List of Tables and Figures. Notes on the Contributors

Transcription

1 Contents List of Tables and Figures Preface Notes on the Contributors vii xii xiv 1 Context, Elites, Media and Public Opinion in Referendums: When Campaigns Really Matter 1 Claes H. de Vreese 2 Opinion Formation and Change in Referendum Campaigns 21 Lawrence LeDuc Part I: The Importance of Context and Political Systems 3 The Strategic Context of Referendums on the EU Constitution 47 Simon Hug 4 Referendum Campaigns: Changing What People Think or Changing What They Think About? 63 Michael Marsh 5 Campaign Information and Voting Behaviour in EU Referendums 84 Sara Binzer Hobolt Part II: Turnout and Political Participation 6 The Participation in Swiss Direct-Democratic Votes 117 Hanspeter Kriesi 7 Turnout in Referendums: the Dutch Experience. An Analysis of Voter and Referendum Characteristics that Influence Turnout in Referendums 142 Peter Neijens, Philip van Praag, Willem Bosveld and Jeroen Slot 8 E-voting and Electoral Participation 159 Alexander H. Trechsel v

2 vi Contents Part III: Understanding the Vote Choice and the Outcome 9 Immigration, Identity, Economy and the Government: Understanding Variation in Explanations for Outcomes of EU-related Referendums 185 Claes H. de Vreese and Hajo G. Boomgaarden 10 The Ideological Response: Saying No to the Euro 206 Henrik Oscarsson 11 The Determinants of Voting Choices on Environmental Issues: A Two-level Analysis 234 Pascal Sciarini, Nicholas Bornstein and Bruno Lanz Index of Names 267 Index of Subjects 271

3 1 Context, Elites, Media and Public Opinion in Referendums: When Campaigns Really Matter Claes H. de Vreese 1.1 Introduction Direct democracy is popular. Across the world referendums and citizen initiatives are an increasingly important means of enacting or preventing legislation. This book argues and demonstrates why campaigns in referendums are important and how they matter for changes in public opinion, political participation, and voters choice to say Yes or No. In a referendum campaign, in contrast to a general election campaign where political parties provide relatively clear-cut information cues for voters, the information cues from political parties are often ambiguous. For example, parties may be internally divided over the referendum issue, political parties from opposite sides of the ideological left right spectrum may form unusual coalitions in referendums, and referendums may also give rise to new parties or movements and thereby reshaping the party system (de Vreese and Semetko, 2004; de Vreese, 2006). In addition, despite the apparent simplistic nature of the referendum vote (Yes/in favour or No/against), the referendum issue is often multifaceted and different aspects may trigger diverse perceptions of the issue among voters. The framing of a referendum issue by political actors and key information sources during a campaign may therefore be unusually important to a referendum outcome. Significant swings in public opinion can occur over the course of a referendum campaign. While longer-term factors such as partisanship or ideology have been found to be important in national elections, the shortterm impact of campaign strategies and tactics can make a substantial difference in determining the outcomes of referendums. Perceptions of the referendum question on the part of the voters, the images they hold of the groups and individuals involved, the salience of political and economic issues, in essence unrelated to the 1

4 2 Context, Elites, Media and Public Opinion in Referendums referendum issue, and voters reactions to the discourse of the campaign can be as important to the voting decision as their opinions and beliefs on the issue itself. In contrast to regular elections, no candidates or parties appear on the ballot thus, voters must decide among alternatives that may be unfamiliar. Given these characteristics one may therefore expect a greater degree of volatility and uncertainty in vote preferences in a referendum campaign compared to a general election campaign. The chapters in this book demonstrate which conditions result in the largest campaign volatility and which citizens are most affected by campaigns vis-à-vis existing predispositions. The book focuses on the key actors in a referendum (the political elites/parties, the media and citizens) and is centred around themes such as campaign effects, electoral mobilization and turnout, as well as vote choice. The book is international and comparative in nature, but rather than being structured in terms of country experiences, the chapters address different themes in a comparative perspective. Relatively few campaign studies have been carried out in the context of (national) referendums, even though different forms of direct democracy are becoming increasingly common (Bowler and Donovan, 1998; Bowler et al., 1998; Butler and Ranney, 1994). The book is designed to start filling a gap in the current literature on referendums. There is a genuine lack of work that specifically addresses the role of campaigns and their impact in the light of the considerations outlined above for why campaigns in a referendum may matter more than campaigns in other electoral contests. 1.2 Referendum: one term, several meanings Before turning to the specific actors engaged in referendum campaigns, the referendum as a phenomenon is addressed. The term referendum is a catch-all term that covers several different forms of democracy. Referendums can be initiated by governments and legislators or by citizens. Referendums can be binding or advisory. Table 1.1 summarizes the different types of referendums. In the interpretation of campaign dynamics, we do not distinguish between the different types of referendums. National referendums are addressed most elaborately, but the actors involved and the dynamics underlying campaigns have more similarities than differences across the different types of referendums. However, one of the key parameters for understanding the dynamics of a referendum campaign is the subject matter. In addition to the variation in type and legal framework of referendums, the subject matter also varies. LeDuc (2003: 33) introduced a

