strongly centralized, ﬁne-tuned, and controlled campaign in which a select insider group gives guidance and directs most campaign activities. A second important dimension is the proliferation in electoral and market research. Political parties increasingly use surveys and focus groups in the preparation stages of a campaign. While image building is essential at the outset of a campaign, so too is the permanent monitoring of public responses to the images of public ﬁgures. Focus groups have become popular for providing a very rich input in terms of associations that voters have with certain politicians, images, and slogans. A third dimension of contemporary political campaigning is the increasing importance of new information and communication technologies. Through these politicians are provided with the opportunity to communicate and interact with voters, thereby circumventing the intermediary journalists. Voters are empowered with the opportunity to acquire information about candidates and parties on their conditions, i.e. via channels that they choose and at time slots that suit them. A ﬁnal component of contemporary campaigning is the emergence of sophisticated publicity strategies. Parties and politicians have a better understanding of journalism that enables them to package the political message to ﬁt the formats,

The Yes and No camps

47

needs, and demands of journalistic products. Candidates are increasingly involved in pro-active news management. The goal of this is two-fold: ﬁrst, to gain access to the news, and second, to control the publicity. Access is fueled if the message meets the professional criteria for news selection. Controlling the spin of the news story, that is the way in which journalists frame the story, is a second-step challenge for which news management and bypass strategies (refrain from participating) can be used as instruments.

Professionalization of politics around the world Examples of professionalization of politics can be found in countries around the world. The most prominent example is the United States where campaigning is a considerable industry, also outside the ofﬁcial campaign time. Pollsters, speechwriters, media advisors, and fundraisers are permanently employed and spend the periods in opposition or ofﬁce to work on how to proﬁle candidates and issues in the next contest. The extensive degree of professionalization of American politics makes it a valuable benchmark to assess campaigns in other contexts. In Europe, marketing jargon entered politics along with the successful image-building of Britain’s Iron Lady, Margaret Thatcher. “Political campaigning was transformed during the Thatcher decade and her contribution was fundamental. The hiring of Saatchi and Saatchi in 1978 as a full-time agency for a political party represented a quantum leap in the marketing of Margaret Thatcher and the Conservative party” (Scammell 1995: 2). The 1987 general election campaign saw a boost in the campaign expenditure and Britain moved towards the permanent campaign. A new landmark in British campaigning came with the Tony Blair campaign and his 1997 landslide election victory which marks a second major leap in techniques of controlled electioneering (Butler and Kavanagh 1997: 243). From 1983 to 1997 the general election campaign expenditure went up from 2.2 to 26.0 million British pounds sterling for Labour and from 3.6 to 28.3 million for the Conservatives (Norris et al. 1999: 39). Generally speaking Britain has largely moved to a third phase of campaigning. Advances in technologies, heavy use of market research, awareness of the “rules” of campaigning through both mass media and direct, targeted communication channels are characteristics of modern campaigning and these all apply to the British case (Scammell 1995). The election of Tony Blair as Prime Minister in 1997, his 2001 re-election as well as his years as Prime Minister are testament to an increasingly professional attitude towards communication from politicians to voters through the media. Germany also displays the trend towards centrally run, professionally designed and executed campaigns. In 2002, the German Chancellor Gerhard Schröder was re-elected after a ﬁerce electoral battle. Observers agree that a skillful media campaign was instrumental to the re-election of

48

The political parties’ campaigns

the SPD party in a government (Semetko and Schönbach 2003). In Spain, the post-Franco era developments in campaigning were primarily propelled by television (Gunther et al. 2000), the medium that continues to be the most important source of political information for most Spaniards and particularly in elections in which a not insubstantial portion of the electorate now holds off on making a decision about which party to support until the ﬁnal weeks of the campaign (Diez-Nicolas and Semetko 1995). A “mediatized” and modern mode of campaigning and television was a key feature of the PP’s campaigns in 1996 and 2000, and the party’s leader, former Prime Minister Aznar’s style. In Italy, Silvio Berlusconi’s position as a major media industrialist and his path to electoral ofﬁce is seen by many as evidence of the (too) close relations between media, campaigning, and governance (Mazzoleni 1995, 2003; Mazzoleni and Campus 2003). With the exception of the “political earthquake in the polder” in the May 2002 national elections, in which a new party, List Pim Fortuyn (LPF) – named after its maverick leader – came to power after a campaign in which Fortuyn’s personality, style, and negative views on the incumbent government and the country’s Muslim immigrants were the major focus, most national election campaigns in the Netherlands had been relatively quiet and reﬂected the strong consensus-biased democracy (Cherribi 2003a: 1). Various media (magazines, newspapers, broadcasting, and book publishers) had pro-actively contributed to establish Pim Fortuyn as a brand name well in advance of the launch of his political party, and routine reporting of local and world events in 2000 and 2001 along with “news-making” reporting by key Dutch public broadcasting programs provided visibility for a key issue [the threat of Islam in Dutch society] that fueled his political success (Cherribi 2003a, 2003b). The Fortuyn phenomenon shifted what had often been a focus in previous campaign coverage on the organization and campaign strategies of the various parties (van Praag and van der Eijk 1998), to the successes of politicians as campaigners and political debaters. This focus in the news on campaign style was continued in the January 2003 Dutch national elections when the Labour (PvdA) leader Wouter Bos was the focus of much campaign reporting. On Election Day, the PvdA reclaimed all the seats it had lost in the May 2002 elections and took more, and this success was attributed to the (media) qualities of the new leader. Dutch election campaigns before May 2002 had been known for their comparative civility and attention to the issues in the news (Semetko 1998). There was some evidence of professionalized campaigning as early as 1988/89, prior to the 1989 general elections, when the Labour Party (PvdA) used Saatchi and Saatchi for its marketing and PR activities, but this cooperation led to friction over who was in charge of content and style of the campaign (Simons 1998: 16). A decade later at the Dutch 1998 elections, there was a considerably higher campaign budget, more electoral research initiated by the parties, an increase in

The Yes and No camps

49

the strategic use of local media, and strategic news management (van Praag 2000: 16). The degree of professionalization is rather conditional upon political parties themselves. This makes it particularly important to establish parameters that are indicative of the different phases in political campaigning so as to assess a particular campaign and place this in the perspective of developments in campaigning in general. We stress the importance of recognizing that not all countries and electoral contexts are characterized by the third age of campaigning. One essential precondition for substantiating such observations is to put the largely theoretical and descriptive accounts of developments in campaigning to empirical tests in different contexts, including referendum campaigns.

The media response: strategic interpretative reporting The developments in campaigning go hand in hand with a number of changes in politics, media, and journalism. At least three structural developments have contributed to today’s mode of campaigning: 1) changes in the political parties, 2) changes in the media landscape, and 3) changes in journalism. First, political parties have gone through great changes themselves. A campaign no longer implies the mobilization of 10,000 volunteers to distribute pamphlets. Instead political parties have increasingly become streamlined organizations with distinct divisions of responsibilities and tasks. This professionalization of the political parties coincides with dramatic changes in the European media landscape over the past three decades. The institutional association of newspapers with political parties and the European public service broadcasting monopoly are history. Professional newspapers cover news without a direct association with a political party, replacing the once strong press–party alignments. In television, a transition took place from the early 1980s where only Italy, Luxemburg, and the UK had introduced competition to the national public service broadcaster to a situation today with no public service monopoly in Europe. Along with this signiﬁcant change in the market parameters, journalists no longer see it as their job to cover a campaign by merely bringing daily images from the activities of the political parties. Campaign journalism is increasingly about selecting events, placing these in a perspective, and analyzing and interpreting these events and issues (de Vreese 2001a). Recent developments that go beyond the introduction of competition in the television market also include the increasing fragmentation of the media landscape with multiple suppliers providing news and information not only via traditional media but also, for example, via Internet. These developments in the media market imply easier access for politicians to free publicity, but also decrease the reach to the electorate through the traditional media.

50

The political parties’ campaigns

While these changes alter the conditions on which campaigns are conducted it is at least noteworthy that as late as 1987 it was concluded that “the study of election campaigns, as opposed to elections, is a major gap in the literature” (Harrop and Miller 1987: 240). Within the past ﬁfteen years, however, a number of studies from individual countries and a few comparative enterprises have contributed to our understanding of campaigns around the world and shed new light on the interplay between political actors, the media, and the electorate. Despite this proliferation in the interest in campaigns, we still know strikingly little about the media in referendum campaigns. Recent volumes dedicated to the study of referendums (including, for example, Farrell and Schmitt-Beck 2002; Hug 2003; Mendehlson and Parkin 2001; and LeDuc 2003) only mention the media in rather descriptive and peripheral terms. We therefore know little about how political parties and candidates campaign and how journalists report during a referendum.

Political campaigning in Denmark Political campaigns in Denmark have also gone through the transition from a locally organized campaign relying on unpaid volunteers to a centralized, professional high-expenditure campaign. However, Danish political campaigning is still characterized by many of the features from the second phase of campaigning. During the campaign for the 1994 European parliamentary elections some evidence suggested that politicians designed their campaigns to include spectacular and visually compelling elements to meet television’s need for images. In 1994, for example, a politician drove through the streets of Copenhagen in a horse-drawn open carriage to provide a positive image but the attempt backﬁred and was ridiculed and criticized in the Danish news media. There now appears to be a reluctance among political parties towards planning such “stunts.” Similarly during the campaign for the European elections in 1999, a Social Democratic candidate who campaigned using an old sailing boat with which he visited a number of Danish towns provided the only extraordinary campaign moment. Professionalization of campaigning does not seem to come from such developments. Generally speaking, the 1999 European election campaign consisted of limited advertising, information material distributed in shopping malls, and poorly organized political meetings. The latter were not well visited at all and several meetings were even cancelled because of lack of interest. One new element in campaigning, utilized broadly from the 1998 national election campaign and onwards, is the use of Internet. Denmark along with the other Scandinavian countries is among those with the highest proportion of households using the Internet in the world. In Denmark (as in most other European countries) political campaigns are essentially organized by the central party organizations. Previously

The Yes and No camps

51

local afﬁliates and units of the party were more autonomous, now there are central candidates. In the late 1980s/early 1990s Denmark was on the level of Italy and the UK in terms of average expenditure per voter in national elections (van Praag 1995: 235). The average campaign expenditure per citizen was higher than in, for example the Netherlands, but less than, for example Austria and Germany (where campaigns are heavily subsidized by the state).

Key questions The existing literature can be translated into a number of parameters that represent the different phases of political campaigning. Traditional campaigning is characterized by activities such as rallies, party meetings, canvassing, and the distribution of information material. Modern campaigns include the use of advertising, direct mail, and PR and marketing agencies. The most sophisticated campaigns also include extensive use of the Internet, electoral research, party-initiated surveys and focus group tests. Based on the literature discussed above we can expect a considerably professional mode of campaigning in a national referendum in a Western European democracy in the year 2000. We identify a number of indicators of campaign investment, efforts, and strategy. We analyze the party agenda and we assess the campaigns of individual parties as well as the Yes and No camps as a whole in the light of the literature. Towards the end of the chapter we also look in more detail at the arguments put forward by the different political parties and movements. We analyze the extent to which the Yes and No camps campaigned together. As discussed in Chapter 2, a referendum can lead to new alliances in the political spectrum. In this chapter we identify the conditions under which the (extreme) right and left sides of the political spectrum operated strategically together.

Campaign efforts and party strategies In our overview of the efforts, strategies, and expenditure of the different political groups, we ﬁrst provide a descriptive analysis of the types of activities that these actors in the campaign engaged in. We include in this overview elements of traditional campaigning such as canvassing and the organization of party rallies and meetings. As indicators of the so-called modern campaign we map the use of advertising and involvement of PR agencies in the campaign. Finally we also assess the use of the Internet and party-initiated electoral research such as surveys and focus groups. Table 3.1 provides an overview of the various activities by the political parties and movements in the referendum campaign. All parties, without exception, engaged in traditional activities such as locally organized meetings, distribution of information material, and canvassing on the street.

Yes Some Yes Yes Some Yes Yes

No parties and movements Socialist People’s Party June Movement Danish People’s Party People’s Movement Freedom 2000 Christian People’s Party Unity List Yes Yes Yes Yes Yes Yes Yes

Yes Yes Yes Yes Yes

Info material

Yes Yes Yes Yes No Yes Yes

Yes Yes Yes Yes Yes

Street happenings/ canvassing

Yes Yesb No Yesb No Yes Yes

Some local No Some local No Yes

Direct mail

Yes Yes Yes Yes Yes Yes Yes

Yes Yes Yes Yes Yes

Web/ Internet

Some Few Some Some Few Few Few

Yes Yes Yes Some Yes

Ads

No No Yes No Yes No No

Yes No Yes Yes Yes

PR ﬁrm

No No No No No No No

Yes No Noc Noc Yes

Polls/ Focus groups

1,700,000 (227,882) 2,300,000 (308,311) 3,500,000 (469,169) 3,000,000 (402,145) 3,800,000 (107,239) 3,628,000 (84,182) 3,600,000 (80,429)

6,000,000 (804,289) 1,500,000 (201,072) 5,800,000 (777,480) 1,400,000 (187,668) 9,200,000 (1,233,243)

Budget (in kroner and eurosa)

Notes a Budget in euros based on conversion rate at time of referendum. b June Movement and People’s Movement mailed out 800,000 info packages in a joint project. c Access to material by Europe Movement was made available to the Conservatives and the Center Democrats. All information stems from interviews with the party and movement campaign leaders.

Yes Yes Some Some Yes

Yes parties Social Democrats Radical Liberals Conservatives Center Democrats Liberal Party

Meetings/ Rallying

Table 3.1 Overview of political parties’ and movements’ activities in the referendum campaign

The Yes and No camps

53

Some parties even organized a substantial part of the campaign activities around happenings and visits to local markets during weekends. The parties varied in the extent to which they used direct mail as a strategic component of their campaign. While some parties, primarily the Social Democratic Party, the Conservatives, and the Liberals, organized this locally, the two Movements along with the Socialist Party distributed a little over 2.1 million pamphlets through direct mail. There was also large variation in the extent to which advertising agencies played a role in the campaign. While, for example, the Social Democrats ran a professionally designed ad campaign in the national daily newspapers, others, such as the June Movement only used newspaper ads marginally. The Socialist People’s Party and the two Movements, along with the smaller parties (the Christian People’s Party and the Unity List, all from the No camp), designed their ad campaign themselves without any use of professional agencies. All parties and political movements participating in the referendum campaign had special links on their homepages about the euro and the campaign. The homepages were updated continuously throughout the campaign and the estimated trafﬁc on the pages was between three and ﬁve times higher than normal, with a peak towards the end of the campaign. The pages contained information about the standpoint of the political party, offered links to most of the party programs and contained factual information about the euro and the third phase of the Economic and Monetary Union. A number of the parties also initiated interactive chat-sessions with key politicians. In terms of campaign budget, a total of 36.4 million Danish Kroner (4.7 million euro) was spent on the campaign by the parties. The Yes side spent 23.9 million of this amount (66 percent) and the No side 12.5 million (34 percent of the total expenditure). In comparison to for example the 1990 general election campaign, these ﬁgures point in two directions. The Labour Party in 1990 had spent 10 million DKK (1.3 million euro) on the general election campaign (Bille et al. 1992), but this was 6 million in the 2000 referendum campaign. The Liberal Party, on the other hand, spent 9.2 million in the referendum compared to only 5 million in the 1990 general election. One conclusion to be drawn from this level of expenditure for the euro referendum campaign is that it is in the range of that spent on a general election, an indication of the political importance assigned to the referendum by the political parties. By way of comparison, however, the Swedish Parliament – the Riksdag – allocated some 120 million Swedish Kroner (SEK) (10 million euro) for campaign activities in the run-up to the 2003 Swedish referendum on the euro. And 90 million SEK was divided between the Yes and the No campaigns through specially established campaign organizations. 48 million SEK was allocated to the No campaign and 42 million SEK to the Yes campaign. The remaining 30 million SEK was divided among the parties represented in the Riksdag.

54

The political parties’ campaigns

In addition to the publicly available funding, the Swedish Yes campaign also established the “Association for the Yes campaign.” A total of 21 million SEK was distributed from this Association, amongst others to the Social Democrats for Europe, the “Sweden in Europe Foundation” and to the Federation of Swedish farmers. The Swedish No campaign established the “Association for the No campaign” and this association allocated 21 million SEK to the umbrella organizations “No to the EMU” and “Together against the EMU.” By this comparison the Danish referendum campaign was less well funded. Stakeholders in Sweden in 2003, both political, social, and from civic society, invested more in the Swedish referendum campaign on the euro.

Advertising strategies There was considerable variation across the Danish political parties in terms of their strategies and budget reserved for political advertising. It should be noted that the ad strategy is not aimed at television because Denmark has rather restrictive legislation which prohibits broadcasting political commercials. The comments in this section therefore pertain in particular to ads in the press. The Yes camp parties advertised extensively and their target group was broadly deﬁned as the general electorate. As shown in Table 3.2 they all Table 3.2 Advertising strategies by political party Strategy Yes parties Social Democrats Radical Liberals Conservatives Center Democrats Liberal Party No parties and movements Socialist People’s Party June Movement Danish People’s Party People’s Movement against EU Freedom 2000 Christian People’s Party Unity List

Ads in national and local newspapers with prominent party and union members advocating Yes Ads in national newspapers Ads in national newspapers with a Yes editorial position (JyllandsPosten and Berlingske Tidende) Ads on billboards on train stations Extensive ad use in national newspapers Some ads in national newspapers Few ads at speciﬁc moments Ads primarily placed in local and only few in national newspapers at the end of the campaign Ads primarily in local and regional radio spots. Some ads in newspapers Ads in weekly women’s magazines, cinema ads, and free publicity post cards Limited use of ads, few placed in newspapers Very limited use of ads, only one in newspaper

The Yes and No camps

55

used national newspapers as their vehicle and the Social Democrats made use of strong message endorsers from within the party and the labor unions. The Liberal Party centered its newspaper ads on the party leader, Anders Fogh Rasmussen. The Conservatives targeted its core voters by advertising in newspapers primarily read by Liberal and Conservative voters. The No camp made less use of newspaper ads and when these were used this happened in a focused way and not utilizing the more extensive across-the-board strategy of the Yes camp. The Socialist People’s Party, the June Movement and the Danish People’s Party all had some ads, but these were limited and placed at strategically important moments, in particular towards the end of the campaign. The Danish People’s Party and the People’s Movement against the EU used local and regional newspapers as their primary vehicle for print ads. The small No party Freedom 2000 did not use newspaper ads at all, but opted for ads in women’s weekly magazines and distributed free publicity postcards. Looking at the content of the different ad campaigns, a number of aspects stand out. First, the Social Democrats were – as usual – challenged in their campaign. The party advocated a Yes knowing that many of its voters were either undecided and against the introduction of the euro. This affected the ads that often listed or at least acknowledged known No arguments. Second, the Yes ads often used celebrity endorsers, i.e. wellknown public ﬁgures recommending a Yes. Third, the Danish People’s Party made use of a highly professional ad campaign that was particularly targeted towards undecided voters. The party’s core voters were likely to vote against the euro anyway and the ad campaign was therefore aimed at voters that had not already made up their mind at the outset of the campaign. The ads played up traditional Danish society values and positioned the euro as representing change and uncertainty for Danes. Finally, the ad campaign of the remaining No parties can best be characterized as sporadic and rather traditional, also in terms of layout.

Referendum campaigning between traditionalism and professionalism The recent literature has pointed towards an increase in the professionalization of political campaigning not only in the United States, but also in Europe (for example, Farrell 1996; Farrell and Webb 2000; Norris 2000). Phenomena such as spin-doctors, news management, and staged media events have become part of the political communication vocabulary. The expectation for a referendum campaign in the year 2000 would therefore be a highly professional and streamlined campaign. The referendum campaign was characterized by traditionalism and professionalism. Many of the components known to the pre-modern and modern modes of political campaigning were still found along with some “late-modern” tools such as

56

The political parties’ campaigns

on-line, interactive communication. By and large, however, the campaign was more traditional than what may be expected based on recent studies in the realm of political communication.

The party issue agenda Advertising as discussed above is one means of conveying a political party’s agenda and message in a campaign. Ads as a mode of communication are preferred by political parties given the full control they have of the content of the message, despite the fact that they may be a limited means of addressing the wider range of issues that a party may want to “use” during a campaign. In this section the issue agenda of the political parties is analyzed. To identify the most important issues, we draw on two sources: interviews with the campaign managers of the political parties and party broadcasts. As discussed above, commercials on national television are prohibited by law. Political parties are, however, granted free broadcast time in which they can air a short program outlining their agenda and arguments. The dynamics of issue ownership is different in a referendum campaign from a general election campaign. While general election campaigns tend to be waged over issues of health, welfare, and culture by center-left parties and taxes and economy by center-right parties, the overarching issue of a referendum campaign is obviously deﬁned in advance. That said, parties and candidates may deﬁne this key issue differently and thereby try to gain control and get media coverage in relation to sub-issues within the broader referendum topic. The battle over the agenda is an aspect of the dynamics of a referendum campaign that has been alluded to, but not yet addressed in its totality. We ﬁrst deﬁne the agenda of the political parties and movements. In the next chapter we analyze the visibility of the referendum in the media and in Chapter 5 we compare the agenda of the political parties and the media to assess differences and similarities between the two. Finally, we analyze the relationship between the media agenda and the public agenda in Chapter 6. In the ﬁrst instance, we are concerned with how the parties deﬁned the referendum issue. Though the topic of the referendum may at ﬁrst sight appear relatively well deﬁned, previous research has demonstrated how opponents in a referendum can utilize a strategy in order to raise doubts and create confusion about the issue of the referendum. In the words of Magleby (1989: 113): “With sufﬁcient money for advertising, careful research on what parts of the issue voters would not like, and effective commercials playing to these themes, the ‘no’ side can usually prevail.” Such a strategy can indeed be observed in many referendum campaigns on a variety of issues. During the 1986 Irish referendum on divorce, observers noted that stakeholders tried to “change the subject” of the debate by creating doubts about what was actually being proposed. In that case it led to solid support evaporating (D’arcy and Laver 1990).

The Yes and No camps

57

In recent referendums on European integration similar trends could be noted. In the 2003 referendum in Poland on membership in the EU, the Yes side was keen to talk about economy and the democratic traditions of the EU. Voices on the No side, however, put forward an agenda that linked Polish membership of a supranational organization with the Polish relations to the former Soviet Union. In the 2003 Swedish euro referendum a battle over the referendum agenda also took place between the Yes and No camps. The Yes camp deﬁned the referendum as an issue of Swedish economy whereas the No camp addressed issues such as the continuation of the strong Swedish welfare state. In the Danish euro referendum, the Yes parties emphasized the economic aspects of the issue (Table 3.3). The campaign managers identiﬁed a Table 3.3 Party issue agenda I Issue Yes parties Social Democrats Radical Liberals Conservatives Center Democrats Liberal Party No parties and movements Socialist People’s Party June Movement Danish People’s Party People’s Movement against EU Freedom 2000 Christian People’s Party Unity List

The euro as a currency: how does it work and what does it mean. Economic arguments most important Creating public conﬁdence in the euro, the future, and in Europe The euro and Europe means peace and freedom Denmark must have a say in the economic policies, the euro will ensure favorable economic conditions and peace A healthy economy is the most important issue The euro means endangered enlargement plans, welfare cuts and “United States of Europe” The euro is part of a larger process. Democracy and the evolution of the EU. “The euro is a POLITICAL issue” The “Krone” goes before euro, to preserve the existing is important, a No vote is a Danish vote The euro is a threat to Danish welfare and political sovereignty The Nice treaty, the long-term visions, and political consequences are key issues The euro is a risk that threatens democracy and the EU enlargement Welfare is at stake, most opposed to the social dimension of the EMU

Note Based on interviews with campaign leaders.

58

The political parties’ campaigns

viable economy, inﬂuence on economic policies in Europe, and future prosperity as key issues. The No side “diverted” from this economic and ﬁnancial agenda and worked to achieve a broader deﬁnition of the referendum issue. It was in the No camp’s interest to get a broadly deﬁned debate and campaign issue agenda rather than a narrow economic agenda. Drawing on experiences from the 1972 referendum on entering the EEC and the 1992/1993 referendums on the Maastricht Treaty and the Edinburgh amendments, the No side was aware of the beneﬁts of setting a broadly deﬁned agenda, including more fundamental ideological questions such as the long-term future vision of Europe and cooperation between national states on the continent. With this as the agenda, there was much more opportunity for public opinion and the electorate to voice concern, reluctance, and criticism than with a campaign on the economic beneﬁts of the euro. As a second indicator of the agenda of the political parties we turn to the party broadcasts which were aired on national television during the campaign. The short programs were scheduled so that the largest parties’ programs were broadcast closest to the referendum. Looking at the issues of the party broadcasts (Table 3.4) we ﬁrst ﬁnd a high level of consistency between the issues deﬁned by the campaign leaders and the issues addressed in the party broadcasts. The Yes parties emphasized the economy while the No parties deﬁned the referendum issue broadly ranging from EU enlargement, political integration in Europe to sovereignty, national values, and the expenses related to the physical introduction of the coins and notes. Two exceptions are noteworthy. The Social Democrats deﬁned the referendum as containing multiple issues. Their campaign manager identiﬁed the economy as the key topic. The party broadcast utilized well-known party politicians as endorsers of the euro while stressing topics ranging from EU beneﬁts to the environment, the stability of the welfare state, Danish trade and export, and inﬂuence on economic policies in Europe. The second exception to the general consistency in the deﬁnition of key topics came from the Liberal Party. Their campaign manager identiﬁed the economy as the key topic, but the party broadcast was changed immediately prior to the broadcast time in order to address a very speciﬁc target group: the elderly. The party broadcast only addressed the issues of the national pension scheme and retirement beneﬁts. This decision was a reaction to the emergence of the future of pensions as a key theme in the hot phase of the campaign. The issues identiﬁed by the campaign managers and the issues addressed in the party broadcasts provide two key indicators for the agenda of the political parties during a referendum campaign. In the referendum on the euro, the Yes side wanted to “talk economy,” but was forced to respond to social welfare themes effectively initiated by the No side. Contrary to the predictions by Magleby (1989) the No side was

The Yes and No camps

59

Table 3.4 Party issue agenda II Issue Yes parties Social Democrats Radical Liberals Conservatives Center Democrats Liberal Party No parties and movements Socialist People’s Party

June Movement

Danish People’s Party People’s Movement against EU Freedom 2000b Christian People’s Party Unity List

Environment, money for welfare, currency, export, inﬂuence on decision making, welfare Welfare, EU is education, welfare, more than money. But money buys welfare Lots to thank the EU for. For business and peace. Will not inﬂuence taxes, security and DK welfare. OK to doubt, best for children, ruins in Vienna (peace), need good neighbors, East enlargement needs the euro Convincing elderly that the state pension is not endangered by the euroa Pre-pension, health care, elderly: threatened by the EU. East enlargement and ﬂexibility more important issues. Vote is identical to 1993 Don’t believe only either/or arguments. Euro part of European state-formation. Real issue is Treaty of Nice: taxes, social issues, voting rules. Euro makes enlargement with East slower! Sovereignty, Danish values/ traditions threatened, no turning back if Yes Yes to euro: more power out of DK, huge costs to introduce the euro (50 billion) – Enlargement top priority, not the euro. Risky experiment. Goodbye to independent economic policy in Denmark Undemocratic, risky

Notes Table based on party television broadcasts. a denotes that party broadcast was changed at the last moment to address the key issue of pensions. b denotes no party broadcast.

successful in establishing a broader deﬁnition of the referendum issue without strong ﬁnancial resources at their disposal.

Campaign effectiveness: why the Yes side lost Bringing the analyses of the efforts, strategies, and agendas of the political parties and movements together, it is possible to assess who was more successful. Obviously the No side came out as the most effective by winning

60

The political parties’ campaigns

this close-race referendum. As we saw in Chapter 2 (Table 2.2), the Social Democrats and Liberals did not manage to align sufﬁcient voters with the parties’ Yes policy. A number of reasons for this can be found in the campaigning of the political parties. The No side was successful in playing up long-term issues and the potential political implications behind the single currency. In essence they applied the strategy of issue diversion (Magleby 1989) by broadening the deﬁnition of the referendum issue. The Yes side was determined to focus solely on the euro as a currency and the economy. This came across as untrustworthy and it backﬁred as a simplistic “onetrick-only” strategy. The Yes campaign was late and weak in its response to the No side’s more general and ideological discussions. A number of things jeopardized the Yes side campaign. The ﬁrst was a report produced by four economic experts (referred to as “The Wisemen Report”) that was launched in the middle of the campaign and which almost discredited the economic argument entirely by concluding that the differences between a Yes and a No outcome for the Danish economy would, at least on the short run, not be discernible. Second, the strength and stability of the euro as expressed in its exchange rate against the U.S. dollar continued to decrease in a free fall during the campaign. This made arguments about the viability of the European economy untrustworthy. In the end a frustrated European Central Bank was forced to intervene and support the euro which was seen as a sign of weakness for the currency. Taken together these two facts discredited the Yes campaign. The Yes camp was strikingly unable to adjust to a new campaign mode with different issues at the fore. Finally, the Yes side failed because the political parties varied in their intensity of campaigning in the referendum campaign. The government parties (Social Democrats and Liberal Democrats) were active, but the Liberal Party (Venstre), traditionally a strong proponent of liberalism and free market economy, was less active and performed poorly in terms of getting its own voters to follow the advice endorsed by the party.

Lessons learned from the loss of the Yes camp Looking at political campaigns in general we can conclude that the Danish referendum campaign was dominated by strong elements of traditional campaigning. This, however, was concurrent with an increased electoral segmentation-thinking and a strong use of the Internet as a resource and communication tool. The Yes side was comparatively much more professionally run and spent by far the largest amount of money on campaigning. Their research included focus groups and polls and their campaigns were designed by professional ad and PR agencies. The No side was less professional and rather traditional in its approach. However, despite the limited budget, the campaign teams managed to optimize the available means and, for example, placed newspaper ads at strategic moments in the campaign.

The Yes and No camps

61

If we relate these ﬁndings to the more general and theoretical literature on the evolution of political campaigns and the increase in expenses and professionalization of politics, we must draw the conclusion that there is variation in the level of professionalization. While the trend may appear to be going in one direction, this referendum campaign suggested that we need to consider at least three contextual factors when assessing the degree to which politics is managed professionally as prescribed by the doctrines of “permanent campaigning.” First, countries may be very different in terms of campaign styles. This is a function, on the one hand, of the political system and culture and, on the other hand, of the competitiveness of the media market and the journalistic culture. Second, referendum campaigns may be different from general election campaigns. In this example, the level of expenditure was considerable but the political strategic alliances were different and this affected the sharpness and coherence of the campaign. Finally, political parties differ greatly in terms of their professional approach to campaigning. This appears to be a function of means and funding available, but even within comparable budgets the innovativeness of campaigns differed considerably. These observations put in perspective the seemingly sweeping and generally applicable conclusions about the professionalization of politics as they emerge from the literature. Variation stemming from country characteristics, the nature of the electoral context, and the composition of the party political landscape need to be taken into account.

Part III

The campaign in the news Polls, personalities, and the multi-faceted issue of the euro

4

The campaign in the news

Referendums are not always about issues of high salience or great importance to the electorate. Generally speaking, the higher the salience of the issue to the public and the more elite stakeholders involved in the campaign, the higher the visibility of the referendum in the media. Recent examples include the recall referendum in California in the fall of 2003 in which Democrat Governor Grey Davis was voted out of ofﬁce following a media frenzy around Republican candidate Arnold Schwarzenegger, former actor and current Governor of California. Such intense levels of media attention have also been seen in relation to national referendums, such as the Australian vote on sovereignty and the Irish referendum on the introduction of abortion. Referendums on issues of European integration have produced only sporadic accounts of the media coverage of the campaigns (see Jenssen et al. (1998) for an account of the 1994 referendums in Norway, Finland, and Sweden on EU membership and Siune and Svensson (1993) and Siune et al. (1994) for descriptive accounts of the previous Danish referendums in the early 1990s). The lack of systematic knowledge about the media coverage of these key events is an example of the underdeveloped understanding we have of news reporting of European political affairs (Semetko et al. 2000). Until recently there was limited scholarly research available about the way in which issues of importance to European integration were covered in the news. In the past years several studies have started to map the coverage of the EU in the news (Semetko and Valkenburg 2000; de Vreese 2002; de Vreese et al. 2001; Peter et al. 2003), but little is known about the reporting of referendum campaigns in the media. In this chapter we analyze the media coverage of the referendum campaign. We draw on almost 5,000 television news stories and 3,000 newspaper stories during the campaign to analyze the visibility of the referendum issue in the media, the presence and evaluation of key actors in the news, as well as the focus of the media on the electoral content in the form of poll-oriented news coverage. The results show a major increase in the amount of coverage of the referendum in the ﬁnal four weeks of the

66

The campaign in the news

campaign, a bias towards strategic and opinion poll-oriented news coverage, and neutral or negative evaluations of key campaign actors. We identify the differences among speciﬁc news outlets and discuss the relevance of this pattern of visibility to referendum outcomes in other national contexts. We examine how key features of journalism operate and we assess the impact of the journalistic approach on the news coverage.

EUrope in the news Previous research on media coverage of aspects of European integration has largely focused on elections for the European Parliament or key EU events. The study of the ﬁrst European Parliamentary elections in 1979 (Blumler 1983) showed that European topics did not surface on the television agenda before the ﬁnal weeks of the campaign. In relation to subsequent European elections, it has generally been noted that while the ﬁrst elections received some media coverage because of the novelty of the event, already from the second elections in 1984 and thereafter, the campaign was already “nothing special” (Leroy and Siune 1994: 52–53). One comparative study of the 1989 European election campaign concluded that the elections were given low priority in the news and that most of the news coverage was of a domestic nature, with only little reference to the European dimensions of the issues in question (Leroy and Siune 1994). In 1999 national television news programs such as, for example, the BBC and ITN in Britain, NOS and RTL in the Netherlands, TVE and Antenna3 in Spain, and RaiUno and Canale5 in Italy spent between 2 percent (the Netherlands) and 8 percent (Italy) of their news programs on the elections during the ﬁnal two weeks prior to the June 1999 elections (de Vreese et al. 2004). On average, the main national news programs in the EU member states devoted 8 percent of the news to the elections. Belgium, Britain, Germany, Ireland, the Netherlands, and Spain stand out for devoting less than 5 percent of the news to the elections during the campaign. In Austria, Denmark, Finland, France, Greece, Italy, and Sweden, between 8 percent and 13 percent of the news dealt with the European elections in the ﬁnal two weeks before election day. A second group of studies deal with key EU events, such as the recurring EU summit and, for example, the introduction of the euro. One study found that media attention to EU affairs is cyclical which meant that news about the euro vanished in the immediate aftermath of the launch. The coverage of the euro replaced other political and economic issues in the period around the introduction so that the volume of political and economic news stayed constant during the launch period and in a routine period (de Vreese et al. 2001). This suggests that major EU events have the ability to enter the national news agendas, but this does not increase the proportion of political and economic news overall. A similar pattern

The campaign in the news

67

was found in studies of EU summits in which the summit entered the news agenda, dominated it for the period of the summit, and vanished in the immediate aftermath (Semetko and Valkenburg 2000; de Vreese 2001b). A recent study investigated the coverage of EU affairs during a routine period, that is during times when there are no elections or prescheduled important events (Peter and de Vreese 2003). This study showed in most of the countries studied, EU politics is marginal in national television news and EU ofﬁcials are virtually absent from the news. However, when the EU is covered, EU politics tends to be more prominently presented than other political news. It is against this backdrop of previous studies that we can assess the media coverage of a referendum on a European integration topic. We know that there is considerable cross-national variation in terms of the visibility of European news. We know that the coverage tends to be cyclic, that is almost invisible and then, following a short peak, vanishing in the immediate aftermath of an event.

EUrope in the newsroom Journalists play an important mediating role in shaping public opinion about European integration. By choosing to cover some topics and not others journalists and news organizations play an active role in the campaign process. Moreover, inclusion of one perspective above another or the use of certain sources or viewpoints can affect the framing of a news story. In addition, critical comments and sarcastic remarks can affect the valence of the news. Obviously, however, journalists do not operate in total freedom in terms of their choices. They are constrained by a number of factors. A distinction can be made between factors internal and external to journalism that shape and inﬂuence news content (de Vreese 2003). External factors include, for example, the political system of a country – whether political conﬂicts and battles are fought in a two-party system or in a multi-party system. Additionally, in the case of the EU, it may matter whether or not a country is generally pro-European or rather EU-skeptic. Factors internal to journalism include the editorial policy of a news organization, the journalistic effort invested in covering an issue/event in terms of staff and resources, and the application of news selection criteria. Editorial approach. Formally deﬁned organizational and editorial approaches shape news content. In studies of national elections in primarily Britain but also the U.S., a theoretical distinction between a sacerdotal and a pragmatic approach to campaign coverage has been applied (Blumler 1969 in Blumler and Gurevitch 1995; Semetko et al. 1991). The distinction refers to the status attributed to elections and campaign news. In a sacerdotal approach, elections are perceived as the fundament of democracy and campaigns are considered newsworthy per se. The attitude

68

The campaign in the news

towards politicians is respectful, cautious, and reactive. In a pragmatic approach campaign news is evaluated against conventional news selection criteria and is not automatically given special attention. The pragmatic orientation implies that the “amount of time or space allocated to [political events] will be determined by strict considerations of news values, in competition with the newsworthiness of other stories” (Semetko et al. 1991: 6). In one study journalists at the BBC were found to be prudential and cautious, concerned about “ensuring that television journalism was, and would appear to be, politically beyond reproach – perhaps even politically innocuous” (Semetko et al. 1991: 53). In comparison it was suggested that journalists at the U.S. network NBC were more analytic and committed to a conventional journalistic approach “looking for events to report that would be laced with drama, conﬂict, novelty, movement and anomaly” (Semetko et al. 1991: 55). In a study of British, Danish, and Dutch news organizations’ approach to and role in reporting the 1999 European elections, it was found that though all news programs had the intention to cover the elections extensively, this was adjusted considerably during the campaign. Advance planning concerning the initiation of speciﬁc issues was often not implemented in the actual coverage. A commonality between the news programs was the pragmatic approach to the elections – implying that politics (including the elections) was not newsworthy per se, but that political events and issues competed against other topics in the allocation of time in the news. The event was not considered to have sufﬁcient intrinsic importance or interest and the European Parliament was evaluated critically and did not enjoy any privileges as a political authority. In general, the elections were not “up-graded” and events in the campaign were mostly evaluated according to normal news selection criteria (de Vreese 2003). Though all news programs during the 1999 European Parliament elections were pragmatic in their approach, this had different implications. Some news programs chose to neglect the elections due to an editorial assessment of the event as non-newsworthy (for example, Dutch NOS Journaal and RTL Nieuws), while others set their own agenda without paying attention to the political party agenda (for example, Danish DR TV-Avisen). Yet other news programs opted to make the anticipated voter apathy a key theme in the coverage (for example, British BBC and ITN, and Danish TV2). These ﬁndings suggest that the approach taken by television journalists in the coverage of European affairs has changed considerably since the ﬁrst EP elections in 1979. Noël-Aranda (1983) concluded that broadcasters during the 1979 campaign were cautious and adhered largely to the agenda put forward by politicians. This was no longer an appropriate description when assessing the approach taken by broadcasters in the 1999 elections. News selection criteria. A second aspect of journalism that shapes news is the application of news selection criteria. An important indicator is

The campaign in the news

69

whether standard criteria for evaluation of the newsworthiness of events and issues are applied or whether special criteria are applied for evaluation of campaign stories. While news values are not organizationally deﬁned, as suggested by, for example, Galtung and Ruge (1965), the news organization is the place where these are manifest and translated into daily production routines. Shoemaker and Reese (1996) synthesize work on news values and suggest: a) prominence/importance, b) human interest, c) conﬂict/controversy, d) the unusual, e) timeliness, and f) proximity as the most important news values to be distilled from both journalistic practice and literature about journalism. During the 1999 European Parliamentary campaign, the different news programs varied in the degree to which they adapted the political agenda. In Britain, the agenda of the political parties was important for the building of the news agenda. Most notably, the Conservatives received extensive coverage and their slogan “In Europe, not run by Europe” was inﬂuential. Dutch NOS Journaal chose to almost entirely ignore the elections. This was a deliberate editorial choice not to deﬁne the elections as newsworthy. Danish TV-Avisen covered the elections rather extensively and assumed a clear agenda-setting proﬁle in the network’s focus on the fraud issue (de Vreese 2003). The elections were most visible in Denmark and Britain and invisible in the Netherlands. The low visibility of the European elections in the Dutch news is in line with observations made about the 1979 campaign. For the 1979 election, it was suggested that the Netherlands was the only country in which “neither broadcasters nor parties felt very concerned about promoting a European consciousness” through television which is one explanation for the absence in coverage (Noël-Aranda 1983: 92). The argument put forward by Noël-Aranda (1983) suggests that news coverage of an issue on which there is widespread consensus will be only marginal. This seems like a plausible explanation and it is in line with other research on, for example, news values. Studies in different countries have suggested that presence of conﬂict in an issue is an important common criterion for selecting news (for example, Eilders 1997; McManus 1994; Shoemaker and Reese 1996). Editorial approach and the application of news selection criteria – both factors internal to the news organization – inﬂuence the visibility and priority given to an event/issue in the news. The dimensions also give direction to the news agenda. Agenda-setting addresses the relationship between political actors, the news media, and the electorate (see Chapter 6 for an investigation of media agenda-setting in a referendum campaign). Although the majority of studies on agenda-setting have focused on the relationship between the media and the public agenda, an equally important, but relatively neglected area is the dynamics of how the media agenda is formed, agenda-building, and the role news organizations and journalists play in this process. Lang and Lang (1981: 278) concluded that “the

70

The campaign in the news

whole question of how issues originate is side-stepped, nor is there any recognition of the process through which agendas are built.” One reason for the less central role of this strand of research is reluctance on the part of news organizations to open up their newsroom and share experiences and insights with communication scholars. Access is a crucial factor which may have limited the volume and impact of this type of research (Blumler and Gurevitch 1998; Holtz-Bacha 1999; Semetko 1996a). Previous studies of national election campaigns across Europe in the 1970s (for example, Asp 1983; Siune and Borre 1975) and 1980s (for example, Blumler and Gurevitch 1989; Hjarvard 1999; Semetko et al. 1991) suggested that European broadcasters have traditionally covered elections in a respectful and cautious way, adhering largely to the agenda of parties and politicians. While this role may be labeled “agenda-sending,” a reference to the media passing on the priorities of news sources; in the case of elections, these are often candidates, party leaders or party spokespersons. Research based on elections in the 1990s and more recently suggests that this may no longer be the case (see, for example, Blumler and Gurevitch 1998; Hjarvard 1995; Norris et al. 1999; de Vreese 2001a). As in the 1999 European Parliament elections, news organizations have been found to be more pro-active in their coverage of politics and to exert more discretion when choosing which stories to bring and how to cover these issues. This role may be more appropriately labeled “agenda-setting,” a reference to the media’s primary role in shaping the news agenda. The role of journalists and news organizations is unexplored territory in relation to referendums about European integration. We build on previous research in relation to national and European elections to investigate the journalistic approach to a referendum and its effect on news coverage.

