When Campaigns Can Backfire: National Identities and Support for Parties in the 2015 U.K. General Election in Scotland

Abstract

Research on campaign strategies generally assumes that political parties avoid campaigning on issues that are internally divisive. However, this strategy might not always be viable, especially when parties attack each other in high-stake elections. This article provides novel evidence on the effects of campaigning on cross-cutting issues by focusing on the 2015 U.K. general election in Scotland. Results based on an experiment and a nationally representative survey show that the strategy to criticize the Scottish National Party (SNP) with regard to the cross-cutting issue of Scottish independence polarized voters along national identity lines. Among British voters, attack statements and perceived negativity increased support for some of the parties sponsoring the attacks, whereas among Scottish voters they actually increased support for the target of the attacks. In addition, experimental results indicate that attack statements affected mainly ideologically close parties (the Labour Party and the SNP). At the theoretical level, these findings indicate that the strategy to attack opposite parties on divisive issues can lead to both electoral gains and losses depending on voters’ “identification” with such issues.

Keywords

voting behavior multiparty systems national identity negative campaigning U.K. elections

In election campaigns, political parties compete on numerous, contentious issues. Although the classic, Downsian approach assumes that parties would position themselves close to the preferences of the median voter (Downs 1957), other prominent accounts of electoral competition argue that parties compete by selectively emphasizing the issues on which they have an advantage (e.g., Budge and Farlie 1983; De Sio and Weber 2014; Green and Hobolt 2008; Petrocik 1996; Petrocik, Benoit, and Hansen 2003). Common to these approaches is the assumption that parties avoid campaigning on issues that are internally divisive or that divide their supporters. Yet, this strategy might not always be viable, especially during high-stake elections in which intense competition can bring contentious issues to the center stage. When parties attack each other on cross-cutting topics, how do voters react? Do they stick to their preferred party’s positions, or do they rather shift support toward other parties?

Despite extensive research on campaign strategies and recent studies on negative campaigning in multiparty systems (Nai and Walter 2015a), our knowledge of campaign effects in relation to issues that cut across party lines is limited. Examples of such issues include not only disputes over national independence in Catalonia (Rico and Liñeira 2014), Scotland (Johns et al. 2009), and Quebec (Clarke 1983; Nadeau, Martin, and Blais 1999), but also environmental (Carter 2006) and European Union (EU) integration issues (Green 2007) in the United Kingdom or ethical and socioeconomic issues in the United States (Abramowitz 1995; Baldassarri and Gelman 2008).

This study argues that when parties compete on cross-cutting issues, they might activate voters’ sense of “identification” with these issues and either gain or lose electoral support, depending on the stance taken by both voters and parties on these topics. Consider, for example, the case of two voters A and B both holding the same opinion on the issue of gay rights and both sharing party C’s stance on this issue. If party C decides to attack another party D on gay rights, party C might reinforce voters A and B’s identification with this specific issue and ultimately gain electoral support. On the contrary, imagine a scenario in which both voters A and B support party C, but only voter A shares party C’s position on gay rights, whereas voter B’s position on the issue is closer to the position of party D. In this case, party C’s strategy to attack party D on gay rights might succeed in securing voter A’s electoral support, but, at the same time, it might steer voter B closer to party D. This simple example summarizes the idea that the strategy of attacking a party in relation to cross-cutting issues can lead to opposite effects.

The 2015 U.K. general election provided an ideal case to test how voters reacted to a campaign in which parties attacked each other on several contested issues. Contrary to most of the forecasts, this election—in which the Conservative party obtained the absolute majority of seats in Parliament—changed the British political landscape strikingly. One of the major sources of change originated in Scotland, in which the Labour Party and the Liberal Democrats lost, respectively, forty and ten seats out of a total of fifty-nine, to the advantage of the Scottish National Party (SNP) that won all but three seats and became the third strongest party in the United Kingdom. To prevent this largely predicted defeat, the main Westminster parties (the Conservatives, Labour, and Liberal Democrats) adopted an aggressive campaign tactic against the SNP and the issue of Scottish independence, which constituted the SNP’s original raison d’être. The strategy was pursued by Labour in particular, in line with previous elections for the Scottish Parliament (Johns et al. 2010), despite the fact that Labour supporters were fairly split on the issue of independence, as opinion polls conducted right after the 2014 Scottish independence referendum indicated.¹ Did voters reward the parties sponsoring attack messages or did they instead reduce their support for these parties?

To answer this question, I conducted an experiment in which those assigned to treatment read four short texts in which party leaders criticized the SNP mostly with regard to the cross-cutting issue of Scottish independence. All the arguments presented in the texts had been selected from real party leaders’ statements. In addition to the experimental study, I analyze nationally representative data from the British Election Study (BES) Internet Panel (Fieldhouse et al. 2015). These data permit us to test whether respondents’ perceived tone of the campaign affected actual voting in line with the trend identified in the experimental setting.

The results indicate that the strategy of criticizing the SNP with regard to the issue of Scottish independence affected support for parties and voting behavior, but the effects differed across parties and depending on the element of British/Scottish national identity. Both exposure to attack statements in the experimental setting and perceived negativity in the panel survey increased support for some of the parties sponsoring the attacks (the Conservative and the Labour party) among those who identified as British. On the contrary, the same strategy of attack politics increased support for the target of the attacks (the SNP) among those who identified as Scottish. In other words, the strategy of criticizing the adversaries on the cross-cutting issue of Scottish independence proved at the same time beneficial and detrimental for some parties, thus leading to voters’ polarization. In addition, experimental results indicate that attack messages affected mainly ideologically close parties—that is, Labour and SNP—in line with previous research on parties’ ideological proximity (Curini and Martelli 2010; Haynes and Rhine 1998; Walter 2014a).

In the following two sections, I will briefly review some of the existing literature on the effects of negative campaigning on voting behavior. Although the experimental manipulation adopted in this study can be interpreted as a particular type of “competitive campaigning” that does not necessarily imply negativity, it is argued that it would fit within a broad, but commonly accepted definition of negative campaigning—that is, “talking about the opponent—criticizing his or her programs, accomplishments, qualifications, and so on” (Lau and Pomper 2001, 805–806)—and especially within a strand of research that focuses on “issue-based” attacks instead of “trait-based” (personal) attacks (Brooks and Geer 2007; Finkel and Geer 1998; Lau and Pomper 2001; Ridout and Walter 2015; Walter 2014b). In this sense, the literature on negative campaigning can provide us with a suitable theoretical background from which a series of hypotheses can be derived.

Effects of Negative Campaigning: A Mixed Picture

Research on the effects of negative campaigning on vote choice has yielded mixed results. Although experimental research has shown that exposure to negative ads (Kaid 1997) and negative messages (Fridkin and Kenney 2004) increases support for the attacker, other studies have demonstrated that attack advertisements actually reduce support for their sponsors (Haddock and Zanna 1997; Houston, Doan, and Roskos-Ewoldsen 1999; Maier and Maier 2007; Sonner 1998). Additional experimental studies suggest that candidates do not generally benefit from attacking political adversaries (Arceneaux and Nickerson 2010; Min 2004). Although scholars have also pointed to a series of positive effects of attack strategies—because “negativity carries valuable information . . . that voters otherwise would not have received” (Mattes and Redlawsk 2014, 25; see also Geer 2006)—the blunt conclusion of a meta-analysis of 111 studies is that “there is no consistent evidence in the research literature that negative political campaigning ‘works’ in achieving the electoral results that attackers desire” (Lau, Sigelman, and Rovner 2007, 1185).