5 Claes H. de Vreese 3 Table 1.1 Referendums: a typology Type Implies Outcome Applied in (for example) Compulsory/ Referendum held Binding Australia, Denmark, binding to change Ireland, Switzerland constitutional law (mandatory constitutional) Rejective/ Referendum on law Usually binding Austria, Denmark, facultative already passed by Italy, Sweden, legislature (abrogative) Switzerland Initiative/ Referendum on a Usually binding Switzerland, New direct legislation subject brought Zealand, United about by petition States of citizens (citizen initiated) Advisory/ Referendum on a Non-binding France, Britain, plebiscite subject initiated by Canada, Finland, the government or Netherlands, Spain legislature (consultative) Source: Adapted from Suski (1993) and LeDuc (2003). useful distinction between referendums on constitutional issues, referendums on treaties and international agreements, referendums on issues of sovereignty, and referendums on a number of specific policies. The first group covers referendums on amendments to national constitutions and changes in political institutions and forms of governance. This group includes the 1992 Canadian constitutional referendum, the 1992/93 referendums in New Zealand on electoral reform, and the 1991 Swiss referendum to lower the voting age. In the second group, referendums have been held on agreements between national and supranational organizations. Denmark has held referendums to ratify the European Single Act (1986), the Maastricht Treaty (1992) (also put to a referendum vote in France (1992)), the Edinburgh Treaty (1993), and the Amsterdam Treaty (1998). Other examples include the 1998 referendum on the Northern Ireland peace agreement and the wave of referendums in 2005 in the European Union on ratification of the EU Constitutional Treaty. In the third group, referendums have been held on issues of territorial boundaries, self-determination, devolution, and federation. Examples include the 1991 Ukrainian independence referendum, the 1997

6 4 Context, Elites, Media and Public Opinion in Referendums referendums to establish national Welsh and Scottish assemblies, and the 1980 and 1995 referendums on sovereignty in Quebec. The final group covers referendums on specific policy issues, including consultative votes on proposals. Examples include the Swedish (1980) and Austrian (1978) referendums on nuclear power, the 1993 Italian referendum on the decriminalization of drugs and the 1983 and 2002 Irish referendums on abortion. This group also includes numerous citizen initiatives in the United States (including states such as California and Oregon) and more than one hundred separate issues that Swiss voters have considered in referendums. In most research on these referendums little attention has been devoted specifically to the role of the campaign, with some recent exceptions that make specific reference to referendums campaigns and campaign effects (Farrell and Schmitt-Beck, 2002; Mendelsohn and Parkin, 2001). In all types of referendums the campaign matters, but on issues of, for example, sovereignty and territorial definition and change, a campaign is more likely to activate and crystallize already existing notions about national borders, identities and sense of belonging than to fundamentally alter how large segments of the electorate think about these issues. For other issues, however, such as internationally focused issues involving treaties, a campaign may significantly affect the definition of what the referendum is about. International treaties are often technical texts, which implies that summarizing and framing the referendum issue is a key component of the campaign (see also LeDuc in this volume). Under such circumstances the campaign might affect voters considerably. In sum, the relative importance of the campaign in referendums on specific policy issues is dependent on the issue. While individual citizens may have relatively clear opinions about abortion or the use of nuclear power, their knowledge and awareness about tax reform, electoral reform, and international treaties may be subject to larger influences during a campaign. In this book we consider different types of referendums on a variety of issues. One issue plays a prominent role in the book and is also of key importance to our understanding of referendum campaign dynamics in general: European integration. 1.3 European integration and referendums European integration has been the focus of more national referendums than any other issue (see Hug (2003) for an introduction): Between 1972 and 2005, there were no less than 35 national referendums in European countries on issues of European integration. Early national referendums