Key questions This chapter addresses the coverage of a referendum in the media and the journalistic approach applied to this coverage. The ﬁrst part of the analysis investigates the visibility of the referendum issue in the media. There are no comparable studies to assess whether the visibility of this referendum exceeds or falls below the standard amount of coverage that a referendum generates. We do know that other studies suggest that the EU is peripheral in national television news, with occasional peaks (de Vreese 2002; Norris 2000). Nothing is known, however, about the visibility of a referendum campaign on an issue of European integration. The second part of the analysis deals with the presence and depiction of actors in the news. Actors are broadly understood as persons, organizations, institutions, political parties, candidates, etc. The visibility of political actors in the news is a necessary condition for the functioning of political representation in national democracies and this has been addressed in studies of national election campaigns (for example, Kleinni-

The campaign in the news

71

jenhuis et al. 2001; Norris et al. 1999; Schönbach et al. 2001). We know little about the visibility of EU actors (de Vreese 2002). There is only limited evidence of how visible EU actors are or whether they are more or less visible than other (domestic) political actors (see de Vreese et al. (2004) for an EU-wide analysis of the visibility of actors during the 1999 European elections), and we know even less about the presence of political actors in the news during a referendum campaign. We go beyond investigating the mere visibility of the different actors in the news to also assess the tone of the news coverage of these actors. This is potentially crucial for the public evaluation of these actors (see Chapters 8 and 9). Evaluations of politicians have been investigated in content analyses of national politics (for example, Herr 2002; Kepplinger 1998; McCombs et al. 2000; Semetko et al. 1991). Moreover, many of these studies suggest that the news coverage of candidates may signiﬁcantly affect audiences’ perceptions of these candidates (for example, Kiousis et al. 1999; McCombs et al. 2000) and under certain conditions also affect voting preferences and behavior (Herr 2002). Evaluations in the news are, when predominantly and permanently negative, considered to contribute to a decrease of citizens’ participation in democratic processes (for example, Kepplinger 1998). Recent research has focused speciﬁcally on the evaluation of EU representatives in news. It suggested that EU representatives are generally depicted neutrally, but when they are evaluated this tends to be in a negative direction (Peter et al. 2003; de Vreese 2002). This pattern, moreover, is similar to the evaluation of domestic political actors (de Vreese 2002). The third dimension of the content analysis deals speciﬁcally with the role that opinion polls play in the news coverage of the referendum. Election time is poll time. In the content analysis we investigate the visibility of polls and assess the quality poll reporting in the news media coverage of the referendum campaign. We draw on a number of indicators for the quality of poll reporting that pertain to the presence of methodological information such as provision of question wording, population deﬁnition, sample size, interview method, and data collection ﬁeld dates. Finally, we address the organization and journalistic approach to the referendum campaign as discussed above. We address the factors internal and external to journalism that shape and inﬂuence news content that were discussed at the outset of this chapter.

The visibility of the campaign in television and newspapers To gain an initial idea of the volume of the coverage of the referendum during the campaign we ﬁrst take a look at television news. Drawing on our content analysis of 5,000 news stories from the outset of the campaign in March until Election Day in September 2000, we found that the referendum, on average, occupied about 10 percent of the news, albeit

72

The campaign in the news

% of news stories about euro referendum/EU

somewhat less during the summer months in the middle of the campaign (see Figure 4.1). In the ﬁnal month of the campaign the visibility of referendum increases to take up more than one-quarter of the news. The analysis suggests that the 2000 referendum campaign was visible in the news media. Compared to, for example, the 1999 European elections, the referendum was highly visible. During the ﬁnal two weeks leading up to the 1999 European elections, about 10 percent of the news in Denmark was devoted to the elections (de Vreese 2002). In the comparable two weeks leading up to the 2000 referendum, this ﬁgure was higher than 30 percent. Media attention of this magnitude to a referendum suggests the importance of the topic and the potential domestic political implications. To get a general idea of the news environment during the campaign, we turn to the overall television news agenda in the period from March– September. As Figure 4.1. shows, the euro issue increases from about 10 percent of the news early in the campaign to around 25 percent in the ﬁnal month (see also second row in Table 4.1). The “economy” took up between 5–10 percent of the news throughout the campaign and other political news varied from 22 percent (TV1 September) to 40 percent (TV1 July) (see Table 4.1). Did the referendum increase the overall percentage share of news devoted to political and economic topics? Or did other political and economic news get bumped off the screen when the referendum became more visible, so that the overall percentage share the news devoted to political and economic topics remained the same? A previous study found that 30

TV1 (DR) TV2

25 20 15 10 5 0 March

April

May

June

July

August

September

Figure 4.1 Visibility of the referendum in television news. Note Percentage of news stories about referendum on TV1 DR 9 o’clock news and TV2 7 o’clock news. A total of 4,953 stories are analyzed (TV1 n ⫽ 2,578; TV2 n ⫽ 2,375) between March and September 2000.

May

June

July

100.0

9.8 6.3 36.7 0.3 3.1 2.8 3.1 0.0 4.5 16.4 2.8 0.0 7.3 6.6

100.0

8.6 5.4 37.4 0.4 4.3 1.8 0.4 0.0 4.0 15.5 10.4 0.0 7.2 4.7

100.0

9.4 10.2 29.6 1.6 3.1 3.4 2.9 1.8 4.2 11.0 8.9 0.5 7.3 6.0 100.0

8.8 6.4 29.3 1.1 2.8 0.0 1.4 0.0 4.9 16.3 13.1 0.7 9.2 6.0 100.0

11.9 8.1 36.0 2.1 2.9 1.0 0.5 1.0 2.1 13.5 6.8 2.3 7.8 3.6 100.0

8.6 3.7 36.9 3.0 1.7 1.7 1.0 1.7 3.0 14.0 9.0 2.7 8.0 5.3 100.0

10.9 7.7 31.1 2.2 1.5 2.5 1.0 1.7 2.7 19.2 7.7 2.0 7.5 2.2 100.0

7.7 6.3 32.1 0.3 1.1 1.6 1.9 0.3 4.7 20.9 8.2 3.0 7.4 4.4 100.0

7.4 1.8 40.3 1.8 3.8 6.2 0.6 0.3 7.6 13.8 6.2 0.9 8.2 1.2 100.0

5.3 6.6 28.6 2.7 3.4 6.9 1.1 0.3 8.5 14.6 5.8 1.3 8.0 6.9

Note Base is all television news between March and September 2000. Cell entries are percentage of all stories, n ⫽ 4.953).

Total

Euro/EU Economy Politics Unemployment Welfare Environment Immigration Norms/values Accident/disasters Crime Human interest Sport Weather Other

April

August

September

100.0

7.3 5.2 30.7 1.8 3.7 3.4 2.6 4.2 8.7 15.7 5.0 0.8 7.1 3.7

100.0

8.2 3.3 27.4 0.0 2.7 3.0 4.9 0.3 9.6 15.9 8.8 1.1 7.9 6.8

100.0

26.1 7.7 23.9 2.0 2.0 4.0 2.2 1.5 5.0 9.7 2.5 3.2 6.5 3.7

100.0

23.8 9.9 22.0 0.0 1.5 2.5 6.6 0.3 2.5 9.4 3.5 6.1 7.6 4.3

TV1 TV2 TV1 TV2 TV1 TV2 TV1 TV2 TV1 TV2 TV1 TV2 TV1 TV2 (n⫽268) (n⫽278) (n⫽382) (n⫽283) (n⫽385) (n⫽301) (n⫽402) (n⫽364) (n⫽340) (n⫽377) (n⫽381) (n⫽365) (n⫽402) (n⫽395)

March

Table 4.1 Television news agenda during referendum campaign

74

The campaign in the news

when the euro was introduced, the event itself was visible in the news but this news displaced coverage of other political and economic topics, so that the overall share of the program devoted to political and economic news remained constant (de Vreese et al. 2001). In Figure 4.2 we illustrate the percentage share of news about the referendum, other political and economic news, and news about topics such as crime and human interest, on the two television channels in the ﬁnal two months during the campaign. We clearly see that the share of news about the referendum as well as other political and economic news increased from around 40 percent in both news programs in August to about 55 percent in both programs in September. The share of economic and political news remained stable in both programs thus suggesting that the increase was accounted for by the increase in the coverage of the referendum, indicated by the black bars in Figure 4.2. In the 2000 euro referendum then, we see the unusual result that the overall percentage share of political and economic news increased. The press Turning to the content analysis of the newspaper coverage, we identiﬁed 3,201 items dealing with the referendum during the campaign. We distinguish among front-page articles, regular articles, newspaper features and reportages, letters to the Editor, Editorials and Commentaries, advertisements as well as information boxes and cartoons. Table 4.2 summarizes the distribution of these more than 3,000 items dealing with the referendum.

100

Human interest Crime Politics Economy Euro/EU

90 80 70

%

60 50 40 30 20 10 0 TV1 August

TV1 September

TV2 August

TV2 September

Figure 4.2 Share of news topics on television in August and September 2000.

The campaign in the news

75

Table 4.2 Visibility referendum in daily newspapers Politiken Berlingske Tidende (Yes) (Yes) (n ⫽ 842) (n ⫽ 779) Front page article/ bullet Inside the paper Article Reportage/analysis Portrait/interview Letter to the Editor Editorial/commentary Advertisement Information box/quiz Other (e.g., cartoon, bullet, graphs)

Jyllands- B.T. Posten (Yes) (Yes) (n ⫽ 771) (n ⫽ 367)

EkstraBladet (No) (n ⫽ 523)

47

43

49

7

13

256 13 40 150 80 72 78 129

246 31 43 128 81 73 80 84

248 8 43 216 55 62 45 70

72 1 8 159 30 47 13 30

84 1 35 223 77 42 13 35

Note Period is August 28–September 27, 2000. Cell entries are counts.

In total more than 150 references to the referendum were found on the front pages of the ﬁve dailies included in the sample in the ﬁnal month of the campaign. The number of regular news articles totaled more than 800 which means that each newspaper, on average, carried one article about the referendum each day during the entire period from when the referendum was called in March until election day in September. The number of Letters to the Editor (876) and editorials (323) was also very high, which is an indication of the high salience of the topic also to the electorate which engaged in considerable public debate. To gain an initial idea about the topics and the tone of the newspaper coverage, we collected key headlines from newspapers during the ﬁnal month of the campaign (see Table 4.3). The number of front-page headlines varied signiﬁcantly between the different newspapers. Politiken, Berlingske Tidende, and JyllandsPosten are all broadsheet quality newspapers while EkstraBladet and BT are tabloid with a format comparable to the popular British tabloid the Sun. The front pages of the broadsheet newspapers were dominated by the referendum in the entire month leading up to the referendum. The headlines centered around four themes: 1 2 3 4

Contradictory announcements of the costs of a Yes or a No, Human interest headlines about key campaign leaders, The exchange rate and strength of the euro, Disagreement on Yes side.

The announcements of the costs of a Yes or a No were found in several newspapers and spread throughout the campaign. They ranged from

–

–

–

–

–

–

–

–

–

–

–

–

28/8

29/8

30/8

31/8

01/9

02/9

03/9

04/9

05/9

06/9

07/9

–

–

–

–

–

–

–

–

–

–

B.T.

Date EkstraBladet

–

–

–

A Banker says no

Berlingske Tidende

–

Euro takes another dive

–

–

Britons support No campaign

Experts: “Nordic welfare will survive the euro”

–

A Yes is three times more expensive than a No

Dead race on the euro

Lawyers: Information on the – euro is biased

–

Euro hits rock-bottom. – Danish People’s Party lowproﬁle campaign not to scare No voters

–

No will cost 5 billion

Inﬂation too high in Denmark to join the euro

Undecided vote Yes after deliberate poll

Politiken

Table 4.3 Key newspaper headlines during the campaign

–

–

–

Brits row over support to Danish No campaign

–

Leader Danish People’s Party: “Maybe euro one day”

–

EU defense plans: full speed ahead

Euro No will cost billions

–

No camp: “No more Union”

JyllandsPosten

–

–

–

–

–

Minister’s “euro- – bomb”: Pension in danger!

–

–

10/9

11/9

12/9

13/9

14/9

15/9

16/9

17/9

–

–

–

–

–

–

–

–

–

09/9

–

–

08/9

Nyrup (PM): “A Yes secures Danish welfare”

–

Tietmeyer (President German Bank): “Not possible to leave euro once you are in”

–

–

–

Emergency aid to the euro

Duisenberg rejects United States of Europe; Yes parties join in common campaign

Nyrup’s (PM) tries to secure pension

continued

Harsh criticism of Nyrup’s (PM) pension guarantee

Panic on the Yes side

Yes side join forces to secure a Yes

Moeller (Director “A boring campaign” says Denmark’s largest company) political analysts says Yes

–

Gallup: “A No is emerging”

–

–

–

–

Foreign Minister: No side is Pension guaranteed after dishonest; Danish Krone euro under pressure; No overtakes Yes in the polls

–

–

–

Austria case affects Prime Minister

–

–

–

Jelved (Economy Minister) promises strong euro

–

Former Foreign – Secretary: Prime Minster is to blame for No to the euro

Leader Danish – People’s Party: “Thanks for helping out Poul (PM)”

–

19/9

20/9

21/9

22/9

–

Disagreement on the costs of a No

A No will cost 20,000 jobs

Nyrup’s (PM) campaign heavily criticized

Euro politicians Harsh criticism on pension alert: Karen guarantee (domestic affairs minister) worse than Haider!

Member EU Nyrup (PM) demands Parliament: Socialist Party to join in on excess in greed: the pension guarantee 60 days EXTRA vacation

Brutal Prime Minister delivers harsh criticism of Leader Socialist Party

18/9

Politiken

B.T.

Date EkstraBladet

Table 4.3 continued

New strategy in the ﬁnal hot phase: “EU can be used to improve environment”

Total disagreement between Yes parties

A No will cost 20.000 jobs; Nyrup (PM) forced to change strategy

Fogh (Liberal leader): “Nyrup (PM) harms Yes side” ; Where to go after a No

Predictions of clear No

Berlingske Tidende

East European countries: Tired of the Danish No voters

Criticism: “No side campaign based on lies and exaggerations”

Nyrup (PM) gives up on letter to other EU countries to secure pension

A campaign on emotions; Yes side guarantee: “Pension is secure”

Guaranteed pension

JyllandsPosten

–

–

Leader Socialist Party: “Speak the truth Nyrup (PM)”

Nyrup (PM) is unclear about Danish EU policy in case of No

–

Yes candidates preparing for a loss: Ready to face a No

23/9

24/9

25/9

26/9

27/9

28/9

Final poll: 47 percent– 47 percent

Yes side: ﬁnancial cuts if it’s No

Yes gains support: A dead race

Support for weak euro

Yes or No: Today Denmark decides if the Krone is to be replaced by the common currency

Today the Danes decide; Gallup: “Most exciting race in decades”

New euro referendum again soon if it is a No

EU’s competence and power Interest rate will go up if No must be clear

Yes side improves in the polls

Not clear how to proceed in case of a No

Banks intervene to stop euro freefalling

Yes, B.T. Hard ﬁnal debate: “Behave recommends, please!” Denmark MUST be in

–

–

Dead race: Yes side 3 percent behind No side

Nyrup’s chaos management

Exit poll not popular

–

No side: “Maybe euro later” Poll: “Danish Yes will give Swedish Yes”

A No can split the EU in two

Dead race on the euro

Trade Union: A No will be costly; World Bank buys euro to save it

80

The campaign in the news

“A No will cost 5 million” (Politiken, August 30, 2000), via “A No will cost 20,000 jobs” (Berlingske Tidende, September 20, 2000) to “Nordic welfare will survive the euro” (Berlingske Tidende, September 7, 2000). The headlines pertaining to the performance of the campaign leaders were primarily found in the tabloid newspapers and centered around Poul Nyrop Rasmussen, then Prime Minister: “Brutal Prime Minister” (EkstraBladet, September 18, 2000), “Thanks for helping out Poul” (EkstraBladet, September 21, 2000), and “Speak the truth Poul” (EkstraBladet, September 25, 2000). The strength of the single currency and the exchange rate against the U.S. dollar emerged in headlines such as “Euro hits rockbottom” (Politiken, September 1, 2000), “First aid to the euro” (JyllandsPosten, September 8, 2000), and “Bank intervenes to stop euro freefalling” (Politiken, September 23, 2000). Finally, the campaign style and strategy of the Yes camp hit the headlines with stories such as “Yes side joins forces to secure Yes” (JyllandsPosten, September 15, 2000) and “Nyrop’s chaos management” (B.T., September 23, 2000).

The centrality of polls We now turn to a more substantive aspect of the campaign coverage: the use of opinion polls in news. Polls are the essence of campaigning. The question of who is ahead, which camp, which party, which candidate is at the core. News coverage of U.S. election campaigns is saturated with polls and expressions of public opinion and the use and importance of polls has grown explosively (Lavrakas and Traugott 2000). Kerbel (1994) described the network coverage of the 1992 U.S. presidential campaign as a “poll orgy.” Polls are a prominent feature of election news reporting and together with “people-on-the-street”-reporting, they are the two most common expressions of public opinion which drive “horse race” news coverage (Larson 1999, 2000). Patterson (1993) points to the increasing importance of indicators of who is ahead in the political race to give form to news of politics as a strategic game in his analysis of the evolution of presidential campaign news coverage since 1960. Given the centrality of polls and media coverage of public opinion in campaigns, two questions arise: How important are polls to campaign coverage? To what extent are polls reported in a way that enables citizens to judge the quality of the poll and the interpretations derived from it? Many critics claim that poll coverage is superﬁcial and inaccurate. The empirical evidence for this claim, however, is missing as only “very few systematic studies of poll coverage have been published” (Smith and Verrall 1985: 59). In one study of television election coverage Smith and Verrall (1985: 77) concluded that “simpliﬁcation was achieved by omitting methodological details, reporting results in general terms, and committing the errors of overgeneralization, ambiguous comparison, and evaluative description.”

The campaign in the news

81

Brettschneider (1997) echoed this in an investigation of the news media reporting of poll results and comments on public opinion prior to national German elections from 1980 to 1994. Drawing on indicators for both the quantity and formal and substantial quality of the press coverage of the polls, he concluded that the quantity of reports on public opinion had increased over the years, but that the formal quality leaves much to be desired. Similarly, a study of the television and newspaper reporting of pre-election polls in the 1997 Canadian elections examining both the emphasis given to polls and the quality of reporting of methodological information similarly found that the media relied heavily on polls to chart the dynamics of the campaign, but that polls were often reported with little or no information about methodology (Andersen 2000). We investigate the visibility of polls and assess the quality poll reporting in the news media coverage of the referendum campaign. We draw on a number of indicators for the quality of poll reporting. These indicators are partly derived from Smith and Verrall (1985) and pertain to the presence of methodological information in reports of poll results such as provision of question wording, population deﬁnition, sample size, interview method, and data collection ﬁeld dates. The overall conclusion from the analysis of the ﬁve largest and most widely read daily newspapers as well as the two most widely viewed main evening news programs from the public and the private broadcasters is that polls and expressions of public opinion played an important role in the referendum campaign coverage. More than one-third of the stories referred to polls and the Yes and No sides’ standing (see Table 4.4). With many news organizations joining forces with renowned pollsters in Table 4.4 Visibility and quality of reporting of polls Television

Newspapers

TV1 (DR) (n ⫽ 76)

TV2 (n ⫽ 79)

Broadsheet (n ⫽ 160)

Stories mentioning a poll Stories referring to speciﬁc poll

32.9 17.1

36.7 15.2

34.4 23.1

Of the stories dealing with a speciﬁc poll Pollster identiﬁed Population identiﬁed Sample size identiﬁed Data collection method identiﬁed Error margin identiﬁed Field work dates identiﬁed Question wording mentioned

92.3 15.4 15.4 7.6 23.1 38.5 0.0

75.0 16.6 61.5 8.3 25.0 33.3 38.5

94.6 2.7 91.9 83.8 83.8 86.5 86.5

Note Television and Newspapers, August 28–September 27, 2000. n is number of stories about the referendum on TV1 and TV2 and on the front pages of the broadsheet newspapers.

82

The campaign in the news

conducting regular, some even daily, polls, the poll coverage has become a news beat of its own (Kagay 2000). However, while polls may seem to have been a prominent feature of the coverage, comparatively speaking the share of the coverage focused on the horse race aspect of the referendum is much less than what was found in, for example, the 2000 U.S. election campaign where more than 70 percent of the news mentioned the horse race (Patterson 2002). In terms of the quality of poll reporting, newspapers did a better job than television in providing methodological information. This ﬁnding has resonance in previous analyses of the features and quality of poll reporting which also found newspapers to be superior to television (Hardmeier 1999). By applying simple layout tools such as adding high-density information boxes to the poll, Danish newspapers were able to provide elaborate technical information. Comparatively speaking, Danish news media overall, however, seemed to do well in providing information about the poll which is required to evaluate the quality of the information and the accuracy of the conclusions drawn. In most cases all news outlets, both television and the press, provided some methodological information which is a far better record than what an analysis of the Australian television election poll coverage found (Smith and Verrall 1985). There was considerable awareness and anxiety about poll-driven coverage, also on the side of the news organizations. Given the close race of the referendum, minor sample biases and question wording differences at the different pollsters, led to contradictory poll outcomes about who was taking the lead in the referendum. TV1 issued an internal memo to the journalists covering the campaign warning to be cautious when reporting polls. This memo also included a short section on how to critically evaluate polls, including the ones commissioned by the network itself. This caution was echoed at TV2 where there was considerable fear of reporting ﬂawed polls. TV2 played an important role with respect to polls in the referendum. The network decided to release two exit polls during Election Day when the voting stations were still open. This decision is in accordance with Danish law, but it was a breach of the previous “gentleman’s agreement” not to publish exit polls before the closing of all voting booths. The intention to publish these polls led to questions in Parliament and an initiative, in vain, to prohibit the publication. The Editor-in-Chief of TV2, responsible for the decision to go ahead and publish the exit poll, referred to the situation in the U.S. when the exit polls from the East Coast were published prior to closing time on the West Coast. In addition he emphasized the potentially mobilizing effect of publishing exit polls with the polling stations still open: If there is any effect of this [publishing the exit polls], it will only be to animate some people to go out and vote. And that should not be a problem. The idea that people are affected by the poll suggests that

The campaign in the news

83

people are so stupid and easy to inﬂuence that they turn out to vote No, even though they support a Yes, just to adjust the current standing. It does not work like this. The awareness of the pitfalls of poll reporting in a close race did not translate into modesty in terms of the number of used polls. As discussed above, polls were a substantial part of the referendum coverage in the news. While the literature is divided on the issue of whether polls inﬂuence voter mobilization and vote choice (see de Vreese and Semetko (2002b) for a summary of this literature), we investigate in Chapter 7 how the proliferation of polls in the news affected citizens’ perceptions of the campaign.

The importance of personalities: visibility of party leaders in the campaign The literature on campaigning and the media consistently points out the importance of individual politicians and candidates in the news. We therefore analyzed the presence and depiction of party leaders and other candidates in the coverage of the referendum campaign. The presence of political actors is necessary for a viable concept of political representation. While the visibility and evaluation of actors in the news in national election campaigns has been studied often, the presence of political actors during a referendum campaign, however, is unexplored territory. As Table 4.5 shows the Prime Minister and other members of his Social Table 4.5 Political actors in the news Political actor

Frequency actor in the news

Prime Minister Poul Nyrup Other members of the Social Democrats Finance Minister Marianne Jelved Other members of the Liberal Democrats Anders Fogh Rasmussen Other members of the Liberals Bendt Bendtsen Other members of the Conservatives Socialist People’s Party Danish People’s Party Christian People’s Party June Movement People’s Movement EU-level actor

78 105 19 16 31 20 13 22 74 66 26 29 7 30

Total

536

Note Television news, March–September, both channels.

84

The campaign in the news

Democratic Party were the most visible actors in the news. Other highly visible actors included members from the No parties, including the Socialist Party and the Danish People’s Party. As argued above, the potential advantage of visibility in the news for a political candidate is conditional upon the tone of the coverage. After all, it is only an asset if the coverage is not negative. A news story may portray an actor either neutrally (no evaluations), favorably, mixed, or unfavorably. Based on investigations of the evaluation of politicians in content analyses of news about domestic politics, we expect that evaluations are primarily absent (i.e. the news is neutral), but when evaluations are present we expect to ﬁnd that they are mostly negative. For each actor the visibility and average tone was assessed by adding the number of negative evaluations subtracted by the number of positive evaluations divided by the total number of evaluations.1 In the ﬁnal month of the campaign, the analysis of the visibility and the evaluation of the most important political candidates and actors in the news showed that all actors received either neutral or negative coverage (Table 4.6). On the scale ranging from ⫺1 to ⫹1, most actors received a balanced mixture of positive and negative evaluations (resulting in a mean around zero) or a slightly negative overall evaluation (ranging from ⫺0.10 to ⫺0.30). The government as an entity, the two key actors representing the government, Prime Minister Poul Nyrup Rasmussen and Finance Minster, Marianne Jelved, and other members of their political parties all received negative coverage in every media outlet. The most negative coverage of these actors appeared in the press and on TV2 news while the least unfavorable (albeit still negative) evaluations appeared on TV1. Other members of the Yes camp such as Anders Fogh Rasmussen (Liberal) and Bendt Bendtsen (Conservatives) were less visible in the news and received, on average, neutral coverage in all outlets. Members of the No camp were less visible than government actors in the ﬁnal month of the campaign. No camp actors received modestly negative news coverage, but this was signiﬁcantly less negative than the coverage of the Yes camp actors. For example, on TV2 Prime Minister Nyrop Rasmussen received a mean evaluation of m ⫽ ⫺0.26 and Pia Kjaersgaard, leader of the Danish People’s Party and the most visible No camp campaigner received a mean evaluation of m ⫽ ⫺0.13.

Covering the euro referendum: a challenge for political journalism To get an impression of the constraints and challenges that journalists face when covering a referendum, interviews were held with Editors-in-Chief and key political reporters at the two national television news programs. The visibility of an election in the news as well as the choice of topics and

The campaign in the news

85

news sources are, at least in part, attributable to the ofﬁcial policy of news organizations, the daily logistic challenges, the application of news selection criteria, and the approach to politics (Blumler and Gurevitch 2001). Above we outlined how the application of news selection criteria may differ from news outlet to news outlet and between elections. Additionally we summarized recent research from national elections that suggests that broadcasters today are less sacerdotal – that is less cautious and reactive – and more pragmatic – that is selective and proactive in their coverage. We investigated the news production process by analyzing the editorial approach to the referendum, the news organizations’ application of news selection criteria, and the implementation of balance in the reporting of the referendum. We ﬁrst turn to the organization of the news coverage. Organization. In the case of the referendum on the introduction of the euro, none of the news programs implemented a special segment in their bulletin which has been tradition during, for example, national election campaigns. The referendum was covered within an already existing triadic structure of political and economic EU coverage. The studio headquarters and central newsroom cooperated with the political units in the Parliament, Christiansborg in the center of Copenhagen, and the news organizations’ Brussels desk. However, a number of additional initiatives were taken. Both news programs made additional funding available for the referendum news coverage. TV1 conducted a number of surveys that – in addition to identifying public support for the euro – also assessed which issues the public considered important in relation to the referendum. TV2 created a “euro group” that consisted of journalists from the Parliamentary unit and journalists from the domestic desk. News executives at TV1 produced an internal document outlining the guidelines for its journalists in the coverage of the referendum. This document deﬁned the journalistic approach to the referendum and listed a number of criteria for news stories to fulﬁll. One criterion deﬁned by the news executives was a rather classical information-provisional one. According to the Editor-in-Chief of TV1: We acknowledge the rather substantial need for basic information from our audience. We therefore prioritize content more than for example “who is ahead” in the coverage. A similar approach was found at TV2 where there was a short daily item with audience members posing questions about the referendum and the euro to a senior political editor. This rather old-fashioned and pedagogical television item was a conscious editorial choice. Editorial approach. The two news programs differed slightly in their approach to the referendum. TV1 assumed a more pro-active policy. It conducted its own surveys and played an initiating role in contributing to the agenda of the referendum. The Editor-in-Chief:

– 17 27 3 2 0 6 5 3 1 10 0 6 5 1 2

No actors Leader of the Socialist People’s Party, Holger K. Nielsen Other members of the Socialists People’s Party Leader of the People’s Party, Pia Kjaersgaard Other members of the People’s Party Leader of the Christian People’s Party, Jann Sjursen Other members of the Christian People’s Party

⫺0.10 – 0.00 ⫺0.20 0.00 0.00

– ⫺0.18 ⫺0.11 0.00 0.00 – 0.00 0.00 0.00 0.00 17 5 8 7 0 0

– 19 15 8 2 2 4 1 1 5 ⫺0.18 ⫺0.20 ⫺0.13 0.00 – –

– ⫺0.26 ⫺0.20 ⫺0.25 0.00 0.00 0.00 0.00 0.00 0.00

Number of Mean mentions evaluation

Number of Mean mentions evaluation

Yes actors Government Prime Minister, Poul Nyrup Rasmussen Other members of the Social Democrats Minister of Economy, Marianne Jelved Minister of Foreign Affairs, Niels Helveg Petersen Other members of the Radical Liberals Leader of the Liberal Party, Anders F. Rasmussen Other members of the Liberal Party Leader of Conservatives, Bendt Bendtsen Other members of the Conservatives

TV2 (n ⫽ 79)

TV1 (n ⫽ 76)

Table 4.6 Visibility and evaluation of political actors in the ﬁnal month of the campaign

18 7 4 7 2 0

7 27 21 13 7 2 7 6 4 7

⫺0.16 ⫺0.14 ⫺0.25 ⫺0.14 0.00 –

⫺0.43 ⫺0.44 ⫺0.29 ⫺0.23 0.00 0.00 0.00 0.00 0.00 0.00

Number of Mean mentions evaluation

Newspapers (n ⫽ 79)

3 0 1 15 25

European institutions European Central Bank European Parliament European Commission Experts Voters and citizens ⫺0.33 – 0.00 ⫺0.06 0.00

– – ⫺0.50 0.00 0.00 – – 0 0 0 11 47

4 2 5 1 0 0 0 – – – 0.00 0.00

0.00 0.00 ⫺0.20 0.00 – – – 3 2 1 35 2

1 2 3 4 0 2 0

0.00 –1.00 0.00 ⫹0.02 0.00

0.00 ⫹0.50 0.00 ⫹0.25 – ⫺0.50 –

Note Mean evaluation of the actors is based on a scale ranging from –1.00 (consistently negative) 0 (neutral) to ⫹1.00 (consistently positive). An actor in a news story is deﬁned as a person/institution in the story who is mentioned at least twice (newspapers) or mentioned and seen at least once (television).

0 0 2 1 2 0 0

Leader of the Unity List, Frank Aaen Other members of the Unity List Leader of the June Movement, Jens Peter Bonde Leader of the June Movement, Drude Dahlerup Other members of the June Movement Leader of the People’s Movement, Ole Krarup Other members of the People’s Party

88

The campaign in the news We wanted to set an agenda that had as a starting point some of the themes that we know are important to the voters, the Danes, our audience. These themes may not necessarily be at the top of the politicians’ agenda. We have implemented this to some extent, though obviously we have at times followed the political agenda if major issues surfaced on a particular day.

This approach resembles TV1’s approach to the European elections where the news agenda was partially inﬂuenced by the outcome of a survey deﬁning the audience’s priorities in relation to the election (de Vreese 2003a). According to the Editor-in-Chief, the approach was considerably different though: In the European election campaign we chose to put a very critical focus on one particular topic: fraud and corruption. That resulted in a perhaps negative depiction of the system which the electorate was now voting for. For the referendum, our surveys meant that we addressed a wider range of issues, such as the EU enlargement and the welfare state. The two news programs differed in their approach to initiating topics. While TV1 was rather pro-active, there was reluctance towards this at TV2. Both the Senior Political Editor who said, We are not here to create news, but to report it. and the Editor-in-Chief of TV2 agreed on this: We have not deﬁned the agenda. I do not believe it is our task to do so either. [. . .] Our “agenda-setting” efforts are little more than trying to translate what we believe the public wants to hear about. In terms of the application of news selection criteria, the two news programs hardly differed and both devoted extra attention to the referendum by default. TV1 gave the referendum extra attention through, for example, bringing reports from its international network of correspondents. TV2 gave the referendum no extra attention until about a month prior to the election when it was upgraded in the selection. The TV2 Editor-in-Chief commented on this upgrade in the ﬁnal weeks: The use of criteria in election is difﬁcult. The criteria applied daily are dissolved when you enter an election or referendum time. [. . .] Some of that is just news regardless of whether we like it or not. We sometimes bring stories that almost require that people saw the news yesterday. We would not accept that outside the election moment, but now we have to do it.

The campaign in the news

89

Indeed when looking at the visibility of the referendum in both news programs during the campaign (Figure 4.1) there is a sharp increase in the ﬁnal weeks. In the coverage of a referendum news programs are challenged to adapt the notion of impartial and balanced coverage. Potential biases in news during general elections are avoided, for example, by providing equal access to the news of different political parties or access proportional to the representation in Parliament (Semetko 1996a). However, in the case of a referendum where political parties and leaders may run in the campaign without being represented in parliament, this interpretation of the balanced reporting is inappropriate. Under such circumstances, equal access to the Yes and No camp could be a more appropriate measure. In the Danish case, for example, more than 80 percent of the candidates in Parliament favored a Yes which could bring in to question whether a 50–50 access to the news represents balanced coverage. Based on our content analysis during the six-month long campaign, we assessed the visibility of actors in the news from the Yes and No camps. Figure 4.3 shows that Yes actors made up 60 percent of all actors in the news about the referendum while No actors made up 40 percent. Editors and journalists from both news programs expressed anxiety about striking a balance. All agreed that balanced reporting was more likely to be achieved when looking at a number of stories rather than within each individual news story. The Editors-in-Chief both emphasized the political pressure to report impartially. The TV2 Editor said that “the politicians call us after the bulletin” and the TV1 Editor stressed an example of a

40% No actors 60% Yes actors

Figure 4.3 Share of actors from the Yes and No camps. Note Base is total number of actors (n ⫽ 506) in television news, March–September 2000.

90

The campaign in the news

news story about the enthusiasm for the euro in Italy that led politicians from the Yes and the No camp to both feel that their arguments had been disadvantageously presented in the news. In sum, the journalistic approach to the referendum was characterized by elements of both the sacerdotal and pragmatic. The referendum was upgraded and given extra attention in the news which indicates a rather sacerdotal approach. However, the news programs, and most evidently so TV1, were rather pro-active and initiated topics of their own that were unrelated to the agenda of political parties. The results corroborate recent research that suggests that nowadays in news not even the public TV broadcasters can be described, in Blumler and Gurevitch’s words, as reacting to and reporting the campaign in a “sacerdotal” manner (de Vreese 2001a).

Discussion The news agenda was strongly inﬂuenced by the events of the referendum campaign, in particular in the ﬁnal month leading up to the vote. Our content analysis of television news showed that the share of political and economic news rose by about 15 percent in the ﬁnal weeks due to a sharp increase in the amount of news about the referendum. In the last four weeks of the campaign, more than 800 newspaper items were devoted to the referendum. The headlines focused on the contradictory predictions of the costs and gains of either a Yes or a No, human interest stories about key campaign leaders, the fall of the euro against the U.S. dollar during the campaign, and the internal disagreement in the Yes camp. Our content analysis of the coverage also showed a strong presence of polls and reference to public support for the single currency. More than one-third of the stories referred to polls and the Yes and No sides’ standing. This emphasis on polls in the news was not in line with the intentions for the coverage as formulated by the journalists and Editors in our interviews. However, Danish media tended to perform comparatively well in providing methodological information enabling the audience to assess the poll. Turning to our analysis of actors in the news, we saw the importance of key campaign leaders. In the news coverage of the referendum a limited number of key ﬁgures from different political parties dominated the coverage. Visibility, however, is only one dimension of media coverage. Evaluations in the news are at least as important. A pattern emerged from our content analysis that showed that the incumbent government, and in particular the Prime Minister, was highly visible in the news, but at the same time was also seen handling the referendum issue poorly, thus the consistently negative evaluations these actors received in the news. We also identiﬁed a number of differences and similarities in the journalistic approach to the campaign. Both television news programs pre-

The campaign in the news

91

pared coverage of the referendum campaign in advance and allocated extra budgetary means and staff to cover the referendum. The news programs differed in their approach. While TV2 was somewhat more reluctant towards pursuing a pro-active role in setting its own agenda, TV1 was, consciously, more assertive and focused on a number of topics based on polls they had carried out among their viewers. There was little variation in terms of the application of news selection criteria. Both programs devoted extra attention to the referendum, i.e. applied a somewhat sacerdotal set of criteria in which “election time is news time.” None of the programs, however, implemented a special referendum segment in the news on a daily basis which has been common practice during, for example, national elections. There are only limited possibilities to put our ﬁndings about the coverage of the referendum in an appropriate comparative perspective given the lack of data from other referendums. There is evidence from national elections in several countries and some benchmarking can be made against elections for the European Parliament. We found the referendum on the front page of most national newspapers almost daily during the ﬁnal weeks of the campaign and the referendum took up about 25 percent of national television news in the last month leading up to the referendum which, on average, means about four to ﬁve news stories daily. Compared to the 1979 and 1999 European parliamentary election campaign, the 2000 referendum was much more visible in the news (Blumler 1983; de Vreese et al. 2004). Danish news spent about 10 percent of the news in the ﬁnal two weeks leading up to the European elections in 1999 while this was more than 25 percent in the case of the 2000 referendum. This suggests that the referendum was quite visible in the news. This chapter has assessed a number of key features of the news coverage, including the visibility of the referendum, the visibility and evaluation of political actors, and the journalistic approaches taken by broadcasters in covering the campaign. The next chapter shows that visibility of the campaign in the news is only one dimension of the media’s coverage of a referendum campaign, and points to the importance of the interaction between news media and political elites in framing the issue in the news.

5

Elite framing of the issue

The battle between political elites and the news media over the media agenda is one of the key aspects of referendum campaigns, as we discussed in the previous chapter. The framing of the referendum issue is also of great importance. Our study shows how the simple Yes/No issue of the referendum was framed differently by each of the political parties and movements within the two opposing camps. Broadly speaking, the Yes camp framed the referendum issue as an economic one, whereas the No camp framed it as a political one that challenged sovereignty, the welfare state, and national identity. The individual variations and internal contradictions on these general frames are discussed in this chapter, in the context of how the news media responded to the different spins put forward by the parties. The frames used by the two camps in our case study are relevant to the ways in which parties and movements in many other EU countries take stands or positions on issues in referendums. In this chapter we review the literature on frame-building (Scheufele 1999; de Vreese 2002) and we compare news framing to elite framing and identify the way in which a Yes/No referendum issue becomes multi-faceted and complex. We also discuss the importance of studying the relationships between elite framing and the frames that emerge in the news, and how this can be expected to vary in the EU’s different national contexts.

Framing the referendum issue We understand a frame as “an emphasis in salience of different aspects of a topic” (de Vreese 2002: 27). By virtue of emphasizing some elements of a topic over others, a frame provides a way to understand an event or issue by activating or stimulating certain constructs and values. Frames in the news can be examined and identiﬁed by “the presence or absence of certain keywords, stock phrases, stereotyped images, sources of information and sentences that provide thematically reinforcing clusters of facts or judgments” (Entman 1993: 52). In doing so, frames “deﬁne problems,” “diagnose causes,” “make moral judgments,” and “suggest remedies” (Entman 1993).

Elite framing of the issue 93 The notion of framing has historical antecedents in psychology, one obvious predecessor being the series of “Asian decease”-studies by Kahneman and Tversky (e.g., 1983). Though their framing manipulation – altering the wording of a medical scenario – was appropriate to explore the psychological process, this deﬁnition of framing is rather narrow. Simple wording differences that reverse information are not easily compatible with more complex political and economic issues. Therefore a broader deﬁnition of a frame is appropriate, both when looking at politics and news about politics. Indeed most issues – political and social – cannot be meaningfully reduced to two identical scenarios. Political, economic and social events and issues are presented to citizens as alternative characterizations of a course of action (Sniderman and Theriault 2002). When conceiving of, for example, oil drilling, citizens may be presented with frames such as economic costs of gas prices, unemployment, environment, U.S. dependency on foreign energy sources (Zaller 1992). Frames are parts of political arguments, journalistic norms, and social movements’ discourse. They are alternative ways of deﬁning issues, endogenous to the political and social world.

A typology of news frames While newsmakers may employ many different frames in their coverage of an issue, scholars agree that this abundance in choice in how to tell and construct stories can be captured in analyses as certain distinctive characteristics and particular frames recur. In order to synthesize previous research and the different types of news frames that have been suggested, we apply a more general typology with reference to the nature and content of the frame. Certain frames are pertinent only to speciﬁc topics or events. Such frames may be labeled issue-speciﬁc frames. Other frames transcend thematic limitations and can be identiﬁed in relation to different topics, some even over time and in different cultural contexts. These frames can be labeled generic frames (de Vreese 2002). An issue-speciﬁc approach to the study of news frames allows for a profound level of speciﬁcity and details relevant to the event or issue under investigation. This advantage, however, is potentially an inherent disadvantage as well. The high degree of issue-sensitivity makes analyses drawing on issue-speciﬁc frames difﬁcult to generalize and compare. Moreover, such frames have led researchers to “too easily ﬁnding evidence for what they are looking for” and to contribute to “one of the most frustrating tendencies in the study of frames and framing [being] the tendency for scholars to generate a unique set of frames for every study” (Hertog and McLeod 2001: 150–151).

94

The campaign in the news

Issue-speciﬁc news frames Studies of issue-speciﬁc news frames have been in the arenas of international relations, politics, labor disputes, and the economy. One study of issue-speciﬁc news frames focused on U.S. national budget deﬁcits (Jasperson et al. 1998). Drawing on a content analysis of several major newspapers four frames were identiﬁed, and these were labeled “talk,” “ﬁght,” “impasse,” and “crisis.” The four frames reﬂected the chronology of the development of the issue in the news and were issue-sensitive. In the realm of elections, Shah et al. (2002) identiﬁed three frames recurrent in the news during the ﬁnal stages of the Clinton presidency. They identiﬁed “Clinton behavior scandal,” “Conservative attack scandal,” and “Liberal response scandal.” In the same vein, other studies of frames in the news have investigated the presence of frames that pertain to the particular issue in question.

Generic news frames One group of studies of generic frames concentrates on the coverage of politics, in particular election campaigns. A second group focuses on generic news frames that are structural and inherent to the conventions of journalism. In the ﬁrst group, Cappella and Jamieson (1997) investigated the consequences of strategically framed news on political cynicism. Strategic news is deﬁned as news that (1) focuses on winning and losing, (2) includes the language of war, games, and competition, (3) contains “performers, critics and audiences,” (4) focuses on candidate style and perceptions, and (5) gives weight to polls and candidate standings (Jamieson 1992). According to Cappella and Jamieson (1997), strategic news dominates American news coverage of not only election campaigns, but also of policy issues. The focus on winning and losing and polls bears close resemblance to Patterson’s (1993) discussion of the use of “game schema” in election news. Game refers to strategies and (predictions of) electoral success, emphasizing candidates’ positions in the electoral race. Patterson (1993) provides evidence of the historical increase in the use of the game or horse race frame in the press coverage of U.S. elections from 1960 until 1992. A second group of studies link news frames to more general features of news coverage such as journalistic conventions, norms, and news values. In an analysis of the U.S. network coverage of social issues such as poverty, crime, and unemployment from 1981 to 1986 Iyengar (1991) found that daily news coverage was strongly biased towards an episodic interpretation in which news depicts social issues as limited to events only and not placed in a broader interpretation or context (the thematic frame). Iyengar (1991) suggested that norms and standards within news organizations and news production reinforce episodic framing. This practice “simpliﬁes complex

Elite framing of the issue 95 issues to the level of anecdotal evidence” (Iyengar 1991: 136–137) and induces a topical, disorganized, and isolated, rather than general and contextual, understanding of public affairs and social issues. Neuman et al. (1992) in their exploratory study identiﬁed “human impact,” “powerlessness,” “economics,” “moral values,” and “conﬂict” as common frames used by the media and the audience. The human impact frame focused on descriptions of individuals and groups affected by an issue. The powerlessness frame referred to “the dominance of forces over weak individuals or groups” (1992: 67). The economics frame reﬂected “the preoccupation with ‘the bottom line,’ proﬁt and loss” (1992: 63). The moral values frame referred to the often indirect reference to morality and social prescriptions by, for example, including certain quotations or inferences. The conﬂict frame referred to the journalistic practice of reporting stories of clashing interpretation and it was found to ﬁt well with news media’s “game interpretation of the political world as an ongoing series of contests, each with a new set of winners and losers” (1992: 64). These frames were found in relation to different issues which suggest that the frames are more generally applicable than issue-speciﬁc news frames. Semetko and Valkenburg (2000) developed this line of research and identiﬁed an approach to measuring ﬁve news frames: “conﬂict,” “human interest,” “attribution of responsibility,” “morality” and “economic consequences.” The conﬂict frame emphasizes conﬂict between individuals, groups, institutions, or countries. The human interest frame brings a human face, an individual’s story, or an emotional angle to the presentation of an event, issue, or problem. The responsibility frame presents an issue or problem in such a way as to attribute responsibility for causing or solving to either the government or to an individual or group. The morality frame interprets an event or issue in the context of religious tenets or moral prescriptions. The economic consequences frame, ﬁnally, presents an event, problem, or issue in terms of the economic consequences it will have on an individual, group, institution, region, or country. The study found that in political reporting the attribution of responsibility frame was the most commonly used followed by the conﬂict and economic consequences frames, based on an analysis of national print and television news in a European country during a major EU event involving heads of state (Semetko and Valkenburg 2000). These frames are generic in nature.