Furthermore, our knowledge of these effects remains confined mostly to elections for candidates in the United States, although evidence indicates that this phenomenon occurs also in elections for parties in several other countries (Curini and Martelli 2010; de Nooy and Maier 2015; Hansen and Pedersen 2008; Nai and Walter 2015b; Ridout and Walter 2015; Sigelman and Shiraev 2002; Walter 2014b). As argued by Pattie and colleagues (2011, 335), “reactions to negative campaigning may be very different when the perpetrators and victims are parties rather than people.” In the following section, a set of hypotheses will be drawn on the basis of existing findings on negative campaigning in the United Kingdom.

Voters’ Reaction to Negativity in the United Kingdom

Studies of recent past elections provide consistent evidence that campaigns influence turnout and voting behavior in Britain (Clarke et al. 2004; Denver, Hands, and MacAllister 2004; Fisher et al. 2016; Johnston et al. 2013; Pattie, Johnston, and Fieldhouse 1995; Whiteley and Seyd 1994). With regard to campaign tone, research has highlighted a tendency toward increased negativity in the 1997 and the 2001 U.K. general elections (Hodess, Tedesco, and Kaid 2000; Pipkin and Bartle 2002), as well as the 2010 general election in Scotland (Nai and Walter 2015b, 17). Furthermore, recent studies have shown that in the United Kingdom parties tend to go negative, especially when they are losing at the polls (Walter 2014a; Walter, van der Brug, and van Praag 2014). These trends are in line with evidence indicating that candidates with challenger status (Druckman, Kifer, and Parkin 2009; Lau and Pomper 2001) or lagging behind in the polls (Damore 2002) are more likely to adopt attack strategies. In the case of the 2015 U.K. general election in Scotland, the main Westminster parties resorted to attack statements mainly as a response to the “incumbent” status of the SNP in Scotland,² in addition to the predicted landslide victory for that party, according to opinion polls.

A few studies provide theoretical expectations about the effect of negativity on voting behavior in the United Kingdom. With regard to the 1997 British general election, research by Sanders and Norris (1998; see also Norris et al. 1999) shows that positive television campaign messages were more effective than negative messages in shaping British voters’ political perceptions. In a subsequent study on the 2001 general election, the authors find not only that negativity was not effective but also that in some cases “negative campaigning was explicitly counter-productive in the sense that it appears to have actively stimulated sympathy for the target of the attack” (Sanders and Norris 2005, 525).

Such a “reverse effect” finds confirmation also in a study by Pattie and colleagues (2011) on the 2007 election for the Scottish Parliament. In this study, the authors underline that in some cases negativity increases voters’ support for the targets of the attacks. In addition, they demonstrate that this campaign strategy can backfire, as “the more negative a party’s campaign is perceived to have been, the less likely people are to vote for the party” (Pattie et al. 2011, 342). Research from social psychology provides additional evidence that attack messages and negative advertisements can have a backlash effect because they can lead to unfavorable evaluation of the sponsor of the messages (Matthews and Dietz-Uhler 1998; Roese and Sande 1993).

Although reverse effects and backlash effects can be interpreted as two sides of the same coin, they do not automatically occur together, especially in multiparty systems. If voters disapprove of a party that resorts to attack statements, it does not necessarily follow that they will increase their support for the target of the attacks, especially when attack statements concern divisive issues. On the contrary, they might become either less likely to go to vote in general or more likely to vote for any other party competing for the elections. In the context of this study, this leads to the following assumptions:

Hypothesis (H1): A backlash effect: the strategy of attacking the SNP reduces voters’ support for the parties sponsoring the attacks.

Hypothesis (H2): A reverse effect: the strategy of attacking the SNP increases voters’ support for the SNP.

Extant research, however, suggests that campaign strategies can have different effects at both party level and individual level. In the former case, evidence indicates that parties are more likely to attack each other if they are ideologically close (Curini and Martelli 2010; Haynes and Rhine 1998; Walter 2014a). In the context of the 2015 U.K. general election in Scotland, this evidence suggests that Labour should have been more likely not only to attack the SNP (due to the ideological proximity between the parties) but also to suffer from a potential reverse effect of a campaign strategy that focused on the SNP’s main issue, that is, Scottish independence. The assumption is that voters are more likely to support the target of the attacks if this target is ideologically close to the party sponsoring the attacks. Evidence from Pattie and colleagues (2011, 340) partially confirms this assumption by showing that in the 2007 election for the Scottish Parliament, “those who thought that Labour had campaigned negatively came to like the SNP more.” Thus, if a reverse effect is likely to occur, this effect should concern especially the Labour party with regard to the SNP.

Hypothesis (H3): Ideological proximity: the strategy of attacking the SNP increases voters’ support for the SNP, mostly at the expense of the Labour party.

With regard to the individual level, one of the key factors explaining voting behavior in Scotland concerns national identity feelings. Correlations between Scottish identity and support for the SNP have been documented for the 1997 U.K. general election (Brown et al. 1999, 64–65) and for the 2011 election for the Scottish parliament (Carman, Johns, and Mitchell 2014, 40; McCrone and Bechhofer 2015, 197), as well as for the 1997 Scottish devolution referendum (Pattie et al. 1999) and the 2014 Scottish independence referendum (Bond 2015). If we assume that such a correlation applied also in the 2015 U.K. general election in Scotland, it can be argued that voters who identify as Scottish would have been more likely to react to campaign statements that criticized the SNP in relation to Scottish independence, because these attacks pose an indirect threat to their “group identity,” in line with social identity theory (Capozza and Brown 2000; Hogg and Abrams 1988; Tajfel 1981). Experimental research on individual responses to group-identity threats in electoral contexts indicates that, after receiving negative information about their own group, those who identify strongly with the group tend to react defensively and criticize out-group members (Branscombe and Wann 1994; Crocker et al. 1987; Matthews and Dietz-Uhler 1998). Therefore, it can be assumed that those who identify as Scottish will be more likely to react to attack messages against the SNP by both increasing their support for the SNP and reducing support for the parties sponsoring the attacks. In other words, the effects of the treatment should be larger among Scottish participants.

Hypothesis (H4): National identity: the strategy of attacking the SNP leads to a particular backlash effect and a reverse effect among those who identify as Scottish.

Design

The analysis relies on an original experimental study that I carried out during the last two weeks of the campaign in April and May 2015, in addition to observational data from the BES Internet Panel (Fieldhouse et al. 2015).

Experimental Design³

The experiment was conducted online and involved 201 participants recruited by BLUE Lab, University of Edinburgh. The pool of subjects included mostly university students⁴ (86% of students, mean age = 25.3 years, SD = 9.5) who had been living in Scotland already for several years (median number of years = 7). As a crucial requirement for participation in the study, all the participants had to be eligible to vote and be a resident in Scotland.

