Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, POLITICS ( (c) Oxford University Press USA, 2018. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy and Legal Notice).

Subscriber: null; date: 16 December 2018

Do Campaigns Make a Difference?

Summary and Keywords

The study of political campaigns is very varied in the political science literature. On the one hand, campaigns can involve groups of citizens working together on a local issue of concern to them, such as preventing an airport expansion from threatening their community. Only a relatively few people are likely to be actively involved and the goals of such a campaign are fairly clearly defined and limited. At the other end of the scale a campaign can consist of a broad social movement that is trying to influence public opinion and bring about changes in public policies on really big issues like climate change and global warming. Large numbers of people are likely to be involved and the goals are broad and ambitious. In between these two extremes, a whole range of campaigns with different objectives and strategies are to be found in contemporary democracies. This article focuses on election campaigns which are in an intermediate position between these two. Early research suggested that such campaigns were not very important but subsequent research shows that they are influential both in increasing turnout and changing the party choices that individual electors make.

Keywords: political campaigns, Ground War, Air War, electoral choice, elections


This article focuses on the nature and role of campaigns in contemporary democracies. It will not attempt to examine all types of campaigns, but rather focus more narrowly on the role of campaigns in elections. At first sight the argument that election campaigns have effects on voting seems to be fairly obvious and uncontroversial, but this has been contested in the past. Early research cast considerable doubt on the argument that campaigns make a difference (Berelson, Lazarsfeld, & McPhee, 1954; Lazarsfeld, Berelson, & Gaudet, 1944). This work stressed the influence of the long-term attachments of citizens to political parties and also events that generally occurred well before election campaigns took place as influences on voting, leaving little room for campaign effects. That said, Lazarsfeld and his colleagues introduced the notion of “activation,” which proved very useful in later work. This is the idea that latent political attachments of the voters are often “woken up” by campaigns making voters more attentive as elections approach.

The aim of this article is to examine the impact of campaigns in influencing voting behavior, but it emphasizes two relatively neglected issues in this literature relating to the scope of campaign effects and also why they work. The literature on campaigning has tended to focus a lot on identifying the size of effects, and the question as to why they work and what their scope is in the political process have been secondary considerations. We will argue that the neglect of these issues has detracted from the task of understanding their role in modern Democratic politics. There has been a growing emphasis on field experiments in the research on this topic; and while these have many positive features, they narrowly focus attention. Our concern is to understand campaigns as part of a wider process of understanding Democratic politics.

The central argument in this essay is that campaigns have to be seen as one component of a wider model of electoral behavior and also part of the policy-making process in contemporary democracies. Different theories make different predictions about the effectiveness of campaigns, and so we can better understand why they work by focusing on theories of electoral choice. Campaign effects fit more easily with some theories than with others, and so we review the three major theories of voting behavior with the idea of clarifying the role of campaigns in these analyses.

In addition the distinction between the “Air War” and the “Ground War,” described below, means that election campaigns are a central to understanding the process of policy-making in contemporary democracies. Campaigns are just one aspect of party strategies that involve wider issues of policy-making and governance. We begin by examining different dimensions of campaigning that need to be taken into account when trying to understand their role in elections.

The Dimensions of Campaigning

It is useful to categorize different aspects of election campaigning in order to identify more closely the scope of what should be studied. One useful distinction is between the “Air War” or the campaign conducted by politicians and parties via the press and electronic media and the “Ground War” or campaigns organized at the local level by party activists and volunteers. These different types of campaign complement and reinforce each other.

A narrow view of the “Air War” sees it as consisting of campaign speeches, political advertising, and televised debates, as well as campaign events organized at the center by political parties and party leaders. In comparison, the ground campaign concentrates on canvassing people on their doorsteps, fundraising, and mobilizing the voters at the local level. A good example of a successful “Air” campaign was that conducted by Barack Obama during the 2008 presidential election, and it was complemented by an equally successful “Ground” campaign that made extensive use of the Internet to mobilize and activate local volunteers. However, the “Air” campaign is in fact much broader than this, since it is really a continuous feature of governing in modern democracies. Incumbent and opposition parties and leaders are continuously trying to shape the political agenda and gain advantages over opponents during the entire period between elections. In a very real sense, campaigning is central to Democratic governance, as parties try to define agendas, dominate the airways, and steer the policy-making process in their favor on a continuing basis. Understanding the scope of modern election campaigning involves casting the net wide. In contrast, the Ground War is much more confined to the period of the election campaign itself.

Another long-established issue in the study of campaigns is the distinction between priming, mobilization, and conversion. One can win elections by attracting the attention of existing supporters who do not necessarily pay attention to politics in “normal” times. An effective campaign will prime them to the messages being put out. This process of priming can set the agenda and make voters pay attention to one issue rather than another, and if this is more favorable to a particular candidate it can be quite effective. Alternatively one can mobilize or “wake up” the voters, and this often works best among independents who are unattached and get them to participate in an election that they otherwise might ignore. Finally, one can convert supporters from rival candidates and parties. An effective campaign can do all three of these, but by common consent it is easier to prime existing supporters and to mobilize the unattached than to convert opponents.

When Tony Blair’s Labour Party won a spectacular landslide victory over their Conservative opponents in the 1997 general election in Britain, this involved all three processes. Voters were primed to focus on the weak leadership of the then Prime Minister John Major, they were mobilized to support “New Labour” by a popular party leader and many switched parties (Butler & Kavanagh, 1997). Again different theories have different predictions about the effectiveness of these types of campaigns in contemporary elections.

A third important issue is the duration of campaigns (Stevenson & Vavreck, 2000). In most democracies a new election campaign starts immediately after the old one has finished, such is the emphasis on media management in modern government. However, politicians and parties are limited in their ability to get the population to focus on politics and elections during the midterm period. In the absence of outside shocks, a large proportion of the electorate can only really be persuaded to pay close attention to politics as the election approaches (Lau & Redlawsk, 2006). This is because many voters have only a limited interest in politics, and so politicians may try to run continuous campaigns but they are not necessarily successful. In practice, political leaders have to think strategically about shaping the narratives of modern politics, even if they do not always succeed.

The early skepticism about the effect of campaigns referred to earlier arose in part because it became apparent to early researchers that elections are very often predictable well in advance of polling day (Rosenstone, 1983). Given this, it is difficult to argue that campaigns play an important role in influencing the results if we know the outcomes well before the campaign gets underway. Gelman and King (1993) suggested an explanation for this apparent paradox when they argued that campaigns make a difference because they allow the electorate to learn about the fundamental determinants of electoral choice such as the state of the economy and the effectiveness of candidates. These fundamentals drive voting behavior, but the electorate does not fully understand or engage with them in advance of the campaign. This analysis sees campaigns as educational devices that “enlighten” voters (Arceneaux, 2006). Clearly this process is likely to be influenced by the length of the campaign, with longer campaigns having the potential to enlighten more than short ones.

