Most of the research into voting behavior is carried out in the context of elections for parliaments and, particularly in the US, of an executive. These elections tend to focus on parties, and to a variable extent on individuals who will assume responsibility for policy making. However, in many countries voters are also provided on occasion with the chance to vote directly on policy options through a referendum. These have long been common in Switzerland, where citizens can initiate such votes, and in Italy, and have also been common in some US cities and states, but have been less common in most democracies. Major constitutional changes and questions of sovereignty have often been put to a referendum, as have moral issues. The establishment, and particularly the enlargement of the EU, seems boosted to the referendum industry as states have provided opportunities to the electorate to vote on initial membership, on treaty change and even, in Greenland and the UK, on whether or not to remain a member. Several European countries have also held votes on changes to laws, or constitutional provisions on moral issues like divorce and abortion and same-sex marriage, but we have also seen votes on matters as diverse as a new flag, a new electoral system, water privatization or cuts in judges’ pay. The question addressed in this chapter is how far what we know about voting behavior from looking at elections generalizes to voting behavior in referendums.
We can start by considering some of the more widespread findings about parliamentary and presidential elections. Probably the most basic one concerns party loyalties. Critically, voters do not start to make their choice with a blank slate. As studies from the 1950s and onwards showed, voters tend to have partisan loyalties, and these influence vote choice both directly and indirectly, by influencing the selection of and interpretation of information about the election. There are disputes about the stability of these loyalties and certainly there are questions about how they develop in new party systems, but these loyalties are a factor that cannot be ignored. For some, party leaders are becoming as important, if not more important, than the parties themselves, with loyalties to parties weakened by a disliked leader, or attractive leader for some other party. (A strong case is made in Clarke, Kornberg and Stewart 2004. An alternative view is taken by Curtice and Holmberg 2005.) However, when we move on to think about referendums, whether the important factor is the leader or the party is perhaps secondary to the fact that a “party” and its leader may be recommending a yes or no vote.
The second general set of findings about elections focuses on issues. There have been two broad interpretations of how voters and parties think about issues in elections. The first, and most obvious, is that on any issue a voter has a position, and that the party whose stance is closest to that position is most likely to be chosen. One problem here is that of course there are lots of issues and the closest party on one issue might not be the closest on another. This is commonly dealt with by just taking the most important issues, or by dealing with issues in much broader terms, summarizing them in terms of liberal–conservative or left–right. Both approaches make it easier for the analyst, but also recognize the challenge for any voter in developing a position across multiple issues and finding out where each party stands. Much of the research on voting behavior has demonstrated that voters do not have the knowledge to assess parties issue by issue, but something like “left–right” can help by reducing the amount of detailed information that a voter needs to make a sensible decision, and arguably the decision they would make with full information (the most comprehensive discussion of this can be found in Lupia and McCubbins 1998).
A rather different way to look at issues is to recognize that on most of the topics that dominate political debate voters are in broad agreement, and parties do not differ. Peace, security, economic prosperity are what Donald Stokes (1963) called “valence issues,” and voters will pick the party considered most competent to deliver these. Voters arguably do not need a lot of information to judge competence in areas that impinge on them directly. V. O. Key (1966) suggested a voter simply needed to know was he better off than at the last election to judge economic competence, and there is a wealth of research linking voters’ judgments about the economy to support for incumbent parties (a good review is Lewis-Beck and Stegmaier 2000).
Whether we consider issues in terms of position or valence it is possible that the same party would not “win” on all issues. A party strong on economic prosperity might be beaten by another on security; a party with a position close to a voter on health might be beaten by another on education policy. Hence there is an incentive for parties to fight the election on the issues that are most favorable to them. This is not something they can control, given the existence of other parties with other agendas, but we can recognize that the election agenda will favor some parties over others.
Campaigns could be important whether elections are about parties or issues – and of course they are about both of these. In general, campaigns give parties an opportunity to mobilize support on the basis of existing loyalties. To the extent that issues matter, they should matter more – and in predictable ways – following a campaign which raises voter awareness. Gelman and King (1993) suggested that campaigns made voting more predictable, as voters were more likely to vote in ways that we would expect them to, given past loyalties, social background and general political attitudes. Arguably, the deep rooted determinants of electoral behavior are now much less influential in most countries as party attachment wanes (Dalton and Wattenberg 2003) and social structure becomes less important (Franklin, Mackie and Valen 2009), something that should allow for campaigns to become more important.