The experiment consisted of a basic between-subjects design with two conditions. After replying to a few preliminary questions, which included party identification and national identity,⁵ the participants were randomly assigned to either a control or a treatment condition. In the control condition, they were briefly shown a summary table with the positions of the four main parties—the Conservatives, Labour, Liberal Democrats, and SNP—on five issues on which parties mostly agreed on. This table aimed to introduce the participants in a nonconflictual scenario by showing that a certain level of agreement existed across parties on some campaign issues. After this step, they read four short texts containing arguments in favor of each of the main parties (see text below). These arguments had been selected from party leaders’ statements reported in mainstream media in the months of March and April 2015, and the participants were explicitly instructed that the text derived only from publicly available information. Crucially, these texts did not include any mentions of either political leaders or opposing parties; thus, they could be defined as “advocacy messages” (Sanders and Norris 2005, 526) as they were making a positive appeal to voters.

The participants assigned to the treatment condition were also presented with a brief table with party positions on five conflictual issues, which aimed to introduce the participants to a conflictual campaign environment. In the following step, the participants were asked to read the same texts presented in the control condition, but with an additional paragraph in which party leaders attacked the SNP mostly with regard to the issue of Scottish independence, and encouraged voters not to support that party for issue-based reasons. An additional paragraph in which SNP party leaders discouraged voters from supporting all the other main parties was also included in the text supporting the SNP. All the information materials in the treatment conditions were approximately 200 words long. The following is the text with the statements supporting the Labour party—the part in brackets was shown only in the treatment condition:

Labour: “We are the only anti-austerity party in Scotland”

Scotland will decide the outcome of this election. If Scotland votes Labour, then Britain will have a Labour government and we can start the job of rebuilding our country.

We have a radical plan to transform Scotland, so that working-class families get a fair shot at life. A Labour government will introduce a Mansion Tax on properties worth £2 million so we can fund 1,000 more nurses in Scotland’s NHS. And we will call time on exploitative zero hours contracts that cause so much harm particularly to young Scots in work.

If you vote Labour, you’re voting for a completely different approach to the management of the economy, for a party that cares about the fact that inequality has grown, that can show through our record that we’re committed to proper investment in public services.

[A vote for the SNP is one for austerity, because it makes it more likely that the Tories remain in power, and because the SNP remains wedded to a fiscal approach for Scotland which rejects the pooling and sharing of resources across the United Kingdom. Their ultimate goal remains the same. It’s independence. And Labour is not going to put at risk the unity of the United Kingdom in a coalition with the SNP.]

Only Labour can credibly claim to be the anti-austerity party of Scotland. If you vote Labour, you’re voting to end austerity in Scotland.

After completing the reading task, the participants in both conditions replied to a final battery of questions, including the dependent variables, that is, voting intentions and support for parties. The latter variable was measured by asking the participants to rate the “likeability” of the four main parties on a standard eleven-point scale commonly used by the BES.⁶

Panel Survey

Additional data from the BES Internet panel permit tests of the effect of campaign tone on the actual vote within a representative sample of voters. In the following analysis, I will consider three waves of the panel—wave 4 (February/March 2015), preelection wave 5 (April/May 2015), and postelection wave 6 (May 2015)—and only the responses from those who were resident in Scotland.⁷ The dependent variable corresponds to the actual vote as recalled by respondents in wave 6.

Wave 5 included the key independent variable, that is, voters’ perception of negative campaigning. The respondents were asked to rate on different five-point scales to what extent the Conservatives, Labour, Liberal Democrats, and the SNP focused in their campaign either on criticizing other parties or on putting forward their own policies.⁸ To guarantee comparability with the experimental findings, the four scales have been combined in a single additive index of perceived negativity, with values rescaled from 0 (minimum level of negativity) to 1 (maximum level of negativity). Although voters rated each party’s campaign differently depending on party identification, this index provides a cumulative measure of perceived negativity of the general campaign, which can be reasonably compared with the operationalization used in the experimental setting.

The other key independent variables introduced in the analysis are party identification and national identity as reported in wave 4.⁹ The latter variable has been measured with two separate seven-point scales of “Scottishness” and “Britishness.” To facilitate comparability with experimental data, these two scales have been combined in a single index of “Scottish versus British” identity by subtracting the values of the Britishness scale from the Scottishness scale. The final index has been recoded in a dummy variable, with value 0 for those feeling more British than Scottish and value 1 for those feeling either equally Scottish and British or more Scottish than British.

Results

In the following paragraphs, I will present, first, the results on support for parties based on experimental data and, second, the results on voting based on both experimental and panel data.

Support for Parties

Figure 1 presents the distribution of party likeability scores by control and treatment group for the entire pool (first row) and by national identity categories (second and third rows). As the bar graphs in the first row highlight, the only sizeable difference between control and treatment groups concerns support for the Labour party, which decreased in the attack-statement condition compared with the control condition. However, none of the differences between mean values in the first row reach a common level of statistical significance (see t tests in Table A1 in online appendix A).

Figure 1.

Support for parties by control and treatment groups and by identity categories.

When respondents are divided by national identity categories, the bar graphs in the second and third row provide a completely different picture. Among the participants who identified as British, attack statements increased support for all the main Westminster parties and reduced support for the SNP, in comparison with a control condition of positive campaigning. Although confidence intervals overlap in the graph, two-tailed t tests confirm that the differences between control and treatment group are statistically significant for all the three Westminster parties (see Table A1 in online appendix A).

On the contrary, an opposite pattern occurred within the group of those who described themselves as either Scottish or equally British and Scottish. As the bar graphs in the third row clearly illustrate, in this case the experimental manipulation actually reduced support for all the Westminster parties, especially for the Conservative and the Labour party, while it increased support for the SNP. The results in Figure 1, therefore, indicate that attack statements concerning mostly Scottish independence affected support for parties in opposite directions depending on participants’ self-reported national identity.

Table 1 reports additional analysis on party likeability scores using ordinary least squares (OLS) regressions. This type of analysis allows to control for covariates that might influence the size of treatment effects, such as respondents’ party identification as measured before the treatment. The key independent variable “Attack statement” corresponds to those who have been assigned to the treatment group, whereas the reference category corresponds to the control group.

Table 1.

The Effect of Attack Statements on Party Likeability (Experimental Data).