7 Claes H. de Vreese 5 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 referendums in Denmark in 2000 and in Sweden in 2003 (de Vreese and Semetko, 2004; Oscarsson and Holmberg, 2004). In 2001 and 2002, the provisions of the Treaty of Nice were subject to ratification by national referendums in Ireland (Garry et al., 2005; Marsh in this volume), and in 2003 nine of the ten countries that entered the EU by May 2004 held national referendum on the country s position on membership. Malta, Slovenia, Estonia, Slovakia, the Czech Republic, Latvia, Hungary, Lithuania, and Poland all voted in favour of membership. In 2005 a new wave of national referendums took place in Europe. This time the EU Constitutional Treaty was the subject matter. However, after referendums in Spain, France, the Netherlands, and Luxemburg, the ratification process was halted, and the future of the Constitutional Treaty is, at the time of writing (Spring 2007), unknown. Steps in the past and future steps in European integration depend on national referendums and European integration has therefore become a key backdrop for the study of issues of direct democracy. The importance of referendums to the process of European integration is augmented by the so-called democratic deficit of the EU. 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 lack of citizen engagement. Its 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, most notably in the Council. 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 was below 50 per cent in the 2004 elections.

8 6 Context, Elites, Media and Public Opinion in Referendums Table 1.2 National referendums ( ) on issues of European integration Year Country Issue Outcome 1972 Denmark EC membership Yes 1972 Ireland EC membership Yes 1972 Norway EC membership No 1975 Great Britain Continuation of EC membership Yes 1986 Denmark Single European Act Yes 1987 Ireland Single European Act Yes 1992 Denmark Maastricht Treaty No 1992 France Maastricht Treaty Yes 1992 Switzerland EEA Treaty No 1993 Denmark Edinburgh Treaty Yes 1994 Norway EU membership No 1994 Austria EU membership Yes 1994 Finland EU membership Yes 1994 Sweden EU membership Yes 1998 Denmark Amsterdam Treaty Yes 2000 Denmark EMU/ euro No 2001 Ireland Nice Treaty No 2002 Ireland Nice Treaty (amended) Yes 2003 Sweden EMU/ euro No 2003 Lithuania EU membership Yes 2003 Latvia EU membership Yes 2003 Estonia EU membership Yes 2003 Poland EU membership Yes 2003 Czech rep EU membership Yes 2003 Slovakia EU membership Yes 2003 Slovenia EU membership Yes 2003 Cyprus EU membership Yes 2003 Hungary EU membership Yes 2005 Spain EU Constitution Yes 2005 France EU Constitution No 2005 Netherlands EU Constitution No 2005 Luxembourg EU Constitution Yes 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 per cent, except in the cases of the Single Act (1986) and the Amsterdam Treaty (1997) where turnout was 75 per cent. 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 and turnout

9 Claes H. de Vreese 7 was notable: France 71 per cent, Denmark 83 per cent, and Ireland 57 per cent. This was also the case in the French and the Dutch referendums on the EU Constitution with turnout above 60 per cent in both cases. Referendums have the ability to bridge some of the gap between European level governance and European citizens (de Vreese and Semetko, 2004). In advance of the referendums in, for example, France and the Netherlands on the EU Constitution (2005) media coverage and public interest in European affairs increased (Schuck and de Vreese, 2006), without this, however, resulting in an increase in support for the Constitution. 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 (see Farrell and Schmitt-Beck, 2002). In an elaborate study of the 2000 referendum in Denmark on participation in the third phase of the EMU and introduction of the euro, it was demonstrated how the campaign affected the strategies of political elites and influenced public opinion as well as the final vote choice (de Vreese and Semetko, 2004). However, there is an eminent need to identify and understand the actors involved in referendum campaigns and how they affect the dynamics of the campaign under different conditions. 1.4 Context, political elites, media, public opinion and citizens The most important categories of actors involved in a referendum campaign are: (i) the political elites (including parties and candidates); (ii) civil society, interest organizations and lobbyists; (iii) the media and public opinion, and (iv) the electorate. These actors (understood in the broadest sense of the word) interact differently in a referendum campaign than, for example, in a general election campaign and the parameters of this interaction are contingent upon the specific context of the referendum. Figure 1.1 summarizes the reciprocity of the relationship between various actors. In the next section, we review the role of each of these actors in a referendum campaign. Context The specific context and situational factors of a referendum co-determine the dynamics of referendum campaigns. In comparative political science