On the other hand, if one takes the view that there is a continuous campaign then there is a sense that campaigns are really important even if elections are relatively predictable well in advance. This is because the Air War continues throughout the entire period of an administration, and if it frames media debates and, above all, successfully delivers the policies that most people want, this is one aspect of what might be described as the “long campaign.” If effective policy-making in the midterm delivers success that is as much of a campaign effect as much as winning a TV debate shortly before polling day. So the apparent paradox goes away if governance is seen as a continuous campaign.

A fourth feature is the tone of a campaign, which largely focuses on whether or not messages are framed in positive or negative terms (Wattenberg & Brains, 1999). Early research suggested that the tone of the campaign was quite important, with negative campaigning tending to demobilize voters (Ansolabehere, Simon, & Valentino, 1994). But more recent work challenges this idea, showing that negative campaigning can be quite effective in the right circumstances and can mobilize voters, providing the messages have credibility (Lau & Pomper, 2002). The result is that many contemporary election campaigns try to find a balance between positive and negative messages, but are willing to “go negative” if it appears to convey an immediate advantage.

Most of the research on campaigning has focused on the impact of various campaign activities on voting. A related question concerns whether or not their impact is growing or declining over time, or alternatively if campaign effects are entirely dependent on the context. They may be more important in local elections than in general elections as far as the Ground War is concerned since districts are smaller and therefore easier to cover, but it may also depend on what is at stake. In the referendum on Scottish independence in 2014, the campaign had the effect of mobilizing the entire population and produced a very high turnout (84%), something that can be attributed to the fact that voters saw it was as a very important issue for the future of the United Kingdom. On the other hand, the referendum in Britain in 2011 on changing the electoral system for Westminster Parliamentary elections had a turnout of only 42%, reflecting the fact that electoral systems are not very important for most voters (Whiteley, Clarke, & Stewart, 2012).

In light of this discussion of some of the issues that frame the analysis of campaigns, we turn next to an examination of the evidence about the importance of campaigns in practice.

Demonstrating Campaign Effects

The “Michigan” model of electoral choice dominated the picture in the analysis of electoral behavior in the 1950s and 1960s (Campbell, Converse, Miller, & Stokes, 1960). At the core of this model is the concept of partisanship, which was defined as a long-term emotional attachment to a party that is typically acquired by individuals in adolescence or early adulthood. In this view partisanship is seen as an enduring product of socialization processes in the family and community, and once formed it generally strengthens over time. Partisanship is an “unmoving mover” that influences other factors that help to explain voting behavior, but which is not itself influenced by them. To use a modern expression, partisanship is exogenous to the electoral process in this interpretation.

To reiterate an earlier point, one of the implications of this is that it creates a barrier to campaign messages. Republican partisans, for example, will believe campaign messages supportive of Republican candidates but reject those supportive of Democrats (see Zaller, 1992). Moreover this perceptual screen has a strong influence on the reception of information in campaigns. Bartels (2002), for example, showed that many strong Democrats in the United States thought that inflation had increased during Republican Ronald Reagan’s presidency when it had actually declined precipitously.

The implication of this argument is that campaigning may be able to prime or mobilize existing partisans when attachment to parties is strong, but it can do little to convert opponents since they are much less likely to be attentive to critical messages. Furthermore the ability of party campaigns to mobilize their own partisans is also constrained by the strength of partisanship. Most people had relatively strong partisan attachments in the 1960s, and this fact encouraged them to vote anyway without needing much additional prompting from campaigns. They did not really need to be mobilized in order to get them to vote. This means that in a world of strong partisans campaigns in general, particularly the “short campaign” in the run-up to an election, would be marginal at best and irrelevant at worst. This is the main reason why campaigns were discounted in the early literature on voting behavior.

Despite this, it has been apparent for a long time that Ground War campaigns can influence elections, and this was shown to be true well before the partisanship model was developed. In 1924, Harold Gosnell performed two experiments in elections in Chicago, one prior to the presidential election of 1924 and the other a year later during the mayoral election in the city (Gosnell, 1927). He found that direct mail increased turnout by 1% in the presidential election and by 9% in the mayoral election. Later research focused on campaigns in different localities in the United States using aggregate election returns to identify similar effects.

One of the earliest of these studies looked at campaigning in Gary, Indiana (Cutright & Rossi, 1958), and focused on precinct committeemen, that is, prominent party organizers. Their activities were analyzed in a model of voting using aggregate data from election returns and also from the U.S. Census. The results showed that home ownership and the ethnic status were the strongest predictors of democratic voting in the 1956 presidential election, but at the same time campaigning by Democrat committeemen was also important. If a very active democratic committeeman faced a relatively inactive Republican committeeman in a particular district, this increased the Democratic vote share by about 4%. Subsequent research utilizing the same approach while incorporating additional controls showed that similar effects could be found in other parts of the country (see Cutright, 1963).

Kramer (1970) conducted the first study of campaigning at the national level, utilizing American National Election Study data to examine campaign effects for presidential, congressional, and local state candidates. He found that local campaigns had an impact on turnout by mobilizing voters, but they did not appear to have an influence on party choice, suggesting little effect on conversion. On balance, while the empirical evidence was ambiguous in relation to what the effects were because of the difficulties of isolating campaign effects from other factors, it was apparent that campaigning did have a significant impact on electoral choice.

One of the earliest field experiments in political science was conducted in Britain by Bochel and Denver (1972) with the aim of identifying campaign effects in local elections. They worked closely with the local Labour party in a strong Labour district of the City of Dundee in Scotland. The experiment involved selecting two apartment blocks in the district whose residents had very similar social characteristics. They subjected one of these apartment blocks to an intensive campaign of leafleting and canvassing in an election to the local city council, and left the other one untouched to act as a control. Since this was a relatively safe Labour district, the other parties did very little in the way of campaigning and so they did not have much of an influence on the outcome. A post-election survey showed that the campaign made a difference of about 10% in the turnout of Labour voters.

Other evidence of the effects of local election activity on voting in Britain comes from official spending data on local campaigns. Spending on campaigns is tightly regulated in Britain, and it can be used as a surrogate measure for local campaigning (Johnston, Pattie, & Allsop, 1988). Campaign spending includes things like the costs of printing leaflets, advertising in the local media, organizing public meetings, telephone canvassing, arranging transport and accommodation for activists, and a variety of other things. The pattern observed is that parties spend the maximum or close to the maximum in competitive or marginal constituencies and much less in uncompetitive safe seats. One of the effects of this is to raise turnouts in the marginal seats compared with the safe seats and also to influence conversion. This is indirect evidence that campaigning mobilizes voters (Denver & Hands, 1974).