Key questions In this chapter we ﬁrst look speciﬁcally at the news framing of the referendum in terms of conﬂict, consequences, and strategy. These frames have been identiﬁed in several studies as prominent during election campaigns (Cappella and Jamieson 1997) and in relation to the topic of European integration in general (de Vreese 2002) and the euro in particular (de Vreese et al. 2001). Second, we look at the issue-speciﬁc framing and

96

The campaign in the news

frame building process by analyzing the interaction between the news media and the political realm with reference to framing the referendum topic. To do so we ﬁrst turn to the sub-issues within the broader topic of the referendum. We analyze those aspects of the referendum issue that the media focused on and towards the end of the chapter we relate that to the sub-issues emphasized by the political parties (see Chapter 3). By doing so we can assess the success of the parties and movements in spinning the sub-issues in to the media coverage.

The referendum in the news: campaign conduct, consequences, and the strength of the euro In the previous chapter we saw how the visibility of the referendum increased in the media during the campaign. In this section we move beyond our indicator of the visibility of the referendum to ask the pertinent question: which sub-issues within the broader context of the referendum were covered by the media? Turning ﬁrst to the television coverage of the referendum, we see (Table 5.1) that the news about the campaign and how it was conducted by the different political parties and leaders took up, on average, about 50 percent of the news about the referendum. These stories dealt with how the political parties campaigned, analyses of key leaders’ performance, and discussions of campaign conduct. The focus on campaign conduct is a feature of political journalism on the increase. As we discussed in Chapter 3 the professionalization of politics and campaigning has happened concurrently with changes in political reporting styles. This implies that the process of politics, such as how campaigns are conducted, is covered extensively. By means of comparison, we can look at the British elections in the 1990s where between 50 percent and 66 percent of the BBC and ITN news coverage was devoted to the conduct of the campaign. The second most important topic was the potential consequences of the referendum in terms of how it would affect domestic politics. The third most important topic dealt with the euro as a currency and included stories about the strength of the euro as expressed in the exchange rate against the U.S. dollar. In addition, some stories addressed technical aspects of the Economic Monetary Union. This topic was more prominent on TV1 than on TV2. The fourth most important topic dealt with ﬁnancial aspects and the economic implications of the referendum issue. This topic dropped in importance on the news agenda during the summer months following the conclusion by several experts that membership of the EMU would leave the Danish economy unaffected. After publication of these conclusions the topic never re-emerged on the television agenda. Finally, the topic of welfare – which was a typical No camp topic – was absent in the early stages of the campaign but obtained about 10 percent of the news in the ﬁnal month.

100.0

0.0 4.1 14.3 7.1 8.2 0.0 0.0 0.0 55.1 16.3

Note Base is campaign from March–September 2000.

Total

Sovereignty EU evolution Consequences Euro as currency Financial aspects Welfare Power in the EU Immigration Campaign and credibility Other euro/referendum topic 100.0

0.0 0.0 9.1 4.5 11.4 0.0 0.0 0.0 61.4 13.6 100.0

0.0 17.6 23.5 0.0 11.8 0.0 0.0 0.0 47.1 0.0

TV1 (n ⫽ 17)

TV1 (n ⫽ 49)

TV2 (n ⫽ 49)

Summer (June, July)

Spring (March, April, May)

Table 5.1 The referendum issue on television news

100.0

0.0 8.3 8.3 0.0 16.7 0.0 0.0 0.0 50.0 16.7

TV2 (n ⫽ 12)

100.0

0.0 5.0 10.0 15.0 5.0 0.0 0.0 0.0 50.0 15.0

TV1 (n ⫽ 20)

August

100.0

0.0 0.0 0.0 9.5 4.8 4.8 0.0 0.0 57.1 23.8

TV2 (n ⫽ 21)

100.0

1.4 2.9 12.9 14.3 1.4 5.7 0.0 0.0 50.0 11.4

TV1 (n ⫽ 70)

September

100.0

0.0 0.0 7.0 2.8 4.2 9.9 0.0 0.0 47.9 28.2

TV2 (n ⫽ 71)

98

The campaign in the news

Turning to the newspaper coverage in the ﬁnal month of the referendum we see that the tabloid newspapers devoted considerably less attention to the referendum compared to the broadsheet papers. The latter often included extensive enclosures devoted to the referendum and were saturated with information. In terms of topical focus, on average about one-third of the news was devoted to campaign conduct and the credibility of the political parties’ campaigning (Table 5.2). Other important topics in the newspaper coverage were the implications of the referendum, the strength and stability of the currency as well as technical aspects of the new currency cooperation, welfare, and the evolution of the single currency with the EU. In general, the agenda of sub-issues that emerged in the media within the broader referendum topic had a strong focus on the conduct of the campaign. This was far from the agenda deﬁned by the political parties. Quality newspapers devoted 33 percent of their coverage to the conduct of the campaign while this was as high as 50 percent in the television news. The broadsheet papers spent the most time on issues and dealt both with Yes camp topics such as the economy and No camp issues such as welfare and the future evolution of the EU.

Framing in the news: conﬂict, consequences, and strategy The sub-issues discussed above largely pertain speciﬁcally to referendums on issues of European integration. Journalists and news organizations may pick up on these issues to varying degrees, but they also have a profesTable 5.2 The referendum issue in the press Newspaper Politiken Jyllands- Berlingske Posten Tidende (n ⫽ 63) (n ⫽ 65) (n ⫽ 75) Sovereignty EU evolution Consequences Euro as currency Financial aspects Welfare Power in the EU Immigration Campaign and credibility Other euro/referendum topic Total

Ekstra- B.T. Bladet (n ⫽ 16) (n ⫽ 16)

0.0 4.8 4.8 15.9 4.8 9.5 0.0 0.0 34.9 25.4

1.5 12.3 13.8 7.7 4.6 6.2 0.0 0.0 29.2 24.6

1.3 9.3 12.0 6.7 4.0 6.7 5.3 0.0 33.3 21.3

0.0 0.0 0.0 0.0 0.0 6.3 0.0 0.0 56.3 37.5

0.0 6.3 0.0 0.0 6.3 0.0 0.0 0.0 37.5 50.0

100.0

100.0

100.0

100.0

100.0

Note Base is campaign from August 28–September 27, 2000.

Elite framing of the issue 99 sional toolbox themselves for reporting political and economic news. We look in particular at the journalistic framing of news about the referendum in terms of conﬂict, consequences, and strategy. The conﬂict news frame emphasizes disagreement between different candidates, parties, or institutions and often includes reproach towards other issue stakeholders. The consequences frame translates events and issues in to tangible consequences and impact for, for example, countries, groups, or individuals. The strategy news frame reports news so as to interpret political actions as tactical moves to maximize public support and to look favorably in public opinion. These frames have previously been identiﬁed in political and economic news and also in relation to news about European integration speciﬁcally (de Vreese et al. 2001). Figures 5.1 and 5.2 show how the use of both the conﬂict and the consequences frames increased during the campaign as the news coverage of the referendum intensiﬁed. Both frames were prominent in the news and these ﬁndings corroborate earlier studies of news about national elections but also extend these studies to the context of a national referendum. We already showed in Chapter 4 that the question of who was ahead and behind was a prominent feature of the coverage. We now focus on strategic news framing in the main evening news bulletins of the two most widely watched networks and the front page of the ﬁve most widely read daily national newspapers. We draw on data from our content analysis covering the hot phase of the referendum campaign (i.e., the ﬁnal month of the campaign, from August 28–September 27, 2000).1 We draw three dichotomous indicators of strategic news derived from Cappella and Jamieson (1997). All stories were coded “yes” (1) or “no” (0) for the presence of [1] emphasis of performance, style, and perception 70

Two or more sides Reproach

Number of stories

60 50 40 30 20 10 0 March

April

May

June

July

August

September

Figure 5.1 Television news framing of the referendum: conﬂict news framing.

100

The campaign in the news

60

Consequences

Number of stories

50 40 30 20 10 0 March

April

May

June

July

August

September

Figure 5.2 Television news framing of the referendum: consequence news framing.

of the candidate, [2] analysis of candidate actions as part of a consolidation of positions, and [3] language of wars, games, and competitions. Following Valentino et al. (2001a) we do not include polls in our content measure of strategic news. This was discussed in Chapter 4 and in Chapter 7 we discuss citizens’ perception of the use of opinion polls in the referendum campaign. In addition to these speciﬁc measures, we classiﬁed all news stories about the referendum to determine if the topic of the story was focused on strategy (such as, for example, stories about campaign conduct, analyses of the personalities of candidates and the strategies of political actors), a combination or mixture of strategy and policy issues (e.g., stories about rallies and the historical evolution of the EU), or policy issues (e.g., stories about the political and economic consequences of the referendum or stories about the exchange rate of the euro). In the presentation of the results we accordingly distinguish between two indicators of strategic news coverage: (a) a classiﬁcation of the topic of all news stories as either “strategy,” “mixed strategy-issue,” or “issue,” and (b) the presence of different indicators of strategic coverage in the stories. Table 5.3 summarizes the content analysis of the television and newspaper coverage of the referendum campaign. The content analysis shows that during the ﬁnal month of the campaign, the referendum was highly visible in the news environment (see also Chapter 4). Section A in Table 5.3 shows that about 50 percent of the television news and between 25 percent and 45 percent of the broadsheet press news stories dealt with strategy-related topics. The low number of stories in the tabloids is due to format because only one story is on the front-page each day. Section B in Table 5.3 shows that reference was made

32 40 54

B Strategic news indicators Focus on candidate presentation-style Focus on candidate position consolidation Use of language war-games-sports 28 25 46

42 41 16 99 23 33 31

42 27 31 100 22 28 42

30 42 28 100

26 28 46

45 36 19 100

0 13 38

25 50 25 100

(n ⫽ 8)

B.T.a

Notes The analysis covers the period August 28–September 27, 2000. Data entries are percentages. a denotes that these two newspapers are tabloids which generally carry only one story on the front page and often in a headline only style. b denotes ofﬁcial No-vote supportive editorial policy.

54 23 24 101

A Topic Strategic Mix issue-strategy Issue Total topic

(n ⫽ 79)

(n ⫽ 76)

(n ⫽ 52)

Berlingske Tilende (n ⫽ 58)

JyllandsPosten (n ⫽ 50)

TV2

TV1

Politiken

Newspapers

Television

Table 5.3 Strategic news framing of the referendum

23 0 39

69 31 0 100

EkstraBladeta,b (n ⫽ 13)

102

The campaign in the news

to the presentation and style of a politician and/or party in, on average, one-fourth of the stories and that actions of politicians and/or parties were interpreted in terms of strategy and consolidation of their position in, on average, one in three news stories. Finally, the analysis showed that roughly half of the news included terminology from war, games, or sports situations, which is often used to illustrate and emphasize the electoral battle. The share of news framed in terms of strategy neither differed signiﬁcantly between media (newspapers versus television), nor between news outlets such as public and private news programs. While the focus in this chapter is on the content of news framing, it is important to acknowledge that both the conﬂict frame and the consequence frame are crucial cue givers to citizens for understanding politics. News framing has signiﬁcant implications. One study of the conﬂict news frame investigated the ability of newspaper articles to inﬂuence readers’ thoughts about the introduction of the euro (Valkenburg et al. 1999). Another study examined the inﬂuence of the conﬂict frame in television news on how the citizens conceive of the issue of the enlargement of the European Union (de Vreese 2004a). In these studies, it was found that public opinion about European issues such as the euro and the EU enlargement may be inﬂuenced by the ways in which news media frame the news. Both studies demonstrated that news frames employed by journalists when making the news stories could be traced in the thoughts offered by citizens on the issue. For example, if a news story about the enlargement of the European Union with former East European countries focuses on political disagreement between two parties, the disagreements and conﬂictual nature of the issue was found in the audience members’ thoughts on the issue. For news about the enlargement of the European Union, one experimental study conducted with a sample of Dutch voters in cooperation with the NOS Journaal, the national public broadcaster main evening news program comparable to the British BBC, showed that the way in which the news is framed is at least as important to the audience as the core news facts in a story. When retelling the main points of the news story, the news frame was as pertinent to the audience understanding and recollection of the story as were the key news facts presented in the story (de Vreese 2004a). Yet another experimental study demonstrated how the alteration of the wording of news with the inclusion of comments about politicians self-serving agenda (strategy news) affected viewers’ level of political cynicism (de Vreese 2004b). Taken together, this experimental evidence demonstrates how the news media may inﬂuence the short-term processing of information about the European Union and European issues. It remains of course an open question as to the duration of such effects, but news frames are important sources of information that contribute to public opinion formation on European issues, including the introduction of the euro.

Elite framing of the issue 103 So while these frames are prominent in the journalistic coverage of the referendum, we may return to the broader question of the sub-issues. A comparison of the political topics and the media topics can tell us more about elite framing of the issue in this important referendum.

Winning the framing battle Bringing together our analysis of the issue framing by the Yes and No camps on the one hand, and the media on the other, we can assess the extent to which different parties managed to succeed in getting their framing across in the news. The media focused extensively on the campaign dynamics and included analyses of how different parties and candidates campaigned. The free fall of the euro, the implications of the EMU for the welfare state and the political implications of the referendum were other issue frames dominating the media. The Yes camp primarily framed the referendum as an economic issue and framed the trajectory of European integration as one of peace, stability, and prosperity. The No framed the referendum in terms of welfare, sovereignty, and the potential lack of strength of the common currency. It is obvious that the No parties’ issue framing found greater resonance in the media (see Table 5.4). This may seem surprising given that most media ofﬁcially supported a Yes (see Chapter 2). However, the No camp was successful in its strategy of broadening the topic of the referendum to include issues that were technically and juridical beyond the referendum proposal, but emotionally associated with the issue. Moreover, these topics were of strong interest to verbal and inﬂuential segments in the electorate. The tension between Yes and No camps in terms of issue framing is a general feature of referendum campaigns. The arguments listed here, however, are of particular relevance to referendums on EU issues and they have been seen to emerge in several referendums. In the 2003 Swedish referendum on the euro, for example, the Yes camp framed the issue in purely economic cost/beneﬁt terms while the No camp framed the referendum as an issue of sovereignty and national identity. In the British euro debate, the economic argument has also been center stage and the Table 5.4 The framing of the referendum issue by the media and the Yes and No camps Media

Yes parties

No parties

Campaign conduct Strength and exchange rate euro Welfare (pensions) Consequences

Economy Europe ⫽ peace and prosperity

Welfare endangered Sovereignty (euro ⫽ symbol political integration) Weakness euro

104

The campaign in the news

British government has even identiﬁed a number of macro-economic targets to be met before the government will put the euro to a vote in a referendum. The issue framing in the Danish 2000 referendum therefore compares well to other EU referendum contexts. The battle over framing the referendum issue is not only between Yes and No contenders but in interaction with the news media.

Part IV

Campaign effects

Referendums see proposals being adopted or rejected. Despite the fact that most of the referendum literature emphasizes the volatility, dynamics, and importance of the campaign only few studies have formally incorporated the campaign in models of citizen engagement, turnout, and vote choice. While the political science literature stresses the importance of party cues in electoral contests, these cues are, as argued in Chapter 1, potentially less inﬂuential in the context of a referendum characterized by volatility, ambiguity, and unpredictability. The media are key actors in the campaign. Media content may affect citizens in three ways during a campaign. First through editorial endorsement, second through paid campaigning, and third through free campaigning. Editorial endorsement refers to how the press can endorse a particular political viewpoint. This phenomenon is important in Europe where the traditional party-newspaper alignment, albeit much less overtly than previously, still affects whether a newspaper will endorse a speciﬁc political party or viewpoints in its editorials. This was also the case in the 2000 referendum in Denmark where all newspapers took an outspoken editorial stance. Paid campaigning is of less importance in the European situation compared to the United States. In the United States there are fewer budget regulations and the budget magnitude is considerable beyond that of any campaign in Europe. In most European countries, the use of political ads during a campaign is restricted and the use of commercials on television is limited to an absolute minimum (Kaid and Holz-Bacha 1995). In the United States, ads and commercials play a central role in the campaign, though the available evidence on whether ads, and in particular negative ads, have the desired effect is contradictory (see, for example, Freedman and Goldstein 1999). We analyzed the advertising strategy of all political parties and contestants in Chapter 3. None of them used negative or attack messages in their ads. In the European context, free campaigning is the most important feature of the campaign in the media. Free campaigning refers to the journalistic coverage of the electoral contest. In this section of the book we review key theories of campaign effects

106

Campaign effects

and pay particular attention to the role of the media. Vote choice is obviously a key variable in campaign effects research, but the campaign may affect citizens in other equally signiﬁcant ways. As outlined in Chapter 1, we discuss the effects of the campaign on a number of cognitive, affective, and behavioral measures. We ﬁrst deal with how a referendum can affect the public’s perception of contentious issues facing the country. This perspective is referred to as agenda-setting. Second, we deal with how the news coverage of a referendum can affect campaign evaluations and political cynicism. We know from extant research that if news reports about politics as a game of politicians serving self-interests, this may lead citizens to respond negatively to on-going campaigns and to increase political cynicism. We furthermore test in the context of a referendum whether such campaign coverage in the news affects political mobilization in the form of turnout. We also address the dynamics of a referendum campaign with respect to public evaluation of the incumbent government and other key politicians. We investigate whether and how a government’s overall performance rating is inﬂuenced by how the government is perceived to be handling the topic of the referendum. Finally, we bring together the different chapters on campaign effects and model vote choice in a referendum. We do this by testing existing, competing hypotheses on voting behavior in referendums and by in addition exploring the speciﬁc role the campaign played for citizens to make their ﬁnal decisions. The theoretical concepts explored in the context of a referendum are well-known in the campaign literature. However, most of our knowledge is based on national election studies and the theories have not yet been put to a compelling test in a referendum campaign. Moreover, the existing literature on referendums has primarily dealt with the campaign in descriptive terms, but has not modeled the effects of the campaign along with other theories. Our study was speciﬁcally designed to investigate the dynamics of a campaign. We therefore in all the chapters in this section of the book draw on our panel survey data. We discuss developments in the campaign at the aggregate level and investigate change during the campaign for individual respondents. We aim to identify the effects of the campaign both as a mechanism in which citizens’ preferences are reinforced and predispositions crystallize and also as a process that in itself may exert considerable inﬂuence in terms of affecting information processing, issue perceptions, political judgments, as well as turnout and vote choice.

6

Agenda-setting in a referendum

One of the most inﬂuential ideas about media effects is that the media have the ability to inﬂuence the public salience of issues. This is also called the agenda-setting function of the media. Agenda-setting is a heterogeneous ﬁeld of research originating in Lippmann’s (1922) notion of how the media shape our pictures of the world. McCombs and Shaw’s (1972) study of the 1968 U.S. presidential election campaign, where they found a strong correlation between the rank order of the most prominent issues in the news and undecided voters’ ranking of the most important political issues, was a seminal exploration of this relationship. In the past decades, agenda-setting research has proliferated. A review of the literature up until the early 1990s cited more than 250 empirical studies of the media’s ability to inﬂuence what to think about (Dearing and Rogers 1996). Agenda-setting research addresses a reciprocal process involving the interaction and mutual dependency of the political/policy agenda (what actors in the political arena do and want), the media agenda (which issues are prominent in the news), and the public agenda (which issues are considered salient by the public) (Kosicki 1993; Rogers and Dearing 1987). In this book, we dealt with the agenda of the political parties in Chapter 3 and with the media agenda in Chapters 4 and 5. To study the agenda-setting dynamics in its totality, in this chapter we turn to the link between the media agenda and the public agenda. Most agenda-setting research has been conducted in the context of national elections. McCombs and Shaw’s (1972) seminal study focused on the 1968 U.S. Presidential election campaign. Later, Weaver et al. (1981) investigated the agenda-setting of television and the press over the 1980 presidential election year. The evidence of media agenda-setting during national elections is mixed. Most studies are supportive of the phenomenon, albeit at various levels (Dearing and Rogers 1996). In Europe, there have been a number of examples of election studies in which the agendasetting hypothesis did not ﬁnd support, however, such as in Britain in 1987 (Miller 1990) and 1997 (Norris et al. 1999), or Germany in 1990 (Semetko and Schonbach 1994).

108

Campaign effects

Agenda-setting in the context of a referendum The dynamics of agenda-setting has not been investigated in the context of a referendum. A referendum campaign is distinctively different from a general election in that the broader theme of the campaign has been deﬁned a priori. The agenda-setting dynamics may therefore be different as sub-issues emerge within the broader referendum issue. General elections may be waged over multiple issues such as the economy, health care, or national security, but in a referendum campaign the overall issue is deﬁned in advance of the campaign. The news media, as demonstrated in Chapter 5, play a key role in deﬁning and either reinforcing or deﬂating sub-issues as they are put forward by key campaign actors. At the core of agenda-setting is the notion of “issue,” but it remains an open question what an issue is (Lang and Lang 1981). Agenda-setting research draws extensively on a categorization of topics in the news and on the public agenda into broader “issues,” but neither previous research nor empirical techniques exist to determine how issues should be collapsed and merged together (McCombs and Zhu 1995). In fact, some of the close relationships between salience and media content are artifacts of the subject categories under which speciﬁc news items are classiﬁed (Lang and Lang 1981). Consequently, it is an important empirical question whether an “issue” on the media agenda has the same meaning as that same “issue” on the public agenda. Benton and Frazier (1976) investigated “the economy” in more detail and proposed a total of 13 issues related to “the economy.” They compared the salience of the dimensions of the economy in print media, on television news, and on the public agenda and found a high correlation between the salience of the sub-issues in the press and on the public agenda. The study shows that media may also inﬂuence the public salience of sub-issues, but more importantly, it exposed the different dimensions and content of an issue on two agendas. Similarly Chan (1999) demonstrated that the issue “environment” on the media agenda alone consists of several sub-dimensions. Swanson (1988) made a cautionary comment in this respect by suggesting a potential discrepancy between how the public conceives of an issue compared to how the media covers an issue. The question of the “meaning of an issue” is addressed in this chapter. Whether an issue has the same meaning to the media as it does to the public is an important, but often neglected, aspect of agenda-setting research. A referendum campaign provides an opportunity to explore a particular issue in more depth and investigate the multiple dimensions and sub-issues on the media and the public agenda. In this case whether the referendum means the same to the media as it does to the public.

Agenda-setting in a referendum 109

Aggregate level and individual level perspectives on agenda-setting The simplicity of early agenda-setting studies and under-speciﬁed models and conceptualizations have fueled agenda-setting research with inconclusive evidence (Kosicki 1993). This leaves the often-implied across-theboard effects of agenda-setting with the status of a “plausible, but unproven idea” (McQuail 2000: 456). Several scholars have identiﬁed the requirements for a compelling test of causality in the agenda-setting argument (see, for example, Kosicki (1993) and McQuail (2000)). Such a research design would involve (1) media content data documenting the salience of issues in different (news) media outlets, (2) evidence that members of the public are in fact exposed to the content, and (3) a control for extraneous factors. As noted, “such data have rarely, if ever, been produced at the same time in support of the hypothesis of agenda-setting” (McQuail 2000: 455). McCombs (1981) and McCombs et al. (1995) developed a typology of agenda-setting studies with the scope of the agenda and the level of investigation as the two most important dimensions. Type I studies (see Table 6.1) focus on sets of issues and on aggregated data capturing the public and the media agenda. Many studies of this type have been conducted, including the 1972 study of agenda-setting during the U.S. presidential elections (McCombs and Shaw 1972). Type II studies are referred to as automaton studies (McCombs et al. 1995: 285) and examine sets of issues and individual agendas. The assumption in these studies is that the media not only inﬂuences the issues that we think about, but also the relative importance or ranking of these issues. Type III studies investigate a single issue using aggregate level data. Examples include the tracking of environmental issues and the civil rights issue in the news and on the public agenda (for example, Winter and Eyal 1981). Type IV studies are referred Table 6.1 Overview of agenda-setting studies

Multiple issues

Single issue

Aggregate level data

Individual level data

I Comparison (correlation) between rank ordered media agenda and public agenda

II Comparison between individual issue agenda and exposed media agenda

III Comparison between salience of one issue on public agenda and on media agenda

IV Comparison between a single issue’s importance and the importance of the issue in the media, including other predictors

Source: Typology derived from McCombs et al. (1995: 285).

110

Campaign effects

to as “cognitive portrait studies” and investigate single issues drawing on individual level data. In this chapter we investigate the agenda-setting process at both the aggregate and the individual level and compare both the composition of the aggregate public agenda with the media agenda (type I) and the effects of news media on the public salience of a single issue (type IV). The study was speciﬁcally designed to provide a compelling link between the content of speciﬁc news outlets that individuals were actually exposed to and their perception of the salience of a speciﬁc issue. This link is often implied in agenda-setting research, but only seldom empirically tested (Kosicki 1993; McQuail 2000). Among the notable studies that meet the requirements for a compelling test of agenda-setting is the research by Erbring et al. (1980) who matched content analysis of newspapers with individual media use measures and people’s perception of important issues. The study provided little evidence of agenda-setting, ﬁnding only a weak relationship between the volume of news coverage and the public’s ranking of issue salience. Instead they found a stronger relationship inﬂuence of interpersonal communication on issue salience. Another study by Huegel et al. (1989) demonstrated agenda-setting effects of media content to which respondents were actually exposed. They additionally found effects of general media use patterns, interpersonal communication, and issue sensitivity on the salience of issues and concluded that the mass media agenda-setting effect was very sensitive and often overridden by other factors. Roessler (1999) matched survey data including individual media use patterns with a content analysis of newspaper, television, and radio news. At the aggregate level the study suggested a high correspondence between the media and the public agenda, but at the individual level, only little evidence of agenda-setting was found. In fact, at the individual level, the importance of an issue in the media coverage to which the respondent was exposed to exerted only minimal inﬂuence on the respondent’s assessment of issue importance. Factors such as issue involvement, interpersonal communication, and interpersonal network exerted a substantially larger impact.

Key questions In this chapter we address the relationship between the media agenda and the public agenda during a referendum campaign. The question is investigated at both the aggregate and at the individual level. First the composition of the news agenda during the referendum campaign is compared to the public agenda. Agenda-setting theory suggests that the more persons are exposed to news about a particular issue, the more salient this issue becomes. This effect is dependent on a number of contingent and intervening situational variables. Following agenda-setting theory, we test

Agenda-setting in a referendum 111 whether or not exposure to news about the referendum increased the salience of the referendum issue. Given the context of the study – a national referendum campaign – the topic of the referendum is investigated in more detail. While most agendasetting studies have been carried out in the context of general elections, a referendum is distinctively different in that the broader theme has been deﬁned a priori. To address this important, but often neglected aspect of agenda-setting, we investigate whether the referendum has the same meaning to the media as it does to the public. Swanson (1988) suggested a potential discrepancy between how the public conceives of an issue and how the media covers an issue. We therefore ask whether the referendum issue means the same thing on the public agenda and the media agenda.

Aggregate level agenda-setting To gain an initial idea about the relationship between the media agenda and the public agenda, we ﬁrst turn to the composition of the two agendas. The content analysis includes the main evening news bulletins of the two most widely watched networks and the front page of the ﬁve most widely read daily national newspapers (see Chapter 2 for more details about the content analysis). For studying the agenda-setting process at the aggregate level, the television news outlets in the two months preceding the referendum is analyzed. For studying the agenda-setting process at the individual level, the content analysis covers the hot phase of the referendum campaign (i.e., the ﬁnal month of the campaign) and includes both television and the press. The unit of coding and analysis is the individual news story deﬁned as a semantic entity with at least one topic.1 A total of 1,543 television news stories and 235 newspaper stories are included. The issue of immigration dominated the public agenda in both waves (see Table 6.2). In wave one welfare and crime were the most important issues following immigration. In wave two, these were the euro/EU and welfare. At the aggregate level, the euro issue increased in importance on the public agenda.2 Seven per cent of the respondents mentioned the euro/EU as the most important issue in wave one and this went up to 26 percent in wave two. The television agenda was ﬁrst dominated by general political issues with crime being the second most important issue. In the second wave, the focus of the coverage changed and the euro/EU was the most salient issue with general politics and crime (TV1) and the economy (TV2) being the next most important issues in the news. When ranking the topics in the news and on the public agenda the relationships between the agendas can be established. As Table 6.3 shows rank order correlations suggested that the agenda of the two networks was very similar: Spearman’s rho between TV1 and TV2 in August was ⫹0.77 and in September ⫹0.92 (both signiﬁcant). The correlation between TV1 and the public agenda in August was ⫹0.19 and in September ⫹0.32. The

100.0

Total

100.0

8.2 3.3 27.4 0.0 2.7 3.0 4.9 0.3 15.9 6.8 27.4 100.0

7.4 5.9 5.7 1.4 11.9 2.7 40.5 3.6 8.2 12.7 – 100.0

26.1 7.7 23.9 2.0 2.0 4.0 2.2 1.5 9.7 3.7 17.2

100.0

23.8 9.9 22.0 0.0 1.5 2.5 6.6 0.3 9.4 4.3 19.7

TV2 (n ⫽ 395)

Note Data entries are percentage of stories (television news agenda) and answers (public agenda) to the Most Important Problem question.

7.3 5.2 30.7 1.8 3.7 3.4 2.6 4.2 15.7 3.7 32.3

Euro/EU Economy Politics Unemployment Welfare Environment Immigration Norms and values Crime Miscellaneous Accidents, human interest, sports and weather

TV1 (n ⫽ 402)

Public (n ⫽ 1,122)

TV1 (n ⫽ 381)

TV2 (n ⫽ 365)

September

August

Table 6.2 Media agendas and public agendas during the referendum campaign

100.0

25.9 8.1 8.7 1.5 9.8 3.0 31.5 3.7 4.1 3.9 –

Public (n ⫽ 982)

Agenda-setting in a referendum 113 Table 6.3 Rank public and media agendas August

Immigration Euro/EU Welfare Politics Economy Crime Norms and values Environment Unemployment

September

TV1

TV2

Public

TV1

TV2

Public

8 3 6 1 4 2 5 7 9

4 3 6 1 5 2 8 7 9

1 4 2 6 5 3 7 8 9

8 1 6 2 4 3 9 5 6

5 1 7 2 3 4 8 6 9

1 2 3 4 5 6 7 8 9

Note Rank order topics on public agenda and television news media agenda.

correlation between TV2 and the public agenda in August was ⫹0.45 and in September ⫹0.58 (none of the correlations are signiﬁcant). These correlations suggest that there was limited congruence between the issues covered in the news and the issues considered salient by the public.3 In other words, there is no evidence at the aggregate level to support the media agenda-setting hypothesis.

Individual level agenda-setting: effects of exposure to news about the referendum on issue salience Turning to the individual level, we were interested in investigating the effects of exposure to news about the referendum on the public’s perception of the salience of the referendum issue. At the aggregate level, we found an increase in importance of the referendum issue on the public agenda. Seven percent of the respondents mentioned the euro/EU as the most important issue in the August survey and this went up to 26 percent in the September survey (Table 6.2). We are speciﬁcally interested in investigating the inﬂuence of exposure to news about the referendum on the salience of the referendum. Establishing a causal relationship between (news) media coverage of an issue, individual exposure to the coverage, and issue salience requires knowledge of the level of salience of the issue the public brought to the study and being able to accurately characterize the content they actually read and watched. To do this we draw on our panel study with repeated measures assessing the salience of the issue in both wave one and wave two combined with individual level data on news exposure and a content analysis of the actual news media coverage. This provides a compelling link between exposure to news about the referendum and its effects on issue salience.

114

Campaign effects

To estimate the amount of exposure to news about the referendum for each respondent, the survey and content analysis data sets were merged. By matching the panel survey data, including measures of respondents’ exposure to speciﬁc news media outlets and their issue agenda with the content analysis of the presence of news about the referendum, we constructed a more precise measure of the amount of news about the referendum to which each respondent was actually exposed. Based on the content analysis data, the presence of news about the referendum per news outlet was assessed (see Chapter 4). This measure was linked to a respondent’s exposure to that speciﬁc news outlet (measured as number of days an outlet was watched/read). The “exposure to news about the referendum” is therefore a measure of the volume of news about the referendum per news outlet.4 To investigate the effect of news about the referendum on the public’s estimation of the issue’s salience, a number of control variables are included: social-demographics (gender, age, and education), issue importance, need for orientation, interpersonal communication, voting behavior, and EU skepticism (see Appendix for exact question wording). “Issue importance” is an individual’s estimation of the importance and obtrusiveness of the issue. Previous research suggests that the personal importance attached to an issue increases the chance of it becoming “the most important problem” (for example, Weaver et al. 1981). “Need for orientation” refers to an individual’s (potential) information seeking about an issue depending on the person’s interest in and uncertainty about the issue. A high need for orientation leads people to seek information more actively in the news about the issue to reduce uncertainty. The active information seeking increases the susceptibility to media agenda-setting (Weaver 1977, 1980). “Interpersonal communication” refers to whether and how often an individual talks about an issue. Discussing an issue may either enhance or diminish an individual’s perception of the importance of an issue (McLeod et al. 1974). Finally, EU skepticism and voting behavior are included in the model as control variables. General attitudes towards European integration may shape the perception of the speciﬁc issue of the euro (Svensson 2002). Similarly, voting behavior is included in the model as a control since persons who voted No in the referendum and thus “won” the referendum, are less likely to consider the referendum issue a problem (see also Chapter 9). Figure 6.1 presents the results of a logistic regression analysis with the dependent variable as naming the referendum topic (the euro) the most important issue in wave two of the panel survey. The model controls for what the respondent said on this question at wave one of the panel survey. This lag term is the strongest of the predictors and shows considerable opinion stability even at the individual level (Markus 1979; Page and Shapiro 1992). The results nevertheless show that exposure to news about

Agenda-setting in a referendum 115 Talking about politics with others Voting No in the referendum Being skeptical about the EU Exposure to TV1 news about the referendum

⫹0.09 ⫺0.53 ⫹0.36 ⫹0.10

Exposure to news in the No press about the referendum

⫹0.06

Naming the referendum issue as the most important problem facing the country (time 1)

⫹1.47

Naming the referendum issue as the most important problem facing the country

Figure 6.1 Effects of exposure to news coverage about the referendum on the salience of the issue. Note Logistic regression analysis (n ⫽ 672, 2 log likelihood 695.07, pseudo R2 0.14). Positive sign of beta-coefﬁcient indicates positive association with naming the euro/referendum as the Most Important Problem facing Denmark in the second panel survey wave. Only signiﬁcant indicators are shown in the ﬁgure. Controls were made for age, gender, education, issue importance, need for orientation, and exposure to TV2 news and the Yes press. The full regression analysis is provided in the Appendix.

the referendum contributed to perceptions of the salience of the issue among the public. The effects of exposure to news about the referendum on TV1 and in the No press were signiﬁcant, even when controlling for the dependent variable at wave one and a number of other standard sociodemographic characteristics. None of the social-demographic variables was found to be good predictors of considering the euro the most salient issue. EU skepticism and No vote in the referendum were both signiﬁcant negative predictors of issue salience suggesting that persons reluctant towards advanced European integration, who had essentially won the referendum, were less likely to consider the issue salient. Need for orientation was positively associated with considering the euro the most important topic, though this variable did not attain signiﬁcance. These initial results provide evidence in support of the agenda-setting hypothesis at the individual level by suggesting that exposure to news about the referendum contributed to considering the euro the most salient issue, even when controlling for social-demographic variables, a number of intervening personal characteristics, and the salience of the euro issue at the outset of the campaign.

The euro: economy, welfare, identity, or sovereignty? We now turn to the issue of the referendum. The results above suggest that the referendum issue increased in salience on the public agenda over

116

Campaign effects

the course of the campaign and that exposure to news about the referendum contributed to this increase. We now compare the sub-issues within the broad theme of the referendum that the public considered important, on the one hand, to a detailed analysis of the news media coverage of the referendum, on the other. Table 6.4 shows the distribution of topics considered salient by the media and by the public in relation to the euro referendum.5 As discussed in Chapter 5, we coded and categorized all news about the referendum into ten issue categories pertaining speciﬁcally to the euro referendum:6 “Sovereignty,” “Evolution of the European Union,” “Consequences of a Yes/No,” “The euro as a currency,” “Economic and ﬁnancial aspects,” “Welfare in the EU,” “Power in the EU,” “Immigration,” “Campaign and credibility,” and “Other.”7 In addition, all responses in our survey to the Most Important Problem-question were coded using the same categories in order to make comparisons between the media agenda and the public agenda. While the news media devoted most time to the campaign, the conduct of campaigning and the political leaders, the public was more concerned with issues such as sovereignty, the evolution of the European Union, and the technical details of the single currency. A rank order correlation analysis comparing the public euro issue agenda in wave one and wave two suggested a high degree of agenda stability (signiﬁcantly correlating at ⫹0.91). The discrepancy between the public agenda and the media agenda during the ﬁnal month of the campaign is evident in Table 6.4. The Spearman’s Table 6.4 The multi-faceted issue of the euro: media agenda and public agenda referendum issue (euro introduction) Public agenda

News media agenda

August September TV1 TV2 Newspapers (n ⫽ 337) (n ⫽ 715) (n ⫽ 75) (n ⫽ 76) (n ⫽ 235) Sovereignty EU evolution Consequences Euro as currency Financial aspects Welfare Power in the EU Immigration Campaign and credibility Other euro/referendum topic Total

36 18 3 5 9 1 13 1 0 14

22 22 2 16 6 3 17 0 2 10

1 3 12 15 3 5 0 0 49 12

0 0 7 3 5 9 0 0 47 29

1 8 9 9 4 7 2 0 35 26

100

100

100

100

101

Note Data entries are percentage of stories (news agenda) and answers (public agenda) to the question “What is the most important issue in the euro referendum?”

Agenda-setting in a referendum 117 rho rank order correlations between the public agenda in September and the three news outlets were all negative: TV1 ⫺0.26; TV2 ⫺0.53; and the newspapers ⫺0.18.8 In sum, these results suggest that there was signiﬁcant difference between what the media considered salient when covering the euro referendum and what the public considered salient. This is an important ﬁnding because it suggests that the media and the public may both ﬁnd an issue most important at a given point in time, but that this shared perception does not imply that the same aspects of the issue are considered important.

The dynamics of agenda-setting in a referendum campaign As the above analysis of the sub-issues illustrates, the public and the media agenda differed in terms of the ranking of issues. This result, however, is at odds with early studies that suggested a strong relationship between the media issue agenda and the public agenda (for example, McCombs and Shaw 1972). But this ﬁnding dovetails with recent studies that have also found little or no evidence of media agenda-setting at the aggregate level during election campaigns (for example, Miller 1990; Norris et al. 1999; Semetko and Schönbach 1994). One explanation for the lack of support for the agenda-setting hypothesis at the aggregate level is that correlation analysis does not appropriately link the public agenda to the topics in the news. As has been argued elsewhere, there is in fact little reason to expect agendas to mirror perfectly in terms of rank-ordering (McCombs et al. 1995). A second explanation is the fact that immigration was at the top of the public agenda throughout the referendum campaign. Immigration was and has remained a “killer issue” on the public agenda in Denmark since the late 1990s (Andersen et al. 1999). In this study, 30–40 percent of the respondents in the two waves in the panel identiﬁed “immigration” as the Most Important Problem and the issue is highly obtrusive (Zucker 1978). The ability of the media to set the public agenda in terms of salience of issues under the condition of a pervasive and dominant issue (immigration) seems limited. We did ﬁnd, however, that the issue of the referendum increased in importance on the public agenda over the course of the campaign. We therefore investigated the salience of the referendum issue in more detail. By merging survey data with a content analysis of the news media coverage, a measure of actual exposure to speciﬁc news outlets, incorporating the volume of coverage of the referendum in the news, was created. We found that exposure to news about the referendum contributed to an increase in the salience of the euro issue which lends support to the agenda-setting hypothesis at the individual level. Why did exposure to television news (on TV1) about the referendum have an agenda-setting effect, while this effect was not signiﬁcant for the Yes press?

118

Campaign effects

Some research suggests that newspapers have stronger agenda-setting effects compared to television (for example, Benton and Frazier 1976; McClure and Patterson 1976; Tipton et al. 1975; Weaver et al. 1981). Our study suggests that (public) television news produced stronger agendasetting effects, an idea which is echoed in McCombs’ (1977) observation of television’s ability to focus great attention to a few topics which is labeled the “spotlighting” function of television. Similarly, Palmgreen and Clarke (1977) when investigating the relative agenda-setting inﬂuence of television versus newspapers found television to be superior to the press in terms of agenda-setting effects. They note that this hierarchy of effects is consistent with data on the importance of different media as sources of information where television is seen as the most important medium. The same argument can be made in the Danish case where Eurobarometer data suggests that television is the most important and preferred source for receiving information about European issues (Eurobarometer 54 2001). In this study of the euro referendum, more than 95 percent of the respondents in the panel reported having paid attention to information about the referendum on television. The second part of the chapter focused on sub-issues within the broader notion of an “issue” on the public and the media agenda respectively. Swanson (1988) alluded to the idea that the meaning of an issue in the news may not correspond with the meaning of the same issue to the public. We found evidence that the referendum issue did not mean the same to the public as it did to the news media. This is an important consideration in agenda-setting research that goes beyond comparing the rank-ordering of issues to also consider the content and meaning of issues. One example of how important the substantive content of an issue can be, rather than the mere salience of an issue on the media agenda, is provided by Schönbach and Semetko (1992). A detailed media content analysis revealed that while the economy was indeed an important and highly ranked issue in the news in one national widely read newspaper, the Bild, the way in which it was reported encouraged readers to believe that it was a problem that would disappear in the near future; and indeed, the economy did not remain high nor move higher on the agendas of the readers of this newspaper over the course of the campaign. Investigating sub-dimensions of an issue may appropriately be considered an extension of current agenda-setting research, though the term second-level agenda-setting in the literature to date is used to describe the inﬂuence of news on the public perception of, for example, candidate attributions (for example, McCombs et al. 2000). Our study is an initial exploration of different dimensions and sub-issues on the media and the public agenda. The quest for disaggregation in agenda-setting research remains pertinent, especially when investigated at the individual level or when exploring psychological antecedents of susceptibility to agenda-setting. The implication for agenda-setting research is to devote more attention to

Agenda-setting in a referendum 119 the substantive dimensions of the issues on the agenda and to further explicate the conditions under which certain types of issues may emerge on the public agenda and the role played by the media in shaping the subissues. The chapter emphasizes the importance of the battle over a referendum campaign agenda. Magleby (1989) has suggested that the No side in a referendum campaign may often prevail by campaigning with an agenda that concurs with the issues that only a minority of voters consider salient. By “changing the subject” and expanding the campaign agenda to include more dimensions of the referendum issue, actors in the campaign can either generate or suppress support for an issue. This is particularly important in the context of referendums on European integration. National referendums function as a policy and decision-making instrument and they can have signiﬁcant domestic political implications (see Chapters 8 and 9). In Denmark, the No side was effective in broadening the scope of the referendum issue to also include themes as national sovereignty and the welfare state (see Chapter 3). These topics found resonance with the public while the media devoted a lot of attention to the conduct of the campaign itself. Other research has suggested that by introducing new dimensions to an issue, campaigners are able to decrease the abstractness of an issue which in turn increases the chances of the issue being picked up on the public agenda (Yagade and Dozier 1990). In a referendum, winning the battle over the campaign agenda implies not only increasing the salience of the referendum, but also controlling the substantive content and the spin on the issue. This chapter suggested that a referendum issue can gain in importance during a campaign and that exposure to news media reporting about the referendum contributes to this increase in salience, when controlling for other inﬂuences. Moreover, the chapter addressed the multi-dimensional nature of referendum issues. It was demonstrated that though a referendum issue may increase in salience on both the media agenda and the public agenda, the connotation and meaning of the referendum issue on the two agendas may be two very different things. The next chapter investigates how news coverage of a referendum is related to the evaluation of the campaign and political cynicism.