	Like/dislike scores
	(1)Con.	(2)Lib Dem	(3)Labour	(4)SNP	(5)Con.	(6)Lib Dem	(7)Labour	(8)SNP
Attack statement	0.01 (0.03)	0.02 (0.03)	0.01 (0.03)	−0.04 (0.03)	0.08* (0.05)	0.06 (0.04)	0.13*** (0.04)	−0.04 (0.05)
Scottish identity (vs. British)					0.05(0.05)	0.07(0.04)	0.12**(0.05)	0.10*(0.05)
Attack Statement × Scottish Identity					−0.13*(0.07)	−0.08(0.06)	−0.21***(0.06)	0.01(0.07)
Party ID
Conservative	0.47***(0.05)	0.09**(0.04)	−0.10**(0.04)	−0.33***(0.05)	0.44***(0.05)	0.06(0.05)	−0.14***(0.05)	−0.32***(0.05)
Lib Dem	0.15***(0.06)	0.30***(0.05)	0.04(0.05)	−0.22***(0.06)	0.08(0.06)	0.28***(0.05)	0.03(0.05)	−0.18***(0.06)
Labour	0.07*(0.04)	0.07*(0.04)	0.21***(0.04)	−0.08*(0.04)	0.01(0.05)	0.06(0.04)	0.20***(0.04)	−0.05(0.05)
SNP	−0.13***(0.04)	−0.11***(0.04)	−0.05(0.04)	0.27***(0.04)	−0.17***(0.05)	−0.16***(0.05)	−0.07(0.05)	0.26***(0.05)
Constant	0.24***(0.03)	0.40***(0.03)	0.49***(0.03)	0.49***(0.03)	0.26***(0.04)	0.39***(0.03)	0.44***(0.04)	0.53***(0.04)
R ²	.426	.226	.191	.415	.463	.276	.267	.463
N	201	200	201	198	168	168	168	166

OLS regressions, experimental data. Dependent variable: party likeability scores measured on an eleven-point scale, with values rescaled from 0 (strongly dislike) to 1 (strongly like). Standard errors in parentheses. Con. = Conservatives; Lib Dem = Liberal Democrats; SNP = Scottish National Party.

p < .1. **p < .05. ***p < .01.

As expected, in all models, party identification is positively correlated with the rating of the corresponding party, meaning that, for example, those who feel close to the Conservatives give a higher rating to that party compared with those who identify with other parties or with no parties. Although attack statements are not associated with party likeability scores in models 1–4, when national identity is introduced in models 5–8, significant correlations occur between assignment to treatment and support for parties. As regressions coefficients in models 5 and 7 indicate, after reading both advocacy messages and negative statements, those who identified as British gave a significantly higher rating to the Conservatives and Labour compared with a control condition in which the participants read only advocacy messages. On the contrary, the negative interaction coefficients reveal that attack statements significantly reduced support for the Conservatives and Labour among those who identified as Scottish compared with those who felt only British.

The fact that the interaction coefficient is larger in the case of Labour indicates that attack statements affected support especially for this party, in line with predictions. Further analysis with marginal effects presented in Figure 2 confirms that the strategy of criticizing the SNP boosted Labour support among British participants and, at the same time, reduced Labour support among Scottish participants, compared with a control condition of advocacy messages.

Figure 2.

Average marginal effects of attack statements on likeability of Labour party.

With regard to the Labour party, therefore, the strategy to criticize the SNP mostly in relation to Scottish independence generated a polarizing effect, as exposure to treatment made British and Scottish participants more far apart in their support of the party. This polarizing effect finds additional confirmation when party ratings are compared among each other. The results of this analysis reveal that attack statements affected support especially for Labour relative to the SNP (see Table A2 in online appendix A).

The findings presented so far indicate that the experimental manipulation affected support for parties in more complex directions than initially assumed. When only Scottish participants are taken into account, the analysis confirms the occurrence of the hypothesized backlash effect (H1) with regard to the Conservative and the Labour party, as exposure to attack statements reduced support for these parties. When the party rating gap is considered, the analysis suggests that an indirect reverse effect (H2) occurred with regard to the SNP, as participants increased support for this party relative to Labour. This result partially confirms the assumption that reverse effects of negative campaigning are more likely to occur among parties that are ideologically close (H3). Yet, when only British participants are considered, the results show that attack statements actually benefited the sponsor of the attacks, especially when the target (the SNP) and the sponsor (Labour) are ideologically close, thus contradicting the original hypotheses. The fact that British and Scottish participants reacted differently to the treatment supports the assumption that national identity plays a crucial role in voters’ response to campaign statements focusing on the cross-cutting issue of Scottish independence, although only those who identified as Scottish responded in line with theoretical expectations (H4).

Voting Behavior

Following previous analysis of vote choice in the 2011 Scottish parliament election (Carman, Johns, and Mitchell 2014), in the next step I ran multinomial logistic regressions to test the effect of attack statements in the experimental setting and perceived negativity in the panel survey on actual voting. Table 2 presents results from experimental data, in which the dependent variable takes separated values for the intentions to vote for each of the four parties included in the reading task. The models reported in the table estimate whether the participants in the treatment group (conditional on their national identity) have a higher or lower probability of voting for one of the three Westminster parties rather than to vote for the SNP, which constitutes the base outcome.¹⁰

Table 2.

The Effect of Attack Statements on Voting Intentions (Experimental Data).

	Voting intentions
	Model 1(base outcome = SNP)			Model 2(base outcome = SNP)
	Conservative	Lib Dem	Labour	Conservative	Lib Dem	Labour
Attack statement	0.69(1.00)	1.39(1.22)	0.47(0.95)	−1.12(1.81)	2.16(1.66)	2.22*(1.35)
Scottish identity(vs. British)	−1.47*(0.79)	−1.18(1.15)	−1.58**(0.70)	−1.33(1.44)	−0.46(1.42)	−0.09(1.05)
Attack Statement × Scottish Identity	−3.29**(1.51)	−1.78(1.56)	−2.19*(1.17)	−1.02(2.36)	−1.98(2.12)	−3.49**(1.75)
Party ID
Conservative				4.91***(1.61)	−17.06(10,131.2)	−1.04(1.82)
Lib Dem				0.77(1.49)	3.59***(1.29)	0.19(1.05)
Labour				−14.59(2,959.9)	2.66(1.79)	3.69***(1.16)
SNP				−0.55(1.63)	−17.38(4,092.4)	−18.06(2,020.5)
Constant	0.69(0.61)	−0.69(0.87)	1.32**(0.56)	−0.67(0.92)	−2.64*(1.38)	−0.27(0.77)
Pseudo R²	.164			.637
N	123			123

Multinomial logistic regressions, experimental data. Standard errors in parentheses. SNP = Scottish National Party; Lib Dem = Liberal Democrats.

p < .1. **p < .05. ***p < .01.

In model 1, the interaction terms indicate that among Scottish participants (compared with British participants) attack statements significantly reduced the probability of voting for the Conservatives and Labour, relative to the probability of voting for the SNP. When party identification is introduced in model 2, the interaction coefficient remains statistically significant only with regard to the probability of voting for Labour. In line with previous analysis of party likeability, results from model 2 highlight that after exposure to attack messages, British participants became more likely to vote for Labour and less likely to vote for the SNP compared with the control condition. On the contrary, the treatment significantly reduced the intentions to vote for Labour and increased the intentions to vote for the SNP among Scottish participants compared with British, thus partially supporting the hypothesized backlash (H1) and reverse (H2) effects for this group of voters. The fact that a significant trade-off occurs only between the Labour and the SNP further supports the idea that attack messages affect mostly ideologically close parties (H3).