These early studies all focused on aggregate analysis, but one of the first papers to look at individual-level data from the American National Election Study was published by Steve Finkel (1993). Echoing the earlier work by Lazarsfeld and his colleagues, he argued for an “activation” theory of campaign effects. In this view rather than simply reinforcing individuals’ preexisting vote intentions, campaigns serve to activate individuals’ political predispositions and make them more likely to vote. These results showed that campaigns had the potential to exert large electoral effects. Additional work showed that campaigns are important in other respects too. In the long-drawn-out U.S. presidential campaigns, events such as party conventions and television debates are influential in mobilizing the voters (Holbrook, 1996).

Erikson and Wlezien (2012) investigated this idea by examining nearly 2,000 pre-election polls conducted in campaigns from 1952 through to 2008 in order to investigate how voting intentions changed during campaigns. They conclude that the “fundamentals” or the key drivers of electoral support such as the state of the economy, ideology, and partisanship drive electoral choice, but campaign events can “shock” the relationship between voting and these fundamentals, and this can influence the outcome particularly if they take place close to polling day. Other research supports this interpretation (see, for example, Kenski, Hardy, & Jamieson, 2010).

In her work on campaigns, Vavreck (2009) focuses on the economy, which is commonly seen as one of the fundamentals that drive electoral choice. She makes the plausible argument that while the economy is very important it nonetheless requires candidates to draw attention to it in a campaign. Clearly if an incumbent has a good economic record, he or she will put economic performance at center stage in the campaign but will try to avoid the issue if their record is poor. James Carville, the chief campaign strategist for Bill Clinton’s campaign for the presidency in 1992, coined the phrase “It’s the economy stupid!” for the benefit of his campaign activists. The economic slowdown in the United States in the early 1990s meant that it was vitally important for the Clinton campaign to focus on this issue when attacking a Republican incumbent.

In Britain, measures of campaigns were first incorporated into models of turnout and party choice at the aggregate level utilizing survey data from the party activists who actually do the campaigning on the ground (Seyd & Whiteley, 1992). Campaign activism scales built from the data collected in these surveys showed that local campaigning influenced both the turnout and party choice in elections. Later work utilized data from the British Election Study surveys (Clarke, Sanders, Stewart, & Whiteley, 2004). These asked respondents about the efforts of persuasion by friends and family and also if they had been canvassed on the doorstep or had been telephoned by party activists. These individual-level voting choice models showed that both mobilization and conversion were taking place among voters exposed to campaigning of this type.

The modeling also showed that the different political parties had varying capacities to conduct campaigns and to target them where they would be most effective. In the 2001 general election, for example, Labour and the Liberal Democrats were able to target their local campaigns on sympathizers and the uncommitted much more effectively than the Conservatives. The Conservative campaign was not targeted at all well and so ended up mobilizing many of the supporters of their rival parties. These campaigns generally failed to identify their electoral base. Essentially the Conservatives lacked effective party organizations on the ground in many constituencies and so were unsure where their supporters could be found when the election campaign started.

Subsequent work has explored which mode of campaigning is more effective in comparing face-to-face contact, direct mail, or telephone canvassing. Large-scale surveys show that face-to-face contact is most effective, direct mail has positive if limited effects, and telephone canvassing is largely ineffective when it comes to mobilizing or converting voters (Gerber & Green, 2005). However, this claim has been challenged by rival evidence suggesting that telephone canvassing by volunteers has an effect larger than door-to-door canvassing (Nickerson, 2005). Later work established that emails sent by parties or candidates appears to have no effect on the vote, so some things work and other things do not (Nickerson, 2008). Similarly, in a field experiment Arceneaux (2008) showed that telephone canvassing could have a strong effect on voter preferences for parties but only weak effects on preferences for candidates and little or no effect on attitudes to issues. But later work challenged the earlier conclusions about the influence of door-to-door canvassing on issue perceptions (Arceneaux & Kolodny, 2009). So there is some confusion in the literature about the effects of different types of campaigning.

These kinds of conflicting findings have encouraged some researchers to focus exclusively on field experiments in order to isolate effects (see Green & Gerber, 2004). As is well known, these have a number of advantages, principally the use of random allocation of participants to treatment and control groups. This eliminates many of the confounding factors that complicate survey-based studies. On the other hand, it is not always possible to isolate the true effects, even using experiments. As Green and Gerber explain:

A single experiment can establish that a GOV (Get Out the Vote) tactic works in a particular setting; a series of experiments is necessary to show that the experimental hypothesis holds when the political, economic, or demographic conditions are different.

(Green & Gerber, 2004, p. 17)

Green, Aronow, and McGrath (2012) provided a meta-analysis of the large number of studies that have used experimental methods to measure campaign effects. It shows that face-to-face canvassing is effective and telephone canvassing can also influence voters but plays a lesser role than door-to-door canvassing. Direct mail has a small but significant effect and text messaging may be important too, although the effects from one experiment using this appear to be implausibly large (Malhotra, Michelson, Rogers, & Valenzuela, 2011). Overall, this body of work shows that campaigns are effective, but the effects are variable and in general rather modest.

Needless to say, repeat experiments with multiple treatment groups in different settings designed to control for all the confounding factors are a tall order. There is also what might be described as hidden costs of such experiments, that is, the danger of neglecting theoretical considerations that might explain effects. When the focus is heavily on ensuring that the methodological aspects of the research are well designed and conducted, there may be a tendency to forget the details of how things actually work in practice. Campaign effects are rooted in interpersonal relationships, and we know from social psychological experiments that these can be very powerful.

Why Do Campaigns Work?

One of the classic studies of interpersonal influence in psychology was conducted by Solomon Asch (1958). He showed that subjects could be compelled to deny the evidence of their own eyes in an experiment involving interpersonal pressures to conform to a group consensus. A typical experiment involved a room full of subjects announcing to everyone present their estimates of the lengths of a series of measuring rods. Unknown to the one subject in the group, all of the other participants were there to administer the treatment. This consisted of successive individuals announcing lengths that were grossly out of line with those actually observed. Despite this, as far as the subject was concerned the group members all appeared to agree with each other, and videos of the experiment revealed that subjects often appeared confused and surprised by this. However, this did not prevent them frequently falling into line when faced by this apparently strong group consensus; they very often agreed with the group estimate even when it was obvious that it was wrong.