All parties do their best to mobilize those they expect to support them, and ensure such people go out and vote. It is always possible that when turnout is not universal some parties are hurt more than others by the failure of their supporters to vote, with the suggestion in some research that parties of the left, who rely on less educated, poorer and perhaps younger voters, suffer a systematic disadvantage (Pacek and Radcliff 1995). This has been rejected by other findings (Fisher 2007; Bernhagen and Marsh 2007). But even if there are not systematic differences over time and place, turnout can still matter in any election, particularly if it falls far short of 100 percent.
These broad approaches are important in explaining voting behavior in elections. We will now turn to look at each of them in the context of voting choice in a referendum to assess how useful each approach is. We might expect that since a referendum is about an issue rather than selecting a person or party to govern, the “issue” itself would be of primary importance and the relevance of party, and incumbency, questionable. We might also expect campaigns to be of considerable importance as referendums – and in particular referendums on a particular issue – are typically unusual events, in contrast to regular elections. As turnout is often much lower than in general elections at least, there is ample opportunity for differential turnout to be significant.
Before exploring the reasons why people vote as they do in referendums, we should first summarize the extent and nature of referendums in democracies (for general accounts, see Butler and Ranney 1994; Gallagher and Ulieri 1996; LeDuc 2003). There are broadly two kinds of votes. The first, and this is the nature of referendums in most countries that have them, is a vote called by parliament. The results may or may not be binding, but the key is that the vote is a consequence of a decision, usually by the government, that a particular policy is to be followed. In most cases, the policy requires a constitutional change and it is this that necessitates a referendum. Referendums have been most common across Europe on constitutional matters, with membership of the European Union and on issues of public morality such as divorce and abortion in Catholic countries common reasons for constitutional change, but they remain relatively rare events, much less common than elections in all but very few countries: Ireland stands out with three dozen votes and Australia has held two dozen. This first type of referendum can also be subdivided, according to whether or not it is necessary. The UK referendums on the EU in 1975 and 2016 were not required, but were wholly political decisions, whereas those in Ireland are required to make changes in a constitution that is particularly specific and anachronistic on many issues (a referendum was needed to restrict the provision of bail conditions for those facing criminal charges, and another – perhaps very many – would be needed to remove sexist language in that constitution, written as it was in a traditional Catholic society in the 1930s). Referendums may also be binding or not, but typically they can only be non-binding in circumstances when they are not formally required.
The second type of referendum is one called by the public, or at least a section of it. This is the Swiss experience, where there have been hundreds of such votes, and there is also a provision for this in Italy, which has seen more than fifty such votes since the mid-1970s to reject (or not) pieces of new legislation.1 It is also common at the local level across the US in some states, and in the UK local referendums can be forced by a petition on issues such as directly electing a mayor.
The variations in the rules governing referendums are potentially important. As the study of voting developed, researchers were able to use a comparative approach to demonstrate a degree of conditionality in behavior. Different electoral rules and different types of party system, for instance, had an impact on the weight of different factors on vote choice. The same, it has been argued, can be true of referendums, with the required and binding ones providing different incentives for voters than do non-required and non-binding ones, as will be discussed below.
Studies of electoral behavior are now based on very extensive post-election surveys, supplemented increasingly by more surveys through the campaign and beyond, but academic studies of referendums typically have been more limited. While the comparative study of voting behavior is now well developed, properly comparative studies of referendums are much less so, although this literature has been growing, prompted not least by the fact that several countries have sometimes held referendums of EU treaty change at about the same time (the pioneer here was Pierce, Valen and Listhaug 1983; subsequently Franklin, Marsh and McLaren 1994 on Maastricht referendums, Glencross and Trechsel 2011 and Hobolt and Brouard 2010 on the European Constitutional Treaty, and Hug 2002, Hobolt 2005, 2007 and Petithomme 2011 on EU referendums more generally, while Svensson 2002 and Marsh 2015 provide comparative studies of several referendums within the same country).
Much of the research does find that party matters in referendum voting, and often matters a great deal. There are a number of reasons why party loyalties can become very significant factors. The first is that the vote is prompted by a government, and this typically identifies the government party (or parties) with the position of supporting a position in the vote. This brings in a government vs. opposition dimension, allowing the possibility of embarrassing or weakening the government by defeating the referendum. The second is that the issue conflict in the referendum may be reflected in the party system, so it would be natural for (some) opposition parties to campaign against proposals by the government.