10 8 Context, Elites, Media and Public Opinion in Referendums Political elites (parties, interest groups) CONTEXT Media (Visibility, framing, tone) Public opinion (volatility) Citizens (engagement, interest, knowledge, turnout, vote choice) Figure 1.1 The dynamics of referendum campaigns and political communication research, there is increasing acknowledgment of the need to integrate contextual factors in our analyses of crossnational differences (see, for example, LeDuc, Niemi and Norris, 1996; van der Eijk and Franklin, 1996). Important contextual factors include the political and electoral system, a country s democratic tradition (including its experience with referendums), the structure of financing of political activities and the media system s regulatory regime. Political systems in most western democracies are either two-party or multi-party systems. Compared to multi-party systems, voters in a twoparty system are more likely to be accustomed to a campaign resembling the referendum campaign because electoral choice is de facto a binary decision for or against a specific candidate and/or party. A second contextual factor is the electoral system, which explicates how votes are cast and seats are allocated. Across the world, multiple systems are in use (see Blaís and Massicotte, 1996), but in a referendum most voters are faced with an alternative situation compared to electoral contests at local, regional, national or supranational level. To citizens in both systems of plurality (first past the post) and proportional representation, a referendum campaign poses a change to well-known electoral choice and this may affect voters view of the dynamics of a referendum and their choice to participate and how to vote. An additionally important contextual factor is a democratic entity s (for example, a country s) previous experience with referendums. In places where direct democracy is well established (for example, Switzerland and California) or where voting in referendums is used with some regularity

11 Claes H. de Vreese 9 (for example, Denmark), voters have a larger pool of references and experiences to rely on when forming an opinion and making a vote choice. In Denmark, for example, general attitudes towards European integration are influential to the vote choice in referendums on European treaties. Some scholars have argued that citizens in such referendums tend to rely on proxies, most often anchored in domestic politics and satisfaction with the incumbent government (for example, Franklin et al., 1995), while others have argued that attitudes towards European integration matter for vote choice (for example, Svensson, 2002). Given the relative consistency of voting behaviour of, for example, the Danish electorate it is reasonable to assume that an electorate achieves a certain knowledge and familiarity with the notion of referendums which affects the dynamics of the campaign. This argument dovetails with studies of the Swiss and Californian experiences where previous referendums play in to the dynamics of current ones. The financing of political activities and the media system, including the regulatory regime, also affect the parameters of a campaign. Countries differ considerably in their restrictions with respect to the financing of political activities and the degree to which, for example, paid political advertisement is allowed (LeDuc et al., 1996). In contexts where restrictions are minimal, ads tend to play a larger role and drive up the levels of expenditure involved in campaigning. In contexts with a rather restrictive regulatory regime campaign expenditure tends to be lower and campaigners have to rely more on unpaid, mediated communication through, for example, the news media. In previous research there is an implicit understanding of the relevance of context. Most studies are single-country studies and often report elaborately about the specificities of a referendum. There is relatively little work on the impact of context given the lack of comparative analyses and adequate data that would enable a better understanding of when and how contextual factors matter. Political elites and civil society Political parties and candidates contesting in a referendum campaign are confronted with the need to raise money for the campaign and to raise the party profile in the media and in the awareness of potential voters. Contrary to domestic politics, where the parties are positioned in ways fairly well known to most voters, a referendum can pose an important challenge for a party. In several referendums on European issues the traditional left right political continuum has little meaning because centre partiers tend to be pro-european integration and (semi-)extremist parties, at both