If information is provided by figures of authority, then again classic experiments have shown that the effects on behavior can be even more powerful. Stanley Milgram (1974) demonstrated that figures of authority could induce subjects to undertake very distasteful tasks that actually involved the appearance of administering torture to fellow citizens. Clearly, we should not expect election campaigning to have such strong effects as in these experiments, but the general principle that individuals are strongly influenced by others, particularly if they observe a group consensus or are exposed to one-sided information by figures of authority, largely explains campaign effects.

In practice voters face conflicting messages from rival parties and party leaders, from members of their family and immediate circle of friends, as well as from the media in election campaigns. In addition there has been a decline in trust in politicians across the Democratic world, which weakens the effects of any messages put out by political leaders in these contexts (Rothstein, 2005). There is an acknowledgment of these ideas in Green’s review of the experimental evidence on campaigning. He writes, “When social norms are asserted forcefully, the effects tend to be quite large, and even prerecorded phone calls conveying social pressure messages significantly increase turnout” (Green et al., 2012, p. 10).

A lot of research has focused on media effects in campaigns. In his study of TV advertising in presidential contests between 1988 and 1996, Shaw (1999) concludes that advertising coverage did significantly influence the vote, though the effects were not large enough to be decisive in explaining the election outcomes. In contrast, Johnston and his colleagues’ (2004) analysis of the National American Election Study concluded that advertising played a decisive role in the closely fought 2000 U.S. Presidential election. George W. Bush outspent Al Gore during the final week in key battleground states, and he gained significant support, which tipped the balance (see also Huber & Arceneaux, 2007; Kenski et al., 2010). The consensus in the research on this topic appears to be that advertisements can be effective, but the effects tend to be short lived, and they cannot overturn an election in which one side starts with a big lead in the polls over another (Gerber, Gimpbel, Green, & Shaw, 2011; Sides & Vavreck, 2013).

In the United States, context advertising is closely linked to campaign spending, and as a review paper by Jacobson (2013) shows, at the congressional level, campaign spending matters. A key debate in this literature has been about the effects of spending by incumbents compared with challengers. Earlier work suggested that challengers had an advantage over incumbents in relation to the effects of campaign spending (Jacobson, 1980), but this was later challenged for methodological reasons. Campaign spending is appropriate for a game theoretical analysis, since it involves a complex interaction between the actions of incumbents and challengers, making it difficult to estimate reliable effects.

Attempts to untangle this interaction have proved challenging. One group of researchers have supported the earlier findings that challenger spending is more important than incumbent spending (Abramowitz, 1988; Ansolabehere & Gerber, 1994). But others have argued that the effects for challengers and incumbents are about the same, or even that incumbent spending is more important (Erikson & Palfrey, 1998; Goldstein & Freedman, 2000; Green & Krasno, 1988). A third paper suggests that neither challenger nor incumbent spending makes a difference (Levitt, 1994), so there is some confusion about this in the literature.

The evidence for Britain is much more clear cut than in the United States and shows that spending in local campaigns makes a real difference to turnout and party choice, and there is no particular incumbency advantage (Johnston & Pattie, 1995; Whiteley, Seyd, & Richardson, 1994). The tight regulation of spending in Britain means that neither candidates nor parties can buy time for advertising on television. This means that a spending arms race typical of American elections is not possible. Parties concentrate their spending in marginal seats and these produce modest but significant effects.

To summarize the discussion up to this point, campaign effects are rooted in the psychology of persuasion and conformity, and the modeling and experimental work shows that they are real and play an important role in elections even though they are seldom decisive. However, the empirical literature on campaigning tends to be isolated from broader theoretical models of electoral choice. The argument was made earlier that “fundamentals” are affected by campaigns, but this begs the question of what the key fundamentals actually are. They are both ill-defined but at the same time taken for granted in much of the discussion.

An important point that has not been recognized in the literature is that the fundamentals differ in alternative models of electoral choice, and so we are unlikely to fully understand the role of campaigning unless it is seen as part of a wider theoretical analysis of electoral choice. Campaigns need to be judged in the context of a broad theory of voting if they are to be properly understood. Indeed the failure to do this may in part explain some of the conflicting findings referred to earlier. At the same time the evidence on the magnitude of campaign effects throws considerable insight on the issue of which theories of electoral choice are best at explaining what we observe in the world. With this in mind we outline the expectations that the three dominant models of electoral choice have about the role of campaigns.

Campaigns and Models of Electoral Choice

The three dominant models of electoral choice can be described as the cleavages, spatial, and valence models. The oldest of these is the cleavage model, which arises from the work of Lipset and Rokkan (1967) deriving it from a rich historical tradition of political sociology going back to Marx and Weber. Their aim was to identify the major cleavages in the advanced democracies and examine how they interacted with political parties and with individual voters. The term cleavage refers to conflicting groups in society where individuals are united by common identities in opposition to members of other groups. The core idea of the model is that cleavages politicize members of the group, and political parties arise to represent the group interests created by these divisions. In party politics, cleavages represent broad social divisions such as social class, ethnicity, religion, linguistic, and cultural identities.

Lipset and Rokkan (1967) identified the dominant cleavages in party systems and examined their relationship to voting behavior and political representation. Cleavages often represent minorities such as ethnic groups, but they can also represent majorities such as the working class or women. The approach they took was to establish that important cleavages provide a social base for the formation of political parties and social movements, and these in turn structure electoral choice. Elections then ultimately determine the political orientations of governments and the policies they produce.

The well-known “funnel of causality” of electoral choice developed in the 1950s in the United States was grounded in the cleavages model, but it added a social–psychological dimension by introducing partisanship into the analysis (Campbell et al., 1960). The idea was that social and economic cleavages, rooted in social class and income, nurture group identities, which then create partisanship. Partisanship, together with issue perceptions and candidate evaluations then explain voting behavior further down the causal chain. The application of this model to Britain was captured in a well-known quote from the political sociologist, Peter Pulzer: “[i]n British party politics, class is everything, all else is embellishment and detail” (1967, p. 98).

This model of electoral choice did not go unchallenged at the time, however, since early research showed that political divisions across Europe were not on the whole characterized by enduring and stable relationships between parties and voting behavior based on social cleavages. Rather they were more like a shifting set of political allegiances linked to changing policy programs by parties, which in turn led to varying levels of electoral support. As Zuckerman put it, the links between social groups and political organizations were “more like a kaleidoscope than a rocklike strata” (1982, p. 136). There appeared to be a relatively weak association between social cleavages, political divisions, and voting behavior in national elections.