Additional analysis of BES panel data allows us to identify whether perceived negativity of the campaign also affected the actual vote in different directions, depending on respondents’ feelings of “Britishness” or “Scottishness.” A preliminary analysis of voters’ perceived negativity confirms that the campaign conducted by the SNP was perceived as the most positive at both the sample level and after weighting for voters’ party identification (see Table A3 in online appendix B), thus supporting the expectation that parties that are leading in the polls are less likely to engage in attack politics (Walter 2014a; Walter, van der Brug, and van Praag 2014).¹¹

Table 3 reports the results from multinomial logistic regressions that mirror the models previously adopted in Table 2. As mentioned in the design part, the key independent variable is a cumulative index of perceived negativity for the campaigns of the Conservatives, Liberal Democrats, Labour, and the SNP as measured in preelection wave 5, whereas the outcome is the actual vote for these parties as recalled by the participants in postelection wave 6. The aim of the analysis, therefore, is to test whether perceived negativity of the campaign in general affected vote choice in the same directions as the experimental findings indicate.

Table 3.

Perceived Negativity and Vote (BES Internet Panel).

	Actual vote
	Model 1(base outcome = SNP)			Model 2(base outcome = SNP)
	Conservative	Lib Dem	Labour	Conservative	Lib Dem	Labour
Perceived negativity	2.98**(0.95)	0.11(0.96)	2.67**(0.80)	3.17*(1.23)	−0.17(1.20)	2.26*(1.06)
Scottish identity (vs. British)	0.72(0.62)	−0.33(0.60)	0.55(0.49)	2.35**(0.77)	0.92(0.73)	1.45*(0.63)
Perceived negativity × Scottish identity	−3.60**(1.07)	−2.16*(1.10)	−3.41***(0.87)	−4.77***(1.35)	−3.07*(1.34)	−3.60**(1.15)
Party ID
Conservative	3.52***(0.22)	2.41***(0.26)	1.39***(0.23)	2.03***(0.30)	1.35***(0.33)	0.71*(0.30)
Lib Dem	0.32(0.36)	2.68***(0.26)	0.98***(0.26)	−0.03(0.45)	2.44***(0.35)	0.77*(0.34)
Labour	−0.78**(0.23)	0.39^†(0.21)	1.79***(0.13)	−0.48(0.29)	0.43(0.27)	1.25***(0.19)
SNP	−3.71***(0.41)	−3.66***(0.51)	−3.45***(0.28)	−2.16***(0.47)	−2.26***(0.57)	−2.14***(0.32)
Yes vote in future referendum				−2.58***(0.29)	−2.42***(0.28)	−2.30***(0.17)
Trust in MPs				2.18***(0.40)	0.98*(0.41)	0.95**(0.32)
Scottish government approval				−4.63***(0.44)	−4.52***(0.45)	−4.75***(0.38)
Performance of future Labour government				2.73***(0.44)	−1.06*(0.43)	0.96**(0.34)
Interest in elections				−0.05(0.44)	0.54(0.46)	0.77*(0.36)
Constant	−1.92**(0.56)	−0.66(0.54)	−1.13*(0.45)	1.10(0.89)	2.21*(0.86)	0.98(0.73)
Pseudo R²	.410			0.551
N	3,738			3,509

Multinomial logistic regressions, BES Internet Panel (Fieldhouse et al. 2015). Perceived negativity: values from 0 (positive) to 1 (negative). Trust in MPs: from 0 (no trust) to 1 (a great deal of trust). Approval of Scottish government: from 0 (strongly disapprove) to 1 (strongly approve). Performance of future Labour government: replies to the question “If there were a Labour U.K. government today, do you think that [the economy] would be getting better, getting worse or staying about the same?” Answers from 0 (getting a lot worse) to 1 (getting a lot better). Interest in elections: from 0 (not at all interested) to 1 (very interested). Standard errors in parentheses. BES = British Election Study; SNP = Scottish National Party.

†

p < .1. *p < .05. **p < .01. ***p <.001.

The results from model 1 show that, in the case of British respondents, the more negative they evaluated the campaign, the more likely they were to vote for the Conservatives and Labour rather than to vote for the SNP, after controlling for preelection party identification. However, perceived negativity actually reduced the probability of voting for the Conservatives and Labour among Scottish respondents compared with British, as indicated by the negative interaction terms. The same results apply in model 2 after controlling for constitutional preferences and key elements of “performance politics” that previous research has shown to be associated with support for the SNP (Carman, Johns, and Mitchell 2014; Curtice et al. 2009; Johns, Mitchell, and Carman 2013). Results based on panel data analysis, therefore, confirm that campaign tone affected vote choice in the 2015 U.K. general election in Scotland, but voters’ national identity crucially moderated the effect.¹²

Final analysis of marginal effects contributes to better understanding of the role of national identity. Figure 3 illustrates the average marginal effects of campaign strategies on voting for Labour and SNP calculated on full multinomial regression models in Tables 2 and 3. The plots reveal a striking similarity between experimental and panel data and highlight a polarizing effect. Compared with a baseline of positive campaigning, attack statements and perceived negativity increased the probability of voting for Labour by around 13 percentage points among British voters (Figure 3A and 3C), whereas they increased the probability of voting for the SNP by around 10–13 percentage points among Scottish voters (although this effect is statistically significant only in panel data analysis; Figure 3D). This final set of results confirms that both exposure to attack messages and perceived negativity had a similar effect on voting, as they shifted support for Labour and SNP in opposite directions depending on respondents’ feelings of “Britishness” or “Scottishness.”

Figure 3.

Average marginal effects of attack statements and perceived negativity on voting for Labour and SNP by national identity.

Conclusion

This study set out by investigating how voters react to election campaigns in which parties attack each other on divisive issues. Although models of electoral competition generally assume that parties avoid campaigning on such issues, in reality this strategy might not always be viable, especially during high-stake campaigns in which political opponents attack each other on several, contested topics. The 2015 U.K. general election in Scotland provided an ideal case study, as one of the crucial issues of the campaign—namely, the dispute over Scottish independence—divided voters within some of the major parties (especially the Labour party). Findings from an original experiment and a nationally representative survey show that attack messages and campaign tone significantly affected voting behavior and support for parties, but the same strategy of attack politics led to opposite effects depending on voters’ feelings of national identity. While those who identified as British increased their support for some of the parties sponsoring the attacks—that is, the Conservative and the Labour party—those who identified as Scottish, on the contrary, increased their support for the target of the attacks—that is, the SNP. In this sense, the findings suggest that campaigning on an issue that cross-cut the electorate polarized voters along identity lines, as it widened the gap between SNP supporters, on the one hand, and Labour and Conservative supporters, on the other. In line with previous research on parties’ ideological proximity (Curini and Martelli 2010; Haynes and Rhine 1998; Walter 2014a), experimental results show that attack statements affected mainly those parties that are ideologically close, that is, the Labour party and the SNP.

These findings also contribute to explaining why the campaign strategy that was pursued by the main Westminster parties failed to prevent the SNP’s landslide victory in the 2015 British general election. A possible explanation lies in the unequal distribution of voters by national identity feelings. If we consider that a few months before the elections only a fifth of voters in Scotland felt more British than Scottish, while around a half felt more Scottish than British,¹³ the strategy of attacking the SNP on the issue of Scottish independence might well have been more counter-productive than beneficial. As the results in this article suggest, attack politics certainly did not benefit the main Westminster parties with regard to the wide share of voters who identified as Scottish, and, if anything, it brought this share of voters even closer to the SNP, in line with recent evidence on the effect of the Scottish independence referendum on voting behavior in the 2015 U.K. general election (Fieldhouse and Prosser 2018).