7

Polls, strategy news, and political cynicism

Analyses of news media coverage of elections and politics document a change in the balance of news coverage from issue-based stories to strategic coverage emphasizing who is ahead and behind and the strategies and tactics of campaigning necessary to position a candidate to get and maintain a lead position (Jamieson 1992; Patterson 1993). Jamieson (1992) identiﬁed strategic coverage as winning and losing being the central concern, proliferation of language of wars, games, and competition, centrality of performance, style, and perception of candidates, and heavy weighing of polls and candidate ratings. These indicators share common ground with Patterson’s (1993) notion of game coverage as characterized by the game of the campaign providing the plot of the news story, polls, and positioning the electorate as spectators and candidates as performers. Cappella and Jamieson (1997) argue that this mode of coverage invites the attribution of cynical motives to political actors in campaigns and public policy debates and by reporting about politicians and their policies repeatedly framed as self-interest and seldom in terms of the common good – whether such characterizations are correct or incorrect – the public’s experience of their leaders is biased toward attributions that induce mistrust. (Cappella and Jamieson 1997: 142) Other studies have contributed with evidence of negative effects of strategic news on the public’s perceptions of politics. Rhee (1997), for example, investigated how strategic campaign coverage affected individuals’ interpretations of the campaign. Based on two ﬁeld experiments it was found that the strategic frame has broad implications for the interpretation of the campaign. Miller et al. (1979) investigated the relationship between the degree of negative political criticism found in newspapers and their readers’ feelings of trust in government and a sense of their own political effectiveness. Drawing on 1974 American National election survey data and a front-page content analysis of newspapers, they found

Polls, strategy news, and political cynicism

121

that “although newspaper reporting was primarily neutral or positive, readers of highly critical papers were more distrustful of government” (Miller et al. 1979: 67). The studies suggesting negative effects of news coverage of politics on the public perception of politics are supported by, for example, Neustadt (1997: 97) who claimed that the press “is very much an actor in today’s political drama, conveying a steady stream of unambiguously negative cues about government and politics.” Such conclusions are balanced by studies reporting no substantial effects of news exposure on, for example, trust in political institutions. In relation to the 1997 British general election, a series of experiments was carried out with more than 1,100 participants. By manipulating the stories shown to the participants in the various experimental conditions, it was found that voters’ attitudinal responses varied systematically according to the types of television news they were shown, so that television viewers exposed to positive news coverage of the Conservative and Labour Parties displayed a strong improvement in party support (Norris et al. 1999: 149). Hibbing and Theiss-Morse (1998) tested medium speciﬁc effects on emotional reactions versus cognitive evaluations drawing on 1992 national survey data in the U.S. They found that people who primarily obtain their news from television or radio are not any more or less likely to evaluate Congress negatively than are people who primarily obtain their news from newspapers. Similarly people who are exposed to news a great deal do not evaluate Congress more negatively than those who pay little attention to news. (Hibbing and Theiss-Morse 1998: 475)

Strategic news and electoral mobilization Studies have also moved to investigate the effects of negative campaign communication on voter mobilization and turnout. These have, by and large, focused on the effects of negativity in political advertising and commercials. Ansolabehere et al. (1994, 1999) challenged the notion that campaigns boost citizens’ involvement (in terms of interest, awareness, and sense of importance) as suggested in classical studies (for example, Berelson et al. 1954). They found that exposure to negative advertising dropped intentions to vote by 5 percent and concluded that the demobilizing effects of negative campaigns are accompanied by a weakened sense of political efﬁcacy. Voters who watch negative advertisements become more cynical about the responsiveness of public ofﬁcials and the electoral process. (Ansolabehere et al. 1994: 829)

122

Campaign effects

The ﬁndings in this ﬁeld of research are mixed and the Ansolabehere et al. (1994) argument has later been contrasted and disputed by, for example, Finkel and Geer (1998) who, drawing on content analysis of presidential ads and panel survey data, found no demobilizing effects of negative ads. Freedman and Goldstein (1999) even found that exposure to negative ads increased likelihood of voting. Wattenberg and Brians (1999) questioned the demobilizing effects of negative campaigns and attack advertising by arguing that the intent of these campaign messages is to change voters’ minds and voting preferences rather than to demobilize the electorate and depress turnout. In Europe, turnout levels for national elections are, generally speaking, considerably higher than in the U.S. For elections for the European Parliament, however, the average level of turnout in 1999 was very comparable to the turnout in the 2000 U.S. Presidential election (around 50 percent). In referendums turnout can vary considerably. Traditionally, about 80 percent of the Danish voters turn out to vote in referendums on issues of European integration.

Political cynicism Despite the widespread agreement about the importance of scholarly research on (changes in) political cynicism (for example, Cappella and Jamieson 1997; Erber and Lau 1990; Hetherington 2001; Valentino et al. 2001a), there is little agreement on the conceptualization and measurement of cynicism. Cappella and Jamieson (1997) in experimental studies utilized a number of items dealing with both politicians’ motivations and evaluation of the election campaign. A number of other studies in the U.S. have also discussed political cynicism from the perspective of research on political efﬁcacy (for example, Acock and Clarke 1990; Acock et al. 1985; Craig et al. 1990; Niemi et al. 1991). Outside of the U.S., discussions of political cynicism have been related to trust in different social, economic, and political institutions (Mishler and Rose 2001). Missing in most of the previous work is a distinction between the different dimensions of political cynicism. Cynicism as absence of trust has been treated at the level of the institutions of government and the regime as a whole (for example, Miller 1974) while others interpret cynicism as negativism towards and disapproval at the level of candidates and incumbent political leaders (for example, Citrin 1974). This observation led to a distinction between political cynicism directed towards “persons” on the one hand and “issues and institutions” on the other (Erber and Lau 1990). Apart from reactions directed towards “persons” and “issues/institutions,” an additional dimension of cynical perceptions of politics is campaign evaluation. Beyond how people see candidates and political institutions, it is also important to consider evaluations of the campaign. The link between exposure to strategic and negative news coverage and

Polls, strategy news, and political cynicism

123

campaign evaluations has not received much attention. One study that tapped campaign evaluation included survey items such as “the campaign was concerned with policies to meet the city’s need” and “the campaign was concerned with standing in the polls” (Cappella and Jamieson 1997: 266). In addition, an experimental study (Valentino et al. 2001b: 93) concluded “campaign news that emphasizes the strategic motivations for a candidate’s promises increases negative evaluations of the campaign.”

Key questions This chapter deals with changes in political cynicism towards political candidates, as well as evaluations of the referendum campaign, and the impact of strategic news exposure on these opinions. The majority of studies of effects of campaign communication on public perceptions of politics and elections and likelihood of voting is experimental. Moreover, most studies have been conducted in the United States. Given these characteristics of previous research, the aims of this chapter are: ﬁrst, to investigate the effects of strategic news coverage of a referendum on the public perception of politics. We focus on news rather than political advertising since the news media are the most important source of political information (Eurobarometer; Kahn and Kennedy 1999; Zhao and Chaffee 1995). The second aim is to expand the scope of research on the effects of strategic news coverage by investigating a referendum campaign in Europe with a political system and media landscape that is fundamentally different from the U.S. As Agger et al. (1961: 500) argued more than four decades ago, research on political cynicism in more geographical areas is needed for producing generalizations. Research may otherwise entail the danger of “naive universalism” (Gurevitch and Blumler 1990: 308) and an ethnocentric bias by offering generalizations of theoretical propositions which are based on single-country data only. The third aim is to provide a more compelling link between measures of exposure to the speciﬁc content of different news outlets and the effects on perception and evaluation of the campaign. Rather than drawing on an experimental design, we utilize our panel survey data in combination with our media content analysis. To ﬁrst explore the antecedents of political cynicism, we test a model, built on previous research, including partisanship/ideology (Agger et al. 1961), and gender, age, education, and political efﬁcacy (Mishler and Rose 2001; Pinkleton and Austin 2001) as predictor variables for understanding political cynicism. We additionally include “government disapproval” and “EU skepticism” as independent variables as these are powerful predictors of vote choice in the euro referendum (see Chapter 9). Building on this initial exploration of the antecedents of political cynicism, we investigate the effects of strategic news coverage of politics on public perceptions

124

Campaign effects

of the referendum campaign. Based on earlier ﬁndings (Cappella and Jamieson 1997; Valentino et al. 2001a) on the effects of strategic news coverage on the perception of politics, we expected that exposure to strategic news reporting increases the level of political cynicism during a referendum campaign. Second, we consider the effects of strategic news coverage on campaign evaluations. In line with previous studies (Cappella and Jamieson 1997; Valentino et al. 2001b), we expect that strategic news fuels negative perceptions of the campaign. We are also interested in the relationships between the effects of strategic news coverage on cynicism and campaign evaluations, on the one hand, and voter mobilization on the other. In the U.S., strategic news coverage has been said to contribute to diminished engagement with politics (Cappella and Jamieson 1997), but levels of turnout in U.S. elections are on an entirely different order in comparison with elections in Scandinavian countries where it remains above 80 percent. We also know that in Denmark, turnout in referendum campaigns on European issues has remained relatively high and the 2000 referendum was no exception. We therefore conclude with a discussion of the implications of our ﬁndings for mobilization in the referendum context. We ﬁnish the chapter with a discussion of the effects of polls in the news. As mentioned earlier in this chapter, polls and news driven by polls and public opinion expressions is part of the notion of strategic news coverage. We therefore investigate the public perception of and support for utilizing polls in the news.

Political cynicism and negative campaign evaluations For initial insight into political cynicism and public evaluations of the campaign,1 we begin with a discussion of our different measures of political cynicism and campaign evaluation on the aggregate level. We measured political cynicism directed at candidates (Erber and Lau 1990) with two items tapping trust in the motivations of political candidates.2 Four standard items used in previous research (Cappella and Jamieson 1997) were used to build a scale of campaign evaluation.3 We found that at the outset of the ﬁnal month of the campaign, the electorate displayed a fairly high level of political cynicism and strong negative evaluations. Twelve per cent found the candidates “open and honest,” for example, and more than 60 percent considered the campaign more about strategy than content (Table 7.1). We also found that, at the aggregate level, the public’s political cynicism and negative campaign evaluations increased during the campaign. Between the two waves, the conﬁdence and trust in the political candidates dropped and, at the end of the campaign, three of four respondents considered the campaign to be more about strategy than content. To investigate the effect of strategic news coverage on political cynicism

Polls, strategy news, and political cynicism

125

Table 7.1 Aggregate level change on perceptions of the campaign Wave 1 Wave 2 Change (n ⫽ 946) (n ⫽ 935) Political cynicism The candidates are open and honest The candidates are too superﬁcial in arguments Campaign evaluation The campaign is about what is best for Denmark The campaign is too much about Yes and No sides in polls The campaign gives me sufﬁcient information to make choice The campaign is more about strategy than content

12 58

8 61

⫺4 ⫹3

53 70

31 71

⫺21 ⫹1

31

31

–

62

78

⫹16

Note Cell entries are percentages answering “strongly agree” or “agree.” Post stratiﬁcation weights on age, gender, and education are added for descriptive data.

and campaign evaluation, we include a number of control variables, including standard social-demographics (gender, age, and education), political efﬁcacy4 campaign involvement,5 government approval,6 and EU skepticism.7 The composition and question wording of these variables is listed in the Appendix. Figures 7.1 and 7.2 report the results of two multiple Ordinary Least Square (OLS) regression models investigating the antecedents of political cynicism and campaign evaluations. The ﬁgures show a similar pattern predicting political cynicism and negative campaign evaluations. The more Feeling politically efficacious ⫺0.28

Disapproving of the government

⫹0.15

⫹0.26

Skeptical about the EU

Political cynicism

⫺0.06

Women

Figure 7.1 Predicting political cynicism. Note Entries are standardized beta-coefﬁcients (n ⫽ 892, R2 0.27). Positive sign of betacoefﬁcient indicates positive association with political cynicism. Only signiﬁcant indicators are shown in the ﬁgure. Controls were made for age, education, and political ideology. The full regression analysis is provided in the Appendix.

126

Campaign effects Feeling politically efficacious ⫺0.27

Disapproving of the government

⫹0.07

⫹0.27

Skeptical about the EU

Political cynicism

⫺0.14

Age 40–54 years

Figure 7.2 Predicting negative campaign evaluation. Note Entries are standardized beta-coefﬁcients (n ⫽ 892, R2 0.27). Positive sign of betacoefﬁcient indicates positive association with political cynicism. Only signiﬁcant indicators are shown in the ﬁgure. Controls were made for age, education, and political ideology. The full regression analysis is provided in the Appendix.

politically efﬁcacious tended to be less cynical and less negative in their evaluation than other voters. Respondents disapproving of the incumbent government’s performance record tended to display higher levels of cynicism and negative evaluations together with the anti-European integration segments of the electorate who were also more cynical and negative than respondents in favor of enhanced European integration. In addition, women were less cynical about the candidates and younger voters tended to evaluate the campaign more negatively than elderly voters.

The effect of strategic news coverage on political cynicism and campaign evaluations To test the effects of the campaign on political cynicism and campaign evaluations, we ﬁrst reiterate one of the ﬁndings of the media content analysis. In Chapter 4 we showed that the question of who was ahead and behind was a prominent feature of the coverage. In Chapter 5 we drew on different indicators of strategic news to show that about half of the news dealt with strategy-related topics. Reference was made to the presentation and style of a politician and/or party in one-fourth of the stories and the actions of politicians and/or parties were interpreted in terms of strategy and consolidation of their position in one in three news stories. Given this similarity between the news outlets in terms of reporting about the referendum strategically, we use one additive measure for media exposure in the subsequent analyses. To test the effect of the campaign, and in particular campaign commu-

Polls, strategy news, and political cynicism

127

nication, we conducted an analysis to explain cynicism and negative campaign evaluation at the end of the campaign. Here we include a number of campaign speciﬁc predictors engaging in discussions of the referendum issue, campaign involvement, and vote intention. Finally, we include exposure to strategic news coverage: respondents reported the number of days per week that they read a daily newspaper and watched a news program on television. We created an additive index for exposure to news about the referendum. The exposure to each of the seven news outlets was multiplied by the frequency (number of days per week) to derive a “total media exposure” indicator. For other analyses it is important to distinguish medium differences (television versus press) or outlet differences (e.g., TV1 versus TV2), but given the high degree of similarity in terms of reporting strategically about the referendum (see results above), it is appropriate to consider exposure to the campaign in the media as unidimensional in this context. We also control for each respondent’s level of cynicism and evaluation of the campaign prior to the ﬁnal month of the campaign to investigate the speciﬁc contribution of the ﬁnal weeks of the campaign to the degree of political cynicism and negative campaign evaluation displayed at the end of the campaign (Markus 1979).8 We expected that exposure to news reporting about the referendum campaign in terms of strategy would increase the level of cynicism about the campaign. Figure 7.3 shows that exposure to strategic news was positively associated with an increase in the level of cynicism. The hypothesis is supported in the ﬁnding that the more respondents were exposed to news about the referendum framed in terms of strategy, the more cynical

Involvement in the campaign ⫺0.10

Disapproving of the government ⫹0.11

Exposure to strategic news about the referendum

Right political leaning

⫹0.09 ⫺0.06

Political cynicism

⫹0.56

Political cynicism (time 1)

Figure 7.3 Predicting change in political cynicism. Note Entries are standardized beta-coefﬁcients (n ⫽ 752, R2 0.29). Positive sign of betacoefﬁcient indicates positive association with political cynicism. Only signiﬁcant indicators are shown in the ﬁgure. Controls were made for age, education, and political ideology. The full regression analysis is provided in the Appendix.

128

Campaign effects

they became, also when controlling for their initial level of political cynicism at the outset of the campaign. We also expected that exposure to news reporting about the referendum in terms of strategy would increase the level of negative evaluations of the campaign. This hypothesis is also supported (Figure 7.4) as the more respondents were more exposed to news about the referendum framed in terms of strategy, the more negative they became in their evaluation of the campaign. Additionally, the ﬁndings show that respondents who became more disapproving of the incumbent government during the campaign became both more cynical and more negative about the referendum campaign. Respondents with a preference for the political right became less negative about the campaign over the course of the campaign. Finally, the analysis also showed that involvement in the campaign by, for example, attending public meetings about the euro, was negatively associated with an increase in cynicism. In other words, those more involved in the campaign became less cynical during the campaign.

Strategic news and mobilization Using our measure of voting behavior in the second wave of the panel we investigated the relationship between strategic news and electoral turnout. Respondents reported whether they voted Yes, No, intentionally left the ballot blank or did not vote at all. The actual turnout in the referendum was 87.5 percent and our panel reported a 96 percent turnout. Bias in selfreported measures of turnout are well known (Belli et al. 1999). The eight percentage point discrepancy between our reported survey data and actual

Involvement in the campaign ⫺0.07

Disapproving of the government

⫹0.06

⫹0.08

Exposure to strategic news about the referendum

Political cynicism

⫺0.43

Negative campaign evaluation (time 1)

Figure 7.4 Predicting change in negative campaign evaluation. Note Entries are standardized beta-coefﬁcients (n ⫽ 752, R2 0.29). Positive sign of betacoefﬁcient indicates positive association with political cynicism. Only signiﬁcant indicators are shown in the ﬁgure. Controls were made for age, education, and political ideology. The full regression analysis is provided in the Appendix.

Polls, strategy news, and political cynicism

129

turnout is in line with other studies of elections and national referendums in both the United States and Scandinavia (Granberg and Holmberg 1991). This level of turnout, both actual and reported, almost makes the dependent variable in an analysis a constant, which is why no table will be presented to address the question of whether exposure to news media covering the referendum in terms of strategy and political cynicism (de)mobilized the electorate. A logit regression predicting turnout suggested that being “undecided” about vote choice at the start of the hot phase of the campaign had a marginal effect on not turning out to vote, whereas engaging in discussions about the euro issue with friends or colleagues had a negligible mobilizing effect. The analysis suggested that exposure to news about the campaign, framed strategically, had neither a mobilizing nor a demobilizing effect. Our discussion therefore focuses on how, given the conditions found in this campaign, mobilization could remain so high.

Strategic news and political cynicism in a referendum Our investigation of the effects of strategic news coverage of a national referendum on political cynicism and campaign evaluation, and possible consequences for mobilization leads to the following conclusions. First, the level of political cynicism about political candidates was fairly high one month prior to referendum day, the beginning of the “hot phase” of the campaign. Political cynicism and negative campaign evaluations time were less driven by standard demographic predictors such as gender and education. These attitudes were better explained by political efﬁcacy that was negatively associated with political cynicism. This is supportive of Pinkleton and Austin’s (2001) exploration of antecedents of cynicism. Our analysis also showed that more context-bound predictors such as government disapproval and general negative attitudes towards advanced European integration were important for understanding political cynicism and negative campaign evaluations at this stage in the campaign. Second, the “hot phase” of the campaign produced an increase in the level of political cynicism and an increase in negative campaign evaluations. The news media contributed to this increase in cynicism and negativity. Persons who were exposed the most to strategic news about the campaign, even when controlling for a number of other inﬂuences, displayed the strongest increases in cynicism and negative evaluations of the campaign. Our content analysis of the news media coverage of the campaign suggested that the news media emphasized strategy. Only minor differences were found between the different news outlets in terms of the emphasis on strategy, which suggests a fairly homogeneous reporting style. This degree of similarity across the media market in terms of strategic reporting may explain the consistent positive contribution of news exposure to increases in political cynicism and negative campaign evaluations. These results echo ﬁndings by Cappella and Jamieson (1997) who also

130

Campaign effects

suggest that the more strategic news coverage a voter is exposed to, the higher the level of political cynicism. Strategic news coverage – and the cynicism and negativity that it fueled – did not appear to have any detrimental inﬂuence on turnout or mobilization, as has been suggested in previous U.S. studies of the effects of negative advertising (for example, Ansolabehere and Iyengar 1995; Ansolabehere et al. 1999). What might explain these cross-national differences? Since the evidence from the United States on the effects of negative campaigning on turnout is not consistent, we need to address the speciﬁc conditions under which such negative effects may occur. This referendum campaign took place in a country with a fundamentally different political tradition from the United States, one in which voters are generally engaged, and where participation in national elections is high (above 80 percent of the eligible voters). Turnout in the past referendums concerning Europe was also over 80 percent with two low-points with 75 percent turnout. We do not know much about the campaign information environments in these previous referendum campaigns, though we do know that the news media coverage of referendums in the early 1990s was characterized as playing “an informative role during the campaign” (Siune 1993: 102). In these referendums media coverage might have been less focused on strategy and turnout was high. In the 2000 referendum we found substantial strategic news coverage and turnout remained the same. Admittedly these observations do not merit any conclusions about causal relationships but they do point out that strategic news and negative campaigning does not per se demobilize. Any effects of strategic campaign news coverage on a Danish electorate may be very different from effects in a U.S. context. Based on a study of media content and uses in U.S. senate campaigns (Kahn and Kennedy 1999; Valentino et al. 2001b) there is evidence to suggest that voters with different levels of political involvement or knowledge respond differently to negative campaigning, so that the same kinds of messages may serve to mobilize some (the highly involved) and demobilize others (the uninvolved) (see also Chapter 8). We also know that in Denmark, the levels of political knowledge among citizens in general, and in particular in this referendum campaign were quite high and displayed little variance or change during the campaign (Andersen and Hansen 2001).9 All this is meant to suggest that in a context in which most citizens were highly aware of the issues at stake, strategic or negative news while increasing cynicism and negative campaign evaluations may have played little or no role in mobilizing voters to go to the polls on the day. It may very well be that voters have the capacity to distinguish between their cynical views of politics and the importance of participating in an election. Earlier research suggests that there is a difference between political cynicism about political candidates on the one hand and political institutions on the other (Erber and Lau 1990). Voters may be dissatisﬁed,

Polls, strategy news, and political cynicism

131

cynical, and negative, but still mobilized and sufﬁciently engaged to turn out to vote. In fact, experimental evidence points in a very similar direction. In a recent study, it was conﬁrmed that exposure to news framed in terms of strategy activated political cynicism and negative associations with European issues. However, exposure to strategic news did not suppress participants’ willingness to turn out to vote (de Vreese 2004b). Such an explanation ﬁnds resonance in recent work in political science trying to make sense of the Clinton era in American politics. Popular wisdom during the Clinton presidency and later studies suggested that citizens were stunningly capable of distinguishing their rating and evaluation of Clinton as a person, on which he scored poorly, and as a president, for which he received high ratings (see, for example, Cohen and Hamman 2000; Shah et al. 2002; Zaller 1998). The electorate was relatively cynical about the campaign and although the level of cynicism increased over the course of the campaign, partly due to the media coverage of the campaign, this did not result in an apathetic, disengaged, and disinterested electorate. The case study suggests that voters may be cynical and skeptical, the news media may even contribute to this during a campaign, but this may not have the detrimental effects on democracy, at least in terms of mobilization, which is often an (implied) assumption in the “demobilization” literature. Our ﬁndings dovetail with ﬁndings of, for example, Cappella and Jamieson (1997) and Valentino et al. (2001a) on the contribution of strategic news coverage to political cynicism and negative campaign evaluations, but our ﬁndings are not linked in any way to diminishing turnout.

How the public was turned off by polls Having looked at the campaign coverage and its effect on political cynicism in more detail, we return to the role of opinion polls in the news (see also Chapter 4) and their effects. Polls and public opinion-driven coverage is related to strategy news coverage and plays a speciﬁc role in election campaigns. Though polls may not contribute directly to political cynicism, as demonstrated by Valentino et al. (2001a), polls in the news may affect public perceptions of polls themselves. The centrality of public opinion polls in politics today is well documented (Lavrakas and Holley 1990; Lavrakas and Traugott 2000). Most of the research on the use and impact of opinion polls has been conducted during two-party or multi-party election campaigns and is largely focused on the United States where the First Amendment makes it impossible to restrict rights of news organizations to publish information about public opinion (for example, Frankovic 1998; Traugott and Lavrakas 1996). Previous studies have either studied the quantity and quality of news reports about opinion polls (for example, Anderson 2000; Brettschneider 1997; Smith and Verall 1985) or the effects of opinion polls and exit polls during

132

Campaign effects

national and state level elections (for example, Atkin 1969; Ceci and Kain 1982; Sudman 1986). In Chapter 4 we assessed the quantity and quality of poll reporting in the Danish news media. In this chapter we turn to the effects of polls. Denmark remains one of the countries where there are currently no legal restrictions on the publication of polls at election time. There is a longstanding gentleman’s agreement, however, to end campaigning the day before an election and not to publish exit polls before voting has ended (Siune 1987). This corresponds with a number of countries where pollsters, networks, and others engaged in exit polling accept a “self-denying ordinance not to release their ﬁgures until the last vote has been cast” (Butler 1996: 247). As discussed in Chapter 4, however, TV2 released two exit polls during the election day when the voting stations were still open. Previous research has suggested a variety of possible effects of polls in campaigns. There are possible effects on the candidates themselves, their campaign organizations and campaign strategies and, in the U.S. context especially, their ability to continue to raise money which is an essential element of their continued viability (Barber 1978). There is also concern about the potential effects of polls on the reputations of the pollsters themselves and on the media organizations that emphasize polls in their coverage of campaigns. This has been a major concern after the pre-election polls and/or exit polls have turned out to be far away from the actual result. In the aftermath of the British 1992 general election campaign, for example, because the pre-election polls had been so far off from the actual result, a special commission was set up to investigate what went wrong and who was to blame (Curtice and Payne 1995; Waller 1995; Worcester 1995; Wybrow 1995). Similar concerns were raised in the aftermath of the 2000 U.S. presidential elections where major networks claimed a winner based on ﬂawed exit polls (Frankovic 2001). Effects on voting behavior may be direct (on turnout or vote choice) or indirect (via effects on perceptions of candidates, parties, and issues). Direct, speciﬁc effects on mobilization and turnout have been the subject of studies of the reporting of exit polls in U.S. presidential campaigns (see, for example, Epstein and Strom 1981; Sudman 1986). Bandwagon and underdog effects refer to the possible inﬂuence of polls on vote choice, speciﬁcally to jump on the bandwagon of a candidate who appears to be ahead, or to ﬁght for the candidate who is behind, the underdog, in the race. These effects have been a topic of concern since the ﬁrst U.S. election studies in the 1940s (Atkin 1969; Ceci and Kain 1982; Cook and Welch 1940; Glynn and McLeod 1982; Lavrakas and Holley 1990; Marsh 1985; Traugott and Lavrakas 1996). Indirect effects on voting behavior include those poll effects on voters’ perceptions of candidates’ personal qualities and electoral chances, and candidates’ and parties’ stands on issues, and perceptions of the importance of issues, as well as voters’ opinions about the direction of policy. A

Polls, strategy news, and political cynicism

133

review of the literature on media polls and their effects on vote preference and voting behavior concluded: For those who expected the main effects to be large-scale shifts in voter preference or actual voting behavior based on information about what other citizens are thinking or doing, these effects cannot be demonstrated conclusively . . . The inﬂuences of poll reports on voter preference are likely to be small, although not insigniﬁcant in close races, and to offset one another. (Traugott 1992: 145–146)

The perceived importance of opinion polls We investigate the effects of polls and poll reporting that are considered by the public. To answer this exploratory question about the perceived impact of opinion polls, the focus group interviews were used. Participants from all groups (see Chapter 2) expressed frustration about the frequent use of polls in the news coverage of the campaign: All those percentages, it doesn’t grab my attention. I don’t remember it. It is just used for mud-throwing. (No voter, female, 51) It is annoying, confusing with all those polls. (Yes voter, female, 63) I cannot retell the news, it’s two steps ahead and one back in the polls, they do keep you listening. (Undecided, male, 58) What I missed were references to studies telling us who changes their mind and why, not just the numbers. (Undecided, male, 31) Participants from the “no” and “undecided” group more often expressed concerns and thoughts about the use of polls that ﬁnd resonance in the literature on third person perception and bandwagon effects. Many don’t know what to vote. They use the polls: “Oh, now a lot of people say yes, well, then I better vote yes.” (No voter, male, 35) It makes people confused. If they don’t know what to vote they are confused if they hear that one day “no” is ahead and the next day “yes.” (Yes voter, female, 47)

134

Campaign effects

Evaluation of the use of polls during the campaign The impression expressed by participants in the focus group interviews that opinion polls played a too dominant role in the campaign, was reﬂected in the responses in the survey. About 70 percent of the electorate held the opinion that the campaign was too focused on polls. This perception remained stable over the two waves with only a negligible increase at the aggregate level. The opinion was pervasive and shared by a majority with no age or gender differences. Persons who voted “No” in the referendum and thus belong to the successful side were less likely to express a negative impression of the presence of polls in the campaign compared to persons who voted “Yes.” Moreover, we found that media exposure contributed to a negative evaluation of the use of polls in the campaign (see Figure 7.5). Qualitative data from our focus groups suggested two things. First, voters perceive polls as important but also challenging to understand and grasp. Second, voters believe polls to be inﬂuential and powerful, and potentially harmful. The focus groups provide qualitative support for Perloff’s (1993) suggestion that fear of effects are larger when it is seen to be inappropriate to be inﬂuenced by the message. The focus group data also provided evidence that dovetails with survey research on third person effects of the perceived impact of polls (for example, Price et al. 1999) and literature on bandwagon effects (for example, McAllister and Studlar 1991). Participants in the focus groups, without being probed as in survey questions, in their own words formulated these concerns about such effects of polls. The electorate’s evaluation of the use of polls in the campaign indicated Age ⫹0.09

Exposure to television and newspaper coverage of the referendum

⫹0.08

Expressing concern that “the campaign was too dominated by polls”

⫹0.30

Expressing concern that “the campaign is too dominated by polls” (time 1)

Figure 7.5 Predictors of negative change in perception of role played by polls in the referendum campaign. Note Result of OLS regression, entries are standardized beta-coefﬁcients. Only signiﬁcant indicators are shown in the ﬁgure. Controls were made for gender, education, and vote choice in the referendum. The full regression analysis is provided in the Appendix.

Polls, strategy news, and political cynicism

135

a general negative attitude towards the dominance of opinion polls in the campaign. This impression did not change during the ﬁnal month of the campaign, though the content analysis of the media coverage suggested that the amount of polls and poll coverage increased in the ﬁnal phase of the campaign. The negative evaluation of the role of polls in the campaign is in line with ﬁndings from the 1992 U.S. presidential elections where a larger majority of the electorate said that the nation would be “better off” with less opinion polls (Traugott and Kang 2000: 196). As we saw earlier in this chapter, news media coverage of politics in terms of strategy may encourage negative evaluations of politics and campaigns. Comments from participants in the focus groups suggest that a strong and sensationalist focus on polls, which is almost inherent to politics of today (for example, Brettschneider 1997), may generate a negative view of the campaign and these ﬁndings were echoed by respondents in our panel survey. Having examined the effects of the campaign on political cynicism and the perception of the referendum, we now turn to the evaluation of political leaders and the government performance rating during the referendum campaign.

8

Primed by the referendum

In a referendum, the key issue is deﬁned in advance but as we have seen in the previous chapters, the way in which the issue is framed by the Yes and No camps, and the news media, may vary considerably. Do political leaders gain or lose from a referendum campaign? The ﬁrst goal of this chapter is to investigate priming in the context of a national referendum. Political priming theory suggests that political leaders are evaluated on the basis of readily accessible information. A referendum campaign potentially has signiﬁcant political implications. The majority of studies on political priming stems from key historical events: the Iran Contra affair (Krosnick and Kinder 1990), the 1991 Gulf War (Iyengar and Simon 1993; Krosnick and Brannon 1993), and the later bombings of Iraq (Edwards and Swenson 1997), and major European events such as a summit of European heads of state (Semetko et al. 2003a, 2003b). There is also some evidence of the priming phenomenon in the context of general elections (Gidengil et al. 2002; Mendelsohn 1996), but the process of priming has not yet been investigated in the context of a referendum campaign. The information provided by the media, in particular news media, is a key source of information and cues for citizens. News media highlight key issues that enable citizens to “satisﬁze” rather than “optimize” when making judgments and evaluations. That is to say that they rely on information that is most easily brought to mind (Krosnick et al. 1996). Iyengar and Kinder’s (1987) experimental studies effectively introduced priming in the realm of political communication.1 They concluded that news media, in particular television, by calling attention to some matters while simultaneously ignoring others, inﬂuenced the standards by which the U.S. President is judged. In the same vein, one study showed that when the U.S. news media covered the 1991 Gulf War extensively, President Bush Sr’s performance was evaluated on his warfare abilities. Less than a year later, the news was focused on the state of the American economy and Bush was evaluated on his economic performance (Jasperson et al. 1998).

Primed by the referendum 137

Of presidents and prime ministers Most evidence of priming comes from the United States in the form of explanations for ﬂuctuations in presidential evaluations. Iyengar and Kinder (1987) propose priming to be a general phenomenon, but it is an empirical question whether this assumption is correct. Studies of voting behavior and turnout, for example, have suggested that different political systems (two-party vs. multi-party) and political culture affect political behavior and communication (for example, van der Eijk and Franklin 1996; Semetko 1996b). Most Europeans live in countries with large multiparty systems. This implies a fundamentally different ﬁeld of political competition. Though the Prime Minister is an important political actor, the importance can be negligible in comparison with the signiﬁcance of the Presidency in the United States. Therefore, the question is whether Prime Ministers and other political leaders in a diverse political ﬁeld can be “primed” in a fashion similar to what studies in the North American context has suggested. With the exception of one study that examined the priming effects of the news about a European Union summit of heads of state (Semetko et al. 2003a), we know little about priming effects that might emerge under the conditions of the multi-party parliamentary systems in European democracies, where coalition governments are the norm. Even though attention in the news is focused heavily on the Prime Minister in these European multi-party systems, it is also often spread across a number of other party leaders and political actors. This chapter explores priming effects in a multi-party context and looks at a broader range of political actors, beyond just the President or the Prime Minister.

The importance of the campaign Priming theory posits that citizens express evaluations based on newly acquired and readily accessible information. Citizens, in other words, are primed to use information about speciﬁc issues in general evaluations. Given that the vast majority of citizens get their information about politics via the media, it is important to study media coverage of a referendum campaign. Does the topic of the referendum appear on the news agenda so that citizens can make use of their associations with the referendum issue to formulate general evaluations? Pan and Kosicki (1997) found that the stronger themes were in predicting the overall assessment of the U.S. President, the stronger they turned out to have been in the media. This suggests a “hydraulic” character of media priming in which certain issues dominate and minimize the importance of other issues. To study priming in the context of a referendum it is therefore important to assess the visibility of the referendum topic on the news agenda as we did in Chapter 4.

138

Campaign effects

However, news media do not merely bring certain issues to the public attention. They also offer evaluations of issues and persons. These may affect public evaluations of political leaders and this may in turn affect vote preferences. We may hypothesize that strongly negative news about an issue that is also highly visible in the news bolsters negative evaluations of political leaders. Conversely, exposure to neutral or less negative news should mute negative evaluations of political leaders. Previous priming research has investigated the visibility of issues in the news, but has not been able to assess the evaluation of leaders in the news in order to take this into account in understanding the public evaluation of political leaders. This chapter not only discusses the visibility of a referendum in the news, it also assesses the tone of evaluations of political leaders to better understand the dynamic of public evaluations of political leaders as a function of exposure to speciﬁc news content.

The moderating role of political involvement The considerable variation in levels of political expertise among the public (Converse 1962, 1964; Luskin 1987) may inﬂuence the ways in which people take up and make use of information. The evidence is mixed about the contingent nature of priming. The inconsistency with respect to the potentially moderating role of political knowledge (or sophistication, expertise, and interest, terms that previous research has used interchangeably) can be organized as two dominant perspectives. Krosnick and Kinder (1990) found politically knowledgeable persons to be less susceptible to priming effects. This dovetailed with Iyengar et al. (1983) who found that persons with a high thematic knowledge were less susceptible to priming effects. Iyengar and Kinder (1987) found no systematic differences between political “experts” and “novices” in the magnitude of priming. Krosnick and Brannon (1993) found that political expertise facilitates priming effects and this argument was elaborated by Miller and Krosnick (2000) who found the strongest priming effects among politically knowledgeable participants. Similar to the mixed evidence on the moderating effect of political sophistication and knowledge, the existing literature is also inconsistent with respect to whether political involvement enhances or limits the priming process. The dominant perspective seems to suggest that greater priming effects are associated with lower levels of interest (for example, Krosnick and Brannon 1993). This observation ﬁnds resonance in the work of Lodge et al. (1989). Politically attentive citizens frequently update their evaluation of political leaders. These evaluations are largely based on previously acquired information so new information, such as provided by the news media, affects these individuals minimally. Conversely, citizens that are inattentive to politics are less likely to have evaluations of political leaders readily available. Consequently, such citizens are more likely to

Primed by the referendum 139 draw on recently acquired information to formulate evaluations. In effect, these individuals are more likely to consider only few, recently accentuated, domains when expressing an evaluation. Such information is provided by news media. Indeed, studies have suggested that priming manipulations may have the greatest impact when they occur without people paying much attention to them (Lombardi et al. 1987). In referendum studies, political involvement and political interest has been discussed in the context of turnout and electoral mobilization. Previous studies have suggested that those more interested in the topic of the referendum are more likely to support the proposal (Siune et al. 1994). Neijens et al. (1998), for example, found that the politically interested orient themselves and know more than those who are less politically interested. Interest and knowledge, although acknowledged by most of the referendum studies as leading one to be more likely to vote and more likely to pay attention to and learn over the course of the campaign, nevertheless often have not been integrated into the models of referendum voting, with rare exceptions (Buch and Hansen 2002; Jenssen et al. 1998; Johnston et al. 1996). However, these studies look at turnout and vote choice as the dependent variable and say little about the evaluation of political leaders. Moreover, the impact of political sophistication and involvement is not discussed as moderating the impact of the campaign in previous research.

Key questions This chapter investigates priming in the context of a referendum campaign in a multi-party political context with evaluations of the government and key political leaders as the dependent variables. The study not only investigates the importance of speciﬁc issues for the overall evaluation, it additionally assesses the role of news media in providing evaluative comments that may impact assessments of the incumbent government and political leaders. Finally, the moderating role of political involvement for enhancing or diminishing campaign effects is investigated. A referendum campaign is likely to generate media attention which, particularly in close races, will grow exponentially as Election Day approaches. We demonstrated in Chapters 4 and 7 that the topic of the referendum increased on the news agenda during the campaign. Dependent on the importance of the topic of the referendum, general priming theory posits that the evaluation of how this speciﬁc issue is handled will affect the overall assessment of political leaders such as the U.S. President. This chapter extends previous research to the context of a referendum in a multi-party system where the public’s evaluation of the incumbent Prime Minister and government are potentially primed by the issue of the referendum. We expect that the general public assessment of the government and Prime Minister is contingent on how they are perceived to be handling the referendum issue. Moreover, we also investigate whether a

140

Campaign effects

referendum affects the public evaluations of other political leaders, including members of the opposition. Finally, this chapter addresses the speciﬁc contribution of the campaign to the overall evaluation of political leaders. More speciﬁcally it is hypothesized that the overall evaluation of political leaders is contingent upon the visibility of the referendum issue in the media and the evaluative tone towards political leaders in the media in conjunction with the referendum issue. Based on existing research, we can cautiously hypothesize that exposure to news coverage evaluating political leaders negatively will lead to negative public evaluations and that this effect is likely to be largest among the least politically involved (Krosnick and Brannon 1993).

Aggregate level changes in public evaluations of political leaders To investigate the priming process during a referendum campaign we draw on our content analysis of the news coverage during the referendum and on our two-wave panel survey. We ﬁrst turn to the aggregate-level changes in evaluations of the government and key political leaders during the campaigns. The dependent variables are general evaluation of the government, the Prime Minister, and key political actors at the end of the referendum campaign.2 Table 8.1 shows that the overall government performance decreased signiﬁcantly during the campaign (from m ⫽ 3.3 to m ⫽ 3.0). The evaluTable 8.1 Change in overall and issue-speciﬁc government performance rating during the campaign Wave 1

Wave 2

Change during the campaign

Overall government performance rating (n ⫽ 912)

3.3a

3.0b

⫺0.3

Issue-speciﬁc government rating Euro (n ⫽ 878) Immigration (n ⫽ 908) Health care (n ⫽ 915) Environment (n ⫽ 894) Social issues (n ⫽ 908) Food safety (n ⫽ 882) Crime (n ⫽ 908) Elderly (n ⫽ 982) Unemployment (n ⫽ 884)

2.8a 2.3a 2.4a 3.5a 2.8a 3.2a 2.2a 2.4a 3.2a

2.1b 2.4b 2.7b 3.6b 2.9a 3.2a 2.5b 2.5b 3.2a

⫺0.7 ⫹0.1 ⫹0.3 ⫹0.1 ⫹0.1 – ⫹0.3 ⫹0.1 –

Note Performance is evaluated on a ﬁve-point scale ranging from 1 (very bad) to 5 (very good). Row values with different subscripts differ signiﬁcantly from each other at p ⬍ 0.01. Aggregate data weighted on social-demographic characteristics.

Primed by the referendum 141 ation of government performance on different issues showed the biggest aggregate-level change in the evaluation of the euro. This decreased signiﬁcantly (from m ⫽ 2.8 at wave 1 to m ⫽ 2.1 at wave 2). The performance rating on all other issues suggested either stability (i.e., no change) or a slight increase in the evaluation of performance on speciﬁc issues such as health care and crime. Table 8.2 shows that the evaluation of the political leaders changed signiﬁcantly during the campaign. The evaluation of political leaders endorsing a Yes fell, ranging from a 0.5 drop for the Prime Minister (Poul Nyrup Rasmussen) and a 0.2 drop for the Finance Minister (Marianne Jelved) to a 0.1 drop for the leader of the Conservative Party (Bent Bendtsen). The evaluation of the Liberal Party leader (Anders Fogh Rasmussen) was the exception to this pattern and his rating increased by 0.2 during the campaign. The evaluation of leaders endorsing a No increased slightly during the campaign (ranging from a 0.1 increase for the leaders of the popular movements against EU integration to a 0.2 increase for the leader of the Danish People’s Party, Pia Kjaersgaard). The exception to this pattern was the evaluation of the leader of the Socialist People’s Party (Holger K. Nielsen) whose sympathy rating dropped during the campaign.