However, it is not at all unusual to find that the issue conflict is not one that underlies the party system, and in those circumstances it might be expected either that a party is divided on the position to take, or that a party takes a back-seat. In each of these cases, party loyalties will not be mobilized effectively and so partisanship will not be such a strong factor in the vote. It may be more difficult for a governing party that has initiated a vote to do this. This is not to say that supporters of a government party will always be more likely to support that party’s position than those of any other party, but it does suggest that governing parties will normally do a better job of maximizing potential support among its own supporters for its position, other things being equal. A good illustration of this is support in recent referendums on EU Treaty change in Ireland. There was a Fianna Fail (FF) led government in place for four of the last five such votes, and a Fine Gael (FG) led government in place for the most recent one. Expert opinion would see FG as slightly more pro-European than FF, although both are center-right parties that have always actively supported Ireland’s membership.2 In the four referendums where FF led the government, FF voters were more likely to vote “yes” than those of FG. In the most recent vote in 2012, FG voters were more inclined to vote “yes” than those of FF (Marsh 2015). A study of one of the Norwegian and British votes on EU membership suggested that, when parties were divided, so were their followers (Pierce, Valen and Listhaug 1983). Government office is a good incentive to minimize division. But it would be wrong to think that government party voters are always more likely to support their party’s position since that is not the case. Of course in the 2016 “Brexit” referendum in the UK, divisions in the ruling party prompted the referendum in the first place, and so ensured that “party” would not be a unifying factor for Conservative voters.
It was argued by Franklin, Marsh and McLaren (1994; see also Franklin, van der Eijk and Marsh 1995) that referendum voting could become little more than a vote on the popularity of the government. The argument drew on interpretations of European Parliament elections as “second-order” votes (Reif and Schmitt 1980). Although the European Parliament might have a different function and issue agenda to a national parliament, the second-order argument is that voters will pay little attention to that but use the vote as a “referendum” on the current government of the country (similar arguments have been made about sub-national elections). An important condition here must be how salient to voters are the issues raised by the referendum. The “second-order” argument requires voters to have little interest in the ostensible issue per se. As we will see below, there is ample evidence that issues can and do matter in referendums, including those on EU-related matters. Even so, most research on EU referendums have found evidence of an – admittedly sometimes small – anti-government effect where those dissatisfied with the current administration are more likely, other things being equal, to reject the government’s proposal.
It is not uncommon in referendums for parties on both sides (or neither) to cede some of the work to non-party or cross-party campaign organizations. In some countries, such structures to promote the yes or no side are essentially a requirement for funding and media access, but they also provide an opportunity for both sides to remove some of the partisan edge from the debate. This can be useful for the government parties. Commenting on the “success” of Ireland’s second referendum on the Lisbon Treaty The Irish Times noted the important role of civil society activists “whose arguments made it possible to disconnect the treaty as an issue in the minds of voters from the performance of the government” (quoted in Laffin 2015). Indeed, in this referendum, government popularity was not a significant factor at all, despite its remarkably low rating in the wake of the public bailout of Ireland’s banks (Marsh 2015).
There is some evidence that the importance of parties and government status can be conditional on the type of referendum. Hug and Sciarini (2000; see also Hug 2002) find from their study of fourteen referendums on the EU across Europe that government supporters will be mobilized more effectively when the referendum is not required and when the outcome is binding, arguing that in such circumstances the vote is essentially one of support for the government (the “Brexit” referendum provided a notable exception to this, for reasons already given). In contrast, where the decision to hold a vote is not really the choice of the government and, in particular, where the result is not binding, voters are freer to follow their issue preferences.
In many referendums, parties are relatively uninvolved as the impetus for the vote comes from outside the party system. The Italian case is interesting as such votes need a quorum in order to defeat legislation, and while one party may be critical in gathering the signatures necessary to provoke the vote, others – particular parties in government – may do as little as possible so as to suppress turnout and so defeat the proposed abrogation (Ulieri 2002).
Parties matter but issues often matter more is a general theme of referendum voting research. What is most interesting here is what is meant by “issues” in this context, and how the campaign frames the vote in terms of an issue or issues. If issues are important, how are issues related to the vote? If general elections are commonly fought on “valence” issues, what about referendums? Do they revolve around the decision of what will best ensure the “good life,” or is there more room for “positions” to matter, as voters can be expected to have very different attitudes to some of the questions that come up in referendums? Certainly some research has looked at issues in these positional terms, explaining choice in a referendum in terms of attitudes to what might be thought of as the broader issue (some examples are discussed below). In the case of EU-related votes, this means attitudes to European integration. The EU as an issue is not necessarily aligned with the major underlying dimensions of party politics. In most countries, there is a strong economic left–right dimension, often reinforced by a religious–secular one. Europe does not fit naturally into that. Although there was a tendency in some countries in earlier years for the left to be more skeptical about a union based on a free market, currently opposition to the EU is stronger on both the left and right margins of party systems and support is greater in the center, including center-right and center-left. This makes the task of assessing the pros and cons of EU-related referendums more difficult for voters.
However, Danish research has consistently argued that attitudes to the EU are fairly stable and do predict votes in EU-related referendums. This applies to votes on the Maastricht Treaty as well as to those on things like the Euro, where the conditionality of their support led Danes to vote to stay out of the common currency. Svensson (2002: 748) argues strongly that the Danish experience shows how “consistent values may be developed on a salient issue and may become the basis for voting behavior in one or more referendums.”
A positional approach has been taken by Hobolt (2007). She treats each voter as having an “ideal point” with regard to an issue, one that can be compared with the status quo, and the position that would hold if the referendum change were approved. Hence, a voter would have to decide whether the referendum would bring policy closer to her ideal than the status quo. The problem with such an approach, as with positional voting in elections, is identifying the underlying issue dimension. This may be easier when the issue dimension is well structured and relatively stable within a country, but is more problematic when it is not.
The way in which the subject of the referendum is framed can be crucial (de Vreese and Semetko 2004; see also Dekavella 2016 on how the referendum on Scottish independence was framed in the media). Just as with elections, yes and no sides will seek to place the vote on a terrain which is favorable to their own side. Issues still matter in this situation, and they may be positional, but the issues may differ across groups. For instance, in votes on the European Treaty in the Netherlands, those concerned about “identity threats” voted no (de Vreese and Boomgaarden 2007); in various treaty referendums in Ireland, those worried about “neutrality” voted no (Marsh 2015; Garry, Marsh and Sinnott 2005). It may be that the issue is in fact not affected by the referendum, but the important point is that some people are persuaded that an issue matters in a particular vote. Atikcan (2015) shows how pro-change campaigners in rerun referendums sought to focus their efforts better so as to ensure the campaign would be fought on more favorable grounds.
On some votes, the question of what best ensures economic prosperity can come to the fore, as in most elections. As in elections, each side generally seeks to argue that its recommendation provides the best path. This is more of a valence issue, which comes down to the voter’s trust in the advice of one side rather than another. In elections, voters can be seen to be influenced by recent economic conditions; that is, people may judge economic competence by recent economic circumstances. This is harder to generalize to a referendum. Arguably, the economy feeds into government satisfaction, and as discussed above, this is a common influence, but it could also be that bad conditions also influence the way voters think about the possible change that a referendum could bring, making them more fearful, or feeling they have little to lose (for a discussion and some evidence, see Sattler and Urpelainen 2012).
There is a substantial literature on how much voters know when it comes to elections, but in any one country the level of knowledge cannot be expected to vary hugely from election to election. However, in the case of referendums huge variation can be expected, as the salience of the topic, and the attention of parties, media and civic organizations is far from constant. It has been argued that in lower salience referendums voters are more inclined to take their cues from parties, with “second order” considerations more to the fore (Hobolt 2005, 2007).
Referendums always involve some change. Most commonly this is a change to the constitution to enable new legislation, but in all cases there is the expectation of some change to the status quo. A common argument amongst campaigners against such change to voters unsure of the merits of the proposal is to vote “no” (i.e., for the status quo) if you don’t know, and there is evidence that those who feel they do not know enough are inclined to follow this advice. The onus is on those campaigning for a “yes” to demonstrate that the future will be better than the status quo. Bowler and Donovan (1998) find that the least informed are inclined to reject proposal for change and this is also being found in many other studies (see, for example, Nadeau, Martin and Blais 1999; Clarke, Kornberg and Stewart 2004; Whiteley et al. 2012; Schuck and de Vreese 2008; and see also Morisi 2016).
If election campaigns tend to result in making vote choice more predictable, referendum campaigns can be far less clear in their consequences. Hobolt’s argument (2007) that strong campaigns reduce “second-order” effects does not always lead to predictability, even if a strong campaign increases issue voting, because campaigns may frame the debate in unexpected ways (see also Dvořák 2013). LeDuc argues on the basis of a broad comparative study that opinion changes most substantially in cases when “there is little partisan, issue or ideological basis on which voters might tend to form an opinion easily” (LeDuc 2003: 207). Campaigns in these cases involve opinion formation. In contrast, those cases when “the nature of the issue itself or the circumstances of the referendum generate strong cues based on partisanship, ideology or pre-existing opinions” show least evidence of instability (LeDuc 2003: 208). LeDuc describes a third case which should resemble the second type, but where the campaign successfully shifted the bases of decision making. A good illustration is an Australian vote in 1999 on removing the British monarch as head of state, which was lost when the “yes” side divided over the nature of the replacement, whether a president would be elected or appointed. This third type is itself unpredictable. Several votes on EU treaties across Europe were expected to pass on the basis that public opinion favored the EU, as did the major parties, but in the course of the campaign opinion moved against a “yes” vote, for change, as “no” campaigns moved the bases of decision. Hobolt and Brouard (2010) shows how French concerns focused on threats to the “social model” by EU liberalism, while Dutch concerns involved threats to identity (see also Lubbers 2008).
Certainly, polls far in advance of the vote can be very poor guides indeed, and even polls a few weeks before the vote can be very wide of the mark, as the public in many cases have yet to engage with the debate. On the basis of a systematic study of polls in advance of a wide range of referendums in Europe (outside Switzerland) Fisher found clear evidence of a “status quo” effect, with support for change tending to decline over the course of a campaign (Fisher 2016).
Turnout does vary a lot in referendums, and is often very low, but can vary very significantly from referendum to referendum. Butler and Ranney (1994) observed that turnout is typically less than in a general election. For instance, in Ireland, which has had thirty-nine referendums since 1937, turnout has varied from a low of 29 percent on University Representation in the Upper House, changes to adoption laws and changes to Bail provisions, to a high of 76 percent, adopting the new constitution in 1937, and 68 percent on accession to the EU in 1972. While half of all votes attracted a turnout of between 43 and 62 percent, a quarter were below 42 percent and only a quarter above 62 percent (general election turnout in the same period averages 72 percent, with a low of 63 percent). In the US, proposals tend to be placed on the ballot along with the choices in a variety of elections. Magleby (1984: 90–95) observes that more voters participate in candidate choice than indicate support or opposition to propositions. This increases the chances differential turnout can matter to the result. Nevertheless, as LeDuc (2003: 172) points out, referendum turnout can be as high or higher than in an election on some occasions.
Individual determinants of turnout in referendums are broadly the same as in elections. That is, we expect older, more educated, more politically interested and wealthier electors to vote. Partisanship, and the activities of parties, can serve to reduce the impact of these individual level factors, but parties typically are less involved in getting out the vote in referendum campaigns. The intensity of the campaign and the familiarity of the subject are two factors highlighted by Kriesi (2005) as contributing to participation rates. He argues on the basis of Swiss evidence that high intensity campaigns in themselves do little to alter the participation differential between aware and unaware citizens, whereas votes on familiar topics see a smaller differential. Kriesi’s work notwithstanding (see also Lutz 2007), there has been little systematic work on the impact of lower turnout on referendum outcomes. One argument is that low turnout favors those against change, on the basis that those against something new are more committed, but equally plausible is that proponents of change are more committed. (Of course in referendums where turnout has to be above a threshold for it to have any effect, abstention might well be favored by those against the proposal. A recent example was the referendum in Hungary on migrant quotas. Rejection of the proposal [which was EU policy] was almost universal but only 44 percent voted so the result was invalid.) There was evidence that low turnout did help those in the first Irish vote on the Nice Treaty who were opposed to it. There was a significant increase in turnout in the second vote and this went overwhelming to the “yes” side (Sinnott 2003) but the same effect was far less striking in the two votes on Lisbon, where the increase in the pro-Treaty vote was not down to higher turnout (Sinnott and Elkink 2010). Certainly there is no good evidence from the Irish case to indicate that differential turnout always helps the side opposed to change. Another expectation is that low turnout might benefit the position favored by the right – just as some argue left-wing parties are disadvantaged by low turnout. Lutz (2007) – studying the extensive Swiss experience – found low turnout in fact tended to hurt the rightwing position, but argued that a more informed electorate tended to be more left-wing, so if voters became more informed and so more motivated to vote, the bias might not be so clear. The referendum of Scottish independence in 2014 was notable for the very high level of turnout, over 84 percent, compared with below 70 percent in most recent general elections, but there is no evidence that this boosted the vote for the status quo. “No” voters were in any case more prevalent in those groups where turnout would normally be higher: older, more middle class and living in more affluent areas (Curtice 2014).
This chapter has summarized much of the work done in recent years on voting in referendums. While there is far less written on this topic than has been produced on elections, the body of work has been growing rapidly, not least because referendums have become rather more common in recent times. While referendums are ostensibly about particular policy issues or decisions and so differ from elections, in which people are voted in to office and make policy and take decisions across a very wide range of matters, there are some broad themes which run across research both referendums and elections. The role of parties and their leaders in providing a basis for the voter to make their referendum choice, just as this plays an important role in most elections, is one such theme. A second is the place of issues in such votes; and as in elections, research on referendums finds that not everyone views the same issues as important, and shows that for some voters the issue may not be connected directly to the immediate vote. As in elections, the way in which the vote is framed is important. Because referendums are much more irregular than elections, and even when they are common, the topics may vary hugely, campaigns are typically much more important than they are in elections, as more people decide, and decide in unpredictable ways, during the last few weeks and days before the vote. One conclusion then is that what we know about elections generalizes to referendums in as much as the processes underlying decisions are similar but the context can be important, the actors can be different and the weight given to certain factors can be very different. While these themes do run across the growing body of work on voting behavior in referendums, there is also considerable diversity across these studies. In part this is because the study designs, the measures used and the theoretical approaches adopted vary considerably. While election studies have become relatively institutionalized in many countries, allowing both cross-time and cross-country comparisons, referendums are still treated for the most part as one-off events and most of the surveys used to study referendums are quite separate from the more normal election studies. Of course the referendums themselves differ enormously, in terms of topics, the role of parties and other actors in campaigns, and the institutional basis of the vote, but until such variables can be built into particular studies, we are not able to assess properly quite how important that variation is to explaining differences in results obtained by different studies. Just as much of what we are coming to know about electoral behavior shows that institutional context is often a critical conditioning factor, so with referendums, conditionality is perhaps even more important and future research should be designed with this in mind if it is to properly develop our understanding of voter choice in referendums.
1Italy also provides for constitutional referendums, triggered by proposed changes to the constitution. Unlike the popular referendums in Italy, these are not subject to a quorum.
2Using the Chapel Hill series of expert surveys, this author’s analysis shows Fine Gael averages 6.5 and Fianna Fail 5.8 on a 10-point pro-EU scale. On the expert surveys, see Bakker et al. (2015). On Ireland, see also Benoit (2009).
Atikcan, E. Ö. (2015) “The Puzzle of Double Referendums in the European Union,” Journal of Common Market Studies, vol. 53, no. 5, September, 937–956.
Bakker, R., de Vries, C., Edwards, E., Hooge, L., Jolly, S., Marks, G., Polk, J., Rovny, J., Steenbergen, M. and Vachudova, M. A. (2015) “Measuring Party Positions in Europe: The Chapel Hill Expert Survey Trend File, 1999–2010,” Party Politics, vol. 21, no. 1, January, 143–152.
Benoit, K. (2009) “Irish Political Parties and Policy Stances on European Integration,” Irish Political Studies, vol. 24, no. 4, December, 447–466.
Bernhagen, P. and Marsh, M. (2007) “The Partisan Effects of Low Turnout: Analyzing Vote Abstention as a Missing Data Problem,” Electoral Studies, vol. 26, no. 3, September, 401–413.
Bowler, S. and Donovan, T. (1998) “Two Cheers for Direct Democracy,” Representation, vol. 35, no. 4, Winter, 247–254.
Butler, D. and Ranney, A. (eds.) (1994) Referendums Around the World, London: Macmillan.
Clarke, H. D., Kornberg, A. and Stewart, M. (2004) “Referendum Voting as Political Choice: The Case of Quebec,” British Journal of Political Science, vol. 34, no. 2, April, 345–355.
Curtice, J. (2014) “So Who Voted Yes and Who Voted No?,” Available online: http://blog.whatscotland-thinks.org/2014/09/voted-yes-voted/, [accessed February 20, 2017].
Curtice, J., and Holmberg, S. (2005) “Leaders,” in Thomassen, J. (ed.) The European Voter: A Comparative Study of Modern Democracies, Oxford: Oxford University Press: 125–166.
Dalton, R. and Wattenberg, M. (eds.) (2003) Parties Without Partisans: Political Changes in Advanced Industrial Societies, Oxford: Oxford University Press.
De Vreese, C. H. and Boomgaarden, H. G. (2007) “Immigration, Identity, Economy and the Government: Understanding Variations in Explanations for Outcomes in EU-Related Referendums,” in de Vreese, C. H. (ed.) The Dynamics of Referendum Campaigns: An International Perspective, Basingstoke: Palgrave Macmillan: 186–205.
De Vreese, C. H. and Semetko, H. A. (2004) Political Campaigning in Referendums: Framing the Referendum Issue, Abingdon: Routledge.
Dekavella, M. (2016) “Framing Referendum Campaigns: The 2014 Scottish Independence Referendum in the Press,” Media, Culture and Society, vol. 28, no. 6, September, 793–810.
Dvořák, T. (2013) “Referendum Campaigns, Framing and Uncertainty,” Journal of Elections, Public Opinion and Parties, vol. 23, no. 4, November, 367–386.
Fisher, S. D. (2007) “(Change in) Turnout and (Change in) the Left Share of the Vote,” Electoral Studies, vol. 26, no. 3, September, 598–611.
Fisher, S. (2016) “Do People Tend to Vote Against Change in Referendums?,” Available online: https://constitution-unit.com/2016/06/22/do-people-tend-to-vote-against-change-in-referendums/, [accessed June 17, 2017].
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, vol. 18, no. 3, 101–117.
Franklin, M., Mackie, T. T. and Valen, H. (eds.) (2009) Electoral Change: Responses to evolving Social and Attitudinal Structures in Western Countries, Colchester: ECPR Press.
Franklin, M., Marsh, M. and McLaren, L. (1994) “Uncorking the Bottle: Popular Opposition to European Unification in the Wake of Maastricht,” Journal of Common Market Studies, vol. 32, no. 4, December, 455–472.
Gallagher, M. and Ulieri, P. V. (eds.) (1996) The Referendum Experience in Europe, London: Macmillan.
Garry, J., Marsh, M. and Sinnott, R. (2005) “ ‘Second-Order’ Versus ‘Issue Voting’ Effects in EU Referendums: Evidence from the Irish Nice Treaty Referendums,” European Union Politics, vol. 6, no. 2, June, 201–221.
Gelman, A. and King, G. (1993) “Why Are American Presidential Election Campaign Polls So Variable When Votes Are So Predictable?,” British Journal of Political Science, vol. 23, no. 4, October, 409–451.
Glencross, A. and Trechsel, A. (2011) “First or Second Order Referendums? Understanding the Votes on the EU Constitutional Treaty in Four EU Member States,” West European Politics, vol. 34, no. 4, July, 755–772.
Hobolt, S. (2005) “When Europe Matters: The Impact of Political Information on Voting Behaviour in EU Referendums,” Journal of Elections, Public Opinion and Parties, vol. 15, no. 1, February, 85–109.
Hobolt, S. B. (2007) “Campaign Information and Voting Behaviour,” in de Vreese, C. H. (ed.) The Dynamics of Referendum Campaigns: An International Perspective, Basingstoke: Palgrave Macmillan: 84–116.
Hobolt, S. B. and Brouard, S. (2010) “Contesting the European Union? Why the Dutch and the French Rejected the European Constitution,” Political Research Quarterly, vol. 64, no. 2, June, 309–322.
Hug, S. (2002) Voices of Europe: Citizens, Referendums and European Integration, Lanham: Rowan and Littlefield.
Hug, S. and Sciarini, P. (2000) “Referendums on European Integration: Do Institutions Matter in the Voter’s Decision,” Comparative Political Studies, vol. 33, no. 1, February, 3–36.
Key, V. O. (1966) The Responsible Electorate: Rationality in Presidential Voting 1936–60, Cambridge, Harvard University Press.
Kriesi, H. P. (2005) “The Participation in Swiss Direct-Democratic Votes,” in de Vreese, C. H. (ed.) The Dynamics of Referendum Campaigns: An International Perspective, Basingstoke: Palgrave Macmillan: 117–141.