In interpreting the implications of these results, it is important to take into account the limitations of the research design. While mimicking the context of a real campaign in which political parties attack each other simultaneously, the experimental design presented in this study does not allow us to disentangle the effect of each party’s campaign strategy. Thus, it is possible that attack messages from only one of the parties included in the treatment led to the observed effects. In addition, the use of party leaders’ real statements as information stimuli limits the possibility to disentangle the effects of the different components of the messages, such as the “pure” attack dimension and the mentioning of Scottish independence.¹⁴ Furthermore, we need to keep in mind that the experimental measure of negative campaigning differs from the observational measure, as the former is an exogenous manipulation, whereas the latter is a self-reported measure of perceived negativity. Thus, we cannot infer that the panel respondents who evaluated the campaign more negatively were necessarily exposed to more negative messages. What we can observe, however, is that both exposure to attack statements and perception of campaign tone influenced support for parties and voting behavior in similar directions.

At the theoretical level, these findings indicate a possible mechanism that can contribute to understanding voters’ reaction to high-stake campaigns in which parties attack each other on contentious topics. This mechanism concerns the activation of voters’ “identification” with such topics. If voters’ positions on an issue and party identification largely overlap—such as in the case of a single-issue party that supports the interests of a specific ethnic group or social class—attacking the out-group in relation to such issue might ultimately reinforce in-group membership and result in electoral benefits. However, the reverse should also hold, when supporters of the same party are divided along a crucial issue that is at the center of a campaign. In this case, the strategy of attacking a party on such a divisive issue can push a share of voters away from the party sponsoring the attacks, especially if voters’ identification with the issue overcomes party identification. This mechanism seems to explain why a share of voters who identified as Scottish moved away from Labour when this party indirectly threatened their Scottish identity by attacking the SNP especially with regard to the issue of Scottish independence.

A final question concerns the extent to which these findings can “travel” beyond the Scottish borders. There is no doubt that the increasing relevance of the issue of independence has given a peculiar dimension to Scottish politics in recent years. It is possible, therefore, that attack statements targeting the SNP had a particular influence on voting, due to heated national identity sentiments in the electorate. However, we cannot exclude that similar dynamics would apply also in other contexts, especially in other countries in which cross-cutting feelings of (sub)national identities divide the electorate.¹⁵ Furthermore, we cannot exclude that the findings of this study would apply also to other types of cross-cutting issues related, for example, to ethnic, gender, or socioeconomic differences. As the findings from this study suggest, when voters’ issue positions cut across party lines, attacking political opponents on such issues can trigger unexpected reactions in the electorate and lead to both gains and losses on election day.

Supplemental Material

Morisi_PRQ-2017-0042.R2_-_Supplemental_Material_online_supp – Supplemental material for When Campaigns Can Backfire: National Identities and Support for Parties in the 2015 U.K. General Election in Scotland

Supplemental material, Morisi_PRQ-2017-0042.R2_-_Supplemental_Material_online_supp for When Campaigns Can Backfire: National Identities and Support for Parties in the 2015 U.K. General Election in Scotland by Davide Morisi in Political Research Quarterly

Footnotes

Acknowledgements

The author thanks Diego Gambetta for valuable feedback and for inspiring the design of this study. The author also thanks Stefaan Walgrave, three anonymous reviewers and the participants at the European University Institute (EUI) Colloquium on Political Behavior, the 2015 Political Psychology Conference (Amsterdam), the 2016 MPSA Conference (Chicago), and the 2016 EPSA Conference (Brussels) for useful comments.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding for this work was provided by the European University Institute, whereas valuable technical support was provided by Michèle Belot and Ivan Salter at BLUE Lab, University of Edinburgh.

Notes

Supplemental Materials

Supplemental materials for this article are available with the manuscript on the Political Research Quarterly (PRQ) website. Replication data are available at .

References

Abramowitz

Alan I.

1995. “It’s Abortion, Stupid: Policy Voting in the 1992 Presidential Election.” Journal of Politics 57 (1): 176–86.

Arceneaux

Kevin

Nickerson

David W.

2010. “Comparing Negative and Positive Campaign Messages: Evidence from Two Field Experiments.” American Politics Research 38 (1): 54–83.

Baldassarri

Delia

Gelman

Andrew

. 2008. “Partisans without Constraint: Political Polarization and Trends in American Public Opinion.” American Journal of Sociology 114 (2): 408–46.

Bond

Ross

. 2015. “National Identities and the 2014 Independence Referendum in Scotland.” Sociological Research Online 20 (4): 1–13.

Branscombe

Nyla R.

Wann

Daniel

. 1994. “Collective Self-esteem Consequences of Out-group Derogation When a Valued Social Identity Is on Trial.” European Journal of Social Psychology 24:641–657.

Brooks

Deborah Jordan

Geer

John G.

2007. “Beyond Negativity: The Effects of Incivility on the Electorate.” American Journal of Political Science 51 (1): 1–16.

Brown

Alice

McCrone

David

Paterson

Lindsay

Surridge

Paula

. 1999. The Scottish Electorate: The 1997 General Election and beyond. London: Macmillan Press.

Budge

Ian

Farlie

Dennis J.

1983. Explaining and Predicting Elections: Issue Effects and Party Strategies in Twenty-three Democracies. London: George Allen and Unwin.

Capozza

Dora

Brown

Rupert

, eds. 2000. Social Identity Processes: Trends in Theory and Research. London: SAGE.

10.

Carman

Christopher

Johns

Robert

Mitchell

James

. 2014. More Scottish Than British. New York: Palgrave Macmillan.

11.

Carter

Neil

. 2006. “Party Politicization of the Environment in Britain.” Party Politics 12 (6): 747–67.

12.

Clarke

Harold D.

1983. “The Parti Québécois and Sources of Partisan Realignment in Contemporary Quebec.” Journal of Politics 45 (1): 64–85.

13.

Clarke

Harold D.

Sanders

David

Stewart

Marianne C.

Whiteley

Paul

. 2004. Political Choice in Britain. Oxford: Oxford University Press.

14.

Crocker

Jennifer

Thompson

Leigh L.

McGraw

Kathleen M.

Ingerman

Cindy

. 1987. “Downward Comparison, Prejudice, and Evaluations of Others: Effects of Self-esteem and Threat.” Journal of Personality and Social Psychology 52:907–16.

15.

Curini

Luigi

Martelli

Paolo

. 2010. “Ideological Proximity and Valence Competition. Negative Campaigning through Allegation of Corruption in the Italian Legislative Arena from 1946 to 1994.” Electoral Studies 29 (4): 636–47.

16.

Curtice

John

McCrone

David

McEwen

Nicola

Marsh

Michael

Ormston

Rachel

. 2009. Revolution or Evolution? The 2007 Scottish Elections. Edinburgh, UK: Edinburgh University Press.

17.

Damore

David F.