The visibility and tone of the euro issue in the media during the campaign To have an indication of the visibility of the referendum in the media, we turn to our content analysis of television news from the outset of the campaign (March) to the end of the campaign (September 2000), including a total of 4,953 news stories.3 We found that the campaign accounted for about 7–10 percent of the main evening news during the ﬁrst ﬁve months of the campaign. The visibility of the referendum campaign increased dramatically and constituted about 25 percent of all news in in the ﬁnal month of the campaign (see Figure 4.1). In addition to our measure of the visibility of the referendum in the news, we also analyzed the evaluative tone of the news. A news story may portray an actor either neutrally (no evaluations), favorably, mixed, or unfavorably. For each actor the visibility and average tone was assessed by adding the number of negative evaluations subtracted by the number of positive evaluations divided by the total number of evaluations (see Chapter 4). In this chapter we focus on a selection of key political leaders. The analysis of the visibility and the evaluation of the most important political candidates and actors in the news showed that all actors received either neutral or negative coverage in the ﬁnal month of the campaign (Table 8.3). On the scale ranging from ⫺1 to ⫹1, most actors received a balanced mixture of positive and negative evaluations (resulting in a mean around zero) or a slightly negative overall evaluation (ranging from ⫺0.10 to

G G O O O O O NA NA

Y Y Y Y N N N N N

Prime Minister Finance Minister Leader Liberals Leader Conservatives Leader Socialists Leader Danish People’s Party Leader Christian Party Leader June Movement Leader People’s Movement

Title

Poul Nyrup Rasmussen Marianne Jelved Anders Fogh Rasmussen Bendt Bendtsen Holger K. Nielsen Pia Kjaersgaard Jann Sjursen Drude Dahlerup Ole Krarup

Name

5.2a 4.8a 6.2a 6.1a 5.8a 3.8a 5.5a 4.6a 4.3a

Wave 1

4.7b 4.6b 6.4b 6.0a 5.5b 4.0b 5.5a 4.7b 4.4b

Wave 2

⫺0.5 ⫺0.2 ⫹0.2 ⫺0.1 ⫺0.3 ⫹0.2 – ⫹0.1 ⫹0.1

Change during the campaign

Notes Performance is evaluated on a ten-point scale ranging from 1 (very unsympathetic) to 10 (very sympathetic). Row values with different subscripts differ signiﬁcantly from each other at p ⬍ 0.05. NA: The “June Movement” and “People’s Movement” are not represented in national Parliament and only campaign on European issues. Aggregate data weighted on social-demographic characteristics.

Government (G) or opposition(O)

Yes (Y) or No(N)

Table 8.2 Change in evaluation of political leaders during the campaign

10 0 6 5

No actors Leader of Socialist People’s Party, Holger K. Nielsen Other members of the Socialist People’s Party Leader of the People’s Party, Pia Kjaersgaard Other members of the People’s Party ⫺0.10 – 0.00 ⫺0.20

– ⫺0.18 ⫺0.11 0.00 0.00 0.00 0.00 0.00 17 5 8 7

– 19 15 8 4 1 1 5

⫺0.18 ⫺0.20 ⫺0.13 0.00

– ⫺0.26 ⫺0.20 ⫺0.25 0.00 0.00 0.00 0.00

18 7 4 7

7 27 21 13 7 6 4 7

⫺0.16 ⫺0.14 ⫺0.25 ⫺0.14

⫺0.43 ⫺0.44 ⫺0.29 ⫺0.23 0.00 0.00 0.00 0.00

Number of Mean mentions evaluation

Newspapers (n ⫽ 79)

Note Mean evaluation of the actors is based on a scale ranging from ⫺1.00 (consistently negative) 0 (neutral) to ⫹1.00 (consistently positive). An actor in a news story is deﬁned as a person/institution in the story who is mentioned at least twice (newspapers) or mentioned and seen at least once (television).

– 17 27 3 6 5 3 1

Number of Mean mentions evaluation

Number of Mean mentions evaluation

Yes actors Government Prime Minister, Poul Nyrup Rasmussen Other members of the Social Democrats Minister of Economy, Marianne Jelved Leader of the Liberal Party, Anders F. Rasmussen Other members of the Liberal Party Leader of Conservatives, Bendt Bendtsen Other members of the Conservatives

TV2 (n ⫽ 79)

TV1 (n ⫽ 76)

Table 8.3 Visibility and evaluation of selected actors in the campaign

144

Campaign effects

⫺0.30). The government, the Prime Minister (Poul Nyrup Rasmussen), the Finance Minster (Marianne Jelved), and other members of their political parties all received negative coverage in every media outlet. The most negative coverage of these actors appeared in the press and on TV2 news while TV1 carried less unfavorable (albeit still negative) evaluations. Other members of the Yes camp (such as Anders Fogh Rasmussen (Liberal) and Bendt Bendtsen (Conservatives)) were less visible in the news and received, on average, neutral coverage in all outlets. Members of the No camp were less visible than government actors. No camp actors received modestly negative news coverage, but this was signiﬁcantly less negative than the coverage of the Yes camp actors.

Effects of the referendum campaign and news media coverage on the evaluation of government and political leaders We modeled the effects of the referendum campaign along with the following predictor variables: issue-speciﬁc government evaluations, ideology, and exposure to news about the referendum. The issue-speciﬁc evaluations of topics such as the referendum (the euro), unemployment, health care, and immigration were measured on ﬁve-point scales. Ideology was tapped by a scaled measure of left–right placement.4 Exposure to the campaign is measured by four variables of news media exposure. Each variable indicates, on a scale from 0 to 7 days per week, the exposure to speciﬁc media outlets: public television news, commercial network news, Yes press, and No press. By utilizing the content analysis we are able to link the news exposure measures with precise information about the news content to which each respondent was actually exposed. Political involvement was measured by a four-point item tapping political interest. This variable was recoded into a categorical variable of “low political involvement” and “high political involvement.” As control variables, we included a series of social and demographic characteristics: gender (female), age (in years), and education (number of years in school). The model also includes the lagged term for the evaluation of the government and the political actors at wave 1 (see Markus (1979) for discussion of the use of lagged speciﬁcations in panel data). This makes it possible to control for initial evaluations and to examine change over the course of the campaign. The speciﬁc wording of all items can be found in the Appendix. The change in evaluation of the incumbent government and key political leaders during the campaign is investigated both at the aggregate and the individual level. The aggregate level change is assessed by comparing the overall evaluations in wave 1 and 2. To investigate change at the individual level, OLS regression models were speciﬁed. Turning to the individual level, Figure 8.1 shows that the change in

Primed by the referendum 145 How the government handles the referendum issue How the government handles unemployment

⫺0.25

⫺0.09

⫹0.09

Exposure to news on TV1 about the referendum

Overall government evaluation

⫹0.37

Overall government evaluation (time 1)

Figure 8.1 Explaining government evaluation, low political involvement citizens. Note Result of OLS regression, entries are standardized beta-coefﬁcients (n ⫽ 429, R2 ⫽ 0.61). Only signiﬁcant indicators are shown in the ﬁgure. Controls were made for gender, age education, issue-speciﬁc evaluation of other policy issues, exposure to other media outlets, and political ideology. The full regression analysis is provided in the Appendix.

overall government performance rating was largely driven by the negative change in the evaluation of the handling of the referendum issue. This ﬁnding is signiﬁcant when controlling for government performance rating earlier in the campaign and thereby including other predispositions driving government performance rating. For the low-involvement group, the evaluation of the handling of unemployment was a signiﬁcant negative How the government handles the referendum issue How the government handles immigration

⫺0.17 ⫺0.11

How the government handles social issues

⫺0.13

How the government handles food safety

⫺0.12

Right ideological preference

⫹0.44

⫹0.13

Overall government evaluation

Overall government evaluation (time 1)

Figure 8.2 Explaining government evaluation, high political involvement citizens. Note Result of OLS regression, entries are standardized beta-coefﬁcients (n ⫽ 326, R2 ⫽ 0.64). Only signiﬁcant indicators are shown in the ﬁgure. Controls were made for gender, age education, issue-speciﬁc evaluation of other policy issues, exposure to media outlets, and political ideology. The full regression analysis is provided in the Appendix.

146

Campaign effects

predictor for overall performance evaluation. For the high-involvement group, the evaluation of the handling of immigration and social issues were negative predictors for the overall performance evaluation. For the low-involvement group, exposure to news on TV1 was a signiﬁcant positive predictor for overall government performance. Table 8.3 showed that the government, including the PM and other members of the Social Democrats and its coalition partner receive either neutral coverage or negative evaluations on TV1. However, these evaluations were consistently less negative than the evaluations the same actors received on the news on TV2. Turning to the change in evaluation of the Prime Minister, the Finance Minister, the Yes camp actor, Anders Fogh Rasmussen and the central No camp actor, Pia Kjaersgaard, the results showed that the Prime Minister and the Finance Minister experienced the most dramatic drop in evaluations during the period of the referendum campaign. Table 8.4 shows that the evaluation of the handling of the euro was, with the exception of the lagged term for evaluation of the actors at time 1, the strongest negative predictor of the evaluation of the political leaders. This pattern is consistent for both highly and lowly politically involved respondents. The effects of exposure to different news media during the campaign were stronger on the low-involvement group, which supports our expectations. For the low-involvement group, the evaluation of the PM was positively affected by exposure to the Yes press. The evaluation of the Finance Minister was negatively affected by exposure to TV2 and positively affected by exposure to the Yes press. The evaluation of the No camp leader Pia Kjaersgaard was positively affected by exposure to TV2 and the No press while exposure to the Yes press contributed negatively to the evaluation. The evaluation of the leader of the Liberal Party was negatively affected by exposure to TV1.

Primed by the referendum This chapter investigated the dynamics of a referendum campaign and the ramiﬁcations for public evaluation of political leaders. The priming hypothesis was tested in the context of a national referendum on an issue of European integration. The chapter extended previous research to a European, multi-party context and investigated the impact of a referendum on public evaluations of the incumbent government, the Prime Minister as well as leading opposition leaders. Our study conﬁrms the basic hypothesis of political priming theory by showing that the overall evaluation of political leaders was strongly dependent upon an issue that was highly salient, readily available, and at the forefront of citizens’ minds. Drawing on the content analysis of news media we saw that the topic of the referendum (the introduction of the euro) became more visible in the

Primed by the referendum 147 media during the campaign. Utilizing our panel data we found that as the importance of the referendum issue increased in the media, its importance for the public for formulating general evaluations of political leaders increased as well. The incumbent government was seen by the public to handle the referendum topic poorly. Subsequently, the government, including its most prominent ﬁgures and key campaigners the Prime Minster and the Finance Minister, was evaluated on the basis of their performance on the referendum issue. These ﬁndings dovetail with extant priming research and conﬁrms an assumption in previous studies, namely that the phenomenon applies to political leaders more generally. In addition the ﬁndings suggest that not only contentious issues such as the Gulf War(s), but also issues of national referendums (such as on the introduction of the euro), can drive general evaluations of political leaders. Moreover, the ﬁndings suggest that a change in the news coverage is important to understand priming. While we saw in Chapter 6 that the issue of immigration was the most salient issue to the public, and it has been so for a long period of time, our analyses in this chapter showed that immigration did not affect the general public evaluation of the government. The issue-speciﬁc evaluation of immigration remained stable at the unfavorable end of the scale. This issue did not account for changes in the general evaluation of the government. A closer look at the media content during the campaign gives an explanation for the negligible importance of the immigration issue. Immigration was hardly a topic in the news during the campaign. The issue of the referendum, however, featured prominently and was the vehicle in the dynamics causing the decline in the overall government evaluation. Previous research has not been able to assess the effects of evaluations of issues and persons in the news media. Other research has shown that evaluations of political leaders affect public perceptions of these leaders (for example, McCombs et al. 2000). In this study we found support for our hypothesis that negative news about an issue that was also highly visible in the news bolstered negative evaluations of political leaders. Based on existing research (Krosnick and Brannon 1993), we cautiously hypothesized that politically less involved citizens exposed to news evaluating political leaders negatively would fuel negative public evaluations. The results supported this and showed the strongest media effects during the campaign among the least politically involved. Taken together, the results stress the necessity of, on the one hand, testing what is assumed to be a generally applicable process in other settings than the U.S. political context, and on the other hand, considering the campaign and the speciﬁc content of the media to understand ﬂuctuations in public opinion during a referendum campaign. The importance of the campaign in referendums is stressed in general terms (for example, LeDuc 2002, 2003) or at the descriptive level (for example, Siune and Svensson 1993). However, this chapter points to the importance of systematically

Evaluation health Evaluation social issues Evaluation environment Evaluation food safety Evaluation crime Evaluation unemployment

Evaluation immigration

Issue-speciﬁc evaluations Evaluation euro

– – – – – –

⫺0.21a (0.01) – – – – – –

⫺0.14b (0.01)

⫺0.14a (0.01) ⫺0.14a (0.01) – – – – – ⫺0.14b (0.01) – – – – – –

⫺0.12b (0.01) –

High

Low

Low

High

Finance Minister Marianne Jelved

PM Poul Nyrup Rasmussen

Table 8.4 Evaluation of political actors

– – – – – –

–

–

Low

– – – – – –

–

–

High

Opposition leader Pia Kjaersgaard

– – – – – –

–

–

Low

– – – – – –

–

–

High

Opposition leader Anders Fogh Rasmussen

0.55 323

n.s. 0.52a (0.04) 0.52 430

n.s. 0.54a (0.04)

⫺0.10b (0.05) 0.08c (0.00) –

–

0.56 324

n.s. 0.54a (0.05)

0.13b (0.01) –

–

–

0.43 430

0.08c (0.05) ⫺0.09c (0.01) 0.13b (0.01) n.s. 0.51a (0.04)

–

0.44 321

n.s. 0.59a (0.05)

–

–

–

–

0.39 428

n.s. 0.56a (0.04)

–

–

⫺0.08c (0.04) –

0.55 322

⫺0.08a (0.01) n.s. 0.67a (0.04)

–

–

Notes a p ⬍ 0.001. b p ⬍ 0.01. c p ⬍ 0.05. For presentational reasons only coefﬁcients with p ⬍ 0.05 are reported. None of the control variables (gender, age, education, and ideology) approached signiﬁcance and are therefore omitted from the table for presentational reasons. Entries in brackets are standard errors.

R2 N

0.54 430

n.s. 0.48a (0.04)

Control variables Evaluation actor (time 1)

–

–

0.06c (0.01) –

Exposure Yes press

Exposure No press

–

–

Exposure TV2

–

–

Campaign variables Exposure TV1

150

Campaign effects

analyzing media content and formally modeling the exposure to different media outlets to advance our understanding of a referendum campaign. Previous studies of voting behavior in national referendums suggest that a popular government can win the vote while an unpopular one is more likely to lose: “Referenda conducted in the context of national party politics, with the government of the day urging ratiﬁcation of a treaty they have themselves negotiated, will inevitably be contaminated by popular feelings about the government” (Franklin et al. 1995: 102). Our study suggests that a (moderately) popular government can easily take a fall during a referendum campaign. This bad news for the incumbents may go beyond looking unfavorable in the polls because the overall performance rating may even affect the outcome of the referendum. Indeed, as suggested by studies of previous referendums (Franklin et al. 1995; Franklin 2002), the (un)popularity of the government (and the decrease in performance rating during the campaign) was a signiﬁcant predictor of voting No also in the Danish 2000 euro referendum (see Chapter 9). In a broader perspective, this study highlights the importance of national referendums for domestic politics. As referendums become increasingly popular in use (see Butler and Ranney 1994), in particular in the case of European integration (Hug and Sciarini 2000), understanding public opinion dynamics becomes center stage. This chapter demonstrated that incumbent governments face signiﬁcant challenges when campaigning in a referendum. Lessons can be learned from the Danish 2000 example. The Yes camp consisted, as is often the case, of centrist, center-left and center-right political parties. While these parties share little common ground in most domestic political matters, they presented a uniﬁed Yes camp. This chapter suggests that the performance rating of these parties and their leaders must be considerably favorable at the outset of the campaign and that internal ﬁghting in the Yes camp must be avoided so as to ensure reasonable ratings on the handling of the referendum issue. If incumbents and their partners in the campaign are seen to mismanage the issue of the referendum, media priming effects may result in the incumbent having to pay a considerable domestic political price.

Voter stability We are interested in the domestic political ramiﬁcations of a referendum. We therefore turn to a measure of party preference from each of the waves in our panel data.5 Did the negative evaluations of the government and the political leaders result in changes in party preference during the course of the referendum campaign? Table 8.5 shows that the decrease in government performance rating and leader evaluation described in this chapter did not lead to major swings in vote intention. However, there was a clear trend in the electorate during the campaign. About 10 percent of the voters switched voting

100.0

0.0 0.0 0.0

0.9 0.0 0.4

100.0

2.5 90.0 0.0 2.5 2.5 0.0 2.5 0.0

92.9 0.4 0.9 0.0 2.2 1.3 0.9 0.0

100.0

0.0 0.0 0.0

2.2 2.2 88.9 0.0 0.0 1.1 4.4 1.1

100.0

4.2 0.0 0.0

0.0 0.0 0.0 91.7 0.0 0.0 4.2 0.0

100.0

0.0 0.0 0.0

4.8 1.2 0.0 0.0 92.9 1.2 0.0 0.0

100.0

0.9 0.0 0.0

2.6 0.0 0.9 0.0 1.7 86.2 6.9 0.0

100.0

1.1 0.0 0.0

0.7 0.4 1.5 0.0 1.1 1.5 93.5 0.4

100.0

0.0 0.0 0.0

0.0 0.0 0.0 0.0 0.0 11.1 11.1 77.8

Social Radical Conservatives Center Socialist Danish Liberal Progress Democratic Liberal Democrats People’s People’s Party Party Party Party Party

Vote intention wave 1

Note Pre- and post-referendum waves, n ⫽ 904.

Total

Social Democratic Party Radical Liberal Conservatives Center Democrats Socialist People’s Party Danish People’s Party Liberal Party Progress Party Christian People’s Party Unity List Others

Vote intention wave 2

Table 8.5 Change in voting intention in general election

100.0

88.2 0.0 5.9

0.0 0.0 0.0 0.0 0.0 5.9 0.0 0.0

100.0

0.0 86.4 0.0

0.0 0.0 0.0 0.0 4.5 0.0 4.5 4.5

Christian Unity People’s List Party

152

Campaign effects

preference during the referendum campaign. In electoral systems where coalition governments are the norm, such shifts in the vote are often sufﬁcient to have a different coalition with a different ideological leaning. In fact, in 2001, just one year after the referendum, the then incumbent Social Democrat-led government lost the general election. A center-right minority coalition government led by the Liberal Party and the Conservatives replaced the Social Democratic government. The ﬁndings of this chapter suggest that a referendum, in this case on a key issue pertaining to future European integration, can and did have signiﬁcant domestic political implications.

9

Campaign effects of the vote

Referendums are often characterized by close races, an undecided or volatile electorate, and greater importance attached to campaign events because the dynamics of traditional party politics are potentially less important in explaining turnout and voting behavior (see Chapter 1). The information environment in a referendum campaign is thus potentially important for the vote because party attachments often mean less in these campaigns. The arguments, aspects, or framing of the multi-faceted issue by political actors and by key information sources during a referendum campaign, therefore, may be unusually important to a referendum outcome, especially in a close race (see Chapters 5, 6, and 8). Most research on electoral behavior has been conducted in the context of presidential or parliamentary, national, and state-level legislative elections. Comparatively fewer studies of voting behavior have been carried out in the context of (national) referendum campaigns and citizen initiatives, even though this form of direct democracy is becoming increasingly common (Bowler and Donovan 1998; Bowler et al. 1998; Butler and Ranney 1994; LeDuc 2003). In this chapter we investigate factors inﬂuencing the vote in a national referendum.

The campaign and the vote: time of vote decision We ﬁrst take a look at the dynamics of the campaign. In our case study, 20 percent of the electorate reported being undecided on how to vote one month prior to referendum day. Given the expected close race, this made the campaign particularly important and decisive. Table 9.1 shows that 10 percent of the electorate decided during the ﬁnal weeks of the campaign and another 10 percent decided in the ﬁnal days or even on the day of the referendum. The results also show that the undecided voters were evenly distributed among the Yes and No camps in their ﬁnal decisions.

154

Campaign effects

Table 9.1 Time of vote decision Total More than a month prior Within ﬁnal two weeks Within ﬁnal days On referendum day In the booth Total

Voted Yes

Voted No

81.1 8.9 6.0 3.1 1.0

80.1 9.3 6.9 2.2 0.8

81.2 8.5 4.9 4.0 1.2

100.1

99.3

99.8

Note Cell entries are percentage of respondents, n ⫽ 940.

Inﬂuences on the vote in a referendum Theories of voting behavior generally emphasize the importance of party support or attachment (Campbell et al. 1960; Budge and Farlie 1983), as well as contextual factors such as economic perceptions (Lewis-Beck 1997), perceptions of issues (Nie et al. 1979) and evaluations of leaders or the top candidates (Wattenburg 1992). Some of the hypotheses about what drives party or candidate choice are also common in the literature on referendum voting (LeDuc 2002, 2003). We identiﬁed a number of hypotheses about inﬂuences on referendum voting in a review of the literature on referendums and citizen initiatives. Among the variables that explain support for a referendum measure are: ideology or party identiﬁcation, evaluations of or feelings about political elites, government performance or government approval, perceptions of the economy, issue-related contextual evaluations, political cynicism, political efﬁcacy, political knowledge and political interest, and demographic and regional characteristics. In the following sections we review the different explanations for voting behavior in referendums and explore the bivariate relationship between the different variables and voting behavior. We then discuss the importance of the campaign, which we integrate into an encompassing multivariate model of voting behavior.

Partisanship and ideology Studies of national elections repeatedly identify partisanship as a key factor for understanding voting behavior, although its measurement and relevance has been questioned in Europe (Budge and Farlie 1983; van der Eijk and Franklin 1996). While partisanship is still important in referendum campaigns, as voters also turn to their political party and leaders for cues on the referendum issue, partisanship is only an important predictor of vote choice if the political parties take a clear stand on the issue in question (Franklin 2002; Pierce et al. 1983; Sinnott 2002). On the issue of Euro-

Campaign effects of the vote 155 pean integration, however, there are numerous examples of disagreements between political elites within the same party. In a referendum campaign on a European issue, therefore, ideology and political leaning rather than political party afﬁliation can be a more useful predictor of the vote for two reasons. The ﬁrst reason is because a number of parties may be divided on the referendum issue and thus do therefore not serve as a natural reference point (Pierce et al. 1983). A second reason is because a number of European party systems, including Denmark’s, evolves in such a way that both the extreme left and right are opposed to advanced integration (although for very different reasons) and moderate left, centre, and moderate right are in favor of integration (see Andersen (1998) and Jakobsen et al. (2001) for a discussion of the two most recent national referendums on European issues). Looking at Table 9.2 we see that the No share is largest in the left and right groups and smallest in the center group. Plotting left–right against anti- and pro-positions on furthering European integration would thus result in the party system resembling an upside down U shape.1

Government approval, elite evaluations and economic evaluations Most of the research on referendum voting also includes some combination of variables in the triangle involving: 1 2 3

evaluations of political elites, evaluations of government performance, and economic assessments.

Referendum studies have modeled this triangle of variables differently when predicting the vote, usually excluding one on methodological or theoretical grounds, or combining two of the three in some fashion.2 Another approach has been to consider the indirect effects of one or another of these variables on a dependent variable that can be important for the vote. Research has explored the inﬂuence of national and personal Table 9.2 Vote by left–right political ideology

Right (n ⫽ 272) Center (n ⫽ 401) Left (n ⫽ 236)

Yes

No

54.0 57.4 36.9

46.0 42.6 63.1

Notes Row-wise percentage of respondents, n ⫽ 916. Ideology is measured by 10-point left–right self-placement scale. Left is 1–3, middle 4–6, and right 7–10. Post-stratiﬁcation weights are added.

156

Campaign effects

economic assessments, social market position, and campaign exposure on assessments of the economic consequences of EU membership, for example, in the referendums in Norway, Sweden, and Finland on the question of EU membership (Jenssen 1998).3 But taken as a whole, referendum voting studies have not addressed sufﬁciently the issue of the triangulated relationship mentioned above. Although there is a distinction between government support and assessments of government performance, and these indicators are often used interchangeably with little consistency in terms of operationalization and question wording,4 all empirical evidence on the importance of government support and/or performance for referendum outcomes seems to point in the same direction. Franklin et al. (1994, 1995) put the government performance hypothesis in the form of a probabilistic tendency. The proposition is that: “While a popular government might expect to see its referendum proposals approved, an unpopular government will often see its proposals turned down” (Franklin et al. 1995: 106). In a later article, this thesis was modiﬁed to apply in particular to referendums on issues that are of low salience to the electorate (Franklin 2002). In the case of the Danish euro referendum, both strands of theory receive support at the bivariate level. As Table 9.3 shows, almost threequarters of the respondents who evaluated the government negatively voted No, while two-thirds of the respondents who were favorable in their assessment of the government voted Yes. Conforming to the expectations, respondents who assessed their personal as well as the national economic expectations positively were much more likely to vote Yes (Table 9.4). Table 9.3 Vote by government performance rating

Unfavorable government performance rating (n ⫽ 213) Neutral performance rating (n ⫽ 290) Favorable government performance rating (n ⫽ 398)

Yes

No

27.7 47.2 64.3

73.2 52.8 35.7

Table 9.4 Vote by economic expectations Yes

No

National economic expectations Optimistic (n ⫽ 122) Neutral (n ⫽ 673) Pessimistic (n ⫽ 134)

94.3 52.5 2.2

5.7 47.5 97.8

Personal economic expectations Optimistic (n ⫽ 443) Neutral (n ⫽ 426) Pessimistic (n ⫽ 58)

83.5 21.1 13.8

16.5 78.9 86.2

Campaign effects of the vote 157 Conversely, respondents who assessed their personal as well as the national economic expectations negatively were much more likely to vote No.

Issues and arguments The multi-faceted nature of the issue in referendum campaigns means that the arguments put forward by those for and against often include related issues or topics that feature in one’s reasoning on the referendum issue. Johnston et al. (1996) therefore distinguish between “general arguments” and “speciﬁc elements” in their analysis of the vote in the 1992 Canadian sovereignty referendum, and measure speciﬁc elements with attitude questions that emerge as having an inﬂuence on support for the referendum. Clarke et al. (2000) also include issue-related contextual variables on the referendum provisions and referendum process to analyze the vote in the 1992 Canadian referendum. In the context of European referendum campaigns, issue-related contextual variables such as predispositions on the question of European integration, sometimes referred to as one’s pre-existing level of EU skepticism or EU attitudes, also has been considered an important variable in predicting the vote (Siune and Svensson 1993; Siune et al. 1994; Svensson 2002). Siune et al. (1994: 113) conclude “there is an unambiguous pattern in Danish voters’ orientation towards EC integration” and ﬁnd that voters vote according to the consistency of their attitudes. In the case of the referendum on the euro, our seven-item index of attitudes towards European integration (see Table 9.5) also showed that proEuropean citizens were much more likely (95 percent) to vote Yes than anti-integration-oriented citizens that were more likely (81 percent) to vote No (Table 9.6). In the middle group who were neither particularly EU skeptical nor enthusiast, almost 70 percent voted Yes. Table 9.5 Attitude towards European integration: seven items on EU skepticism scale “EU is not a threat for small countries such as Denmark” “European integration is being pushed too fast” “I would be willing to make a sacriﬁce to help a less strong EU country” “Denmark should maintain the exceptions from the Edinburgh Treaty” “EU has more disadvantages than advantages for people like me” “EU should be expanded with former Eastern Bloc countries, e.g. Poland and Baltic states” “EU should be expanded with southern European countries such as Cyprus and Turkey”

158

Campaign effects

Table 9.6 Vote by EU skepticism

Low EU skepticism (n ⫽ 131) Medium EU skepticism (n ⫽ 345) High EU skepticism (n ⫽ 242)

Yes

No

94.7 68.1 18.2

5.3 31.9 81.2

Note Base is respondents who answered all questions, n ⫽ 718.

Political disaffection Another set of predispositions that may inﬂuence support for a referendum concern political disaffection. Based on a cross-national comparative analysis of the referendums in Finland, Norway, and Sweden on entering the EU, for example, Jenssen et al. (1998) suggest that voters displaying a higher level of trust in the political system and institutions are more likely to support a government-initiated proposal in a referendum. Political efﬁcacy emerged in a multivariate model as the most important explanation of support for the California tax revolt in the late 1970s, and led Lowery and Sigelman (1981: 969) to conclude that “the feeling that one is cut off from the political decision-making processes” was the best predictor of support for tax limitation. Support for property tax limitation was greatest among those who were most cynical about political leadership. That said, measures of political disaffection have either not been included or have not featured prominently in the majority of studies on national referendum voting (which excludes the United States). Despite this, we might expect that political efﬁcacy and political trust are potentially important to national referendum outcomes. Those who are politically disillusioned would probably be less inclined to support proposals in referendums, as well as less inclined to vote in referendums.

Political interest Previous studies have also suggested that those more interested in the topic of the referendum are more likely to support the proposal (Siune et al. 1994). Politically interested orient themselves and know more than those who are less politically interested (Neijens et al. 1998). This in turn suggests that interest is an antecedent of knowledge (Johnston et al. 1996), which is a condition for making informed decisions (Price and Neijens 1997) and for supporting candidates’ proposals (Bartels 1986). Interest and knowledge, although acknowledged by most of the referendum studies as leading one to be more likely to vote, and more likely to pay attention to and learn over the course of the campaign, nevertheless often have not been integrated into the models of referendum voting, with rare exceptions (Jenssen et al. 1998; Johnston et al. 1996).

Campaign effects of the vote 159

Socio-demographic inﬂuences Another group of explanatory variables are socio-demographics and regional differences. In Canadian referendum studies, region and regional ethnicity play an important role in the referendum vote (Clarke et al. 2000; Johnston et al. 1996; LeDuc and Pammett 1995). In smaller European societies that are often more homogeneous, region within a country is of less importance. In previous referendums on European issues in Denmark, for example, a pattern has emerged suggesting that in particular women, persons born between 1955 and 1970, and those with less education vote No (Andersen 1998). This inﬂuence of social structure on the vote was partly replicated in three national referendums in Norway, Finland, and Sweden in 1994 on joining the EU. There it was found that men were more likely to vote Yes than women, that the Yes vote increased with education almost in a linear fashion, and that age in general was positively related to voting Yes, though not consistently (Jenssen et al. 1998: 177–190). In the Danish case in 2000, there were no discernible differences in vote choice between men and women or age (Table 9.7). In terms of education, only highly educated citizens with university degrees had a lower propensity to vote No. This pattern dovetails with longitudinal survey data suggesting that the explanatory power of socialdemographics for understanding vote choice in referendums is decreasing, which is in line with evidence from national election studies (Franklin et al. 1992).5 Table 9.7 Percentage No voters by gender, age, and education No Women (n ⫽ 496) Men (n ⫽ 461)

52.8 46.1

18–25 (n ⫽ 122) 26–39 (n ⫽ 217) 40–54 (n ⫽ 268) 55–69 (n ⫽ 189) 70⫹(n ⫽ 139)

51.6 50.2 49.3 48.1 49.6

Education (⬍ 16 years in school) (n ⫽ 487) Education (16–19 years in school) (n ⫽ 351) Education (19 ⬍ years in school) (n ⫽ 97)

52.0 51.6 30.9

Note Only respondents who voted in the referendum are included here. Post-stratiﬁcation weights on age, gender, and education are added.

160

Campaign effects

Bringing it all together: the role of the campaign One commonality across the referendum voting studies in Europe and Canada is that they say little about the ways in which the contents and uses of campaign information served to crystallize opinion about the pros and cons of the key issue in the referendum, based on any analysis of media contents or citizens’ uses of information sources during the campaign. Even studies that have asked respondents about the importance of various sources of information about the campaign (see Borg and Esaiasson 1998; Pesonen 1998; Jenssen 1998), have not been integrated into a model of referendum voting. While some evidence is available on the importance of the media during a referendum campaign, none of the studies has distinguished the differential effects that may emerge from exposure to speciﬁc contents in different media outlets. Our study was speciﬁcally designed to test campaign information effects in comparison with the aforementioned other hypothesized inﬂuences on the vote. The data allow us to test at the aggregate and individual levels and to assess the relative contribution of campaign information effects in understanding the vote, in the context of these other possible inﬂuences. In doing so, we apply the distinction between effects of direct and personal campaign experiences and effects of indirect and mediated campaign experiences. Direct and personal campaign experiences include the extent to which voters engage in discussions with family, friends, or colleagues about politics and the issue of the referendum. Mediated campaign experiences refer primarily to use of mass media (see Chapter 1).

Key questions Given the mixed or general lack of empirical evidence for the effects of a referendum campaign on voting behavior we explore in general terms the effects the campaign had on vote choice. In the process of addressing this general research question, we have a number of expectations for the 2000 euro referendum. Based on the review of the literature, we expect the following to have a positive effect on voting No: 1 2 3 4 5

ideological preference for “left,” government disapproval at the time of the referendum, pessimistic economic expectations, general skepticism towards European integration, and political disaffection in the form of political cynicism and political disinterest.

The ﬁrst and second expectations stem from the fact that the Social Democratic Party in government was in fact split over the question of how

Campaign effects of the vote 161 to vote in the referendum even though its leaders and the incumbent Prime Minister were outspoken about their support for voting Yes. We saw in the previous chapter how the overall assessment of the government decreased during the campaign, especially due to the perceived mishandling of the referendum issue. Opinion polls as well as our own panel data showed that about 20 percent of the electorate was either undecided or did not hold a ﬁrm position on the issue one month prior to the referendum (Reuters, August 28, 2000). This suggests that the campaign was the crucial period in which many voters made up their minds. Voters with a ﬁrm opinion on an issue have been found to be more likely to expose themselves to campaign information (Katz 1971; Kraus and Davis 1978), but the effects of exposure to (information about) the campaign are potentially larger for voters with a less ﬁrm opinion. Zaller (1992) argues that information effects are typically most prominent among “the middle-awareness group,” but also suggests that in the case of a saturated information environment, such as the Danish case, those in the low-awareness group may be inﬂuenced.

Vote intention Table 9.8 sets out the distribution of change in vote intention over the ﬁnal weeks of the campaign. This shows that although most of those who were decided on how to vote at the ﬁrst wave remained committed to their views over the ﬁnal weeks, the not insubstantial portion of undecideds at the ﬁrst wave ultimately split fairly evenly between the Yes and No camps. The referendum campaign was aimed primarily at these undecideds, though the parties also sought to raise doubts in the minds of decided voters. Only a small portion of decided voters actually switched their vote intention from Yes to No or visa versa over the last few weeks of the campaign. In this referendum, however, in which there was a six percentage point spread between the Yes and No camps in the ﬁnal vote, the undecideds and the small group of vote switchers could make the difference. Table 9.8 Vote intention (wave 1) and actual voting behavior (wave 2) Vote intention (wave 1)

No (n ⫽ 333) Yes (n ⫽ 380) Undecided (n ⫽ 249)

Actual voting behavior (wave 2) No (n ⫽ 447)

Yes (n ⫽ 494)

Did not vote (n ⫽ 21)

96.4 3.4 45.4

3.0 95.3 49.0

0.6 1.3 5.6

Notes Cell entries are percentages. Data weighted with post-stratiﬁcation weights on gender, age, and education.

162

Campaign effects

News and voting behavior Turning to the news and information environment during the campaign, our content analysis of the news media coverage of the two most widely watched main evening television news programs showed that more than 25 percent of the news was devoted to the referendum in the ﬁnal month of the campaign (see Chapter 4). The news environment was saturated with information about the referendum. Looking at the general tone of the coverage we found that the coverage of the actors in the news about the referendum was generally neutral or with a negative slant (see Chapter 8). However, there were differences between the news outlets and in the evaluation of Yes and No actors (see Table 8.3). If we look at the different news outlets we identiﬁed the average evaluation for “Yes” actors and for “No” actors, and then estimated the difference between the two by subtracting the number of negative evaluations from the number of positive evaluations, divided by the total number of evaluations. We use this difference measure of tone for each TV news outlet in our model. Speciﬁcally we found that the public broadcasting news reported, on average, more negatively about Yes actors (m ⫽ ⫺0.11) than No actors (m ⫽ ⫺0.08) in our measure ranging from ⫹1 to ⫺1. We also found that there were more Yes actors on TV1 than No actors. On TV2, however, there were more No actors than Yes actors. On TV2, the Yes actors were evaluated twice as negatively (m ⫽ ⫺0.16) as No actors (m ⫽ ⫺0.08).

Predicting the vote Below we present the full change model predicting voting No, controlling for the expressed intention to vote No at the beginning of the campaign. Our model of the vote choice is designed to test the effects of the hypothesized predictors of voting behavior that have been identiﬁed in previous referendum studies, as well as the effects of the campaign. We included a seven-item index of EU skepticism (see items above, Table 9.5) based on respondents’ perception of the extent of future EU integration. This is a more elaborate measure than has been used in previous research. Two scaled indicators of personal and national economic expectations, which are standard questions in most election studies, were included. A scaled measure of government approval (recoded to express disapproval) was used. We also included a scaled measure of ideology (in terms of left–right placement)6 and political cynicism about the campaign was measured by a six-item index. As control variables, we included a series of social and demographic characteristics: gender (female), age (in years), and education (number of years in school). Two measures of political interest, one on politics in general and another in the issue of the euro in particular, are also

Campaign effects of the vote 163 included in the model. As a measure of political efﬁcacy, we include a six-item index tapping both internal and external political efﬁcacy. We use a lagged term for the intention to vote No at wave 1, and for the status of undecided at wave 1 (see Markus (1979) for discussion of the use of lagged speciﬁcations in panel data). This enables us to control for initial vote intention and to examine change over the course of the campaign. We additionally included a number of measures of personal engagement in the campaign and exposure to mediated coverage of the campaign. Direct personal engagement in the campaign was measured by a three-item index tapping into activities and involvement in the campaign as well as the frequency of political discussion. Exposure to the campaign is measured by four variables of weighted news media exposure. These variables were constructed by bringing together data on exposure to speciﬁc news media outlets with a content analysis of the volume and tone of news about the referendum in each news outlet. Each variable indicates frequent exposure to a speciﬁc media outlet: public television news, commercial network news, Yes press, and No press. We add a weight for the exposure to these outlets, incorporating the amount and tone of the news about the referendum.7 The model conﬁrms the hypothesized effects of EU skepticism, government disapproval, economic expectations, ideology, and political cynicism as important inﬂuences on the No vote. They all had a positive and signiﬁcant effect on voting No, although political cynicism did not attain signiﬁcance. Optimistic personal economic expectations were signiﬁcantly negatively associated with voting No, in other words they contributed to voting Yes. General political interest had a negative effect on the likelihood of the No vote, though not signiﬁcant. These ﬁndings conﬁrm the direction of the expected effects in our ﬁve hypotheses. Our model suggests the importance of mediated exposure to the campaign through the news media in the ﬁnal weeks of the campaign. The results indicate that exposure to public television news and the Yes press was signiﬁcantly and negatively related to voting No, in other words, it contributed to crystallizing opinion in the form of a Yes vote. Exposure to private television and the No press was positively associated with voting No, but did not attain signiﬁcance. Political discussion or direct involvement in the campaign had no signiﬁcant effect on the vote. In addition to our change model of the vote in Figure 9.1, we also ran two separate logistic regression models predicting a No vote for undecided and decided voters. In these models we excluded the lag term for vote intention. Both models (for undecided and decided voters) produced similar effects and a consistent pattern in the direction of the effects (in terms of the sign of the beta-coefﬁcients). In both the decided and undecided models, EU skepticism, government disapproval, pessimistic economic expectations, left political leaning, and exposure to private

164

Campaign effects Being skeptical about the EU

⫹1.06

Disapproving of the government

⫹0.36

Pessimistic personal expectations

⫺0.87

Left ideological preference Frequent exposure to news on TV1

⫹0.84 ⫺1.70 ⫺0.49

Frequent exposure to the Yes press Intention to vote No (time 1)

Voting No in the referendum

⫹3.78 ⫹1.02

Not decided about vote choice (time 1)

Figure 9.1 Inﬂuences on the likelihood of voting No. Note Logistic regression, entries are standardized beta-coefﬁcients (n ⫽ 765, R2 ⫽ 0.80). Positive beta-coefﬁcient indicates positive association with voting No. Only signiﬁcant indicators are shown in the ﬁgure. Controls were made for gender, age education, political interest, political efﬁcacy, political cynicism, and other campaign variables. The full regression analysis is provided in the Appendix.

television were positively and signiﬁcantly associated with voting No. Political leaning right, exposure to public television news, and exposure to the Yes press was positively and signiﬁcantly associated with voting Yes. By introducing the lag term (Markus 1979) in Figure 9.1 to investigate change during the campaign, the models are replicated in terms of direction of effects and the explained variance is naturally increased.8 The analyses suggest that the model of the vote is both strong and stable. The model is very inclusive and the correct classiﬁcation percentage of 91.5 points to the full speciﬁcation. The model accounts for 80 percent of the explained variance, which is high for a model of voting behavior testing competing variables (Lowery and Sigelman 1981).

Predicting the referendum vote Our ﬁndings point to the conclusion that information in the ﬁnal weeks of the campaign mattered to how one voted in the referendum. This was a very close race, with a not insubstantial portion of the electorate still undecided in the ﬁnal weeks. Other studies show that in referendums 20–40 percent of the electorate often report being undecided (LeDuc 2003) and the campaign is therefore important for these voters to make up their minds. In our study, mediated sources of information in the ﬁnal weeks of the campaign exerted a signiﬁcant inﬂuence on crystallizing individual opinion on the vote, even after controlling for all other possible inﬂuences. That said, there was no uniform impact of information. Our study found that information effects were dependent upon the characteristics of the

Campaign effects of the vote 165 news coverage in the various information outlets, and this varied between public and private television news and the Yes and No press. We found that the electorate’s attitudes on the issue of European integration, economic expectations, ideology, and the popularity of the incumbent government were all of importance to the vote. The issuerelated contextual variable in this study – EU skepticism – was in fact the most important predictor of the vote, and this is in line with previous research that emphasizes the importance of issue-related contextual inﬂuences on the vote (Clarke et al. 2000; Johnston et al. 1996; Siune et al. 1994; Svensson 2002). Previous research also emphasizes the importance of ideology for the vote. We found that ideology closely followed EU skepticism in importance. Those on the left were more likely to vote No, and those on the right were more likely to support the referendum. The effect of left ideology was stronger than the effect of right ideology, which we attribute to the stronger consensus among left-wing elites on this issue. This dovetails with Zaller’s (1992) explanation of the consequences for public opinion of elite consensus on an issue, in this case though the consensus was somewhat stronger within one camp (see Chapter 1). The right-wing elites were even more divided, with the moderate right fully supporting a Yes vote and the extreme right supporting a No vote. These ﬁndings are in line with research on the 1999 European elections that shows moderate right parties are more supportive of European integration compared to extreme right parties (Hooghe et al. 2002). Economic expectations were the third most important predictor of the vote. Evaluations of the economy have also been found to play an important role in referendum voting, although indirectly by driving support for the government (Clarke et al. 2000). We found a direct effect of personal economic expectations, with optimistic expectations enhancing the likelihood of voting Yes. Although national economic expectations were in the same direction, they did not attain signiﬁcance. Disapproval of the government was the next most important predictor of the vote. Though the evaluation of the government was an important predictor, it was less important as an explanation. The effect size of the government disapproval rating was about half that of the effect size of the EU skepticism variable, for example. The government approval in the Danish case was primarily a function of feelings about the Prime Minister (see Chapter 8). Last but by no means least were the campaign variables, speciﬁcally exposure to certain news outlets. These measures showed that news mattered in this campaign. Exposure to public television news and exposure to the Yes press were signiﬁcant Yes vote predictors, even when controlling for the aforementioned variables along with socio-demographics and vote intention at the ﬁrst wave. In close races, these apparently small campaign effects may be responsible for shaping razor-thin majorities, as Zaller (2002) has argued. In this referendum, exposure to certain news outlets

166

Campaign effects

contributed to voting Yes. Our study moreover revealed a difference between the public service channel (TV1) and the private channel (TV2), with exposure to public broadcast news contributing signiﬁcantly to a Yes vote and exposure to private broadcast news having a different sign and therefore potentially contributing to a No vote. Why might this be the case? Our content analysis shows that the private channel’s news program contained more negative evaluations of Yes actors than the news on the public service channel. Moreover, the private channel’s news gave more room to No actors than was the case on the public service channel. Both these observations may fuel No sentiments. Our study suggests that it is not only important to pay attention to campaign information sources in modeling the vote, but also to the actual content and tone of the coverage. Some studies of national election campaigns have addressed this (see, for example, Kleinnijenhuis and Fan 1999; Kleinnijenhuis and De Ridder 1998). Other referendum studies to date, however, have either been descriptive in a discussion of news events and contents, or acknowledged the importance of the information outlets in referendum campaigns but ignored these both in terms of news content and in modeling the vote. Yet in close races, these types of information effects can make the difference. In light of the close race and small margin (53–47 percent) by which the introduction of the euro was rejected in Denmark, what could have been done differently to change the outcome? If the Danish Yes campaign, on the one hand, had been more successful in setting an economic agenda for the referendum, and the No campaign had been less successful in broadening the thematic agenda of the referendum campaign to also include issues of sovereignty, identity, and the evolution of the EU, then the referendum outcome might have looked rather different. However, the No camp was successful in broadening the scope of the issue (see Chapter 5), which mobilized anti-European integration sentiments in the political arena, and thus contributed to the No vote. Our post-referendum survey wave substantiated this perception of the campaign (see below). Public responses to the Prime Minister may also have been crucial to the outcome of this referendum campaign (see Chapter 8; Marcussen and Zoelner 2001; Qvortrup 2003). As we have seen, government support in this referendum is largely a function of feelings about the Prime Minister, and on more than one occasion during the campaign he evoked strong negative feelings among the public. Our panel data showed that his mean approval rating went down from slightly positive to slightly negative. In between, he is remembered for a major gaffe during a television interview in which he guaranteed the level of the Danish pension if the Yes side won, only to retract this statement days later, shortly before the vote.

Campaign effects of the vote 167

Motivations for voting Yes and No Why did the Danes reject the euro in the referendum in 2000? Citizens’ main motivations for voting No were the feeling that the development towards more Union must be stopped and that Denmark should keep its sovereignty in determining the country’s economic policy (see Table 9.10). Citizens suggested that the need for Denmark to be present where key decisions are taken, a strengthened European cooperation and the potential repercussions for the Danish economy of a No were the main reasons for voting Yes (Table 9.9). The tables below suggest that, with the beneﬁt of hindsight, the campaign of the Yes side should not have been reduced to a simple “it’s just the economy” message. This argument proved fundamentally weak in the attempt to convince EU skeptics about the value of joining the euro. A broader ideological discussion recognizing the intertwined nature of the euro issue as on the borderline between domestic economic politics and foreign policy would have been a better strategy for the Yes campaign. Table 9.9 Reasons for voting Yes Reason

Percent

Denmark has to be present where the key decisions are taken A No will harm the Danish economy A Yes will strengthen the European cooperation Private economy, e.g. that the interest rate will go up Our welfare state best served by a Yes Other Don’t know Total

58.3 12.1 18.2 1.8 5.3 3.0 1.2 100.0

Note Fixed-response categories, percentage of respondents, n ⫽ 494.

Table 9.10 Reasons for voting No Reason

Percent

Denmark must keep its sovereignty in determining its economic policy Denmark will have more economic policy freedom by saying No The development towards more Union must be stopped The attempt to make Denmark join in this election period, i.e. before next general election The possibility to make the incumbent government leave The welfare state will come under pressure if it is a Yes Other Don’t know Total

31.3 6.3 47.9 2.0

Note Fixed-response categories, percentage of respondents, n ⫽ 447.

1.3 3.1 5.8 0.4 98.1

168

Campaign effects

This chapter pulled together the existing literature on voting behavior in referendums. The chapter juxtaposed hypotheses on behavior and tested these in an integrated model also investigating the effects of the campaign. The results, based on an extensive content analysis of the referendum campaign and the two-wave panel survey, suggested that the vote, at the end of the day, was the end-sum of a number of individual predispositions as well as the campaign. In the ﬁnal chapter, the broader implications and lessons to be learned for understanding the dynamics of a referendum campaign are discussed.

Part V

The consequences of referendums

10 Conclusions

The referendum is an increasingly popular instrument in political decision making. This book focuses on the dynamics of a referendum campaign, in particular the role played by mediated communication in this process. This perspective is virtually absent in previous studies of referendums, which at times have included only descriptive impressions of the media coverage. It was therefore a natural starting point for us. As we argued at the outset of the book, behind the apparent simplistic nature of a referendum vote – Yes or No – a referendum issue is often multi-faceted. The arguments and framing of the multi-faceted issue by political actors and the media are key sources during a referendum campaign. Given the potential low public involvement, lack of knowledge, and ambiguity of elite cues, campaign communication may therefore be unusually important to a referendum outcome. In this chapter we discuss the broader implications of our study for campaign and referendum research in light of the theories about information in campaigns and elections. The referendum on the euro had several unique features, but also a range of aspects that apply to other referendums, not only on issues of European integration. We review the results with respect to campaign effects on political institutions, the media system, and citizens. We discuss the speciﬁc implications of our study for referendums and the use of national referendums in the process of advanced European integration. We focus on the combination of factors involving news values and media systems, the visibility of anti-EU parties and movements or prominent individuals, and other key points of variation in national contexts that can inﬂuence EU referendum outcomes. We conclude with a discussion of the strategic issues that emerge for the European Union and national governments when citizens say No.

How a referendum campaign affects political institutions A referendum challenges political parties and movements. On issues of European integration, a referendum (including the Danish referendum on the euro) affects the traditional, domestically anchored party system so

172

The consequences of referendums

that political parties at the (extreme) left and right ends of the ideological spectrum may join forces in a No camp, leaving the centrist political parties to campaign for a Yes. The reservations of the ideologically strongly proﬁled parties towards European integration relate to issues such as loss of national sovereignty, undermining of national welfare state facilities as well as nationalistic considerations about immigration and integration. With the political spectrum turned upside down ideologically, and add to that a number of Yes parties that are ofﬁcially in the Yes camp but in fact do not offer a ﬁrm recommendation on how to vote, it is easy to see how the parties in the political system eventually end up signaling ambiguity to the voter (Chapter 2). The political ambiguity in a referendum may be fueled by the inconsistency within political parties or camps in the referendum. In our case study, the Yes side was campaigning on a single message: the economy. However, when this argument was discredited by several expert reports signaling no discernible difference of being either inside or outside the single currency, the Yes campaign shattered and no alternative strategy was in place. By failing to stay “on message” a primary strategic goal in a referendum campaign – to reduce the ambiguity in cues – was not reached. And voters responded promptly by favoring the political alternative. Part of the success of the No camp was the uncertainty among the Yes parties about the No camp’s strategy. In national politics, parties “owe” certain issues, and although this may change over time there is predictability in European free-market liberals campaigning on tax cuts and European social democrats and socialists on welfare state values. Issue ownership in referendums, however, is by no means clear and the Yes campaign was beaten by its own arguments falling apart and the No camp’s successful broadening of the referendum issue. A referendum campaign can be viewed in the light of the growing body of literature that addresses the professionalization of politics and campaigning (Chapter 3). Judging from this literature, we should expect, ceteris paribus, a highly professional and streamlined mode of campaigning. Our study, however, revealed a mode of campaigning which in some ways was quite far from sophisticated and targeted campaign communication and much more of a traditional mode of campaigning. This suggests that when analyzing referendum campaigns and trying to grasp the dynamics, it is important to assess the efforts that were invested in the campaign from stakeholders in the political arena. Looking beyond the outcome of a referendum, it is evident that these contests can have considerable domestic political implications. Political parties position themselves strategically in relation to the long-term framing of the issues related to a referendum. A referendum, especially in the case of a No outcome, is not the end of the road for the debate about policy on this and related issues. The parties on the winning side of the

Conclusions 173 referendum also have a strategic opportunity to control the framing of the issue in the future. Moreover, political parties and candidates are therefore forced to consider the domestic political implications of a referendum. A referendum places a topic prominently on the political agenda and attitudes towards political leaders may be shaped more by the referendum issue than other issues during and in the aftermath of a referendum campaign. This implies that a successful campaign can bolster the popularity of political leaders while a campaign perceived to be less successful has the potential to damage the domestic political reputation of a party or politician. In our case study, the Social Democratic government that lost the 2000 referendum on the euro also lost the general election a year later to a Liberal–Conservative coalition government.

How a referendum campaign affects the media A referendum forces media organizations to prepare and plan how to report the campaign (Chapter 4). Some of this advance planning shares similarities with preparation for other electoral contests, but in a number of ways, media organizations have to respond to characteristics of the referendum contents. Prior to any electoral contest, news organizations typically decide on funding available for the coverage and the allocation of staff and resources. This is no different in a referendum campaign and these measures are basic indicators of the priority given to a campaign. In addition, news organizations typically prepare using background research that results in canned items for television and feature articles in the press. These preparations are made in anticipation of certain issues that are likely to appear in the campaign and when they do, background information for a story as well as potential interviewees and key facts are already on ﬁle to make it possible to report in a timely manner. Some news organizations, such as TV1 in our study, may utilize a number of additional tools in the planning, such as commissioning a survey to identify topics of interest to its viewers. While these elements have more similarities than differences with a national election campaign, there are distinct differences too. The notion of balance is altered in the context of a referendum campaign. While the “typical” interpretation of political balance in television news involves giving access to political parties across the ideological spectrum, often in some ratio of the size of the party (Semetko 2003), a referendum challenges these existing conceptions of balanced news reporting. As the vote in a referendum is neither party nor candidate based, the balance issue becomes a question of hearing the Yes and No camps. This can have the unintended effect of a magnifying glass being placed on smaller parties or individuals on either side of the issue. We found clear evidence of this magniﬁcation effect. The interviews with journalists and Editors also revealed that at more occasions news organizations were approached after

174

The consequences of referendums

a program by parties supporting a Yes or a No who both felt that they had been treated disadvantageously in the news. We found considerable attention to the referendum campaign in the news. The amount of news devoted to the referendum was more than the attention devoted to the European Parliamentary elections, but less than what is typically devoted to a national election campaign. Our analysis of how the news media framed the referendum showed that news frames such as conﬂict and strategy dominated. Our analysis also showed that the No side was more successful than the Yes side in framing the referendum issue in the news.

How a referendum campaign affects citizens The third section of the book was devoted to campaign effects on voters. Broadly speaking, research in political science and political communication takes two perspectives on the contribution of media to the electoral process and the public perception of the political system. One strand of literature contends that the media and political journalism contribute to political alienation, political inefﬁcacy, and a decline in participation in elections (for example, Ansolabehere and Iyengar 1995; Cappella and Jamieson 1997; Patterson 1993). Another strand of literature suggests that the media–public interface is fruitful and contributes to knowledge gains and political participation (for example, Holtz-Bacha 1990; Newton 1997; Norris 2000). Finally, others point out that effects of news coverage on political attitudes and engagement are contingent on various conditions such as political sophistication or expertise (for example, Moy and Pfau 2000; Pinkleton and Austin 2001), watching public broadcasting or commercial news (Aarts and Semetko 2003) or contextual factors (for example, Peter et al. 2003). Our study of the media’s ability to inﬂuence the public salience of issues (agenda-setting) was the ﬁrst to investigate agenda-setting in the context of a referendum (Chapter 6). We explored how a referendum can enter the public agenda (at the aggregate level) and how some citizens were more likely than others to consider the referendum an important topic (at the individual level). We expected a national referendum campaign to be distinctively different from a general election in that the broader theme of the campaign is deﬁned a priori. The agenda-setting dynamics was therefore more about sub-issues emerging within the broader issue. We found limited support of agenda-setting at the aggregate level. That is to say that the composition of the news agenda and the public agenda differed in terms of the ranking of issues. The referendum, however, increased in salience on the public agenda during the campaign and we found that exposure to news about the referendum contributed to an increase in the salience of the referendum issue. This supports the agenda-setting hypothesis at the individual level.

Conclusions 175 We also explored the meaning of the referendum issue and found evidence that the referendum did not mean the same to the public as it did to the news media. An issue may in other words be important to both the media and the public, but looking beyond this we may find different substantive aspects of that issue being more important to one group than the other. Our study emphasizes the importance of the battle over a referendum campaign agenda (Chapter 5). The No side in a referendum campaign may prevail by campaigning with an agenda that appeals to only a minority of voters. This is particularly important in the context of referendums on European integration. National referendums function as a policy and decision-making instrument and they can have signiﬁcant domestic political implications. In Denmark, the No side was effective in broadening the scope of the referendum issue to also include themes as national sovereignty and the welfare state. These topics found resonance with the public while the media devoted a lot of attention to the conduct of the campaign itself. In a referendum, winning the battle over the campaign agenda implies not only increasing the salience of the referendum, but also controlling the substantive content and the spin on the issue. Our study extended previous research on the effects of strategically framed news to a European setting and the context of a referendum (Chapter 7). We found a fairly high level of political cynicism about political candidates. During the ﬁnal month of the campaign the levels of political cynicism and negative campaign evaluations increased. The news media contributed to this increase in cynicism and negativity. Persons who were exposed the most to strategic news about the campaign, even when controlling for a number of other inﬂuences, displayed the strongest increases in cynicism and negative evaluations of the campaign. Despite the cynicism and negativity, we did not ﬁnd any detrimental inﬂuence on turnout or mobilization. The negative consequences of strategic news and political cynicism on political engagement and mobilization have been suggested in previous studies in the United States. While such negative effects may emerge in different political systems, we can conclude that in a context in which most citizens were highly aware of the issues at stake, strategic or negative news while increasing cynicism and negative campaign evaluations played no role in mobilizing electors to go to the polls. We interpret this as an indication that voters can be cynical and engaged. Voters may be dissatisﬁed, cynical, and negative, but still mobilized and sufﬁciently engaged to turn out to vote. An important yet unanswered question was whether a referendum campaign matters to the public’s evaluations of political leaders and the incumbent government (Chapter 8). Priming theory suggests that new information can affect the parameters by which we evaluate political leaders. Our study conﬁrmed the basic hypothesis of political priming theory by showing that the overall evaluation of political leaders was

176

The consequences of referendums

strongly dependent upon an issue that was highly salient, readily available, and on the top of citizens’ mind. Drawing on our content analysis of news media we saw that the topic of the referendum became more visible in the media during the campaign. Utilizing our panel data we found that as the importance of the referendum issue increased in the media, its importance for the public for formulating general evaluations of political leaders increased as well. The incumbent government was seen by the public to handle the referendum topic poorly. Subsequently, the government, including its most prominent ﬁgures and key campaigners the Prime Minster and the Finance Minister, was evaluated on the basis of their performance on the referendum issue. These ﬁndings dovetail with existing priming research and conﬁrms an assumption in previous studies, namely that the phenomenon apply to political leaders more generally (Semetko et al. 2003a, 2003b). In addition the ﬁndings suggest that not only contentious issues such as the Gulf War(s), but also issues of national referendums (such as on the introduction of the euro), can drive general evaluations of political leaders. Previous research has not been able to assess the effects of evaluations of issues and persons in the news media on public evaluations of political leaders. In our study we found that negative news about an issue that was also highly visible in the news bolstered negative evaluations of political leaders. The strongest impact of this negative news was found among politically less involved citizens. The evaluation of (incumbent) political leaders matters because voting behavior in national referendums can be a function of the popularity of a government. A popular government is more likely to win a referendum and an unpopular government is more likely to loose. We found that a (moderately) popular government can easily take a fall during a referendum campaign. Indeed the (un)popularity of the government (and the decrease in performance rating during the campaign) was a signiﬁcant predictor of voting No, also in our case study. Previous studies offer several and inconsistent explanations on the outcome of national referendums. We identiﬁed a number of hypotheses about inﬂuences on referendum voting and in addition we paid particular attention to the impact of the campaign on the vote given that a volatile electorate may draw more on and be more susceptible from cues in the news and information environment during a campaign. Our ﬁndings point to the conclusion that information in the ﬁnal weeks of the campaign mattered to how one voted in the referendum. This was a very close race, with a not insubstantial portion of the electorate still undecided in the ﬁnal weeks. Mediated sources of information in the ﬁnal weeks of the campaign exerted a signiﬁcant inﬂuence on crystallizing individual opinion on the vote, even after controlling for all other possible inﬂuences. That said, there was no uniform impact of information. Our study found that information effects were dependent upon the characteristics of the news

Conclusions 177 coverage in the various information outlets, and this varied between public and private television news and the Yes and No press. In addition to the campaign-driven effects, we found the following predictors to drive the vote choice in the referendum on the euro. First, EU skepticism was the strongest predictor in our case of voting No. The importance of issue-related contextual inﬂuences on the vote was signiﬁcant so that general sentiments about the EU and European integration drove the vote on the speciﬁc topic of the euro. Second, ideology was a signiﬁcant predictor and left ideology in particular emerged as a strong predictor of voting No. Third, we found that negative economic evaluations and expectations contributed to a No vote. Fourth and ﬁnally, public evaluations of the incumbent government were important so that voters disapproving of the government were more inclined to vote No. Our study was the ﬁrst that was able to test the different (competing) hypotheses about referendum voting simultaneously while also drawing on panel survey data.

The dynamics of a referendum campaign: the contingency of campaign effects We set out to investigate the dynamics of referendum campaigns arguing that the key characteristics of referendums – volatile electorates, uncertainty in elite cues, and issue complexity – mean that these campaigns can really matter. We distinguished between two levels of effects: the individual and the institutional. At the institutional level we analyzed how political parties and the media were affected by the referendum. At the individual level a referendum campaign may affect citizens’ awareness of an electoral contest by emerging on the public agenda. Our study demonstrates that the campaign may 1 2 3 4 5

lead citizens to evaluate the campaign positively or negatively, induce political cynicism, dependent upon the media content to which citizens are exposed and pay attention, affect citizens’ evaluations of domestic political leaders, serve to crystallize opinion on the topic of the referendum, and inﬂuence some voters on how to vote.

The unpredictability of a referendum should not be exaggerated though. Despite the potential low citizen involvement, the presence of ambiguous elite cues, and the importance of mediated (mass) communication, the campaign is only part of the story about political engagement, electoral mobilization, evaluations of key political leaders, and vote choice. Very importantly, we stress that campaign effects are conditional. We neither expected nor found large across-the-board effects of our campaign variables. As Zaller (1992, 2002) has argued, effects of the media in

178

The consequences of referendums

a campaign are not likely to be enormous, but in close electoral races they can be signiﬁcant. Even in races that appear a foregone conclusion, the campaign can serve to further widen the gap between electoral contesters. We add to these observations that campaign effects driven by the media are contingent upon the context and individual characteristics. In our study of agenda-setting during a referendum campaign, we found that different groups changed to consider the referendum the most important problem, but that this effect was particularly large for respondents who frequently viewed television news during the campaign. The study of political cynicism and evaluations of the referendum campaign demonstrated that both cynicism and negativity increased at the aggregate level over the course of the campaign, but this increase was strongest for heavy media users. These two examples show that campaign effects are conditional upon the medium and the amount of news consumed about a referendum. Our study also demonstrated how negative evaluations of the government handling of a referendum increased over the campaign. This in turn fueled negative overall government evaluations. The media that reported most negatively about political leaders had the largest impact on negative evaluations of key political leaders, in particular for persons with low political involvement. This suggests that the susceptibility to effects of the media in a referendum campaign is moderated by the level of political involvement so that those least involved were more susceptible. Finally, we found evidence that – in addition to other hypotheses about vote choice – the media also mattered. The campaign served to crystallize opinions about the referendum and the government which inﬂuenced the vote. In addition exposure to news that reported less negatively about the referendum increased the propensity to vote Yes. This ﬁnding emphasizes the importance of considering the content of media and even the differential ways of reporting an election when understanding campaign effects.

Referendums and the integration of Europe The referendum instrument is of great importance for the trajectory of European integration. More than 25 referendums have been held in the past two decades in several European countries on various issues of integration and this makes EU integration the most voted-on issue in the world. As we ﬁnish this book (December 2003), Europe is experiencing a true “referendum wave.” Nine of the ten new member states that joined the EU in 2004 ratiﬁed the membership in national referendums. And several countries are set to hold a national referendum on the emerging EU Constitution. The referendums in 2003 and those in the future tie in well with the evidence and case discussed in this book. Some of the characteristics of our case study generalize to other EU countries. We found, for example, that attitudes towards European integration was a key predictor of the

Conclusions 179 vote. Denmark is a country in the EU with crystallized opinions about European integration (for example, Svensson 2002). In other countries the level of crystallization is also likely to be of importance in a referendum. In Britain and Sweden, for example, membership of the EU is subject to public and political debate and the question of participation in the single currency is very much driven by diffuse and general sentiments towards “Europe.” Likewise, countries such as Spain and Italy are characterized by crystallized positive sentiments towards European integration. Attitudes towards the EU are therefore likely to be of importance in referendums in these countries. In countries that have less articulated and stable public opinion about the EU, the importance of these opinions in a referendum is less predictable. In the Netherlands, for example, public opinion on the aggregate level towards the EU is rather positive (Eurobarometer), but this opinion has been demonstrated to be highly volatile (Saris 1997). The way that general opinions about the EU will play in the Netherlands, for example, or countries that are former members of the Eastern Bloc in the case of a referendum on a single European integration issue remains an open question. Other aspects of our study travel more generally across most country contexts. The transformation of the political party system in the case of a referendum on a European issue applies to several countries. Moreover, the curvilinear relationship between support for European integration and ideology even ﬁnds increasing support beyond the referendum context. Hix and Lord (1997) report a two-dimensional space of political parties and mass publics (Hix 1999) with respect to ideology and support for the EU. Likewise, the dynamics of the interaction between the political elites, the (news media) and citizens of getting referendum on the agenda and “winning the battle over framing the referendum issue” are generally applicable. Looking back, a number of the 2003 referendums on entry to the EU were fought over arguments about economy and prosperity on the one hand and nationalism and cultural identity on the other. Looking ahead, the upcoming referendums on the future Constitution of the EU are already now covering arguments and disagreement over such diverse topics as immigration, the power balance in Europe, religion (Christianity versus Islam) and human rights. Our study is a ﬁrst attempt to identify and explain the dynamics of a referendum campaign and the effects of the campaign on public opinion and the vote. Future research needs to further complement, develop, and specify some of the ideas introduced here. The issue of European integration is a key issue for direct democracy and for the study of referendums. Future EU governance is contingent upon national referendums and to the extent that the results of the referendums are not forgone conclusions, it is important to understand the dynamics of referendum campaigns.

180

The consequences of referendums

While previous studies of referendum campaigns have paid little or no attention to studying in a systematic fashion the information available in the campaign, and modeling its effects on public opinion and the vote, we argue that no study of the dynamics of a referendum campaign on an issue of European integration is viable without considering the information environment.

Referendums and EU’s democratic deﬁcit Referendums not only play a central role in the development of the European Union. A general feature of referendums is of particular importance in the context of European integration. On the one hand, referendums are generally perceived to have the capacity to mobilize an electorate and to increase citizen engagement in politics. Some political theorists celebrate a “referendum democracy” for the high level of civic participation which is not achieved by general elections only (for example, Sartori 1987). On the other hand, the EU is notorious for its democratic deﬁcit and lack of citizen engagement. The democratic deﬁcit is embedded in the limited political power of the European Parliament – which is the only directly elected institution in the EU – and the lack of transparency in the decision-making procedures. Moreover, citizen engagement in European politics is not increasing despite the growing number of competences and power transferred to the EU. In fact, one of the key indicators of citizen engagement – turnout in the European Parliamentary elections – dropped as low as 30 percent in two countries on the Continent (Finland, a relatively new member, and the Netherlands, one of the original members), and as low as 26 percent in the UK in 1999. Referendums have the potential to alleviate some of this apathy. The turnout level in several referendums on issues of EU membership, key EU treaties, and EU policies has been encouraging. In all ﬁve Danish referendums on EU matters, turnout has been above 80 percent, except in the cases of the Single Act (1986) and the Amsterdam Treaty (1997) where turnout was 75 percent. In the 1994 referendums on EU membership in Austria, Finland, Norway, and Sweden turnout was consistently high, both in countries opting to join the EU and in countries that voted No. Key treaties such as Maastricht were subject to referendums in France (71 percent), Denmark (83 percent), and Ireland (57 percent). Referendums have the ability to bridge some of the gap between European level governance and European citizens.

When citizens say No On more than one occasion, electorates in different countries in Europe have voted No in a referendum, thereby opting not to ratify a key Treaty (e.g., Denmark, Ireland), not to introduce the common currency

Conclusions 181 (e.g., Denmark and Sweden) or not to join the EU (e.g., Norway and Switzerland). What happens to the political arena and the trajectory of European integration when citizens say No? In the case of the referendum on the euro, most citizens perceived the outcome of the referendum as a rather insigniﬁcant status quo. In our post-referendum survey we asked all respondents to assess the consequences of the No outcome Immediately after the referendum the Danish electorate displayed negligible anxiety about the consequences for the economy (see Table 10.1). There was some anxiety that Denmark would have less autonomy in maintaining its monetary policies and most respondents also believed that domestic business would not beneﬁt from the outcome. The most signiﬁcant perceived consequence, however, had to do with the future of EU integration. More than two-thirds believed that the No outcome would imply having another referendum on the single currency within the next three to ﬁve years. A similar share of the electorate also believed that the No outcome would result in an EU developing in two tempi, with those countries more in favor of integration moving along faster than countries with some reluctance and reservations about advanced integration. This is an interesting outcome in the larger picture of the European integration process. Thus far, key treaties, such as Maastricht, Amsterdam, and Nice, and also the upcoming Constitution, have been dependent upon Table 10.1 Perceived consequences of the No result

The Danish currency will be threatened Denmark will not be able to maintain its currency stability policy The Danish Krone will be de-valued The interest rate will increase A general election will be held Taxes will go up My own economic situation will remain the same or improve The economy of Danish businesses will improve The Danish economy will improve Unemployment will go up The euro will break down A new referendum will be held in the next 3–5 years The EU will develop in two tempi, those countries more in favor of integration will move forward The enlargement of the EU will be delayed Note Percentage of respondents, n ⫽ 956.

Agree

Disagree

Don’t know

22.3

55.3

22.4

14.1 6.7 50.6 42.8 39.6

62.0 69.3 26.9 32.3 32.2

23.9 24.0 22.5 24.9 28.2

36.7

31.3

32.0

29.8 17.1 23.6 12.2

43.7 45.8 46.4 52.4

26.6 37.1 29.9 35.4

69.9 64.3

13.3 14.5

16.8 21.2

44.7

24.2

31.0

182

The consequences of referendums

unanimous ratiﬁcation by the EU member states. Only if all countries ratiﬁed, the EU was able to implement the contents of the Treaty. The signal given by the electorate after the 2000 referendum dovetails with current developments in the restructuring of EU institutions and their decisionmaking procedures. A possible outcome is that a number of countries continue the integration process faster and tighter than other countries. Such a scenario might also address the problem of “impasse” in the situation when a country does not ratify a treaty. The EU institutions by then might be institutionally set up to accommodate a No outcome. The outcome of the referendum was also perceived as a postponement of the inclusion of Denmark in the single currency. We stressed in Chapter 1 that in the case of a No outcome, a referendum is also important to position winners and losers on the topic of the referendum as this is likely to re-emerge on the political agenda. In Denmark – as well as in Sweden – the question of participation in the single currency has not reached its ﬁnality by the No outcome of the referendum. Given the long-term implications of the issue, we polled a representative sample of the Danish electorate one year after the referendum, in 2001, right before the actual introduction of the currency. By that time, public support had shifted slightly (see Figure 10.1) so that 54 percent of the respondents reported to vote Yes (of which 38 percent ﬁrmly) and 42 percent reported to vote No (of which 32 percent ﬁrmly). We additionally speciﬁed a reduced model explaining support for the single currency one year after the referendum. As Table 10.2 shows, 4%

38% 32%

Yes Leaning Yes Leaning No No Undecided

10% 16%

Figure 10.1 Public support for the euro one year after the referendum. Note Opinion poll in August 2001, n ⫽ 1,292.

Conclusions 183 Table 10.2 Predicting support for the euro one year after the referendum (OLS) Standard beta-coefﬁcient Demographics Male (1) Age (in years) Education Predispositions Left political leaning Right political leaning Government approval n Adjusted R 2

Standard error

0.15 0.07 0.14a

0.07 0.00 0.04

⫺0.15b 0.05 0.16c 1,189 0.08

0.11 0.09 0.04

Source: Opinion poll in August 2001. Notes Only respondents answering all questions are included in the model. a p ⬍ 0.05. b p ⬍ 0.01. c p ⬍ 0.001.

education and government approval were positive predictors of supporting the euro while left ideological preference was a signiﬁcant negative predictor. The positive relationship between incumbent government support and support for issues of European integration found again a year after the referendum dovetails partly with Ray’s (2003) ﬁndings. He proposes that the positive relationship between incumbent support and proEU attitudes is conditional upon when referendums are held on issues of European integration. While he ﬁnds that the positive relationship is mostly found in years when referendums are held or during European Parliament election years and then primarily in terms of support for the current EU, we also found higher propensity for supporters of the incumbent government to favor further integration, also in the aftermath of the referendum.

Referendums and democracy The referendum instrument itself is subject to scrutiny. Its use as an instrument in political decision making – either by choice or necessity given constitutional requirements – is debated and balances between arguments in favor of citizen mobilization and political legitimacy on the one hand and public disinterest in and ignorance about politics on the other. Referendums are not here to replace but rather complement party democracy, but understanding the dynamics of a referendum campaign is essential not only to interpret the outcome but also to evaluate the process. Not before campaign communications is a well-understood part of the referendum dynamics can we move on to assess the democratic

184

The consequences of referendums

implications. A referendum can mobilize an electorate and increase civic participation. And the campaign and the news reporting of the events, parties, leaders, and aspects of the referendum issue that it brings to prominence, can crystallize opinions and even affect the outcome in a close race.

Appendices

Appendix Chapter 2 Introduction to Danish political parties. Party

Description

Social Democratic Party

Biggest party in Danish politics and Parliament. In power since 1993 under Prime Minster Poul Nyrup Rasmussen, between 1998 and 2001 in minority coalition with Radical Liberals. Ofﬁcally pro-EU and pro-EMU, but party is split.

Radical Liberals

Small centrist party in center-left coalition government (1998–2001). Deputy PM and Economy Minister Marianne Jelved is one of the staunchest backers of the euro.

Liberal Party

Major center-right opposition party (in government after 2001 elections with Conservatives). Most solid pro-EU party in Parliament. Party supports euro membership, but many party members are skeptical.

Conservative Party

Small, moderately right-wing party (in government since 2001). Supports Danish euro membership. Conservatives are wary of moves towards European integration, but take a pragmatic view of euro membership.

Socialist People’s Party

Far-left opposition party, opposed to Danish euro membership and partially to the EU. Leader Holger K. Nielsen was active anti-euro campaigner.

Danish People’s Party

Far-right, anti-immigration, anti-EU party led by Pia Kjaersgaard. Totally opposed to Danish membership of euro which is seen as a step towards European federalism threatening Danish identity and independence.

Center Democrats

Small, centrist opposition party formed in 1970s by disenchanted Social Democrats. Strong pro-EU and supports Danish membership of euro. continued

186

Appendices

Introduction to Danish political parties – continued Party

Description

Unity List

Ultra-left group of reds, greens, and former communists. Opposes euro membership and sees EU as a vehicle of big business interested in markets rather than citizens and democracy.

Christian People’s Party

Tiny, centrist Christian Democrat party which is divided on Danish membership of euro and advocated No, “wait-and-see.” Fears joining euro would reduce national freedom on monetary policies.

Freedom 2000

Small ultra-right group adamantly opposed to euro membership. Sees it as step towards loss of Denmark’s independence to pursue sovereign economic policies.

June Movement

Leading anti-EU grassroots movement represented in European Parliament but not national parliament. Campaign leaders include MEP Jens-Peter Bonde and Drude Dahlerup. Sees euro as impossible project, worries about single currency’s erosion of democracy.

People’s Movement against the EU

Main anti-euro grassroots movement with the June Movement. Sees single currency as worrying move towards European political uniﬁcation, harmonization of tax and labor market and social policies.

Note Table is based on party description by the Reuters News Agency, September 28, 2000.

Appendix Chapter 6 Age: In years. Gender: Female ⫽ 1; male ⫽ 0. Education: Four levels of education. Issue importance: “How important do you consider the Most Important Issue?” (in this case the euro). Five-point scale ranging from 1 “not at all important” to 5 “very important.” Need for orientation: Refers to a person’s information seeking about an issue as a function of a person’s interest in and uncertainty about an issue: “How interested are you in the issue of the euro?” Four-point scale ranging from 1 “not at all interested” to 4 “very interested.” Interpersonal communication: “How often in the past week did you discuss the referendum or the euro with friends, family or colleagues?” A fourpoint scale ranging from 1 “never” to 4 “often” discussing the euro issue.

Appendices 187 Voted No in referendum: No ⫽ 1; Otherwise ⫽ 0. EU skepticism: Seven-item index (Cronbach’s ␣ ⫽ 0.82) tapping attitudes towards European integration. Question wording: We would like to know how you feel about European integration. Please tick one box per statement. Strongly agree, Agree, Neither agree nor disagree, Disagree, Strongly disagree. 1 2 3 4 5 6 7

European integration is being pushed to fast. I would be willing to make a sacriﬁce to help a less strong country. Denmark should maintain its exceptions from the Treaty of Edinburgh. The EU has more disadvantages than advantages for people like me. The EU should be enlarged with former Eastern Bloc countries such as Lithuania and Poland. The EU should be enlarged with southern European countries such as Cyprus and Turkey. The EU and European integration is moving towards becoming a European “super state.”

Exposure to news about the referendum: Number of days watched/read news outlet (0–30/30) ⫻ volume of coverage (n of stories) about the referendum per news outlet. Euro MIP t1: Mentioned euro as Most Important Problem in wave 1 ⫽ 1; Otherwise ⫽ 0.

188

Appendices

Appendix to Figure 6.2 Effects of exposure to news coverage about the referendum on salience of the issue Beta-coefﬁcient Social demographics Age Gender (female) Education

0.08 0.03 ⫺0.08

(0.10) (0.19) (0.12)

Personal characteristics Issue importance Need for orientation Interpersonal communication Voted No in referendum EU skepticism

⫺0.04 0.24 0.09a ⫺0.53a ⫺0.36a

(0.12) (0.11) (0.13) (0.25) (0.17)

0.10a 0.01 0.06a 0.00 1.47b

(0.00) (0.01) (0.04) (0.00) (0.31)

Exposure to news about the referendum TV1 TV2 No press Yes press Euro Most Important Problem t1 ⫺2 log likelihood Nagelkerke Pseudo R2 Percentage cases correctly classiﬁed n

695.07 0.14 76.20 672

Notes Logistic regression analysis. Positive sign of beta-coefﬁcient indicates positive association with mentioning the euro/referendum as the Most Important Problem facing Denmark in the second panel survey wave. Standard errors are in parentheses. a p ⬍ 0.05. b p ⬍ 0.01.

Appendix Chapter 7 Variables and question wording independent variables. Gender: Female ⫽ 1; male ⫽ 0. Age: In years. Education: Four levels of education. Discuss euro: One (never discuss euro) to four (often discuss euro) scale. Left political leaning: Self-placement on left–right scale where 1 equals left and 10 right; recoded as 1–3 ⫽ left ⫽ 1; otherwise ⫽ 0. Right political leaning: Self-placement on left–right scale where 1 equals left and 10 right recoded as 7–10 ⫽ right ⫽ 1; otherwise ⫽ 0. Efﬁcacy: Six standard items. Items recoded to form an index of political efﬁcacy (Cronbach’s ␣ ⫽ 0.69).

Appendices 189 1 2 3 4 5 6

At times, politics can be so complex that people like me don’t understand what is going on. People like me don’t have any say in what the government does. I think that I am better informed about politics than others. MPs want to keep in touch with the people. Parties are only interested in people’s vote, not their opinions. There are so many similar parties that it does not matter who is in government.

Government disapproval rating: 1 (very bad) to 5 (very good) performance rating scale. Reverse coded. In Appendix to Figures 7.3 and 7.4, the change in government rating between panel wave 1 and wave 2 is used. EU skepticism index: Seven-item index tapping general attitudes about European integration, Cronbach’s ␣ ⫽ 0.82. 1 (strongly agree) to 5 (strongly disagree) scale. 1 2 3 4 5 6 7

European integration is being pushed too fast. I would be willing to make a sacriﬁce to help a less strong country. Denmark should maintain its exceptions from the Treaty of Edinburgh. The EU has more disadvantages than advantages for people like me. The EU should be enlarged with former Eastern Bloc countries such as Lithuania and Poland. The EU should be enlarged with southern European countries such as Cyprus and Turkey. The EU and European integration is moving towards becoming a European “super state.”

Intention to vote No: Intention to vote No in wave 1 ⫽ 1; otherwise ⫽ 0. The panel contained very few cross-overs which is why No vote intention is almost equal to actual vote Undecided in wave 1: Undecided what to vote wave 1 ⫽ 1; otherwise ⫽ 0. Campaign involvement: Three-item index tapping engagement in various campaign-related activities. The three item all correlate positively and form a scale of campaign involvement, Cronbach’s ␣ ⫽ 0.65. 1 (never) to 4 (often) scale. 1 2 3

Take part in a public meeting about the euro. Visit an Internet website about the euro. See an advertisement or commercial about the euro.

190

Appendices

Appendix to Figures 7.1 and 7.2 Predicting political cynicism and negative campaign evaluation (OLS regressions) Political cynicism

Negative campaign evaluation

Demographics Female Age 15–25 Age 26–39 Age 40–54 Age 55–69 Education

⫺0.06a 0.02 0.02 0.05 0.01 0.06

(0.04) (0.16) (0.08) (0.07) (0.07) (0.03)

0.08 0.08 0.09a 0.14b 0.08 0.04

(0.04) (0.13) (0.07) (0.06) (0.06) (0.02)

Predispositions Left political leaning Right political leaning Political efﬁcacy Government disapproval EU skepticism

0.04 ⫺0.04 ⫺0.28c 0.15c 0.26c

(0.06) (0.05) (0.04) (0.03) (0.03)

0.04 ⫺0.01 ⫺0.27c 0.07c 0.27c

(0.05) (0.05) (0.03) (0.02) (0.03)

n Adjusted R 2

892 0.27

889 0.25

Notes Cell entries are standardized beta-coefﬁcients and standard errors (in parentheses). Only respondents answering all questions are included in the model. a p ⬍ 0.05. b p ⬍ 0.10. c p ⬍ 0.001.

Appendices 191 Appendix to Figures 7.3 and 7.4 Effects of strategic news coverage on political cynicism and campaign evaluation Political cynicism

Negative campaign evaluation

Demographics Female Age Education

0.00 0.02 ⫺0.02

⫺0.23 0.00 0.01

Predispositions Left political leaning Right political leaning Government disapproval

0.01 (0.05) ⫺0.06a (0.04) 0.11c (0.02)

⫺0.01 (0.06) 0.04 (0.05) 0.06a (0.03)

Campaign variables Discuss euro issue Campaign involvement Exposure to strategic news Political cynicism (time 1) Campaign evaluation (time 1)

0.03 ⫺0.10b 0.09b 0.56c –

0.04 ⫺0.07a 0.08a – 0.43c

Adjusted R n

2

(0.04) (0.02) (0.02)

(0.03) (0.03) (0.01) (0.03)

0.29 752

(0.05) (0.02) (0.03)

(0.03) (0.04) (0.01) (0.03)

0.20 757

Notes Cell entries are standardized beta-coefﬁcients and standard errors (in parentheses). Only respondents answering all questions are included in the model. a p ⬍ 0.05. b p ⬍ 0.10. c p ⬍ 0.001.

Appendix to Figure 7.5 Predictors of negative change in perception of role played by polls in the referendum campaign The campaign was too dominated by polls Age (in years) Gender (female) Education Voted for loser in the referendum (“Yes”) Media exposure The campaign is too dominated by polls (time 1)

0.09a ⫺0.00 ⫺0.03 ⫺0.05 0.08a 0.30b

R2 n

0.11 818

(0.03) (0.06) (0.03) (0.06) (0.01) (0.03)

Notes Result of OLS regression predicting a perception of the campaign as being too dominated by polls. Entries are standardized beta-coefﬁcients, standard errors in parentheses. a p ⬍ 0.05. b p ⬍ 0.001.

192

Appendices

Appendix Chapter 8 Dependent variables Government approval rating: A scale from 1 to 5 where 1 equals very good, 3 neither good nor bad, and 5 very bad. Political leader evaluation: A scale from 1 to 10 with 1 being very unsympathetic, 5 neither nor, 10 very sympathetic. Independent variables Gender: Female ⫽ 1; male ⫽ 0. Age: In years. Education: ⬍ 16 years: Primary school and high school ⫽ 1; otherwise ⫽ 0. Education long: ⬎ 19 years: University degree ⫽ 1; otherwise ⫽ 0. Issue-speciﬁc evaluation: Evaluation of topics such as the euro, unemployment, health care, and food safety on ﬁve-point scales ranging from very well to very bad. Political involvement: A four-point scale ranging from 1 to 4 where 1 equals not at all interested in politics and 4 very interested. Ideology: Left political leaning: Self-placement on left–right scale where 1 equals left and 10 right between 1 and 3 ⫽ 1; otherwise ⫽ 0; Right political leaning: Self-placement between 7 and 10 ⫽ 1; otherwise ⫽ 0. Exposure to public television news: Watching TV1 news (0–7 days per week). Exposure to private television news: Watching TV2 news (0–7 days per week). Exposure to No press: Reading EkstraBladet (0–7 days per week). Exposure to Yes press: Index of reading one or more Yes press newspapers daily (Yes papers are: Politiken, Berlingske Tidende, JyllandsPosten, and B.T., 0–28).

Appendices 193 Appendix to Figures 8.1 and 8.2 Explaining government evaluation (OLS regression) Government evaluation low political involvement

Government evaluation high political involvement

Standardized Standard beta error

Standardized Standard beta error

Issue-speciﬁc evaluations Evaluation euro Evaluation immigration Evaluation health Evaluation social issues Evaluation environment Evaluation food safety Evaluation crime Evaluation unemployment

⫺0.25a ⫺0.11 ⫺0.01 ⫺0.03 ⫺0.02 ⫺0.01 ⫺0.01 ⫺0.09c

0.04 0.04 0.05 0.06 0.06 0.06 0.05 0.05

⫺0.17a ⫺0.11c ⫺0.01 ⫺0.13c ⫺0.03 ⫺0.13c ⫺0.08 ⫺0.05

0.04 0.05 0.05 0.06 0.06 0.06 0.06 0.05

Campaign variables Exposure TV1 Exposure TV2 Exposure Yes press Exposure no press

0.09c ⫺0.00 ⫺0.04 ⫺0.01

0.02 0.02 0.00 0.00

0.06 ⫺0.03 0.02 0.02

0.02 0.02 0.00 0.00

Control variables Left ideology Right ideology Government evaluation (time 1)

n.s. 0.03 ⫺0.05 0.37a

n.s. 0.09 0.09 0.05

n.s. 0.02 ⫺0.12b 0.44a

n.s. 0.09 0.09 0.05

R2 n

0.61 429

0.64 326

Notes Control variables gender, age, and education are omitted from the Table for presentational reasons. All variables were non-signiﬁcant. a p ⬍ 0.001 b p ⬍ 0.01. c p ⬍ 0.05.

Appendix Chapter 9 Description of variables in Figure 9.1 and item wording in questionnaire. Dependent variable No vote: Voted No ⫽ 1; voted Yes ⫽ 0 Independent variables Socio-demographics Gender: Female ⫽ 1; male ⫽ 0. Age: In years. Education: ⬍ 16 years: Primary school and high school ⫽ 1; otherwise ⫽ 0.

194 Appendices Education long: ⬎ 19 years: University degree= 1; otherwise ⫽ 0. Predispositions Interested in euro: A four-point scale ranging from 1 to 4 where 1 equals not at all interested in the euro and 4 very interested. Interested in politics: A four-point scale ranging from 1 to 4 where 1 equals not at all interested in politics and 4 very interested. Efﬁcacy: Six-item index forming a scale of efﬁcacy, Cronbach’s ␣ ⫽ 0.69. Question wording: People have different opinions about politics. We would like to hear your opinion about politics in general. Please tick one box per statement. Strongly agree, Agree, Neither agree nor disagree, Disagree, Strongly disagree, Don’t know. 1 2 3 4 5 6

At times, politics can be so complex that people like me don’t understand what is going on. People like me don’t have any say in what the government does. I think that I am better informed about politics than others. MPs want to keep in touch with the people. Parties are only interested in people’s vote, not their opinions. There are so many similar parties that it does not matter who is in government.