12 10 Context, Elites, Media and Public Opinion in Referendums the left and right ends of the continuum, tend to be apposed to European integration (de Vreese and Semetko, 2004). In planning a national referendum campaign, political parties are therefore faced with internal strategic choices about who will be visible in the campaign. Externally, political parties are faced with potential strategic alliances within the Yes or No camps. These alliances involve a range of actors including employers organizations, labour unions, religious actors, NGOs, businesses, which makes alliance building rather unpredictable. Moreover, these decisions may be beyond the control of the political parties or the candidates. Political parties often appoint lead campaigners, but in many cases these responsibilities fall upon the prime minister s or another incumbent, who mostly come with their own connotations, be they positive or negative. 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 or other actors may aim strategically to mobilize while others aim simply to demobilize voters. The absence of knowledge about what opposing parties are doing is a continuing source of uncertainty. 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 liberation, 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). An example of the consequences when elites are not successful in reducing ambiguity is found in the 2000 Danish national referendum on the euro. Here political parties at the (extreme) left and right ends of the ideological spectrum joined forces in a No camp, leaving the centrist political parties to campaign for a Yes (de Vreese and Semetko, 2004). The reservations of the ideologically strongly profiled parties towards European integration related to issues such as the loss of national sovereignty, the undermining of national welfare state facilities as well as nationalistic considerations about immigration and integration. With the political spectrum turned upside down ideologically, and a number of official Yes parties in fact not offering a firm recommendation on how to vote, it was easy to see how the parties in the political system eventually ended up signalling ambiguity to the voter. The political ambiguity in the referendum

13 Claes H. de Vreese 11 was fuelled by the inconsistency in the campaign messages. In the Danish case, the Yes side was campaigning on a single message: the economy. However, when this argument was discredited by several experts, the Yes campaign was 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 (see also Kriesi in this volume). Whereas general election campaigns happen on a more or less 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 of 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. In a similar way a referendum campaign can affect segments of the electorate. Citizens 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. A national referendum campaign may also have implications for more general levels of trust in political institutions and leaders. Again taking the example of referendums on issues of European integration we may find in the context of multi-level governance, certain spill over effects from domestic levels of governance to the European level and vice versa. For

14 12 Context, Elites, Media and Public Opinion in Referendums example, political cynicism or negativity towards European integration may spill over into general cynicism towards domestic politics. This is an unexplored dynamic in the literature. Previous research 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 (see Steenbergen and Marks (2004) for a discussion of the salience of European issues for domestic politics). Such a benchmarking with the European level may lead the national system to appear favourable, if it is seen to operate better and more efficiently than the European level, but it may also affect the perception of national system in a negative way, if the European level of governance is seen as a more positive example. In referring to the importance of considerations about European integration for domestic politics and elections, van der Eijk and Franklin (2004) refer to a sleeping giant and discuss the electoral potential in the way of euro-skepticism waiting to be exploited (see also Hooghe and Marks, 2007). 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. Media and public opinion A referendum campaign challenges 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. Prior to a first-order election 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 facts are already on file to make it possible to report in a timely manner (de Vreese, 2001). During campaigns, news programmes 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

15 Claes H. de Vreese 13 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, for example, 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. The media coverage of politics is closely related to (change in) public opinion. As Dalton (2002: 39) concludes: campaigns are now largely media events. The centrality of the media for public opinion is also highlighted in Page and Shapiro s (1992: 386) argument in which 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. As referendums are often held in contexts of information complexity and uncertainty, the media are particularly important, especially in the case of low-involvement issues and under the condition of mixed or conflicting opinions among elites. The dynamics of public opinion is, however, not uniform across all campaigns. LeDuc (2002) identifies three types of referendum campaigns: (i) 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; (ii) opinion reversal campaigns in which reasonably well known issues take new directions over the course of the campaign; and (iii) uphill struggle campaigns in which opinion is much firmer and less subject to rapid change or sudden reversal. 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 referendum 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; de Vreese and Semetko, 2004).