The cleavage model has implications for role of campaigns. It implies that campaigns are entirely about activation and mobilization as the earlier discussion of partisanship explained. Campaigning is likely to have a very limited impact on changing group identities such as partisanship, since if these are strong and stable they will be relatively impervious to short-term forces. On the other hand, since voter issue perceptions and candidate evaluations are very close to vote choice in the funnel of causality, they might be affected by campaigns even if partisanship is not. But the key role played by partisanship tended to marginalize campaigns as significant factors in explaining the vote with the exception perhaps of having a modest effect on mobilization.

The second dominant model of electoral choice is the spatial model, which has its origins in economic theory and was pioneered in political science by Anthony Downs (1957). As is well known, the spatial model is based on the proposition that voters choose a party that is in close agreement with their own views on divisive issues in society and politics. Key assumptions are that voters are rational actors who are distributed along a “left–right” ideological scale that bundles together a broad set of issues. These issues commonly relate to questions of personal security, taxation, redistribution, and public spending. Typically parties on the Left favor high levels of taxation and public spending, and parties on the Right favor the opposite set of policies.

Parties are strategic actors in the model and will maneuver in this ideological space in order to capture as many voters as possible. This model supports the “median” voter theorem, which argues that in two-party systems both parties will try to “capture the middle ground of politics” by locating themselves at the median of the left–right dimension, thereby maximizing votes (Downs, 1957). Spatial theory inherited the assumption that voters have exogenously determined issue preferences from neoclassical economic theory. In the latter, consumer preferences are assumed in order to identify effects for estimation purposes and the same is true in the spatial model. In fact, experimental evidence shows that this assumption is a poor representation of reality since parties and leaders are able to mold such preferences (Sanders, Clarke, Stewart, & Whiteley, 2008). But this assumption nonetheless anchors the theory as individuals attempt to “maximize utility” by supporting a party closest to them on the “left–right” ideological dimension. Although spatial models have been imaginatively elaborated in various ways, they retain the core assumption that salient position issues drive the choices of utility-maximizing voters.

The scope for campaign effects is very limited in this model. If voters are rational utility-maximizing individuals, they will seek out information prior to making their decision to vote. But information from parties and candidates will be heavily discounted since it will correctly be seen as being biased and therefore potentially misleading. Rather voters will rely on the performance of these parties in the past and the records of particular politicians in delivering on their promises. If information search is viewed as a rational action, then campaigning is going to have very limited effects because rational actors will acquire enough information on their own without being pressured into supporting parties by self-serving and often unreliable promises from politicians. In the cleavage model, campaigning helps people to process the information they have already acquired, but in the spatial model the rationality assumption severely restricts the credibility of such information.

An additional point is that if the views of voters are ideologically coherent in a way that allows them to be distributed along a single left–right scale, then they will be resistant to any information that tries to change their views. Ideological coherence strengthens resistance to new information because of the cognitive dissonance occurring when challenges occur to highly interrelated attitudes (Zaller, 1992). Again it has long been recognized that this assumption is also a poor representation of reality with many individuals having no consistent ideological views at all (Converse, 1964). But it is nonetheless central to the spatial model and implies that campaign effects are likely to be weak or nonexistent.

The third model is the valence model of electoral choice, introduced originally by Donald Stokes in the first systematic critique of the spatial model (Stokes, 1963). In this model issues play an important role just as they do in the spatial model, but there are no assumptions about voters relying on ideology in order to judge parties and candidates. In the valence model voters support parties which they perceive as being best able to deliver on a very limited number of issues they care about, and over which there is a broad consensus in society about what should be done (Clarke et al., 2004, 2009). The classic valence issue is the economy, because it is very salient and also because there is an overwhelming consensus that prosperity is preferred to poverty, growth to stagnation, and employment to joblessness. Thus voters in general will support a party that can deliver “good times,” and they will abandon one which is unable to do this. The model represents a generalization of an extensive body of work on economic voting that has emerged over the years (see Lewis-Beck & Stegmaier, 2007).

Alongside issues, there are two additional variables that play an important role in the valence model. First, partisanship is important but it works differently from that of the Michigan model discussed earlier. In this model partisanship is conceptualized as a “running tally” or cumulative evaluation of party performance of the delivery of valence issues over time, with more recent performances weighted more heavily than earlier ones (see Fiorina, 1981). This makes partisanship a dynamic rather than a static phenomenon with a successful party acquiring additional identifiers over time, while an unsuccessful one will lose them if it fails to perform in line with expectations.

The second variable in the valence model is leadership evaluations. Since leaders are key political decision makers, voter perceptions of their actual or anticipated performance is an important indicator of the ability of a party to deliver on the issues which count. The more highly a voter thinks of a particular leader, the more likely they are to support that leader’s party. In this account voters are using leaders as “fast and frugal” heuristics or easily applied rules of thumb that they can use to evaluate political parties (Gigerenzer, 2008). Such heuristics are particularly important in a world characterized by significant complexity and uncertainty.

Unlike the cleavages and spatial models there is ample scope for campaign effects in the valence model. Firstly, parties will seek to put as favorable a gloss on their own performance as possible and attack that of their rivals. In addition, party leaders who are effective communicators will be able to influence large numbers of voters, who have limited abilities to judge the effectiveness of policies in practice. There is no assumption that voters are rational information seekers who optimize their search; rather they are “satisficers” who collect as much information as they think they need and often this is quite limited. This means that voters can be fooled by campaign messages, although voter perceptions that an incumbent party is delivering an effective economic performance have to be grounded in reality as we will observe below (Clarke, Stewart, & Sanders, 2013; Whiteley et al., 2013). Voters cannot be persuaded that the economy is doing well if it is not, since an important component of economic evaluations consists of voter judgments of their own personal financial circumstances.

Overall, the conclusion from this section is that the cleavages and spatial models allow only a limited scope for campaigns to influence the voters, though this is true for different reasons. In contrast the valence model puts campaigning at the center stage of electoral choice, although claims by politicians and parties have to be supported by reality. To illustrate this argument we examine a case study: the election campaign in Britain in the 2015 general election that amply illustrates the role of campaigns in explaining voting behavior.

Campaigning and Governance: The 2015 General Election in Britain

This section gives a brief review of the campaigns leading up to the 2015 general election in Britain. This example reinforces the point that campaigns cannot be understood by merely examining the Ground War, or by taking a narrow view of the Air War. The entire economic strategy of a government has to be seen as part of its reelection campaign, which means that policy-making is central to understanding campaigns. This is a broad agenda for looking at campaigns and is illustrated in this example.

The 2015 general election was the culmination of a period of five years of a coalition government that included the Conservatives and Liberal Democrats as the senior and junior partners. The two parties were an unlikely match, since historically they had different ideologies, policy goals, and social bases of support. But the arithmetic of the 2010 general election imposed this since there was no viable alternative government available at the time (Clarke, Kellner, Stewart, Twyman, & Whiteley, 2016). The agreement between the coalition parties published shortly after the election provided a road map for their partnership (HM Government, 2010). It constructed a narrative that Britain needed a strong, stable, and prudent government to deal with what had become a protracted economic crisis following the financial crash and Great Recession. The new government moved quickly to deal with the growing budget deficit produced by the recession and the enormous sums of money committed to bailing out banks following the financial crash (see Clarke et al., 2016).

The new government made large-scale cuts during its first year, particularly following an emergency budget that planned cuts of £113 billion to be made by 2014–2015, a very large fiscal squeeze. The partners claimed that their policies would ensure that public-sector borrowing would decline to 1.1% of GDP by 2015–2016, the structural deficit would be eliminated by 2014–2015 and a surplus of 0.8% of GDP would emerge by 2015–2016. In other words fixing the deficit became the central strategic focus of the new government.

We have argued elsewhere that the austerity strategy was a key component of the electoral campaign in the previous election of 2010 and it was repeated again in 2015 (Clarke et al., 2016). In the earlier campaign the Conservatives made much of the argument that Britain faced imminent bankruptcy and suggested that the country was in a similar position to Greece and other cash-strapped EU member states when the new government took over in 2010. This rhetoric was driven by the need to provide a radical alternative narrative to that of the Labour Party during the campaign and it allowed both coalition partners to blame the previous Labour government for the financial crisis and the Great Recession.

The Chancellor of the Exchequer, George Osborne’s approach of focusing on the size of the budget deficit as the root cause of the problem during the 2010 election was necessary as a campaign strategy because the economy was recovering at that time. Economic growth had rebounded to a respectable 4.1% in the last quarter of 2009, and this had much to do with Prime Minister Gordon Brown’s leadership in working with other political leaders in the G7 countries to coordinate policies to combat the recession. The enormous size of the deficit made it easy for Conservative strategists to focus on the bad news of debt and ignore the good news of growth.

Subsequently, the coalition government abandoned the original austerity strategy in the midterm of the administration in 2012, largely because the economy had failed to recover at that point. It did not of course announce this change of policy. The continuing high level of unemployment and the growing threat of deflation changed the government’s mind and it decided to embark on what amounted to a disguised fiscal stimulus. It was acutely aware that economic recovery had to be given time to work if electoral benefits were to be achieved by 2015. In the event, the economy did start to rebound after the midterm, and this brought down unemployment, which in turn stimulated economic optimism, as can be seen in Figure 1. The figure compares the level of unemployment with the level of economic pessimism about the state of the economy among the voters during the period 2010 to 2015.1 The figure shows that the two variables were closely associated and as unemployment fell so did economic pessimism.

Do Campaigns Make a Difference?Click to view larger

Figure 1. Unemployment and Economic Pessimism 2010-2015

Source: The Office of National Statistics and the Essex Continous Monitoring Survey

Since inflation had been running at low levels since the start of the Great Recession, unemployment is the best objective measure of the actual state of the economy. The relationships in Figure 1 indicate that it is important for a government to show progress in managing the economy rather than just claiming this in an election campaign. So campaigns have to work within the grain of what is actually happening in reality. Figure 2 shows the electoral payoff of the change of strategy, since the percentage of voters who thought that the Conservatives were best at managing the economy compared with Labour grew consistently wider over this period.

Do Campaigns Make a Difference?Click to view larger

Figure 2. Perceptions of Which of the Major Parties is Best at Managing the Economy 2010-2015

Source: The Essex Continous Monitoring Survey

It is evident in Figure 2 that in early January 2011 the continuing economic stagnation meant that perceptions of which party was best at managing the economy were level pegging. But after 2012 the Conservatives started to move ahead of Labour as the austerity policies were quietly abandoned and the economy began to grow again. By the time of the 2015 election, the party dominated on the issue of economic competence. Just prior to that general election the Conservative lead over Labour on this issue was more than 20% points.

In relation to the other key variables in the valence model the Conservatives also had a considerable lead over Labour on public evaluations of their respective leaders, David Cameron and Ed Miliband. Respondents were asked which leader would make the best Prime Minister and by the time of the general election of 2015 the Prime Minister was well ahead of his rival. About twice as many people thought that David Cameron would make the best Prime Minister compared with Ed Miliband. In the case of partisanship, the Conservatives had only a relatively modest advantage over Labour by the time of the general election, but it was nonetheless a lead.

Overall, the Air War went on continuously throughout the five years of the coalition government and because electoral considerations loomed large well ahead of the 2015 vote, economic policy-making became part of the election campaign. The campaign did not just consist of making speeches and promises to the voters, but also by an incumbent changing direction in order to engender a pre-election recovery from recession. The Conservatives won the Air War in the “long campaign,” and this cannot be ignored when evaluating campaign effects.

Who Won the Ground War in 2015?

We can measure the effect of the Ground War in the 2015 general election in Britain using the pre-post-election panel survey of the Essex Continuous Monitoring Survey. This consisted of just over 3,000 respondents in Great Britain surveyed on the Internet shortly before the election and then again immediately after polling day. The post-election survey revealed that 56% of respondents reported being contacted by one or more political parties during the campaign.

Labour was marginally ahead of the Conservatives in prosecuting the Ground War since the party contacted some 43% of the electorate during the campaign while their main rival contacted 40%. The Liberal Democrats were marginally ahead of UKIP and the Greens, and contact by the Scottish Nationalists and Plaid Cymru (the Welsh Nationalists) was much lower because they only campaigned in their own countries. Contact took many forms, including parties delivering leaflets (34%), and letters (17%), sending emails (8%), canvassing people at home (7%), telephone canvassing (2%), talking to people in the street (2%), contacting them on Facebook (2%), and on Twitter (1%).

Do Campaigns Make a Difference?Click to view larger

Figure 3. The Ground War – Party Vote Shares and Party Campaigning in 2015

Source: Essex Continuous Monitoring Survey pre-post-election panel.

Figure 3 shows the relationship between campaign contact and party vote shares in the general election of 2015. It shows that the correlation between the vote shares and the amount of campaigning in the election was very high (r = 0.86). The message is that parties that campaign a lot do well when it comes to winning votes, although the figure provides only an illustration of relationships, since a full analysis requires a multivariate model (see Clarke et al., 2016). The fact that Labour was marginally ahead of the Conservatives in the Ground War gave the party a modest advantage in the election. But this narrow advantage was certainly not enough to offset the lead the Conservatives had in the “Air” War that had been going on for five years by the time of the election.