EU skepticism: Seven-item index tapping general attitudes and opinions about (the extent of) European integration. The seven items form a scale of EU skepticism, Cronbach’s ␣ ⫽ 0.82. 1 2 3 4 5 6 7

European integration is being pushed too fast. I would be willing to make a sacriﬁce to help a less strong country. Denmark should maintain its exceptions from the Treaty of Edinburgh. The EU has more disadvantages than advantages for people like me. The EU should be enlarged with former Eastern Bloc countries such as Lithuania and Poland. The EU should be enlarged with southern European countries such as Cyprus and Turkey. The EU and European integration is moving towards becoming a European “super state.” All variables were recoded so that disagreement reﬂects higher skepticism.

Government disapproval rating: A scale from 1 to 5 where 1 equals very good, 3 neither good nor bad, and 5 very bad. National economic expectations: Scaled item measure of economic expectations. Question wording: If Denmark joins the common currency,

Appendices 195 how do you think this will affect the ﬁnancial situation of Danish businesses? Will it become? Much better, Better, Neither better nor worse, Worse, Much Worse. Personal economic expectations: Scaled item measure of economic expectations. Question wording: If Denmark joins the common currency, how do you think this will affect your own ﬁnancial situation? Will it become? Much better, Better, Neither better nor worse, Worse, Much Worse. Left political leaning: Self-placement on left–right scale where 1 equals left and 10 right between 1 and 3 ⫽ 1; otherwise ⫽ 0. Right political leaning: Self-placement on left–right scale where 1 equals left and 10 right between 7 and 10 ⫽ 1; otherwise ⫽ 0. Campaign variables: Interpersonal communication: A four-point scale ranging from 1 to 4 where 1 equals never discussing the euro issue during the campaign and 4 equals often discussing the euro issue. Campaign involvement: A three-item index tapping the frequency of engaging in various campaign-related activities, Cronbach’s ␣ ⫽ 0.65. Question wording: Within the last month how often did you – (please tick one box per statement) Often, Sometimes, Almost never, Never. 1 2 3

Take part in a public meeting about the euro. Visit an Internet website about the euro. See an advertisement or commercial about the euro.

High exposure to public television news: Watching TV1 news regularly (4 or more days a week) ⫽ 1; otherwise ⫽ 0. This is linked to the amount and tone of the coverage. Amount refers to the number of stories/the number of news programs between the two waves. Tone refers to the difference in tone between evaluations of Yes actors and No actors. The variable therefore consists of high exposure ⫻ 76/30 (amount of news) ⫻ 0.03 (difference in negative evaluations between Yes and No actors) ⫽ 0.08. High exposure to private television news: Watching TV2 news regularly (4 or more days a week) ⫽ 1; otherwise ⫽ 0. This is linked to the amount and tone of the coverage. Amount refers to the number of stories/the number of news programs between the two waves. Tone refers to the difference in tone between evaluations of Yes actors and No actors. The variable therefore consists of high exposure ⫻ 79/30 (amount of news) ⫻ 0.08 (difference in negative evaluations between Yes and No actors) ⫽ 0.21. High exposure to No press: Reading Ekstra Bladet regularly (4 or more days a week) ⫽ 1; otherwise ⫽ 0.

196

Appendices

High exposure to Yes press: Index of reading one or more Yes press newspapers daily (Yes papers are: Politiken, Berlingske Tidende, JyllandsPosten, and B.T.). Political cynicism: Six-item index tapping cynical feelings about the campaign forming a scale of cynicism, Cronbach’s ␣ ⫽ 0.65. Question wording: People have different opinions about the campaign before the euro referendum. We would like to hear what you think. Please tick one box per question. Strongly agree, Agree, Neither agree nor disagree, Disagree, Strongly disagree. 1 2 3 4 5 6

The campaign was more about strategy than content. The campaign gave me sufﬁcient information to make a qualiﬁed choice. The candidates were straightforward and honest in their campaigning. The campaign was about what is best for Denmark. The campaign was too concerned with the Yes and No sides’ standing in the polls. The candidates were too superﬁcial in their argumentation.

Intention to vote No wave 1: Firm intention to vote No in wave 1 ⫽ 1; otherwise ⫽ 0. Undecided wave 1: Don’t know and undecided with leaning but no ﬁrm opinion in wave 1 ⫽ 1; otherwise ⫽ 0.

Appendices 197 Appendix to Figure 9.1 Inﬂuence of campaign effects on changes in likelihood of voting No. Logistic regression Coefﬁcient

Effect exp (B)

Constant

⫺1.02

(2.43)

Socio-demographics Gender (female) Age Education (⬍ 16 years in school) Education (⬎ 19 years in school)

⫺0.01 ⫺0.10 ⫺0.19 ⫺0.07

(0.30) (0.14) (0.32) (0.55)

0.99 0.91 0.83 0.93

Predispositions Interest in euro issue Interest in politics Efﬁcacy EU skepticism Government disapproval rating National economic expectations Personal economic expectations Political leaning left Political leaning right

0.10 ⫺0.20 ⫺0.39 1.06c 0.36a ⫺0.56 ⫺0.87b 0.84a ⫺0.44

(0.20) (0.18) (0.27) (0.29) (0.15) (0.43) (0.28) (0.37) (0.36)

1.11 0.82 0.68 2.90 1.43 0.58 0.41 2.30 0.64

Campaign variables Interpersonal communication Campaign involvement High exposure public TV news High exposure private TV news High exposure No press High exposure Yes press Cynicism about campaign Intention to vote No wave 1 Undecided wave 1

0.07 0.28 ⫺1.70a 0.94 0.60 ⫺0.49b 0.21 3.78c 1.02c

(0.21) (0.40) (0.38) (0.16) (0.60) (0.17) (0.28) (0.41) (0.33)

1.07 1.32 0.01 2.55 1.82 0.61 1.24 44.01 2.78

⫺2 log likelihood Nagelkerke Pseudo R2 Cases correctly classiﬁed n

359.57 0.80 91.50 765

Notes Positive beta-coefﬁcient indicates positive association with voting No. Standard errors are in parentheses. a p ⬍ 0.05. b p ⬍ 0.01. c p ⬍ 0.001.

Notes

2 The context of the campaign 1 The technique involves 1 2 3 4 5

deﬁning an appropriate sample of content to be analyzed, establish categories of referents relevant to the inquiry, deﬁne a “unit of analysis” in the form of for example words, sequences or news stories, examine the number of times a category appears per unit of analysis, report the distribution of occurrences of categories in the content sample (see Riffe et al. 1996).

2 The focus groups were conducted in cooperation with the Audience Research Department and News Department of the Danish Broadcasting Association [DR] between August 21 and August 24, 2000. 3 While focus groups are not designed to obtain a representative sample, the participants in the focus group interviews largely reﬂected the composition of the general electorate. The participants in the focus groups all completed the questionnaire used in the two-wave panel study. A comparison of attitudes, level of political knowledge and media use revealed a similar distribution within the focus group participants compared to approximately 1,000 respondents in the survey (de Vreese et al. 2000). 4 The campaign in the news 1 The intercoder reliability for this measure was 95 percent. 5 Elite framing of the issue 1 In this period, we analyzed 140 front pages and 843 television news stories to identify news about the referendum. The inter-coder reliability test, conducted on a randomly selected sub-sample of 20 newspaper and 20 television news stories, resulted in a pair-wise inter-coder agreement of 84 percent–97 percent for the measures used in this chapter. 6 Agenda-setting in a referendum 1 Each television news story was coded for main topic to identify the composition of the news agenda. Each euro referendum story in the television news and in the newspapers was additionally categorized in sub-issues. The inter-coder relia-

Notes

2 3

4

5

6 7 8

199

bility, computed on a randomly selected sub-sample of 23 news stories, was 91 percent and 84 percent respectively for the two measures. To measure the public agenda, we used the standard “Most Important Problem” question (see Dearing and Rogers 1996). In the investigation of aggregate level agenda-setting effect, an additional time lag is not included. The media agenda in August is seen as predicting the public agenda in late August (wave 1). Some argue that the typical time span for the translation of news about an issue into a high degree of salience for that issue on the public agenda is 5–7 weeks (McCombs et al. 1995). To test this assumption, the media agenda in August was compared to the public agenda at the end of September (wave 2) which includes a 4–8-week time lag. However, the correlations between the public agenda in September and the media in August were lower than the ones reported above and do therefore not support utilizing the longer time lag. This measure combines the volume of news about the referendum per news outlet ⫻ exposure (the number of days watched/read). For example: if respondent A watches TV1 four days a week and TV1 carried 76 news stories about the referendum in the ﬁnal 30 days of the campaign, this respondent would achieve the score (4 ⫻ 76/30) ⫽ 10.1 as an indicator of exposure to news coverage about the referendum by TV1. The public’s “euro agenda” was measured in one month prior to Election Day and in the immediate aftermath of the referendum. The news coverage of the referendum was analyzed in the ﬁnal four weeks of the campaign and is compared to the public agenda in wave two. The public referendum agenda was measured by asking respondents to identify the issue they considered most important in relation to the referendum. These categories come from a monitoring project of the Danish press coverage during the campaign conducted by the weekly news magazine Monday Morning (www.mm.dk). “Other” is also a residual category and excluded from further analyses. Excluding the category “campaign” from the rank order analysis does not alter the (negative) direction of the relationship between the media agenda and the public agenda.

7 Polls, strategy news, and political cynicism 1 For this we use data from our panel survey with a sample of 962 respondents participating in the pre- and post-referendum waves. For more information about the survey, see Chapter 2. 2 These items are derived from Cappella and Jamieson (1997) and Pinkleton and Austin (2001). Respondents were asked the extent to which they agreed (on a ﬁve-point agree–disagree scale) with the following statements: “The candidates are open and honest in their campaigning,” and “The candidates are superﬁcial in their arguments.” The two items formed a reliable measure of political cynicism ranging from one to ﬁve (Pearson bivariate correlation ⫽ 0.50; Cronbach’s ␣ ⫽ 0.67). The mean values were: wave 1: M ⫽ 3.55, SD ⫽ 0.76; wave 2: M ⫽ 3.70, SD ⫽ 0.72. 3 Given the contextual bias of the items used by Cappella and Jamieson (1997), the wording was slightly adjusted to the Danish referendum context. The items were (a) “The campaign was more about strategy than content,” (b) “The campaign gave me sufﬁcient information to make a qualiﬁed choice” (reverse coded), (c) “The campaign was about what is best for Denmark,” and (d) “The campaign was too concerned with the Yes and No sides’ standings in the polls.” All statements were answered using a ﬁve-point Likert agree–disagree scale.

200

4 5 6 7 8

9

Notes

The order of the items was randomized in the two waves of the survey and included both positively and negatively framed items. The items formed an acceptable scale (Cronbach’s ␣ ⫽ 0.59) (Stevens 1996) and were converted to a one (positive campaign evaluation) to ﬁve (negative campaign evaluation) scale. The mean values were: wave 1: M ⫽ 3.26, SD ⫽ 0.64; wave 2: M ⫽ 3.57, SD ⫽ 0.60. Six-item index, Cronbach’s ␣ ⫽ 0.69. Three-item index, Cronbach’s ␣ ⫽ 0.65. One-scale item, inversely coded. Seven-item index, Cronbach’s ␣ ⫽ 0.82. We do not include political efﬁcacy and EU skepticism in this model explaining political cynicism and campaign evaluations at the end of the campaign. Both of these indicators were measured in the ﬁrst wave of the panel. The two measures explain a substantial part of the variance in political cynicism in wave 1. Political cynicism at time 1 is included as a control variable in the change model, which then also accounts for some of the impact of political efﬁcacy and EU skepticism. Figures 7.3 and 7.4 report OLS regressions with the level of cynicism in wave 2 as the dependent variable. As suggested by Aiken, West, and Reno (1991) we also ran the model with the net change between waves 1 and 2 as the dependent variable. This analysis yielded a similar result to what is reported in the ﬁgures. This impression is supported by out data. We included knowledge measures in our survey, and most respondents displayed a high level of knowledge.

8 Primed by the referendum 1 Priming is also well established in other ﬁelds than political communication, e.g., in investigations of entertainment and violence (see Peter 2002; McLeod et al. 2002). 2 The overall government performance was measured by a ﬁve-point scale The scale ranged from 1 to 5 where 1 equals very good, 3 neither good nor bad and 5 very bad. The evaluation of key political actors was measured using ten-point sympathy scales for each of the political actors. The scale ranged from 1 to 10 with 1 being very unsympathetic, 5 neither/nor, 10 very sympathetic. 3 The two most widely watched evening television news programs were analyzed from March–September. A more detailed content analysis also including ﬁve dailies was carried out in the month prior to the referendum (see also Chapter 2). The inter-coder reliability test showed 95 percent inter-coder agreement for the measures relevant to in this chapter. 4 Respondents with a left political leaning are classiﬁed as those who placed themselves as 1–3 on the scale, right political leaning are classiﬁed as those who placed themselves as 7–10. The effective reference category includes those in the political center. 5 The question was: If there were a general election tomorrow which party would you vote for? 9 Campaign effects of the vote 1 Ideology has also been used instead of party attachment, to show associations between one camp (left or right) and the vote. In a study of the 1994 referendums in Norway, Sweden, and Finland on entering the EU, for example, rightwing ideology was more associated with a Yes vote (Aardal et al. 1998). 2 See, for example, the arguments put forward in Clarke et al. (2000: Chapters 4 and 6) modeling the vote in the Canadian 1992 Constitutional Referendum and 1995 Quebec Sovereignty Referendum. See also Clarke and Kornberg (1994).

Notes

201

3 Jenssen (1998: 211) concluded: “people with strong market positions were likely to be optimistic about the economy (past and future, personal and national) and responsive to the campaign message from the Yes side.” Other research on U.S. national elections points out the importance of the media in shaping such economic assessments (Hetherington 1996). 4 See, for example, Table 3 in Franklin et al. (1994), which reports different indicators of government performance and approval. 5 Differences in support for European integration by gender, age, and level of education seem to decrease, which was also found by Jakobsen et al. (2001) in their analysis of the impact of social background on the 2000 vote in the Danish referendum. 6 Respondents with a left political leaning are classiﬁed as those who placed themselves as 1–3 on the scale, right political leaning are classiﬁed as those who placed themselves as 7–10. The effective reference category includes those in the political center. 7 Such a measure was not necessary for the press because all newspapers committed publicly to either the Yes or No standpoint in the campaign and they are therefore classiﬁed accordingly in our analysis (see Curtice and Semetko (1994) for a similar example with the British partisan press). For television, we illustrate our measure with an example: The result for the tone measure for the public service television channel TV1 was that Yes actors were evaluated 0.03 more negative than No actors. Respondent A watches TV1 four days per week. TV1 DR carried 76 news stories about the referendum in the ﬁnal 30 days of the campaign. This respondent would be scored (1 (high exposure) ⫻ 0.03 (difference in tone) ⫻ 76/30 (amount of news)) ⫽ 0.08 as an indicator of exposure to news about the referendum on TV1. This is an individual-level measure that takes into account the exposure to speciﬁc news outlets, the actual tone of the content, and the volume of this content. 8 Additional comparative analyses of the composition of voter groups revealed more similarities than differences between undecided voters and the group of voters holding a ﬁrm opinion on the euro issue. The distribution of undecided voters on key social demographic variables and pre-dispositions was similar to that of the decided voters. An additional logistic regression model predicting the likelihood of being “undecided” at the outset of the campaign using the social demographics and predispositions listed in Figure 9.1 (fully speciﬁed in the Appendix) as explanatory variables resulted in only 75.2 percent correctly classiﬁed cases and a pseudo R2 of 4 percent. Our analysis suggested that undecided and decided voters only differed marginally. Elderly voters, voters with less political interest, those who were less efﬁcacious and persons with pessimistic personal economic expectations, were somewhat more likely to be undecided at the outset of the ﬁnal weeks of the campaign.

References

Aardal, B., Jenssen, A. T., Oscarsson, H., Sankiaho, R., and Saynassalo, E. (1998). Can ideology explain the EU vote: 235–268. In A. T. Jenssen, P. Pesonen, and M. Gilljam (eds), To Join or Not to Join: Three Nordic Referendums on Membership in the European Union. Oslo: Scandinavian University Press. Aarts, K. and Semetko, H. A. (2003). The divided electorate: Effects of media use on political involvement. The Journal of Politics, 65 (3): 759–784. Achen, C. H. (1975). Mass political attitudes and the survey response. American Political Science Review, 69: 1218–1231. Acock, A. and Clarke, H. D. (1990). Alternative measures of political efﬁcacy: Models and means. Quality and Quantity, 24, 87–105. Acock, A., Clarke, H. D., and Stewart, M. C. (1985). A news model for old measures: A co-variance structure analysis of political efﬁcacy. Journal of Politics, 47: 1062–1084. Adams, W. J. (2000). How people watch television as investigated using focus group techniques. Journal of Broadcasting and Electronic Media, 44: 78–93. Agger, R. E., Goldstein, M. N., and Pearl, S. A. (1961). Political cynicism: Measurement and meaning. The Journal of Politics, 23 (3): 477–506. Aiken, L. S., West, S. G., and Reno, R. R. (1991). Multiple Regression: Testing and Interpreting Interactions. Newbury Park, CA: Sage. Andersen, J., Borre, O., Andersen, J. G., and Nielsen, H. J. (1999). Vaelgere med omtanke. En analyse af folketingsvalget 1998 [Reﬂexive voters: An analysis of the 1998 parliamentary elections]. Copenhagen: Systime. Andersen, J. G. (1998). Danskerne og Europa. Analyse af EU-afstemningen den 28. maj 1998 [The Danes and Europe. Analysis of the EU referendum May 28, 1998]. Report published by Ugebrevet Mandag Morgen. Copenhagen: Mandag Morgen. Andersen, O. E. (1995). Medieudbud og medieforbrug i Danmark 1983–1994 [Media availability and media use in Denmark, 1983–1994]. Copenhagen: Ministry of Domestic Affairs. Andersen, V. N. and Hansen, K. M. (2001). Revitalisering af demokrati – den danske folkehoering om euroen [Revitalization of democracy – the Danish deliberative poll on the euro], pp. 169–196. In E. O. Eriksen and M. Skivenes (eds), Deliberasjon og demokrati [Deliberation and democracy]. Bergen, Norway: LOS-Senteret. Anderson, R. (2000). Reporting public opinion polls: The media and the 1997 Canadian election. International Journal of Public Opinion Research, 12: 285–298.

References 203 Ansolabehere, S. and Iyengar, S. (1995). Going Negative. How Political Advertisements Shrink and Polarize the Electorate. New York: The Free Press. Ansolabehere, S., Iyengar, S., and Simon, A. (1999). Replicating experiments using aggregate and survey data: The case of negative advertisement and turnout. American Political Science Review, 93 (4): 891–900. Ansolabehere, S., Iyengar, S., Simon, A., and Valentino, N. (1994). Does attack advertisement demobilize the electorate? American Political Science Review, 88 (4): 829–838. Appleton, A. (1992). Maastricht and the French party system: Domestic implications of the treaty referendum. French Politics and Society, 10: 1–18. Asp, K. (1983). The struggle for the agenda. Party agenda, media agenda, and voter agenda in the 1979 Swedish election campaign. Communication Research, 10: 333–355. Atkin, C. (1969). The impact of political poll reports on candidate and issue preferences. Journalism Quarterly, 4: 515–521. Barber, J. D. (1978). Race for the Presidency: The Media and the Nominating Process. New York: Prentice Hall. Bartels, L. M. (1986). Issue voting under uncertainty: An empirical test. American Journal of Political Science, 30: 709–728. Bartels, L. M. (1993). Messages received: The political impact of media exposure. American Political Science Review, 87 (2): 267–285. Belli, R. F., Traugott, M. W., Young, M., and McGonagle, K. A. (1999). Reducing vote over-reporting in surveys: social desirability, memory failure and source monitoring. Public Opinion Quarterly, 63: 90–108. Bennett, W. L. and Entman, R. (2001). Mediated Politics. Communication and the Future of Democracy. Cambridge: Cambridge University Press. Benton, M. and Frazier, P. J. (1976). The agenda-setting function of the mass media at three levels of “information holding.” Communication Research, 3 (3): 261–274. Berelson, B., Lazarsfeld, P. F., and McPhee, W. N. (1954). Voting. A Study of Opinion Formation in a Presidential Campaign. Chicago: University of Chicago Press. Bille, L., Eklit, J., and Jacobsen, M. V. (1992). Denmark: The 1990 campaign, pp. 63–81. In S. Bowler and D. Farrell (eds), Electoral Strategies and Political Marketing. New York: Macmillan. Blumler, J. G. (1969). Television in Politics. Its Uses and Inﬂuence. Chicago: Chicago University Press. Blumler, J. G. (ed.) (1983). Communicating to Voters. Television in the First European Parliamentary Elections. London: Sage. Blumler, J. G. and Gurevitch, M. (1989). The Earnest vs. the Determined: Election newsmaking at the BBC, 1987. In I. Crewe and M. Harrop (eds), The General Elections in Britain, 1987. Cambridge: Cambridge University Press. Blumler, J. G. and Gurevitch, M. (1995). The Crisis of Public Communication. London: Routledge. Blumler, J. G. and Gurevitch, M. (1998). Campaign journalism at the BBC, 1997. In I. Crewe, B. Gosschalk, and J. Bartle (eds), Why Labour Won the General Election of 1997. London: Frank Cass. Blumler, J. G. and Gurevitch, M. (2001). “Americanization” reconsidered: U.K. – U.S. campaign communication comparisons across time, pp. 380–406. In W. L.

204

References

Bennett and R. M. Entman (eds), Mediated Politics. Communication in the Future of Democracy. Cambridge: Cambridge University Press. Borg, S. and Esaiasson, P. (1998). Exposure to the campaigns: 85–101. In A. T. Jenssen, P. Pesonen, and M. Gilljam (eds), To Join or not to Join: Three Nordic Referendums on Membership in the European Union. Oslo: Scandinavian University Press. Bowler, S. and Donovan, T. (1998). Demanding Choices: Opinion, Voting and Direct Democracy. Ann Arbor: University of Michigan Press. Bowler, S., Donovan, T., and Tolbert, C. J. (eds) (1998). Citizens as Legislators: Direct Democracy in the United States. Columbus: Ohio State University Press. Brettschneider, F. (1997). The press and the polls in Germany, 1980–1994. Poll coverage as an essential part of election campaign reporting. International Journal of Public Opinion Research, 9, 248–265. Buch, R. and Hansen, K. M. (2002). The Danes and Europe: From EC 1972 to Euro 2000 – elections, referendums and attitudes. Scandinavian Political Studies, 25, 1–36. Budge, I. and Farlie, D. J. (1983). Explaining and Predicting Elections: Issue Effects and Party Strategies in Twenty-Three Democracies. London: Allen & Unwin. Butler, D. (1996). Polls and elections, pp. 236–253. In L. LeDuc, R. G. Niemi, and P. Norris (eds), Comparing Democracies. Elections and Voting in Global Perspective. Thousand Oaks, CA: Sage. Butler, D. and Kavanagh, D. (1997). The British General Election of 1997. London: Palgrave. Butler, D. and Ranney, A. (eds) (1994). Referendums Around the World. The Growing Use of Direct Democracy. Washington, DC: Macmillan. Butzer, M. and Marquis, L. (2002). Public opinion formation in Swiss federal referendums, pp. 163–182. In D. Farrell and R. Schmitt-Beck (eds), Do Political Campaigns Matter? Campaign Effects in Elections and Referendums. London: Routledge. Campbell, A., Miller, W. E., and Converse, P. E. (1960). American Voter. Chicago: University of Chicago Press. Cappella, J. N. and Jamieson, K. H. (1997). Spiral of Cynicism. The Press and the Public Good. New York: Oxford University Press. Ceci, S. K. and Kain, E. L. (1982). Jumping on the bandwagon with the underdog: The impact of attitude polls on polling behavior. Public Opinion Quarterly, 46 (2): 228–242. Chan, K. (1999). The media and environmental issues in Hong Kong 1983–95. International Journal of Public Opinion Research, 11: 135–151. Cherribi, O. (2003a). Outing the Dutch: (In)tolerance, populism and the media. Paper presented at the Annual Meetings of the International Communication Association, San Diego, May. Cherribi, O. (2003b). Olanda (s)velata: intolleranza, populismo e media. Comunicazione politica, IV/2: 141–165. Citrin, J. (1974). The political relevance of trust in government. American Political Science Review, 68: 973–988. Clarke, H. D. and Kornberg, A. (1994). The politics and economics of constitutional choice: Voting in Canada’s 1992 national referendum. Journal of Politics, 56 (4): 940–962.

References 205 Clarke, H. D., Kornberg, A., and Wearing, P. (2000). A Polity on the Edge: Canada and the Politics of Fragmentation. Peterborough, Ontario: Broadview Press. Cohen, J. E. and Hamman, C. (2000). Presidential image and job performance in the public mind. Paper presented at the 2000 Annual Meeting of the American Political Science Association, Washington, DC. Converse, P. E. (1962). Information ﬂow and stability of partisan attitudes. Public Opinion Quarterly, 26 (4): 578–599. Converse, P. E. (1964). Nature of belief systems in mass publics, pp. 206–261. In D. Apter (ed.), Ideology and Discontent. New York: Free Press. Converse, P. E. (1970). Information ﬂow and the stability of partisan attitude. In A. Campbell, P. E. Converse, W. Miller, and D. E. Stokes (eds), Elections and the Political Order. New York: Wiley. Cook, S. W. and Welch, A. C. (1940). Methods of measuring the practical effects of polls of public opinion. Journal of Applied Psychology, 24: 449–460. Craig, S. C., Niemi, R. G., and Silver, G. E. (1990). Political efﬁcacy and trust: A report on the NES pilot study items. Political Behavior, 12: 289–314. Curtice, J. and Payne, C. (1995) Forecasting the 1992 election: the BBC experience, pp. 213–224. In I. Crewe and B. Gosschalk (eds), Political Communications: The General Election Campaign of 1992. Cambridge: Cambridge University Press. Curtice, J. and Semetko, H. A. (1994). Does it matter what the papers say?, pp. 43–64. In A. Heath, R. Jowell, and J. Curtice (eds), Labour’s Last Chance: The 1992 Election and Beyond. Aldershot: Dartmouth. Dalton, R. J. (2000). The decline of party identiﬁcation. In R. Dalton and M. Wattenberg (eds), Parties Without Citizens. Oxford: Oxford University Press. Dalton, R. J. (2002). Citizen Politics. Public Opinion and Political Parties in Advanced Industrial Democracies. New York: Chatham House. D’arcy, R. and Laver, M. (1990). Referendum dynamics and the Irish divorce amendment. Public Opinion Quarterly, 54 (1): 1–20. Dearing, J. W. and Rogers, E. M. (1996). Agenda-setting. Thousand Oaks, CA: Sage. Diez-Nicolas, J. and Semetko, H. A. (1995). La television y las elecciones de 1993 [Television and the 1993 elections], pp. 243–304. In A. Munoz-Alonso and J. Ignacio Rospir (eds), Communicacion Politica. Madrid: Editorial Universitas, S.A. Edwards, G. C. and Swenson, T. (1997). Who rallies? The anatomy of a rally event. Journal of Politics, 59: 200–212. van der Eijk, C. and Franklin, M. (1996). Choosing Europe? Ann Arbor: University of Michigan Press. van der Eijk, C. and Franklin, M. (2003). Potential for contestation on European matters at national elections in Europe. In G. Marks and M. Steenbergen (eds), European Integration and Political Conﬂict New York: Cambridge University Press. Eilders, C. (1997). Nachrichtenfaktoren und Rezeption. Eine empirische Analyse zur Auswahl und Verarbeitung politischer Information [News factors and reception. An empirical analysis on the selection and processing of political information]. Opladen: Westdeutscher Verlag. Entman, R. B. (1993). Framing: Toward clariﬁcation of a fractured paradigm. Journal of Communication, 43: 51–58.

206

References

Epstein, K. and Strom, G. (1981). Election night projections and west coast turnout. American Politics Quarterly, 9: 489–491. Erber, R. and Lau, R. R. (1990). Political cynicism revisited: An information-processing reconciliation of policy-based and incumbency-based interpretations of changes in trust in government. American Journal of Political Science Review, 34: 236–253. Erbring, L., Goldenberg, E. N., and Miller, A. H. (1980). Front-page news and real-world cues: A new look at agenda-setting by the media. American Journal of Political Science, 24: 16–49. Eurobarometer 51–59. Public opinion in the EU. Brussels: Directorate X, European Commission. Farrell, D. (1996). Campaign strategies and tactics, pp. 160–183. In L. LeDuc, R. G. Niemi, and P. Norris (eds), Comparing Democracies. Elections and Voting in Global Perspective. Thousand Oaks, CA: Sage. Farrell, D. and Webb, P. (2000). Political parties as campaign organizations. In R. Dalton, J. Russel and M. P. Wattenberg (eds), Parties Without Partisans. Oxford: Oxford University Press. Farrell, D. M. and Schmitt-Beck, R. (eds) (2002). Do Political Campaigns Matter? Campaign Effects in Elections and Referendums. London and New York: Routledge. Finkel, S. E. and Geer, J. G. (1998). A sport check: Casting doubt on the demobilizing effect of attack advertisement. American Journal of Political Science, 42 (2): 573–595. Fiske, S. T. and Taylor, S. E. (1991). Social Cognition. Second edition. New York: McGraw-Hill. Franklin, M. (2002). Learning from the Danish case: A comment on Palle Svensson’s critique of the Franklin thesis. European Journal of Political Research, 41 (6): 751–758. Franklin, M., van der Eijk, C., and Marsh, M. (1995). Referendum outcomes and trust in government: Public support for Europe in the wake of Maastricht. West European Politics, 18: 101–117. Franklin, Mark, Mackie, T. T., and Valen, H. (1992). Electoral Change: Responses to Evolving Social and Attitudinal Changes in Western Countries. Cambridge: Cambridge University Press. Franklin, M., Marsh, M., and Wlezien, C. (1994). Attitudes toward Europe and referendum votes: A response to Siune and Svensson. Electoral Studies, 13 (2): 117–121. Frankovic, K. A. (1998). Public opinion and polling, pp. 150–170. In D. A. Graber, D. McQuial, and P. Norris (eds), The Politics of News. The News of Politics. Washington, DC: Congressional Quarterly Press. Frankovic, K. A. (2001). Fixing the U.S. election polling problems means ﬁxing election problems, Newsletter of the World Association for Public Opinion Research, 1: 1–5. Freedman, P. and Goldstein, K. (1999). Measuring media exposure and effects of negative campaign ads. American Journal of Political Science, 43 (4): 1189 – 1208. Galtung, J. and Ruge, M. H. (1965). The structure of foreign news. The presentation of the Congo, Cuba, and Cyprus crises in four Norwegian newspapers. Journal of Peace Research, 1: 64–91.

References 207 Gidengil, A., Blais, A., Nadeau, R., and Nevitte, N. (2002). Priming and campaign context: Evidence from recent Canadian elections, pp. 76–91. In D. Farrell and R. Schmitt-Beck (eds), Do Political Campaigns Matter? Campaign Effects in Elections and Referendums. London: Routledge, 2002. Glynn, C. J. and McLeod, J. (1982). Public opinion communication processes and voting decisions, pp. 542–549. In M. Burgon (ed), Communication Yearbook, Vol. 6. Beverly Hills: Sage. Graber, D. A. (2001). Processing Politics: Learning from Television in the Internet Age. Chicago: University of Chicago Press. Granberg, D. and Holmberg, S. (1991). Self-reported turnout and voter validation. American Journal of Political Science, 35: 448–459. Gross, A. and Kaufmann, B. (2002). IRI Europe country index on citizenlawmaking 2002. Amsterdam: Initiative and referendum institute Europe. Gunther, R., Montere, J. R., and Wert, J. I. (2000). The media and politics in Spain: From dictatorship to democracy, pp. 28–84. In R. Gunther and A. Mughan (eds), Democracy and the Media. A Comparative Perspective. Cambridge: Cambridge University Press. Gurevitch, M. and Blumler, J. G. (1990). Comparative research: The extending frontier, pp. 305–325. In D. L. Swanson and D. Nimmo (eds), New Directions in Political Communication. A Resource Book. Newbury Park: Sage. Hardmeier, S. (1999). Political poll reporting the Swiss print media: Analysis and suggestions for quality improvement. International Journal of Public Opinion Research, 11: 257–274. Harrop, M. and Miller, W. (1987). Elections and Voters. London: Macmillan. Herr, J. P. (2002). The impact of campaign appearances in the 1996 election. Journal of Politics, 64: 904–913. Hertog, J. K. and McLeod, D. M. (2001). A multiperspectival approach to framing analysis: A ﬁeld guide. In S. D. Reese, O. H. Gandy, and A. E. Grant (eds), Framing Public Life. Mahwah, NJ: Lawrence Erlbaum, pp. 139–162. Hetherington, M. (2001). Declining trust and a shrinking policy agenda: Why the media scholars should care. In R. Hart and D. R. Shaw (eds), Communication in U.S. Elections. New Agendas. Lanham, MA: Rowman and Littleﬁeld, pp. 105–121. Hetherington, M. J. (1996). The media’s role in forming voters’ national economic evaluations in 1992. American Journal of Political Science, 40: 372–395. Hibbing, J. R. and Theiss-Morse, E. (1998). The media’s role in public negativity towards Congress: Distinguishing emotional reactions and cognitive evaluations. American Journal of Political Science, 42 (2): 475–498. Hix, S. (1999). The Political System of the European Union. New York: St. Martin’s Press. Hix, S. and Lord, C. (1997). Political Parties in the European Union. New York: St. Martin’s Press. Hjarvard, S. (1999). Nyheder i konkurrence [News in competition]. Koebenhavn: Samfundslitteratur. Holtz-Bacha, C. (1990). Ablenkung oder Abkehr von der Politik? Mediennutzung im Geﬂecht politischer Orientierungen [Distraction or withdrawal from politics? Media use and political attitudes]. Opladen: Westdeutscher Verlag. Holtz-Bacha, C. (1999). Mass media and elections: An impressive body of research. In H-B. Brosius and C. Holtz-Bacha (eds), German Communication Yearbook. Cresskill, NJ: Hampton Press, pp. 39–68.

208

References

Hooghe, L., Marks, G., and Wilson, C. J. (2002). Does left/right structure party positions on European integration? Comparative Political Studies, 35: 965–989. Hovland, C. (1959). Reconciling conﬂicting results derived from experimental and survey studies of attitude change. American Psychologist, 8–17. Huegel, R., Degenhardt, W., and Weiss, H.-J. (1989). Structural equation models for the analysis of the agenda-setting process. European Journal of Communication, 4: 191–210. Hug, S. (2003). Voices of Europe. Citizens, Referendums and European Integration. Oxford: Rowman & Littleﬁeld Publishers. Hug, S. and Sciarini, P. (2000). Referendums on European integration. Do institutions matter in the voter’s decision? Comparative Political Studies, 33 (1): 3–36. Iyengar, S. (1991). Is Anyone Responsible? Chicago: Chicago University Press. Iyengar, S. and Kinder, D. R. (1987). News that Matters. Chicago: University of Chicago Press. Iyengar, S. and Simon, A. F. (1993). News coverage of the Gulf crisis and public opinion: A study of agenda setting, priming and framing. Communication Research, 20 (4): 365–383. Iyengar, S., Kinder, D. R., Peters, M. D., and Krosnick, J. (1983). The evening news and presidential evaluations. Journal of Personality and Social Psychology, 46: 778–787. Jakobsen, M. S., Reinert, R., and Thomsen, S. R. (2001). Afstemningen om den faelles moent – social baggrund og holdninger [The referendum on the common currency – social background and opinions]. Politica, 33 (1): 66–88. Jamieson, K. H. (1992). Dirty Politics. New York: Oxford University Press. Jasperson, A. E., Shah, D. V., Watts, M., Faber, R. J., and Fan, D. P. (1998). Framing the public agenda: Media effects on the importance of the federal budget deﬁcit. Political Communication, 15: 205–224. Jensen, K. B. (1997). Dansk Mediehistorie, bind 1–3 [Danish media history, volume 1–3]. Koebenhavn: Samlerens Forlag. Jenssen, A. T. (1998). Personal economies and economic expectations, pp. 194–214. In A. T. Jenssen, P. Pesonen, and M. Gilljam (eds), To Join or Not to Join: Three Nordic Referendums on Membership in the European Union. Oslo: Scandinavian University Press. Jenssen, A. T., Pesonen, P., and Gilljam, M. (eds) (1998). To Join or Not to Join: Three Nordic Referendums on Membership in the European Union. Oslo: Scandinavian University Press. Johnston, R., Blais, A., Gidengil, E., and Nevitte, N. (1996). The Challenge of Direct Democracy: The 1992 Canadian Referendum. Montreal and Kingston: McGill-Queen’s University Press. Kagay, M. R. (2000). Continuing evolution in the use of public opinion polls by the New York Times: The 1996 presidential election experience, pp. 23–67. In P. J. Lavrakas and M. W. Traugott (eds), Election Polls, the News Media, and Democracy. New York: Chatham House. Kahn, K. F. and Kennedy, P. J. (1999). Do negative campaigns mobilize or suppress turnout? Clarifying the relationship between negativity and participation. American Political Science Review, 93 (4): 877–889. Kahneman, D. and Tversky, A. (1983). Choices, values, and frames. American Psychologist, 39 (4): 341–350.

References 209 Kaid, L. L. and Holtz-Bacha, C. (1995). Political Advertising in Western Democracies. Parties and Candidates on Television. London: Sage. Katz, E. (1971). Platforms and windows: Broadcasting’s role in election campaigns. Journalism Quarterly, 38: 304–314. Kepplinger, H.-M. (1998). Die Demontage der Politik in der Informationsgesellschaft [The dismantling of politics in the information society]. Freiburg: Alber. Kepplinger, H.-M. (2002). Mediazation of politics: Theory and data. Journal of Communication, 52: 973–986. Kerbel, M. R. (1994). Edited for Television: CNN, ABC, and the 1992 Presidential Campaign. Boulder, CO: Westview Press. King, A. (1977). Britain Says Yes. The 1975 Referendum on the Common Market. Washington, DC: American Institute for Public Policy Research. Kiousis, S., Bantimaroudis, P., and Ban, H. (1999). Candidate image attributes. Experiments on the substantive dimension of second level agenda setting. Communication Research, 26: 414–428. Kleinnijenhuis, J. and Fan, D. (1999). “Media coverage and the ﬂow of voters in multiparty systems: The 1994 national elections in Holland and Germany. International Journal of Public Opinion Research, 11 (3): 233–356. Kleinnijenhuis, J. and De Ridder, J.A. (1998). Issue news and electoral volatility: A comparative analysis of media effects during the 1994 election campaigns in Germany and the Netherlands. European Journal of Political Research, 33 (3): 413–437. Kleinnijenhuis, J., Maurer, M., Kepplinger, H. M., and Oegema, D. (2001). Issues and personalities in German and Dutch television news: Patterns and effects. European Journal of Communication, 16: 337–359. Klingeman, H.-D. and Fuchs, D. (eds) (1995). Citizens and the State. Oxford: Oxford University Press. Kosicki, G. M. (1993). Problems and opportunities in agenda-setting research. Journal of Communication, 43 (2): 100–127. Kraus, S. and Davis, D. (1978). The Effects of Mass Communication on Political Behavior. University Park: Pennsylvania University Press. Krosnick, J. A. and Brannon, L. A. (1993). The impact of the Gulf War on the ingredients of presidential evaluations: Multidimensional effects of political involvement. American Political Science Review, 87: 963–975. Krosnick, J. A. and Kinder, D. R. (1990). Altering the foundations of support for the President through priming. American Political Science Review, 8: 497–512. Krosnick, J. A., Narayan, S. S., and Smith, W. R. (1996). Satisﬁcing in surveys: Initial evidence, pp. 29–44. In M. T. Braverman and J. K. Slater (eds), Advances in Survey Research. San Francisco: Jossey-Bass. Lang, G. and Lang, K. (1981). Watergate: An exploration of the agenda-building process. Mass Communication Review Yearbook, 2: 447–468. Larson, S. G. (1999). Public opinion in television election news: Beyond polls. Political Communication, 16: 133–145. Larson, S. G. (2000). Network differences in public opinion coverage during the 1996 presidential campaign. Journal of Broadcasting and Electronic Media, 44 (1): 16–36. Lauf, E. (2001). The vanishing young reader: Sociodemographic determinants of newspaper use in Europe, 1980–98. European Journal of Communication, 16 (2): 233–243.

210

References

Lavrakas, P. and Holley, J. (1990). Polling and Presidential Election Coverage. Newbury Park, CA: Sage. Lavrakas, P. and Traugott, M. W. (2000). Election Polls, the News Media, and Democracy. New York: Chatham House. LeDuc, L. (2002). Referendums and elections: How do campaigns differ? pp. 145–162. In D. M. Farrell and R. Schmitt-Beck (eds), Do Political Campaigns Matter? Campaign Effects in Elections and Referendums. London: Routledge. LeDuc, L. (2003). The Politics of Direct Democracy. Referendums in Global Perspective. Peterborough, Ontario: Broadview Press. LeDuc, L. and Pammett, J. H. (1995). Referendum voting: Attitudes and behaviour in the 1992 constitutional referendum. Canadian Journal of Political Science, 28: 3–34. Leroy, P. and Siune, K. (1994). The role of television in European elections: The cases of Belgium and Denmark. European Journal of Communication, 9: 47–69. Lewis-Beck, M. S. (1997). Economics and Elections: The Major Western Democracies. Ann Arbor: University of Michigan Press. Lijphart, A. (1968). The Politics of Accommodation. Pluralism and Democracy in the Netherlands. Berkeley, CA: University of California Press. Lijphart, A. (1980). Democracy in Plural Societies: A Comparative Exploration. New Haven, CA: Yale University Press. Lippmann, W. (1922). Public Opinion. New York: Collier-Macmillan. Lipset, S. M. (1960). Political Man. New York: Doubleday. Lodge, M., McGraw, K. M., and Stroh, P. (1989). An impression-driven model of candidate evaluation. American Political Science Review, 83: 399–419. Lombardi, W. J., Higgins, E. T., and Bargh, J. A. (1987). The role of consciousness in priming effects on categorization: Assimilation versus contract as a function of awareness of the priming task. Personality and Social Psychology Bulletin, 13: 411–429. Lowery, D. and Sigelman, L. (1981). Understanding the tax revolt: Eight explanations. American Political Science Review, 75: 963–974. Lupia, A. (1992). Busy voters, agenda control and the power of information. American Political Science Review, 86: 390–404. Lupia, A. (1994). Shortcuts versus Encyclopedias: Information and voting behavior in California insurance reform elections. American Political Science Review, 88: 63–76. Lupia, A. and McCubbins, M. D. (1998). The democratic dilemma: Can citizens learn what they need to know? Review of Policy Research, 20 (2): 343–344. Luskin, R. C. (1987). Measuring political sophistication. American Journal of Political Science, 31: 856–899. McAllister, I. and Studlar, D. (1991). Bandwagon, underdog, or projection? Opinion polls and electoral choice in Britain, 1979–1987. Journal of Politics, 53: 720–741. McClure, R. D. and Patterson, T. E. (1976). Setting the political agenda: Print vs. network news. Journal of Communication, 26 (2): 23–28. McCombs, M. E. (1977). Newspaper versus television: Mass communication effects across time, pp. 89–105. In M. E. McCombs and D. L. Shaw (eds), The Emergence of American Political Issues. The Agenda-Setting Function of the Press. St. Paul, MN: West Publishing Co.

References 211 McCombs, M. E. (1981). The agenda-setting approach, pp. 121–140. In D. Nimmo and K. R. Sanders (eds), Handbook of Political Communication. Beverly Hills, CA: Sage. McCombs, M. E. (2004). Setting the Agenda. The Mass Media and Public Opinion. Cambridge: Polity Press. McCombs, M. E. and Shaw, D. L. (1972). The agenda-setting function of mass media. Public Opinion Quarterly, 36: 176–187. McCombs, M. E. and Zhu, J.-H. (1995). Capacity, diversity, and volatility of the public agenda. Trends from 1954 to 1994. Public Opinion Quarterly, 59: 495–525. McCombs, M. E., Danielan, L., and Wanta, W. (1995). Issues in the news and the public agenda: The agenda-setting function, pp. 281–300. In T. Glasser and C. Salmon (eds), Public Opinion and the Communication of Consent. New York: Guilford. McCombs, M. E., Lopez-Escobar, E., and Llamas, J. P. (2000). Setting the agenda of attributes in the 1996 Spanish general election. Journal of Communication, 50: 77–92. McLeod, D., Kosicki, G. M., and McLeod, J. (2002). Resurveying the boundaries of political communication effects, pp. 215–269. In J. Bryant and D. Zillmann (eds), Media Effects. Advances in Theory and Practice. Hillsdale, NJ: Erlbaum. McLeod, J., Becker, L., and Byrnes, J. (1974). Another look at the agenda-setting function of the press. Communication Research, 1: 131–166. McManus, J. H. (1994). Market-driven Journalism. Let the Citizen Beware. Thousand Oaks, CA: Sage. McQuail, D. (2000). Mass Communication Theory. An Introduction. Fourth edition. London: Sage. Magleby, D. B. (1989). Opinion formation and opinion change in ballot proposition campaigns. In M. Margolis and Gary A. Mauser (eds), Manipulating Public Opinion. Belmont: Brooks/Cole Publishing Company. Marcussen, M. and Zoelner, M. (2003). Monetarism and the masses: Denmark and economic integration in Europe. Cooperation and Conﬂict, 38: 101–123. Markus, G. (1979). Analyzing Panel Data. Sage University Paper Series on Quantitative Applications in the Social Sciences. Beverly Hills, CA: Sage. Marsh, C. (1985). Back on the bandwagon: The effect of opinion polls on public opinion. British Journal of Political Science, 15: 50–74. Mazzoleni, G. (1995). Towards a videocracy? Italian political communication at a turning point. European Journal of Communication, 10: 291–319. Mazzoleni, G. (2003). Medienpluralismus in Italien zwischen Politik und Marktwettbewerb. Media Perspektiven, 11: 517–529. Mazzoleni, G. and Campus, D. (2003). With the media, without the media. Reasons and implications of the electoral success of Silvio Berlusconi in 2001. Paper presented at the Annual Meetings of the International Communication Association, San Diego, May. Media Research (2000). Weekly Ratings. Copenhagen: Danish Broadcasting Association. Mendelsohn, M. (1996). The media and interpersonal communications: The priming of issues, leaders and party identiﬁcation. Journal of Politics, 58, 112–125. Mendelsohn, M. and Parkin, A. (eds) (2001). Referendum Democracy: Citizens, Elites, and Deliberation in Referendum Campaigns. New York: Palgrave.

212

References

Miller, A. H. (1974). Political issues and trust in government: 1946–1970. American Political Science Review, 68: 951–972. Miller, A. H., Goldenberg, E. N., and Erbing, L. (1979). Type-set politics: Impact of newspapers on public conﬁdence. The American Political Science Review, 73 (1): 67–84. Miller, J. M. and Krosnick, J. A. (2000). News media impact on the ingredients of presidential evaluations: Politically knowledgeable citizens are guided by a trusted source. American Journal of Political Science, 44: 301–315. Miller, W. L. (1990). How Voters Change: The 1987 British Election Campaign in Perspective. Oxford: Oxford University Press. Mishler, W. and Rose, R. (2001). What are the origins of political trust? Testing institutional and cultural theories in post-communist societies. Comparative Political Studies, 34 (1): 30–62. Morgan, D. (1998). Focus Group as Qualitative Research. Beverly Hills, CA: Sage. Moy, P. and Pfau, M. (2000). With Malice Toward All? The Media and Public Conﬁdence in Democratic Institutions. Westport, CT: Praeger. Neijens, P., Minkman, M., and Slot, J. (1998). Opinion formation in referendum campaigns. Acta Politica, International Journal of Political Science, 33: 300–316. Neuman, W. R., Just, M. R., and Crigler, A. N. (1992). Common Knowledge. News and the Construction of Political Meaning. Chicago: The University of Chicago Press. Neustadt, R. E. (1997). The politics of mistrust, pp. 197–202. In J. Nye Jr., P. D. Zelikow, and D. C. King (eds), Why People Don’t Trust Government. Cambridge: Harvard University Press. Newton, K. (1999). Mass media effects: Mobilisation or media malaise? British Journal of Political Science, 29, 577–600. Nie, N. H., Petrocik, J. R., and Verba, S. (1979). The Changing American Voter. Cambridge: Harvard University Press. Niemi, R. G., Craig, S. C., and Mattei, F. (1991). Measuring internal political efﬁcacy in the 1988 National Election Study. American Political Science Review, 85: 1407–1413 Noël-Aranda, M.-C. (1983). Projecting a European election: The challenge faced by the communicators. In J. G. Blumler (ed.), Communicating to Voters. Television in the First European Parliamentary Elections. London: Sage, pp. 83–100 Norris, P. (2000). A Virtuous Circle. Political Communications in Postindustrial Societies. Cambridge: Cambridge University Press. Norris, P., Curtice, J., Sanders, D., Scammell, M., and Semetko, H. A. (1999). On Message: Communicating the Campaign. London: Sage. Page, B. I. and Shapiro, R. Y. (1992). The Rational Public: Fifty Years of Trends in Americans’ Policy Preferences. Chicago: University of Chicago Press. Palmgreen, P. and Clarke, P. (1977). Agenda-setting with local and national issues. Communication Research, 4: 435–452. Pan, Z. and Kosicki, G. (1993). Framing analysis: An approach to news discourse. Political Communication, 10: 59–79. Pan, Z. and Kosicki, G. M. (1997). Priming and media impact on the evaluations of the President’s performance. Communication Research, 24: 3–30. Patterson, T. (1993). Out of Order. New York: Alfred A. Knopf Publishers. Patterson, T. (2002). The Vanishing Voter: Public Involvement in an Age of Uncertainty. New York: Alfred A. Knopf Publishers.

References 213 Perloff, R. M. (1993). Third person effect research, 1983–1992: A review and synthesis, International Journal of Public Opinion Research, 5: 167–184. Pesonen, P. (1998). Voting decisions, pp. 127–148. In A. T. Jenssen, P. Pesonen and M. Gilljam (eds), To Join or Not to Join: Three Nordic Referendums on Membership in the European Union. Oslo: Scandinavian University Press. Peter, J. (2002). Medien-Priming – Grundlagen, Befunde und Forschungstendenzen [Media priming: foundations, results and research tendencies]. Publizistik, 47: 21–44. Peter, J. and Lauf, E. (2002). The impact of coder characteristics on cross-national assessment of reliability. Journalism and Mass Communication Quarterly, 815–832. Peter, J. and de Vreese, C. H. (2003). In search of Europe: The European Union in national television news. Paper presented at the Annual Meetings of the International Communication Association, ICA, May, San Diego, CA. Peter, J., Semetko, H. A., and de Vreese, C. H. (2003). EU politics on television news: a cross-national comparative study. European Union Politics, 4: 305–327. Peter, J., de Vreese, C. H., and Lauf, E. (2003). Healthy disagreement. A crossnational comparative study of the impact of conﬂict in TV news on satisfaction with EU democracy and engagement in EU politics. Paper presented at the Annual Meetings of The World Association for Public Opinion Research (WAPOR), Prague, Czech Republic. Petersen, V. and Siune, K. (1997). Denmark, pp 35–47. In B. S. Oestergaard (ed.), The Media in Western Europe. The Euromedia Handbook. London: Sage. Pierce, R., Valen, H., and Listhaug, O. (1983). Referendum voting behavior: The Norwegian and British referenda on membership in the European Community. American Journal of Political Science, 27 (1): 43–63. Pinkleton, B. E. and Austin, E. W. (2001). Individual motivations, perceived media importance and political disaffection. Political Communication, 18: 321–334. van Praag, P. (1995). Hoe “Amerikaans” is de Nederlandse verkiezingscampagne? [How “American” is the Dutch campaign?], pp. 228–245. In K. Brants and P. van Praag (eds), Verkoop van de politiek. De verkiezingscampagne van 1994 [Selling politics. The 1994 election campaign]. Amsterdam: Het Spinhuis. van Praag, P. (2000). De professionalisering van campagnes [The professionalization of campaigns], pp. 16–35. In P. van Praag and K. Brants (eds), Tussen beeld en inhoud. Politiek en de media in de verkiezingen van 1998. [Between image and content. Politics and the media in the 1998 elections]. Amsterdam: Het Spinhuis. van Praag, P. and van der Eijk, C. (1998). News contents and effects in a historic campaign. Political Communication, 15: 165–183. Price, V. and Neijens, P. (1997). Opinion quality in public opinion research. International Journal of Public Opinion Research, 9 (4): 336–360. Price, V., Traugott, M., Lavrakas, P., and Presser, S. (1999). The perceived impact of election polls. Paper presented to the Annual Meetings of the International Communication Association, San Francisco, CA, June. Qvortrup, M. H. (2001). How to lose a referendum: The Danish plebiscite on the euro. Political Quarterly, 72 (2): 190–196. Ray, L. (2003). Reconsidering the link between incumbent support and Pro-EU opinion. European Union Politics, 4: 259–279. Reuters (2000). Polls in the Danish Euro Referendum Campaign, August 28, 2000.

214

References

Rhee, J. W. (1997). Strategy and issue frames in election campaign coverage: A social cognitive account of framing effects. Journal of Communication, 47 (3): 26–48. Roessler, P. (1999). The individual agenda-designing process. Communication Research, 26 (6): 666–700. Rogers, E. M. and Dearing, J. W. (1987). Agenda-setting research: Where has it been, where is it going? pp. 555–594. In J. Anderson (ed.), Communication Yearbook 11. London: Sage. Saris, W. E. (1997). Public opinion about the EU can easily be swayed in different directions. Acta Politica, 32: 406–432. Sartori, G. (1987). The Theory of Democracy Revisited. New Jersey: Chatham. Scammell, M. (1995). Designer Politics. London: Macmillan. Scarrow, S. (2001). Direct democracy and institutional design: A comparative investigation. Comparative Political Studies, 30: 450–483. Scheufele, D. (1999). Framing as a theory of media effects. Journal of Communication, 49 (10): 103–122. Schneider, G. and Weitsman, P. A. (1996). The punishment trap: Integration referendums as popularity contests. Comparative Political Studies, 28: 582–607. Schönbach, K. and Semetko, H. A. (1992). Agenda-setting, agenda-reinforcing or agenda-deﬂating? A study of the 1990 German national election. Journalism Quarterly, 69: 837–846. Schönbach, K., de Ridder, J., and Lauf, E. (2001). Politicians on TV: Getting attention in Dutch and German election campaigns. European Journal of Political Research, 39 (4): 519–531. Semetko, H. A. (1996a). Political balance on television: Campaigns in the U.S., Britain and Germany. Harvard International Journal of Press/Politics, 1: 51–71. Semetko, H. A. (1996b). The media, pp. 160–183. In L. LeDuc, R. G. Niemi, and P. Norris (eds), Comparing Democracies. Elections and Voting in Global Perspective. Thousand Oaks, CA: Sage. Semetko, H. A. (1998). Introduction: Poulders, pillars, and understanding Dutch democracy. Political Communication, 14 (2): 139–148. Semetko, H. A. (2003). Political bias in the media. In D. H. Johnston (ed.), Encyclopedia of International Media and Communications. New York: Academic Press. Semetko, H. A. and Schönbach, K. (1994). Germany’s “Unity Election”: Voters and the Media. Cresskill, NJ: Hampton Press. Semetko, H. A. and Schönbach, K. (2003). News and elections: German Bundestag campaigns in the Bild 1990–2002. Harvard International Journal of Press/Politics, 8 (3): 54–69. Semetko, H. A. and Valkenburg, P. M. (2000). Framing European politics: A content analysis of press and television news. Journal of Communication, 50 (2): 93–109. Semetko, H. A., Blumler, J. G., Gurevitch, M., and Weaver, D. H. (1991). The Formation of Campaign Agendas: A Comparative Analysis of Party and Media Role in Recent American and British Elections. Hillsdale, NJ: Lawrence Erlbaum. Semetko, H. A., van der Brug, W., and Valkenburg, P. (2003a). The inﬂuence of political events on attitudes towards the EU. British Journal of Political Science, 33: 621–634.

References 215 Semetko, H. A., van der Brug, W., and Valkenburg, P. (2003b). Media priming in a multi-party context: Politics in the news and evaluations of political leaders. Manuscript under review. Semetko, H. A., de Vreese, C. H., and Peter, J. (2000). Europeanised politics – Europeanised media? European integration and political communication. West European Politics, 23 (4): 121–142. Shah, D. V., Watts, M. D., Domke, D., and Fan, D. (2002). News framing and cueing of issue regimes. Explaining Clinton’s public approval in spite of scandal. Public Opinion Quarterly, 66: 339–370. Shoemaker, P. J. and Reese, S. D. (1996). Mediating the Message. Theories of Inﬂuences on Mass Media Content. Second edition. White Plains: Longman. Simons, J. (1998). Zwevende kiezers en zappende kijkers; politieke beeldvorming op televisie [Floating voters and zapping viewers: political images on television]. Amsterdam: Amsterdam University Press. Sinnott, R. (1995). Irish Voters Decide: Voting Behavior in Elections and Referendums since 1918. Manchester: Manchester University Press. Sinnott, R. (2002). Cleavages, parties and referendums: Relationships between representative and direct democracy in the Republic of Ireland. European Journal of Political Research, 41: 811–826. Siune, K. (1987). The political role of mass media in Scandinavia. Legislative Studies Quarterly, 12: 395–414. Siune, K. (1993). The Danes said No to the Maastricht Treaty. The Danish EC Referendum of June 1992. Scandinavian Political Studies, 16 (1): 93–103. Siune, K. and Borre, O. (1975). Setting the agenda for a Danish election. Journal of Communication, 1: 65–73 Siune, K. and Svensson, P. (1993). The Danes and the Maastricht Treaty: The Danish EC referendum of June 1992. Electoral Studies, 12 (2): 99–111. Siune, K., Svensson, P., and Tonsgaard, O. (1994). The European Union: The Danes said No in 1992 but Yes in 1993: How and why. Electoral Studies, 13 (2): 107–116. Smith, T. J. and Verrall, D. O. (1985). A critical analysis of Australian television coverage of election opinion polls. Public Opinion Quarterly, 49: 58–79. Sniderman, P. M. and Theriault, S. M. (2002). The dynamics of political argument and the logic of issue framing. In W. E. Saris and P. Sniderman (eds), The Issue of Belief: Essays in the Intersection of Non-Attitudes and Attitude Change. Princeton: Princeton University Press. Stevens, J. (1996). Applied Multivariate Statistics for the Social Sciences. Third edition. Mahwah, NJ: Lawrence Erlbaum. Sudman, S. (1986). Do exit polls inﬂuence voting behavior? Public Opinion Quarterly, 50: 331–339. Svensson, P. (2002). Five Danish referendums on the European community and European Union: A critical assessment of the Franklin thesis. European Journal of Political Research, 41: 733–750. Swanson, D. (1988). Feeling the elephant: Some observations on agenda-setting research, pp. 603–619. In J. Anderson (ed.), Communication Yearbook 11. London: Sage. Tipton, L., Haney, R. D., and Baseheart, J. R. (1975). Media agenda-setting in city and state election campaigns. Public Opinion Quarterly, 52: 15–22. Traugott, M. W. (1992). The impact of media polls in the public, pp. 125–149. In T.

216

References

E. Mann and G. R. Oren (eds), Media Polls in American Politics. Washington, DC: The Brookings Institution. Traugott, M. W. and Lavrakas, P. (1996). The Voter’s Guide to Election Polls. Chatham, NJ: Chatham House. Trenaman, J. and McQuail, D. (1961). Television and the Political Image. A Study of the Impact of Television on the 1959 General Election. London: Methuen. Valentino, N. A., Beckmann, M. N., and Buhr, T. A. (2001a). A spiral for of cynicism for some: The contingent effects of campaign news frames on participation and conﬁdence in government. Political Communication, 18: 347–367. Valentino, N. A., Buhr, T. A., and Beckmann, M. N. (2001b). When the frame is the game: Revisiting the impact of “strategic” campaign coverage on citizen’s information retention. Journalism and Mass Communication Quarterly, 78: 93–112. Valkenburg, P. M., Semetko, H. A., and de Vreese, C. H. (1999). The effects of news frames on readers’ thoughts and recall. Communication Research, 26: 550–568. de Vreese, C. H. (2001a). Election coverage – new directions for public broadcasting: The Netherlands and beyond. European Journal of Communication, 16 (2): 155–179. de Vreese, C. H. (2001b). Europe in the news: A cross-national comparative study of the news coverage of key EU events. European Union Politics, 2: 283–307. de Vreese, C. H. (2001c). Le elezioni del parlamento Europeo 1999. Dinamarca: Un’ elezione al margine dell’Europa [The 1999 European parliamentary elections. Denmark: Voting at the rim of Europe], pp. 223–258. In R. Marini (ed.), L’Europa dell’euro e della Guerra [A Europe of euro and war]. Rome: RAI. de Vreese, C. H. (2002). Framing Europe: Television News and European Integration. Amsterdam: Aksant Academic Publishers. de Vreese, C. H. (2003). Television reporting of second-order elections. Journalism Studies, 4 (2): 183–198. de Vreese, C. H. (2004a). The effects of frames in political television news on audience perceptions of routine political news. Manuscript accepted for publication in Journalism and Mass Communication Quarterly (Summer issue). de Vreese, C. H. (2004b). The effects of strategic news on political cynicism, issue evaluations and policy support: A two-wave experiment. Mass Communication and Society, 7 (2): 191–215. de Vreese, C. H. and Semetko, H. A. (2002a). Cynical and engaged: strategic campaign coverage, public opinion and mobilization in a referendum. Communication Research, 29 (6): 615–641. de Vreese, C. H. and Semetko, H. A. (2002b). Public perception of polls and support for restrictions on publication of polls: Denmark’s 2000 euro referendum. International Journal of Public Opinion Research, 14: 410–433. de Vreese, C. H., Lauf, E., and Peter, J. (2004). The media and European parliament elections. Second-rate coverage of a second-order event? In C. van der Eijk and W. van der Brug (eds), European Elections and Domestic Politics. Lessons from the Past and Scenarios for the Future. Book manuscript under review. de Vreese, C. H., Lindegaard, L., and Lauritzen S. V. (2000). TV-Avisens daekning af euro-folkeafstemningen. Resultater fra fokusgruppetest [TV-Avisen’s coverage of the euro referendum. Results from a focus group project]. Research Report. Copenhagen: Danish Broadcasting Association.

References 217 de Vreese, C. H., Peter, J., and Semetko, H. A. (2001). Framing politics at the launch of the euro: A cross-national comparative study of frames in the news. Political Communication, 18 (2): 107–122. Waller, R. (1995). The polls and the 1992 general election, pp. 177–191. In I. Crewe and B. Gosschalk (eds), Political Communications: The General Election Campaign of 1992. Cambridge: Cambridge University Press. Wattenberg, M. P. (1992). The Rise of Candidate-Centered Politics. Cambridge, MA: Harvard University Press. Wattenberg, M. P. and Brians, C. L. (1999). Negative campaign advertising: Demobilizer or mobilizer? American Political Science Review, 93: 891–900. Weaver, D. (1977). Political issues and voter need for orientation, pp. 107–120. In D. R. Shaw and M. E. McCombs (eds), The Emergence of American Political Issues: The Agenda-Setting Function of the Press. St. Paul, MN: West. Weaver, D. (1980). Audience need for orientation and media effects. Communication Research, 7 (4): 361–376. Weaver, D., Graber, D. A., McCombs, M. E., and Eyal, C. H. (1981). Media Agenda-Setting in a Presidential Election. New York: Praeger. Winter, J. P. and Eyal, C. H. (1981). Agenda-setting for the civil rights issue. Public Opinion Quarterly, 45: 376–383. Worcester, R. (1995). The use of panel studies in British general elections, pp. 192–212. In I. Crewe and B. Gosschalk (eds), Political Communications: The General Election Campaign of 1992. Cambridge: Cambridge University Press. Worre, T. (1989). Det politiske system i Danmark [The political system in Denmark]. Copenhagen: Akademisk Forlag. Wybrow, R. J. (1995). The impact of the 1992 general election on the image of public opinion surveys, pp. 247–256. In I. Crewe and Brian Gosschalk (eds), Political Communications: The General Election Campaign of 1992. Cambridge: Cambridge University Press. Yagade, A. and Dozier, D. M. (1990). The media agenda-setting effect of concrete versus abstract issues. Journalism Quarterly, 67 (1): 3–10. Zaller, J. (1992). The Nature and Origin of Mass Opinion. Cambridge: Cambridge University Press. Zaller, J. (1998). Monica Lewinsky’s contribution to political science. PS: Politics Science and Politics, 31 (2): 182–189. Zaller, J. (2002). The statistical power of election studies to detect media exposure effects in political campaigns. Electoral Studies, 21: 297–329. Zaller, J. and Feldman, S. (1992). A simple theory of the survey response: Answering questions or revealing preferences? American Journal of Political Science, 36: 579–616. Zhao, X. and Chaffee, S. H. (1995). Campaign advertisements versus television news as sources of political issue information. Public Opinion Quarterly, 59 (1): 41–65. Zucker, H. G. (1978). The variable nature of news media inﬂuence. In B. D. Ruben (ed.), Communication Yearbook. New Brunswick: Transaction Books.

Author index

Aardal, B. 200 Aarts, K. 174 Achen, C. H. 7 Acock, A. 122 Adams, W. J. 41 Agger, R. E. 123 Aiken, L. S. 200 Andersen, J. 34, 40, 117 Andersen, J. G. 34, 40, 117, 155, 159 Andersen, O. E. 32 Andersen, V. N. 130 Anderson, R. 81, 131 Ansolabehere, S. 121, 122, 130, 174 Appleton, A. 4 Asp, K. 70 Atkin, C. 132 Austin, E. W. 123, 129, 174, 199 Ban, H. 71 Bantimaroudis, P. 71 Barber, J. D. 132 Bargh, J. A. 139 Bartels, L. M. 14, 41, 158 Baseheart, J. R. 118 Becker, L. 114 Beckmann, M. N. 100, 122, 123, 124, 130, 131 Belli, R. F. 128 Bennett, W. L. 9, 14, 137 Benton, M. 108, 118 Berelson, B. 38, 121 Bille, L. 53 Blais, A. 136, 139, 157, 158, 159, 165 Blumler, J. G. 12, 66, 67, 68, 70, 71, 85, 91, 123 Borg, S. 9, 160 Borre, O. 34, 40, 70, 117 Bowler, S. 3, 17, 153 Brannon, L. A. 17, 136, 138, 140, 147

Brettschneider, F. 81, 131, 135 Brians, C. L. 122 Brug, W. van der 17, 136, 137, 176 Buch, R. 139 Budge, I. 18, 154 Buhr, T. A. 100, 122, 123, 124, 130, 131 Butler, D. 3, 17, 47, 132, 150, 153 Butzler, M. 8 Byrnes, L. 114 Campbell, A. 18 Campus, D. 48 Cappella, J. N. 16, 94, 95, 99, 120, 122, 123, 124, 129, 131, 174, 199 Ceci, S. K. 132 Chaffee, S. H. 123 Chan, K. 108 Cherribi, O. 48 Citrin, J. 122 Clarke, H. D. 3, 9, 14, 18, 122, 157, 159, 165, 200 Clarke, P. 118 Cohen, J. E. 131 Converse, P. E. 6, 7, 18, 138, 154 Cook, S. W. 132 Craig, S. C. 122 Crigler, A. N. 95 Curtice, J. 47, 70, 71, 107, 117, 121, 132, 201 D’arcy, R. 57 Dalton, R. J. 7, 9, 14 Danielan, L. 109, 117, 199 Davis, D. 161 Dearing, J. W. 107, 199 Degenhardt, W. 110 Diez-Nicolas, J. 48 Dominick, J. 40 Domke, D. 94, 131

Author index 219 Donovan, T. 3, 17, 153 Dozier, D. M. 119 Edwards, G. C. 136 Eijk, C. van der 18, 23, 24, 48, 137, 150, 154, 156 Eilders, C. 69 Eklit, J. 53 Entman, R. B. 9, 14, 92, 137 Epstein, K. 132 Erber, R. 122, 124, 130 Erbring, L. 110, 120, 121 Esaiasson, P. 9, 160 Eyal, C. H. 15, 107, 109, 114, 118 Faber, R. J. 94, 136 Fan, D. P. 94, 131, 136, 166 Farlie, D. J. 18, 154 Farrell, D. M. 6, 45, 50, 55 Feldman, S. 7 Fico, F. G. 198 Finkel, S. E. 122 Fiske, S. T. 16 Franklin, M. 18, 23, 24, 137, 150, 154, 156, 159, 201 Frankovic, K. A. 131, 132 Frazier, P. J. 108, 118 Freedman, P. 41, 105, 122 Fuchs, D. 9 Galtung, J. 69 Geer, J. G. 122 Gidengil, E. 136, 139, 157, 158, 159, 165 Gilljam, M. 4, 14, 65, 139, 158, 159 Glynn, C. J. 132 Goldenberg, E. N. 110, 120, 121 Goldstein, K. 105, 122 Goldstein, M. N. 41, 123 Graber, D. 10, 14, 15, 107, 114, 118 Granberg, D. 40, 129 Gross, A. 37 Gunther, R. 48 Gurevitch, M. 12, 67, 68, 70, 71, 85, 123 Hamman, C. 131 Haney, R. D. 118 Hansen, K. M. 130, 139 Hardmeier, S. 82 Harrop, M. 50 Herr, J. P. 17, 71 Hertog, J. K. 93 Hetherington, M. J. 122, 201 Hibbing, J. R. 121

Higgins, E. T. 139 Hix, S. 10, 179 Hjarvard, S. 32, 70 Holley, J. 131, 132 Holmberg, S. 40, 129 Holtz-Bacha, C. 70, 105, 174 Hooghe, L. 23, 36, 165 Hovland, C. 42 Huegel, R. 110 Hug, S. 3, 50, 150 Iyengar, S. 94, 95, 121, 122, 130, 136, 137, 138, 174 Jacobsen, M. V. 53 Jakobsen, M. S. 155, 201 Jamieson, K. H. 16, 94, 95, 99, 120, 122, 123, 124, 129, 131, 174, 199 Jasperson, A. E. 94, 136 Jensen, K. B. 30, 31, 32 Jenssen, A. T. 4, 14, 65, 139, 156, 158, 159, 160, 200, 201 Johnston, R. 3, 139, 157, 158, 159, 165 Just, M. R. 95 Kagay, M. R. 82 Kahn, K. F. 123, 130 Kahneman, D. 93 Kaid, L. L. 105 Kain, E. L. 132 Kang, M.-E. 135 Katz, E. 161 Kaufmann, B. 37 Kavanagh, D. 47 Kennedy, P. J. 123, 130 Kepplinger, H.-M. 17, 70, 71 Kerbel, M. R. 80 Kinder, D. R. 16, 136, 137, 138 King, A. 4 Kiousis, S. 71 Kleinnijenhuis, J. 70, 166 Klingeman, H.-D. 9 Kornberg, A. 3, 9, 14, 18, 157, 159, 165, 200 Kosicki, G. M. 14, 107, 109, 110, 137, 200 Kraus, S. 161 Krosnick, J. A. 16, 17, 136, 138, 140, 147 Lacy, S. 198 Lang, G. 16, 69, 108 Lang, K. 16, 69, 108 Larson, S. G. 80

220

Author index

Lau, R. R. 122, 124, 130 Lauf, E. 36, 39, 66, 71, 91, 174 Lauritzen, S. V. 198 Laver, M. 57 Lavrakas, P. 80, 131, 132, 134 Lazarfeld, P. E. 121 LeDuc, L. 3, 8, 11, 50, 147, 153, 154, 159, 164 Leroy, P. 66 Lewis-Beck, M. S. 18, 154 Lijphart, A. 23, 24 Lindegaard, L. 198 Lippmann, W. 15, 107 Lipset, S. M. 23 Listhaug, O. 154, 155 Llamas, J. P. 17, 71, 118, 147 Lodge, M. 138 Lombardi, W. J. 139 Lopez-Escobar, E. 17, 71, 118, 147 Lord, C. 179 Lowery, D. 158, 164 Lupia, A. 3, 8 Luskin, R. C. 138 McAllister, I. 134 McClure, R. D. 118 McCombs, M. E. 15, 17, 71, 107, 108, 109, 114, 117, 118, 147, 199 McCubbins, M. D. 3 McGonagle, K. A. 128 McGraw, K. M. 138 Mackie, T. T. 23, 159 McLeod, D. M. 14, 93, 200 McLeod, J. 14, 132, 200 McManus, J. H. 69 McPhee, W. N. 121 McQuail, D. 15, 29, 109, 110 Magleby, D. B. 56, 58, 60, 119 Marcussen, M. 166 Marks, G. 23, 36, 165 Markus, G. 114, 127, 163, 164 Marquis, L. 8 Marsh, C. 132 Marsh, M. 18, 150, 156, 201 Mattei, F. 122 Maurer, M. 70 Mazzoleni, G. 48 Mendelsohn, M. 50, 136 Miller, A. H. 110, 120, 121, 122 Miller, J. M. 138 Miller, W. E. 18, 50, 154 Miller, W. L. 107, 117 Minkman, M. 139, 158

Mishler, W. 122, 123 Montere, J. R. 48 Morgan, D. 40, 41 Moy, P. 174 Nadeau, R. 136 Narayan, S. S. 136 Neijens, P. 139, 158 Neuman, W. R. 95 Neustadt, R. E. 121 Nevitte, N. 136, 139, 157, 158, 159, 165 Newton, K. 174 Nie, N. H. 18, 154 Nielsen, H. J. 34, 40, 117 Niemi, R. G. 122 Noël-Aranda, M.-C. 68, 69 Norris, P. 45, 47, 55, 70, 71, 107, 117, 121, 174 Oegema, D. 70 Oscarsson, H. 200 Page, B. I. 7, 114 Palmgreen, P. 118 Pammett, J. H. 159 Pan, Z. 137 Parkin, A. 50 Patterson, T. E. 16, 80, 82, 94, 118, 120, 174 Payne, C. 132 Pearl, S. A. 123 Perloff, R. M. 134 Pesonen, P. 4, 14, 65, 139, 158, 159, 160 Peter, J. 29, 39, 65, 66, 67, 71, 74, 91, 95, 99, 174, 200 Peters, M. D. 138 Petersen, V. 30, 32 Petrocik, J. R. 18, 154 Pfau, M. 174 Pierce, R. 154, 155 Pinkleton, B. E. 123, 129, 174, 199 Praag, P. van 48, 49, 51 Presser, S. 134 Price, V. 134, 158 Qvortrup, M. H. 166 Ranney, A. 3, 17, 150, 153 Ray, L. 183 Reese, S. D. 69 Reif, K.-H. 18 Reinert, R. 155, 201 Reno, R. R. 200

Author index 221 Rhee, J. W. 120 Ridder, J. A. de 71, 166 Riffe, D. 198 Roessler, P. 110 Rogers, E. M. 107, 199 Rose, R. 122, 123 Ruge, M. H. 69 Sanders, D. 47, 70, 71, 107, 117, 121 Sankiaho, R. 200 Saris, W. E. 179 Sartori, G. 180 Saynassalo, E. 200 Scammell, M. 47, 70, 71, 107, 117, 121 Scarrow, S. 3 Scheufele, D. 92 Schmitt, H. 18 Schmitt-Beck, R. 6, 50 Schneider, G. 9 Schoenbach, K. 48, 71, 107, 117, 118 Sciarini, P. 3, 150 Semetko, H. A. 13, 17, 29, 47, 48, 65, 66, 67, 68, 70, 71, 74, 83, 89, 95, 99, 102, 107, 117, 118, 121, 136, 137, 173, 174, 176, 201 Shah, D. V. 94, 131, 136 Shapiro, R. Y. 7, 114 Shaw, D. L. 15, 107, 109, 117 Shoemaker, P. J. 69 Sigelman, L. 158, 164 Silver, G. E. 122 Simon, A. F. 121, 122, 130, 136 Simons, J. 48 Sinnott, R. 4, 154 Siune, K. 4, 9, 18, 30, 32, 65, 66, 70, 130, 132, 139, 147, 157, 158, 165 Slot, J. 139, 158 Smith, T. J. 80, 81, 82, 131 Smith, W. R. 136 Sniderman, P. M. 93 Stewart, M. C. 122 Stevens, J. 200 Stroh, P. 138 Strom, G. 132 Studlar, D. 134 Sudman, S. 132 Svensson, P. 4, 9, 18, 65, 114, 139, 147, 157, 158, 165, 179 Swanson, D. 108, 111, 118 Swenson, T. 136 Taylor, S. E. 16 Theiss-Morse, E. 121

Theriault, S. M. 93 Thomsen, S. R. 155, 201 Tipton, L. 118 Tolbert, C. J. 3, 17, 153 Tonsgaard, O. 18, 65, 139, 157, 158, 165 Traugott, M. W. 80, 128, 131, 132, 133, 134, 135 Trenaman, J. 15 Tversky, A. 93 Valen, H. 23, 154, 155, 159 Valentino, N. A. 100, 121, 122, 123, 124, 130, 131 Valkenburg, P. M. 17, 65, 67, 95, 102, 136, 137, 176 Vreese, C. H. de 12, 24, 29, 49, 65, 66, 67, 68, 69, 70, 71, 72, 74, 83, 88, 90, 91, 92, 93, 95, 99, 102, 131, 174, 198 Verba, S. 18, 154 Verrall, D. O. 80, 81, 82, 131 Waller, R. 132 Wanta, W. 109, 117, 199 Wattenberg, M. P. 18, 122, 154 Watts, M. D. 94. 131, 136 Wearing, P. 3, 9, 14, 18, 157, 159, 165, 200 Weaver, D. H. 15, 67, 68, 70, 71, 107, 114, 118 Webb, P. 45, 55 Weiss, H.-J. 110 Weitsman, P. A. 9 Welch, A. C. 132 Wert, J. I. 48 West, S. G. 200 Wilson, C. J. 23, 36, 165 Wimmer, R. 40 Winter, J. P. 109 Wlezien, C. 18, 156, 201 Worcester, R. 132 Worre, T. 24 Wybrow, R. J. 132 Yagade, A. 119 Young, M. 128 Zaller, J. 7, 8, 11, 14, 93, 131, 161, 165, 177 Zhao, X. 123 Zhu, J.-H. 108 Zoelner, M. 166 Zucker, H. G. 117

Subject index

advertising 50–6, 121–3, 130 agenda-building 69–70 agenda-setting 15–16, 20, 69–70, 88, 106, 107–19, 174, 178, 198–9; second-level 118; typology of studies 109–10 Amsterdam Treaty 4, 5, 27, 180 audience frames see framing effects Austria 4, 5, 51, 66, 77, 180 balance 13, 16, 29, 33, 85, 89, 120, 173 BBC 31, 32, 66, 68, 96, 102 Berlingske Tidende 30, 33, 37, 54, 75–6, 78, 80, 98, 101, 192, 196 bias 16, 20, 66, 89, 94, Britain 3, 4, 5, 6, 12, 25, 31, 35, 47, 66, 67, 69, 107, 179 BT 75, 79 campaign: activities 37, 46, 48, 51–4, 163, 189, 195; budget 48–9, 52–4, 60–1, 105; conduct 96–8, 100, 103; effects 9–18, 19, 20, 37, 42, 105–68, 174–8, 197, 200; evaluations 16, 106, 123, 124–8, 129–31, 175, 190, 191, 200; style 19, 48, 61, 80 Canada 3, 14, 160 Clinton, Bill 94, 131 commercial broadcasting 29, 31–2, 144, 163, 174 comparative design 37–42 conﬂict see conﬂict frame content analysis see news content constitution see EU Constitution cynicism see political cynicism Cyprus 5, 157, 187, 189, 194 Czech Republic 5 democratic deﬁcit 25, 180 Denmark 4–6, 14, 24–5, 29, 31, 33–6, 39,

50–1, 54, 57, 59, 66, 69, 72, 76–7, 79, 105, 115, 117, 119, 124–5, 130, 132, 155, 157, 159, 166, 167, 175, 179, 180–2, 186–9, 194–6, 199 DR 31–3, 38, 68, 72, 81, 198, 201 economy 15–16, 56–60, 72–4, 94, 96, 98, 103, 108, 111–13, 115, 118, 136, 154, 165, 167, 172, 179, 181, 201 economic consequences see frames editorial approach 38, 67–70, 85–90 efﬁcacy 18, 40, 121–3, 125, 129, 154, 158, 163–4, 188, 190, 194, 197, 200 EkstraBladet 30, 75–6, 78, 80 Elections: sacerdotal approach to 67–8, 85, 90, 91; pragmatic approach to 67–8, 85, 90 European Union 22, 34, 35, 39, 102, 116 EU (European Council) Summits 17, 66–7, 136–7 EU Constitution 6, 178–9, 181 EU enlargement 4–5, 36, 58, 88, 102, 181 Euro 3, 4, 5, 19, 20, 23, 25, 27, 28, 33–7, 53, 55, 57–60, 63–104, 115–17, 128–9, 140–1, 146–8, 162, 166–7, 176–7, 181–3, 186–9, 191–9, 201 Euro referendum see Euro European Commission 87 European elections 18, 23–4, 26, 34, 50, 66, 68–72, 88, 91, 122, 165, 174, 180 European Parliament 10, 68, 87, 180 evaluation political leaders see actors in EU news fairness in news reporting see balance Finland 4, 5, 14, 40, 65, 66, 156, 158, 159, 180, 200 focus groups 38, 40–1, 42, 46, 51, 60, 133, 134–5, 198

Subject index 223 frame-building 92 frame-setting see framing effects frames: conﬂict 95, 98–9; economic consequences 95, 98–100; episodic/thematic 94; game 94; generic 93, 94–5; human interest/impact 95; issue speciﬁc 93–5; strategy/issue 94–5, 98–104 framing: as a process 92–104, 153, 171, 172–3, 174, 198; in the newsroom 67 framing effects 102, 127–8 France 4, 5, 6, 24, 66, 180 government: popularity of incumbent government 18, 90, 106, 126, 128, 147, 150, 165, 176–7, 183 horse race see frames; game frame Hungary 5 identity 20, 92, 103, 115–17, 166, 179, 185 journalism 29, 46–7, 49, 67–9, 71, 84–90, 94, 96, 174 journalist roles 67–70 JyllandsPosten 31, 33, 54, 75, 80, 192 ideology 4, 9, 18, 21, 24, 123, 125–8, 144–5, 149, 154–5, 162–3, 165, 177, 179, 192–3, 200 information cues 3, 8–9, 14, 17, 136, 176 information ambiguity 11, 171–2 information processing see framing effects; agenda-setting, typology of studies interviews 19, 37–9, 56, 84, 90, 173 Ireland 4, 5, 6, 66, 180 ITN 66, 68, 96 knowledge see political knowledge Latvia 5 Lithuania 5, 187, 189, 194 Maastricht Treaty 4–5, 25, 27, 34, 58, 180–1 Malta 5 media effects see agenda-setting; priming; framing effects media frames see frames media organizations see BBC, DR, ITN, NOS, RTL, TV2

media system 10, 12–13, 19, 21, 23, 29, 171 mobilization 10, 15–16, 83, 106, 121–2, 124, 128–32, 139, 175, 177, 183 Netherlands, the 6, 24, 48, 51, 66, 69, 179–80 news and voting 162 news content: actors in EU news 83–4, 86–7, 89, 141–50; tone of news coverage 71, 84, 141–4, 162–3, 166, 195, 201; visibility EU news 65–80, 89–91, 141–4, 176 news production see editorial approach news selection 47, 67–70, 85, 88, 91 news values 8, 68–9, 94, 171 newspapers: broadsheets 33, 42, 75, 81, 98, 100; tabloids 30, 31, 33, 75, 80, 98, 100–1 Nice Treaty 4, 5, 57, 59, 181, No campaign 45–62, 76, 89, 103, 136, 153, 161, 166, 173 Norway 4, 5, 14, 65, 156, 158, 159, 180, 181, 200 NOS 66, 68, 69, 102 objectivity see balance opinion polls: effect of 131–5; reporting about 71, 80–3 partisanship 4, 9, 29, 123, 154–5 Poland 5, 57, 157, 187, 189, 194 political actors see news, actors in EU news political advertisement see advertising political awareness see political interest political cynicism 12, 14–16, 18, 21, 40, 94, 102, 106, 119, 120–35, 154, 160, 162–4, 175, 177–8, 190–1, 196–7, 199–200 political efﬁcacy see efﬁcacy political institutions 10–12, 121–2, 130, 171–3 political interest 10, 18, 40, 42, 121, 138–9, 144, 154, 158–9, 162–4, 197, 201 political knowledge 10, 14, 18, 45, 130, 138–9, 154, 158, 171, 174, 198, 200 political mobilization see mobilization political parties 3, 10–12, 16, 20, 23–9, 31, 33, 37–8, 43–62, 69–70, 84, 89–90, 92, 96, 98, 150, 154, 171–3, 177, 179, 185–6: professionalization of 45–50, 61

224

Subject index

political sophistication see political knowledge Politiken 30, 33, 75, 76, 78, 80, 98, 101, 192, 196 priming 15–17, 21, 136–52, 175–6, 200 private broadcasting see commercial broadcasting public opinion 6–8, 37, 41–2, 58, 67, 80–1, 99, 102, 124, 131, 147, 150, 165, 179–90 public service broadcasting 29, 33, 48–9, 162, 174 reason for voting Yes and No 153–68 reliability 39, 198, 200 RTL 66, 68 Saatchi & Saatchi 47, 48 salience 15, 20, 65, 75, 92, 107–11, 113–15, 117–19, 156, 174–5, 188, 199 second order elections 18 Slovakia 5 Slovenia 5 sovereignty 10, 20, 34, 57, 58, 59, 65, 92, 97, 98, 103, 115–17, 119, 157, 166, 167, 172, 175, 200 spin-doctors 45, 55 strategic news reporting 16, 99–101, 120–31, 175, 191 survey (panel) 39–42, 106, 113–18, 123, 128, 135, 140, 147, 150, 161, 166, 168, 176–7, 188–9, 198–200

Sweden 4, 5, 14, 54, 65–6, 156, 158–59, 179–80, 182, 200 Switzerland 5, 181 television: television news 13, 20, 32–3, 38–9, 41, 65, 67, 70–3, 81, 83–4, 89–91, 97–101, 108, 111–13, 117–18, 121, 141, 144, 162–5, 173, 177–8, 192, 195, 197–8, 200 TV2 32–3, 38, 68, 72–4, 81–2, 84–6, 88–9, 91, 96–7, 101, 111–13, 116–17, 127, 132, 143, 144, 146, 149, 162, 166, 188, 192, 193, 195 turnout see mobilization United States 3, 16, 19, 47, 55, 105, 123, 129, 130, 131, 137, 158, 175 vote choice 9–10, 14–15, 17, 21, 83, 106, 123, 129, 132, 134, 139, 154, 159–60, 162, 164, 177 voting behavior 17–18, 21, 24, 37, 106, 114, 128, 132–3, 150, 153–4, 160–2, 168 welfare 10–11, 20, 23, 56–9, 73, 76–7, 80, 88, 92, 96–8, 103, 111–13, 115–17, 119, 167, 172, 175 Yes campaign 4, 10, 13, 20, 43–62, 80, 84, 89–90, 92, 98, 103, 136, 144, 146, 150, 153, 161, 166–7, 172–3