16 14 Context, Elites, Media and Public Opinion in Referendums Citizens 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 magnitude of campaign effects is 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 either reinforce already existing views and perceptions or alter and change these. In our understanding of campaign dynamics voter participation (turnout) and vote choice are two obvious key parameters. With respect to vote choice, a referendum, in contrast to a general election, where a considerable, though shrinking, size of the electorate is guided by ideology, partisanship, and tradition (Dalton, 2000), 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 labelled the punishment trap (Schneider and Weitsman, 1996). In the case of referendums on European integration the proposition that these are second-order national elections hold when the issue at stake is of low salience and does not mobilize political cleavages. However, these referendums can appropriate first-order elections when the issue is salient and the political spectrum polarized (see also Hobolt in this volume). Previous studies offer inconsistent explanations about the outcome of national referendums. Theories of voting behaviour generally emphasize 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, Verba

17 Claes H. de Vreese 15 and Petrocik, 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. 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 vote (see, for example, Clarke et al., 2000; Franklin, Marsh and Wlezien, 1994; Franklin, van der Eijk and Marsh, 1995; Siune and Svensson, 1993; Siune et al., 1994). Several of these are addressed in this book, but in addition we focus on the campaign as a dynamic process in which final vote decisions are made. How then may a campaign matter? Campaigns can serve to activate and crystallize already existing predispositions and opinions, but they can also alter and change these. In understanding how campaigns work, we need to consider, at the individual level, two types of campaign effects: effects of direct and personal campaign experiences and effects of indirect and mediated campaign experiences (de Vreese and Semetko, 2004). 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. 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. 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. A comprehensive campaign effect study (de Vreese and Semetko, 2004) demonstrated that a referendum campaign may: (i) lead citizens to evaluate the campaign positively or negatively; (ii) induce political cynicism, dependent upon the media content to which citizens are exposed and pay attention; (iii) affect citizen s evaluations of domestic political leaders; (iv) serve to crystallize opinion on the topic of the referendum; and (v) influence some voters on how to vote. Despite the importance of the campaign, however, the unpredictability of a referendum should not be exaggerated. Above and beyond the potential low citizen involvement, the presence of ambiguous elite cues, and the importance of mediated (mass) communication, the campaign is only

18 16 Context, Elites, Media and Public Opinion in Referendums part of the story about political engagement, electoral mobilization, evaluations of key political leaders, and vote choice. Most campaign effects are conditional. We should not expect to find large across-the-board effects. As Zaller (1992, 2002) has argued, effects of campaigns and media are not likely to be enormous, but in close electoral races they can be significant. Even in races that appear foregone conclusion, the campaign can serve to further widen the gap between electoral contesters. 1.6 Overview of the book With this discussion of the role of the different actors in a referendum campaign, we have built the framework and argument for why referendum campaigns matter and why they may matter more than campaigns for other electoral contests. In the following we outline the composition of the book. In the following chapter (Chapter 2), Lawrence LeDuc provides an overview of opinion dynamics and specifically applies John Zaller s theoretical framework for the study of information flows. Drawing on survey data from three continents it is shown that there is more unpredictability regarding opinion formation in referendums involving constitutional issues or treaties, because these are areas in which predisposed opinions of mass publics tend to be weaker. Part I addresses the impact of context and political systems on the dynamics of campaigns. Chapter 3 (Hug) analyses the importance of timing and institutional context for the outcome of a referendum. With the example of the wave of referendums on the EU Constitution, it is demonstrated how the prospect of a national referendum affects the bargaining outcome during the negotiations over the constitutional treaty. Chapter 4 (Marsh) demonstrates that in addition to timing and institutional context, campaign intensity and salience is important to understand the dynamics of the campaign and eventually, the outcome. Using the two referendums on the Nice Treaty in Ireland, it is demonstrated that an active campaign can render specific attitudes more important in the electorate for subsequent use in voting decisions. Chapter 5 (Hobolt) investigates how campaign context influences opinion formation and voting behaviour in referendums on European integration. She tests the proposition that higher levels of campaign intensity reduce uncertainty about the vote choice. She examines 14 referendums and illustrates how context matters in EU referendums so that the campaign environment can be an informer and mobilizer and influence voting behavior. Part II addresses voter mobilization and political participation in referendums. Participation is a key benchmark for the vitality and legitimacy of