We have reviewed a lot of research on campaigning that spans many years and examined it in two different countries. Unfortunately space considerations make it difficult to extend this to a wider number of countries. At their core campaigns are about one group of people trying to persuade another to support them, and as such campaign effects are rooted in a long-standing literature in psychology on interpersonal influence. When campaigns are viewed from the perspective of the three dominant models of electoral choice, effects are likely to be weaker in the cleavages and spatial models compared with the valence model. This is true for reasons that are embedded in the logic of these rival theories—issues which have been rather neglected in this literature.

Equally, there is a continuous election campaign going on in modern democracies with incumbents and opposition parties trying to outwit each other during the entire period of an administration. The Air War in the short run is about providing effective leadership and making promises that resonate with voters. But in the long run it is also about pursuing a winning policy-making strategy in government and adopting policies, particularly in relation to the economy, which will deliver and resonate with voters. In the future both these important features of campaigning need to be taken into account if the effects of campaigns are to be fully understood.


Abramowitz, A. I. (1988). Explaining Senate election outcomes. American Political Science Review, 82, 385–403.Find this resource:

    Ansolabehere, S., & Gerber, A. (1994). The mismeasure of campaign spending: Evidence from the 1990 U.S. House elections. Journal of Politics, 56, 1106–1118.Find this resource:

      Ansolabehere, S., Iyengar, S., Simon, A., & Valentino, N. (1994). Does attack advertising demobilize the electorate? The American Political Science Review, 88(4), 829–838.Find this resource:

        Arceneaux, K. (2006). Do campaigns help voters learn? A cross-national analysis. British Journal of Political Science, 36(1), 159–173.Find this resource:

          Arceneaux, K. (2008). I’m asking for your support: The effects of personally delivered campaign messages on voting decisions and opinion formation. Quarterly Journal of Political Science, 2(1), 43–65.Find this resource:

            Arceneaux, K., & Kolodny, R. (2009). The effect of grassroots campaigning on issue preferences and issue salience. Journal of Elections, Public Opinion and Parties, 19(3), 235–249.Find this resource:

              Asch, S. (1958). Effects of group pressure upon the modification and distortion of judgments. In Readings in Social Psychology (pp. 174–192). New York: Holt, Rinehart & Winston, 174–182.Find this resource:

                Bartels, L. M. (2002). Beyond the running tally: Partisan bias in political perceptions. Political Behavior, 24(2), 117–150.Find this resource:

                  Berelson, R. B., Lazarsfeld, P. N., & McPhee, W. M. (1954). Voting: A study of opinion formation in a presidential campaign. Chicago: University of Chicago Press.Find this resource:

                    Bochel, J., & Denver, D. (1972). The impact of the campaign on the results of local government elections. British Journal of Political Science, 2, 239–243.Find this resource:

                      Butler, D., & Kavanagh, D. (1997). The British general election of 1997. London: Macmillan.Find this resource:

                        Campbell, A., Converse, P. E., Miller, W. E., & Stokes, D. E. (1960). The American voter. Chicago: University of Chicago Press.Find this resource:

                          Clarke, H. D., Kellner, P., Stewart, M. C., Twyman, J., & Whiteley, P. (2016). Austerity and political choice in Britain. London: Palgrave.Find this resource:

                            Clarke, H. D., Sanders, D., Stewart, M., & Whiteley, P. (2004). Political choice in Britain. Oxford: Oxford University Press.Find this resource:

                              Clarke, H. D., Sanders, D., Stewart, M., & Whiteley, P. (2009). Performance politics and the British voter. Cambridge, U.K.: Cambridge University Press.Find this resource:

                                Converse, P. (1964). The nature of belief systems in mass politics. In D. Apter (Ed.), Ideology and discontent. New York: Free Press.Find this resource:

                                  Cutright, P. (1963). Measuring the impact of local party activity on the general election vote. Public Opinion Quarterly, 27, 372–385.Find this resource:

                                    Cutright, P., & Rossi, P. H. (1958). Grassroots politicians and the vote. American Sociological Review, 23, 171–179.Find this resource:

                                      Denver, D., & Hands, G. (1974). Marginality and turnout in British general elections. British Journal of Political Science, 4, 17–35.Find this resource:

                                        Downs, A. (1957). An economic theory of democracy. New York: Harper and Row.Find this resource:

                                          Erikson, R. S., & Palfrey, T. (1998). Campaign spending and incumbency: An alternative simultaneous equation approach. Journal of Politics, 60, 355–373.Find this resource:

                                            Erikson, R. S., & Wlezien, C. (2012). The timeline of presidential elections: How campaigns do (and do not) matter. Chicago: University of Chicago Press.Find this resource:

                                              Finkel, S. E. (1993). Reexamining the “Minimal Effects” model in recent presidential campaigns. Journal of Politics, 55(1), 1–21.Find this resource:

                                                Fiorina, M. P. (1981). Retrospective voting in American national elections. New Haven, CT: Yale University Press.Find this resource:

                                                  Gelman, A., & King, G. (1993). Why are American presidential election campaign polls so variable when votes are so predictable? British Journal of Political Science, 23(4), 409–451.Find this resource:

                                                    Gerber, A. S., Gimpel, J. G., Green, D. P., & Shaw, D. R. (2011). How large and long lasting are the persuasive effects of televised campaign ads? Results from a randomized field experiment. American Political Science Review, 105, 135–150.Find this resource:

                                                      Gerber, A. S., & Green, D. P. (2005). Do phone calls increase voter turnout? An update. The Annals of the American Academy of Political and Social Science, The Science of Voter Mobilization, 601, 142–154.Find this resource:

                                                        Gigerenzer, G. (2008). Rationality for mortals: How people cope with uncertainty. Oxford: Oxford University Press.Find this resource:

                                                          Goldstein, K., & Freedman, P. (2000). New evidence for new arguments: Money and advertising in 1996 Senate elections. Journal of Politics, 62, 1087–1108.Find this resource:

                                                            Gosnell, H. F. (1927). Getting out the vote: An experiment in the stimulation of voting. Chicago: University of Chicago Press.Find this resource:

                                                              Green, D. P., Aronow, P. M., & McGrath, M. C. (2012). Field experiments and the study of voter turnout. Journal of Elections Public Opinion and Parties, 23, 27–48.Find this resource:

                                                                Green, D. P., & Gerber, A. S. (2004). Get out the vote! Washington, DC: Brookings Institution Press.Find this resource:

                                                                  Green, D. P., & Krasno, J. S. (1988). Salvation for the spendthrift incumbent: Reestimating the effects of campaign spending in House elections. American Journal of Political Science, 32, 884–907.Find this resource:

                                                                    HM Government. (2010). The coalition: Our programme for government. London: HM Government.Find this resource:

                                                                      Holbrook, Thomas M. (1996). Do Campaigns Matter? Thousand Oaks and London: SAGE.Find this resource:

                                                                        Huber, G. A., & Arceneaux, K. (2007). Identifying the persuasive effects of presidential advertising. American Journal of Political Science, 51, 957–977.Find this resource:

                                                                          Jacobson, G. C. (1980). Money in congressional elections. New Haven, CT: Yale University Press.Find this resource:

                                                                            Jacobson, G. C. (2013). The politics of congressional elections (8th ed.). New York: Pearson.Find this resource:

                                                                              Johnston, R., Jamieson, K. H., & Hagen, M. G. (2004). The presidential campaign of 2000 and the foundation of party politics. New York: Cambridge University Press.Find this resource:

                                                                                Johnston, R., & Pattie, C. (1995). The impact of spending on party constituency campaigns at recent British general elections. Party Politics, 1, 261–273.Find this resource:

                                                                                  Johnston, R., Pattie, C., & Allsop, J. G. (1988). A nation dividing? The electoral map of Great Britain 1979–1987. London: Longman.Find this resource:

                                                                                    Kenski, K., Hardy, B. W., & Jamieson, K. H. (2010). The Obama victory: How media, money, and message shaped the 2008 election. New York: Oxford University Press.Find this resource:

                                                                                      Kramer, G. H. (1970). The effects of precinct-level canvassing on voter behavior. Public Opinion Quarterly, 34, 560–572.Find this resource:

                                                                                        Lau, R. R., & Pomper, G. M. (2002). Effectiveness of negative campaigning in U.S. Senate elections. American Journal of Political Science, 46(1), 47–66.Find this resource:

                                                                                          Lau, R. R., & Redlawsk, D. P. (2006). How voters decide: Information processing during election campaigns. Cambridge, U.K.: Cambridge University Press.Find this resource:

                                                                                            Lazarsfeld, P. N., Berelson, R. B., & Gaudet, H. (1944). The people’s choice: How the voter makes up his mind in a presidential campaign. New York: Columbia University Press.Find this resource:

                                                                                              Levitt, S. T. (1994). Using repeat challengers to estimate the effects of campaign spending on election outcomes in the U.S. House. Journal of Political Economy, 4, 777–798.Find this resource:

                                                                                                Lewis-Beck, M. S., & Stegmaier, M. (2007). Economic models of the vote. In R. Dalton & H. D. Klingemann (Eds.), The Oxford handbook of political behavior (pp. 518–537). Oxford: Oxford University Press.Find this resource:

                                                                                                  Lipset, S. M., & Rokkan, S. (1967). Party systems and voter alignments. New York: Free Press.Find this resource:

                                                                                                    Malhotra, N., Michelson, M. R., Rogers, T., & Valenzuela, A. A. (2011). Text messages as mobilization tools: The conditional effect of habitual voting and election salience. American Politics Research, 39, 664–681.Find this resource:

                                                                                                      Milgram, S. (1974). Obedience to authority: An experimental view. New York: Harper Row.Find this resource:

                                                                                                        Nickerson, D. W. (2005). Partisan mobilization using volunteer phone banks and door hangers. The Annals of the American Academy of Political and Social Science, The Science of Voter Mobilization, 601, 10–27.Find this resource:

                                                                                                          Nickerson, D. W. (2008). Is voting contagious? Evidence from two field experiments. American Political Science Review, 102, 49–57.Find this resource:

                                                                                                            Pulzer, P. (1967). Political representation and elections in Britain. London: Allen and Unwin.Find this resource:

                                                                                                              Rosenstone, S. J. (1983). Forecasting presidential elections. New Haven, CT: Yale University Press.Find this resource:

                                                                                                                Rothstein, B. (2005). Social traps and the problem of trust. Cambridge, U.K.: Cambridge University Press.Find this resource:

                                                                                                                  Sanders, D., Clarke, H. D., Stewart, M. C., & Whiteley, P. F. (2008). The endogeneity of preferences in spatial models: Evidence from the 2005 British election study. Journal of Elections, Public Opinion and Parties, 18(4), 413–431.Find this resource:

                                                                                                                    Seyd, P., & Whiteley, P. F. (1992). Labour’s grassroots: The politics of party membership. Oxford: Clarendon.Find this resource:

                                                                                                                      Shaw, D. R. (1999). The effect of TV ads and candidate appearances on statewide presidential votes, 1988–1996. American Political Science Review, 93, 345–361.Find this resource:

                                                                                                                        Sides, J., & Vavreck, L. (2013). The gamble: Choice and chance in the 2012 presidential election. Princeton, NJ: Princeton University Press.Find this resource:

                                                                                                                          Stevenson, R. T., & Vavreck, L. (2000). Does campaign length matter? Testing for cross-national effects. British Journal of Political Science, 30(2), 217–235.Find this resource:

                                                                                                                            Stokes, D. E. (1963). Spatial models of party competition. American Political Science Review, 57, 368–377.Find this resource:

                                                                                                                              Vavreck, L. (2009). The message matters: The economy and presidential campaigns. Princeton, NJ: Princeton University Press.Find this resource:

                                                                                                                                Wattenberg, M. P., & Brains, L. C. (1999). Negative campaign advertising: Demobilizer or mobilizer? American Political Science Review, 93(4), 891–899.Find this resource:

                                                                                                                                  Whiteley, P. F., Clarke, H. D., Stewart, M. C., & Sanders, D. (2012). Britain says NO: Voting in the AV referendum ballot. Parliamentary Affairs, 64(2), 301–322.Find this resource:

                                                                                                                                    Whiteley, P. F., Clarke, H. D., Stewart, M. C., & Sanders, D. (2013). Affluence, austerity and electoral change in Britain. Cambridge, U.K.: Cambridge University Press.Find this resource:

                                                                                                                                      Whiteley, P. F., Seyd, P., & Richardson, J. (1994). True blues: The politics of conservative party membership. Oxford: Oxford University Press.Find this resource:

                                                                                                                                        Zaller, J. (1992). The nature and origins of mass opinion. Cambridge, U.K.: Cambridge University Press.Find this resource:

                                                                                                                                          Zuckerman, A. S. (1982). New approaches to political cleavage: A theoretical analysis. Comparative Political Studies, 15(2), 131–144.Find this resource:


                                                                                                                                            (1.) The data is taken from the Essex Continuous Monitoring Survey, which is a set of monthly surveys conducted on the Internet since 2004 (see Clarke et al., 2016).