Laffin, B. (2015) “Confronting Europe: The Irish Referendums on Lisbon,” in Elkink, J. A. and Farrell, D. M. (eds.) The Act of Voting: Identities, Institutions and Localities, London: Routledge: 116–134.
LeDuc, L. (2003) The Politics of Direct Democracy: Referendums in Global Perspective, Toronto: University of Toronto Press.
Lewis-Beck, M. and Stegmaier, M. (2000) “Economic Determinants of Electoral Outcomes,” Annual Review Political Science, vol. 3, June, 183–219.
Lubbers, M. (2008) “Regarding the Dutch ‘Nee’ to the European Constitution: A Test of the Identity, Utilitarian and Political Approaches to Voting ‘No’ ,” European Union Politics, vol. 9, no. 1, March, 59–86.
Lupia, A. and McCubbins, M. (1998) The Democratic Dilemma: Can Citizens Learn What They Need to Know?, Cambridge: Cambridge University Press.
Lutz, G. (2007) “Low Turnout in Direct Democracy,” Electoral Studies, vol. 26, no. 3, September, 624–632.
Magleby, D. (1984) Direct Legislation: Voting on Ballot Propositions in the United States, London: Johns Hopkins.
Marsh, M. (2015) “Voting on Europe, Again and Again: Stability and Change in the Irish Experience with EU Referendums,” Electoral Studies, vol. 38, June, 170–182.
Morisi, D. (2016) “Voting Under Uncertainty: The Effect of Information in the Scottish Independence Referendum,” Journal of Elections, Public Opinion and Parties, vol. 26, no. 3, 354–372.
Nadeau, R., Martin, P. and Blais, A. (1999) “Attitude Towards Risk-Taking and Individual Choice in the Quebec Referendum on Sovereignty,” British Journal of Political Science, vol. 29, no. 3, July, 523–539.
Petithomme, M. (2011) “Awakening the Sleeping Giant? The Displacement of the Partisan Cleavage and Change in Government-Opposition Dynamics in EU Referendums,” Perspectives on European Politics and Society, vol. 12, no. 1, April, 89–110.
Pacek, A. and Radcliff, B. (1995) “Turnout and the Vote for Left-of-Centre Parties: A Cross-National Analysis,” British Journal of Political Science, vol. 25, no. 1, January, 137–143.
Pierce, R., Valen, H. and Listhaug, O. (1983) “Referendum Voting Behavior: The Norwegian and British Referenda in Membership in the European Community,” American Journal of Political Science, vol. 2, no. 1, February, 43–63.
Reif, K. and Schmitt, H. (1980) “Nine Second-Order National Elections – A Conceptual Framework for the Analysis of European Election Results,” European Journal of Political Research, vol. 8, no. 1, March, 3–44.
Sattler, T. and Urpelainen, J. (2012) “Explaining Public Support for International Integration: How Do National Conditions and Treaty Characteristics Interact with Individual Beliefs?” Journal of Politics, vol. 74, no. 4, October, 1108–1124.
Schuck, A. and de Vreese, C. H. (2008) “The Dutch No to the EU Constitution: Assessing the Role of EU Skepticism and the Campaign,” Journal of Elections, Public Opinion and Parties, vol. 18, no. 1, January, 101–128.
Sinnott, R. (2003) Attitudes and Behaviour of the Irish Electorate in the Second Referendum on the Nice Treaty, Report for the European Commission Representation in Ireland, Institute for the Study of Social Change, University College Dublin.
Sinnott, R. and Elkink, J. A. (2010) Attitudes and Behaviour in the Second Referendum on the Treaty of Lisbon, Dublin: Geary Institute and School of Politics and International Relations.
Stokes, D. E. (1963) “Spatial Models of Party Competition,” American Political Science Review, vol. 57, no. 2, June, 368–377.
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, vol. 41, no. 6, October, 733–750.
Ulieri, P. V. (2002) “On Referendum Voting in Italy: YES, NO or Non–Vote? How Italian Parties Learned to Control Referendums,” European Journal of Political Research, vol. 41, no. 6, October, 863–883.
Whiteley, P., Clarke, H. D., Sanders, D. and Stewart, M. C. (2012) “Britain Says NO: Voting in the AV Ballot Referendum,” Parliamentary Affairs, vol. 65, no. 2, November, 301–322.