19 Claes H. de Vreese 17 direct democracy. In chapter 6 (Kriesi) a range of individual-level determinants of participation in referendums is considered. However, citizens are not alone in making the choice to vote or not. The chapter also considers the role of political elites and argues by controlling the intensity of the campaign, they determine much of the level of participation by the citizens. When the elites do not mobilize, the citizens will not participate, not for lack of civism, but because they are unaware of the stakes and, accordingly, not interested in the vote. In chapter 7 (Neijens et al.) predictors of turnout in local referendums are investigated across a range of referendums on different topics. The authors find a strong relationship between participation in representative elections and direct elections, so that it is most civic-minded citizens who are mobilized to vote in referendums. Chapter 8 (Trechsel) considers new technological possibilities, in particular E-voting, and its consequences for electoral participation in the case of referendums. Based on data from a referendum in 2004, the chapter assesses a number of hypotheses on the role and impact of new means of participation in referendums. In this referendum vote a multi-channel approach for participation was chosen. The chapter analyses the effect of the introduction of e-voting on turnout and it assesses the contribution induced by this new channel of participation on the electorate s political choice at the poll. Part II addresses the fundamental issue of which citizens are likely to endorse a proposal and vote yes or reject a proposal and vote no in a referendum. Chapter 9 (de Vreese and Boomgaarden) outlines competing hypotheses about voting behaviour in referendums about European integration. The chapter argues that given the change in the integration focus, new attitudes have become relevant to understanding public support for integration. The study emphasizes anti-immigration sentiments and feelings of national identity as key variables for understanding reluctance with regard to integration, together with economic considerations and the evaluation of the incumbent government. The chapter also demonstrates the degree of volatility in such referendums and assesses the potential magnitude of campaign effects. Chapter 10 (Oscarsson) reviews several individual-level predictors for vote choice. Looking at the 2003 case of the Swedish No to the euro, political ideology emerges as an important feature. This chapter illustrates that in a referendum on a highly politicized issue, predisposition may emerge as more powerful predictors of voting behavior than events in the campaign. Chapter 11 (Sciarini et al.) focuses on a specific topic, environmental issues, but looks across 27 referendums to analyse the determinants of voting behaviour in referenda on environmental issues. The chapter extends Zaller s work and focuses on the effects of political elites in a two-level analysis of referendum votes.

20 18 Context, Elites, Media and Public Opinion in Referendums References Appleton, A. (1992) Maastricht and the French Party System: Domestic Implications of the Treaty Referendum, French Politics and Society, 10: Bennett, W. L. and R. Entman (2001) Mediated Politics: Communication and the Future of Democracy (Cambridge: Cambridge University Press). Blaís, A. and L. Massicotte (1996) Electoral Systems, in L. LeDuc, R. G. Niemi and P. Norris (eds), Comparing Democracies: Elections and Voting in Global Perspective (London: Sage). Bowler, S. and T. Donovan (1998) Demanding Choices: Opinion, Voting and Direct Democracy (Ann Arbor: University of Michigan Press). Bowler, S., T. Donovan and C. J. Tolbert (eds) (1998) Citizens as Legislators: Direct Democracy in the United States (Columbus: Ohio State University Press). Budge, I. and D. J. Farlie (1983) Explaining and Predicting Elections: Issue Effects and Party Strategies in Twenty-three Democracies (London: Allen & Unwin). Butler, D. and A. Ranney (eds) (1994) Referendums around the World: The Growing Use of Direct Democracy (Washington DC: Macmillan, 1994). Campbell, A., W. E. Miller and P. E. Converse (1960) American Voter (Chicago: University of Chicago Press). Clarke, H. D., A. Kornberg and P. Wearing (2000) A Polity on the Edge: Canada and the Politics of Fragmentation (Peterborough, Ontario: Broadview Press). 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). der Eijk, C. van and M. Franklin (1996) Choosing Europe? (Ann Arbor: University of Michigan Press). der Eijk, C. van and M. Franklin (2004) 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). Farrell D. M. and R. Schmitt-Beck (eds) (2002) Do Political Campaigns Matter? Campaign Effects in Elections and Referendums (London and New York: Routledge, 2002). Franklin, M., C. van der Eijk and M. Marsh (1995) Referendum Outcomes and Trust in Government: Public Support for Europe in the Wake of Maastricht, West European Politics, 18: Franklin, M., M. Marsh and C. Wlezien (1994) Attitudes Toward Europe and Referendum Votes: a Response to Siune and Svensson, Electoral Studies, 13(2): Garry, J., M. Marsh and R. Sinnott (2005) Second Order versus Issue-Voting in EU Referendums, European Union Politics, 6 (2005) Hooghe, L. and G. Marks (2007) Special issue on Euroskepticism, Acta Politica: International Journal of Political Science. Hug, S. (2003) Voices of Europe Citizens, Referendums and European Integration (Oxford: Rowman & Littlefield Publishers). Jenssen, A. T., P. Pesonen and M. Gilljam (eds) (1998) To Join or Not to Join: Three Nordic Referendums on Membership in the European Union (Oslo: Scandinavian University Press).