2002. “Candidate Strategy and the Decision to Go Negative.” Political Research Quarterly 55 (3): 669–85.

18.

De Nooy

Wouter

Maier

Jürgen

. 2015. “When Do Attacks Work? Moderated Effects on Voters’ Candidate Evaluations in a Televised Debate.” In New Perspectives on Negative Campaigning: Why Attack Politics Matter, edited by Nai

Alessandro.

Walter

Annemarie

287–306. Colchester, UK: European Consortium for Political Research (ECPR) Press.

19.

Denver

David

Hands

Gordon

MacAllister

Ian

. 2004. “The Electoral Impact of Constituency Campaigning in Britain, 1992–2001.” Political Studies 52:289–306.

20.

De Sio

Lorenzo

Weber

Till

. 2014. “Issue Yield: A Model of Party Strategy in Multidimensional Space.” American Political Science Review 108 (4): 870–85.

21.

Downs

Anthony

. 1957. An Economic Theory of Democracy. New York: Harper & Row.

22.

Druckman

James N.

Kam

Cindy D.

2011. “Students as Experimental Participants: A Defense of the ‘Narrow Data Base’.” In Cambridge Handbook of Experimental Political Science, edited by Druckman

James N.

Green

Donald P.

Kuklinski

James H.

Lupia

Arthur

, 41–57. New York: Cambridge University Press.

23.

Druckman

James N.

Kifer

Martin J.

Parkin

Michael

. 2009. “Campaign Communications in U.S. Congressional Elections.” American Political Science Review 103 (3): 343–66.

24.

Fieldhouse

Edward

Green

Jane

Evans

Geoffrey

Van der Eijk

Cees

Schmitt

Hermann

Mellon

Jon

Prosser

Christopher

. 2015. The British Election Study Internet Panel 2014-2015. doi:10.15127/1.293723.

25.

Fieldhouse

Edward

Prosser

Christopher

. 2018. “The Limits of Partisan Loyalty: How the Scottish Independence Referendum Cost Labour.” Electoral Studies 52:11–25.

26.

Finkel

Steven E.

Geer

John G.

1998. “A Spot Check: Casting Doubt on the Demobilizing Effect of Attack Advertising.” American Journal of Political Science 42 (2): 573–95.

27.

Fisher

Justin

Fieldhouse

Edward

Johnston

Ron

Pattie

Charles

Cutts

David

. 2016. “Is All Campaigning Equally Positive ? The Impact of District Level Campaigning on Voter Turnout at the 2010 British General Election.” Party Politics 22 (2): 215–26.

28.

Fridkin

Kim Leslie

Kenney

Patrick J.

2004. “Do Negative Messages Work? The Impact of Negativity on Citizens’ Evaluations of Candidates.” American Politics Research 32:570–605.

29.

Geer

John G.

2006. In Defense of Negativity: Attack Ads in Presidential Campaigns. Chicago: The University of Chicago Press.

30.

Green

Jane

. 2007. “When Voters and Parties Agree: Valence Issues and Party Competition.” Political Studies 55 (3): 629–55.

31.

Green

Jane

Hobolt

Sara B.

2008. “Owning the Issue Agenda: Party Strategies and Vote Choices in British Elections.” Electoral Studies 27 (3): 460–76.

32.

Haddock

Geoffrey

Zanna

Mark P.

1997. “Impact of Negative Advertising on Evaluations of Political Candidates: The 1993 Canadian Federal Election.” Basic and Applied Social Psychology 18:205–23.

33.

Häkli

Jouni

. 1999. “Cultures of Demarcation: Territory and National Identity in Finland.” In Nested Identities: Nationalism, Territory, and Scale, edited by Henrik Herb

Guntram

Kaplan

David H.

, 125–50. Lanham: Rowman & Littlefield.

34.

Hansen

Kasper M.

Pedersen

Rasmus Tue

. 2008. “Negative Campaigning in a Multiparty System.” Scandinavian Political Studies 31 (4): 408–27.

35.

Haynes

Audrey A.

Rhine

Staci L.

1998. “Attack Politics in Presidential Nomination Campaigns: An Examination of the Frequency and Determinants of Intermediated Negative Messages against Opponents.” Political Research Quarterly 51 (3): 691–721.

36.

Hodess

Robin

Tedesco

John C.

Kaid

Lynda Lee

. 2000. “British Party Election Broadcasts: A Comparison of 1992 and 1997.” The Harvard International Journal of Press/Politics 5 (4): 55–70.

37.

Hogg

Michael A.

Abrams

Dominic

. 1988. Social Identifications: A Social Psychology of Intergroup Relations and Group Processes. London: Routledge.

38.

Houston

David A.

Doan

Kelly

Roskos-Ewoldsen

David

. 1999. “Negative Political Advertising and Choice Conflict.” Journal of Experimental Psychology 5:3–16.

39.

Johns

Robert

Denver

David

Mitchell

James

Pattie

Charles

. 2010. Voting for a Scottish Government: The Scottish Parliament Election of 2007. Manchester: Manchester University Press.

40.

Johns

Robert

Mitchell

James

Carman

Christopher J.

2013. “Constitution or Competence? The SNP’s Re-election in 2011.” Political Studies 61:158–78.

41.

Johns

Robert

Mitchell

James

Denver

David

Pattie

Charles

. 2009. “Valence Politics in Scotland: Towards an Explanation of the 2007 Election.” Political Studies 57 (1): 207–33.

42.

Johnston

Ron

Pattie

Charles

Fisher

Justin

Cutts

David

Fieldhouse

Edward

. 2013. “The Long and the Short of it: Local Campaigning at the British 2010 General Election.” Political Studies 61:114–37.

43.

Kaid

Lynda Lee

. 1997. “Effects of the Television Spots on Images of Dole and Clinton.” American Behavioral Scientist 40:1085–94.

44.

Lau

Richard R.

Pomper

Gerald M.

2001. “Effects of Negative Campaigning on Turnout in U.S. Senate Elections, 1988–1998.” Journal of Politics 63 (3): 804–19.

45.

Lau

Richard R.

Sigelman

Lee

Rovner

Ivy Brown

. 2007. “The Effects of Negative Political Campaigns: A Meta-analytic Reassessment.” The Journal of Politics 69 (4): 1176–1209.

46.

Maddens

Bart

Billiet

Jaak

Beerten

Roeland

. 2000. “National Identity and the Attitude towards Foreigners in Multi-national States: The Case of Belgium.” Journal of Ethnic and Migration Studies 26 (1): 45–60.

47.

Maier

Jürgen

Maier

Michaela

. 2007. “Audience Reactions to Negative Campaigns Spots in the 2005 German National Elections: The Case of Two Ads Called ‘the Ball’.” Human Communication 10:329–44.

48.

Mattes

Kyle

Redlawsk

David P.

2014. The Positive Case for Negative Campaigning. Chicago: The University of Chicago Press.

49.

Matthews

Douglas

Dietz-Uhler

Beth

. 1998. “The Black-Sheep Effect: How Positive and Negative Advertisements Affect Voters’ Perceptions of the Sponsor of the Advertisement.” Journal of Applied Social Psychology 28 (20): 1903–15.

50.

McCrone

David

Bechhofer

Frank

. 2015. Understanding National Identity. Cambridge: Cambridge University Press.

51.

Min

Young

. 2004. “News Coverage of Negative Political Campaigns: An Experiment of Negative Campaign Effects on Turnout and Candidate Preference.” The Harvard International Journal of Press/Politics 9 (4): 95–111.

52.

Moreno

Luis

. 1988. “Scotland and Catalonia: The Path to Home Rule.” In The Scottish Government Yearbook 1998, edited by McCrone

David

Brown

Alice

, 166–81. Edinburgh: University of Edinburgh’s Unit for the Study of Government in Scotland.

53.

Moreno

Luis

Arriba

Ana

Serrano

Araceli

. 1998. “Multiple Identities in Decentralized Spain: The Case of Catalonia.” Regional and Federal Studies 8 (3): 65–88.

54.

Mullinix

Kevin J.

Leeper

Thomas J.

Druckman

James N.

Freese

Jeremy

. 2015. “The Generalizability of Survey Experiments.” Journal of Experimental Political Science 2:109–38.

55.

Nadeau

Richard

Martin

Pierre

Blais

André

. 1999. “Attitude towards Risk-taking and Individual Choice in the Quebec Referendum on Sovereignty.” British Journal of Political Science 29 (3): 523–539.

56.

Nai

Alessandro

Walter

Annemarie

, eds. 2015a. New Perspectives on Negative Campaigning: Why Attack Politics Matter. Colchester, UK: European Consortium for Political Research (ECPR) Press.

57.

Nai

Alessandro

Walter

Annemarie

. 2015b. “The War of Words: The Art of Negative Campaigning.” In New Perspectives on Negative Campaigning: Why Attack Politics Matter, edited by Nai

Alessandro

Walter

Annemarie

, 1–34. Colchester, UK: European Consortium for Political research (ECPR) Press.

58.

Norris

Pippa

Curtice

John

Sanders

David

Scammell

Margaret

Semetko

Holli A.

1999. On Message: Communicating the Campaign. London: SAGE.

59.

Pattie

Charles

Denver

David

Johns

Robert

Mitchell

James

. 2011. “Raising the Tone? The Impact of ‘Positive’ and ‘Negative’ Campaigning on Voting in the 2007 Scottish Parliament Election.” Electoral Studies 30 (2): 333–43.

60.

Pattie

Charles

Denver

David

Mitchell

James

Bochel

Hugh

. 1999. “Partisanship, National Identity and Constitutional Preferences: An Exploration of Voting in the Scottish Devolution Referendum of 1997.” Electoral Studies 18:305–22.

61.

Pattie

Charles

Johnston

Ron

Fieldhouse

Edward

. 1995. “Winning the Local Vote: The Effectiveness of Constituency Campaign Spending in Great Britain, 1983–1992.” American Political Science Review 89 (4): 969–83.

62.

Petrocik

John R.

1996. “Issue Ownership in Presidential Elections, With a 1980 Case Study.” American Journal of Political Science 40 (3): 825–50.

63.

Petrocik

John R.

Benoit

William L.

Hansen

Glenn J.

2003. “Issue Ownership and Presidential Campaigning, 1952-2000.” Political Science Quarterly 118 (4): 599–626.

64.

Pipkin

Robert

Bartle

John

. 2002. “The Party Election Broadcasts: A Sleeping Giant or an Old Pair of Shoes?” In Political Communications: The General Election Campaign of 2001, edited by Bartle

John

Atkinson

Simon

Mortimore

Roger

, 87–108. London: Frank Cass.

65.

Rico

Guillem

Liñeira

Robert

. 2014. “Bringing Secessionism into the Mainstream: The 2012 Regional Election in Catalonia.” South European Society and Politics 19 (2): 257–80.

66.

Ridout

Travis N.

Walter

Annemarie

. 2015. “Party System Change and Negative Campaigning in New Zealand.” Party Politics 21 (6): 982–92.

67.

Roese

Neal J.

Sande

Gerald N.

1993. “Backlash Effects in Attack Politics.” Journal of Applied Social Psychology 23 (8): 632–53.

68.

Sanders

David

Norris

Pippa

. 1998. “Does Negative News Matter? The Effect of Television News on Party Images in the 1997 British General Election.” In British Elections and Parties Review, Volume 8: The 1997 General Election, edited by Denver

David

Fisher

Justin

Cowley

Philip

Pattie

Charles

, 150–70. London: Frank Cass.

69.

Sanders

David

Norris

Pippa

. 2005. “The Impact of Political Advertising in the 2001 UK General Election.” Political Research Quarterly 58 (4): 525–36.

70.

Sigelman

Lee

Shiraev

Eric

. 2002. “The Rational Attacker in Russia? Negative Campaigning in Russian Presidential Elections.” Journal of Politics 1:45–62.

71.

Sonner

Brenda S.

1998. “The Effectiveness of Negative Political Advertising: A Case Study.” Journal of Advertising Research 38:37–42.

72.

Tajfel

Henri

. 1981. Human Groups and Social Categories. Cambridge: Cambridge University Press.

73.

Tambini

Damian

. 2014. Nationalism in Italian Politics: The Stories of the Northern League, 1980-2000. London: Routledge.

74.

Unwin

Tim

. 1999. “Place Territory and National Identity in Estonia.” In Nested Identities: Nationalism, Territory, and Scale, edited by Henrik Herb

Guntram

Kaplan

David H.

, 151–74. Lanham, MD: Rowman & Littlefield.

75.

Walter

Annemarie

. 2014a. “Choosing the Enemy: Attack Behavior in a Multiparty System.” Party Politics 20 (3): 311–23.

76.

Walter

Annemarie

. 2014b. “Negative Campaigning in Western Europe: Similar or Different?” Political Studies 62 (1): 42–60.

77.

Walter

Annemarie

van der Brug

Wouter

van Praag

Philip

. 2014. “When the Stakes Are High: Party Competition and Negative Campaigning.” Comparative Political Studies 47 (4): 550–73.

78.

Whiteley

Paul

Seyd

Patrick

. 1994. “Local Party Campaigning and Voting Behavior in Britain.” Journal of Politics 56 (1): 242–51.

79.

Yavuz

Hakan M.

2001. “Five Stages of the Construction of Kurdish Nationalism in Turkey.” Nationalism and Ethnic Politics 7 (3): 1–24.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.41 MB

When Campaigns Can Backfire: National Identities and Support for Parties in the 2015 U.K. General Election in Scotland

Abstract

Keywords

Effects of Negative Campaigning: A Mixed Picture

Voters’ Reaction to Negativity in the United Kingdom

Design

Experimental Design 3

Panel Survey

Results

Support for Parties

Voting Behavior

Conclusion

Supplemental Material

Morisi_PRQ-2017-0042.R2_-_Supplemental_Material_online_supp – Supplemental material for When Campaigns Can Backfire: National Identities and Support for Parties in the 2015 U.K. General Election in Scotland

Footnotes

Acknowledgements

Declaration of Conflicting Interests

Funding

Notes

Supplemental Materials

References

Supplementary Material

Experimental Design³