21 Claes H. de Vreese 19 King, A. (1977) Britain says YES: The 1975 Referendum on the Common Market (Washington DC: American Institute for Public Policy Research). Kriesi, H. (2007) The Participation in Swiss Democratic Votes, in C. H. de Vreese (ed.), Dynamics of Referendum Campaigns (Basingstoke: Palgrave Macmillan), pp Klingeman, H-D. and D. Fuchs (eds) (1995) Citizens and the State (Oxford: Oxford University Press). LeDuc, L., R. G. Niemi and P. Norris (eds) (1996) Comparing Democracies: Elections and Voting in Global Perspective (London: Sage). LeDuc, L. (2002) Referendums and Elections: How do Campaigns Differ?, in D. M. Farrell and R. Schmitt-Beck (eds), Do Political Campaigns Matter? Campaign Effects in Elections and Referendums (London: Routledge), pp LeDuc, L. (2003) The Politics of Direct Democracy: Referendums in Global Perspective (Peterborough, Ontario: Broadview Press). LeDuc, L. (2007). Opinion Formation and Change in Referendum Campaigns, in C. H. de Vreese (ed.), Dynamics of Referendum Campaigns (Basingstoke: Palgrave Macmillan), pp Lewis-Beck M. S. (1997) Economics and Elections: The Major Western Democracies (Ann Arbor: University of Michigan Press). Marsh, M. (2007) Referendum Campaigns: Changing What People Think or Changing What They Think About?, in C. H. de Vreese (ed.), Dynamics of Referendum Campaigns (Basingstoke: Palgrave Macmillan), pp Mendelsohn, M. and A. Parkin (eds) (2001) Referendum Democracy: Citizens, Elites, and Deliberation in Referendums Campaigns (New York, Palgrave). Nie, N. H., J. R. Petrocik and S. Verba (1979) The Changing American Voter (Cambridge, Harvard University Press). Oscarsson, H. and S. Holmberg (eds) (2004) Kampen om euron [The Battle over the Euro] (Göteborg University: Department of Political Science). Page, B. I. and R. Y. Shapiro (1992) The Rational Public: Fifty Years of Trends in Americans Policy Preferences (Chicago: University of Chicago Press). Sartori, G. (1987) The Theory of Democracy Revisited (New Jersey: Chatham). Schneider, G. and P. A. Weitsman (1996) The Punishment Trap: Integration Referendums as Popularity Contests, Comparative Political Studies, Schuck, A. and C. H. de Vreese (2006) Framing the EU Enlargement: News Media Content and Effects, European Journal of Communication, 21(1): Semetko, H. A. (2003) The UK Media System, in D. H. Johnston (ed.), Encyclopaedia of International Media and Communications (New York: Academic Press). Sinnott, R. (1995) Irish Voters Decide: Voting Behavior in Elections and Referendums since 1918 (Manchester: Manchester University Press). Siune, K. and P. Svensson (1993) The Danes and the Maastricht Treaty: The Danish EC referendum of June 1992, Electoral Studies, 12(2): Siune, K., P. Svensson and O. Tonsgaard (1994) The European Union: The Danes said NO in 1992 but YES in 1993: How and why, Electoral Studies, 13(2): Steenbergen, M. R. and G. Marks (2004) Introduction: Models of Political Conflict in the European Union, in G. Marks and M. Steenbergen (eds), European Integration and Political Conflict (Cambridge: Cambridge University Press) 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: