ArticlePDF Available

Abstract and Figures

Scotland’s future within the European Union (EU) played a prominent role in the 2014 independence referendum. The story goes that latent supporters of independence voted to stay within the UK to maintain EU access. Defeated, Scottish leaders declared the referendum a once-in-a-life-time event only repeated if conditions substantially changed. With the UK now facing a chaotic exit from the EU, proponents of Scottish independence have suggested that a second referendum may occur after Brexit negotiations are completed. Faced with a consensus among Scottish party leaders in supporting EU membership, those hoping for a second independence referendum, we argue, looked to alternate sources of information that saw Brexit as an opportunity to create the conditions that would spur a second referendum. Using panel data from the British Election Study, we examine whether Scottish voters voted tactically to leave the EU. We argue that Scottish National Party voters were likely to interpret statements on the conditions for a second independence referendum as an implicit signal to vote “Leave.” The results have important implications for the role of referendums in representative democracy, strategic voting, and the importance of intra-party division on individual vote choices.
Content may be subject to copyright.
Reading between the Lines: Party Cues and SNP Support for Scottish Independence and
Brexit
Zachary Greene
University of Strathclyde
zacgreene@gmail.com
Jae-Jae Spoon
University of Pittsburgh
spoonj@pitt.edu
Christopher Williams
University of Arkansas at Little Rock
cjwilliams@ualr.edu
Please cite:
Greene, Zachary, Jae-Jae Spoon and Christopher Williams. 2016. “Reading between the
Lines: Party Cues and SNP Support for Scottish Independence and Brexit.” Presented at the
Impact of the EU Referendum Conference at the University of Newcastle, Newcastle UK
(September 16).
Abstract
Scotland’s future within the EU played a prominent role in the 2014 independence
referendum. The story goes that latent supporters of independence voted to stay within the
UK to maintain EU access. Defeated, Scottish leaders declared the referendum a once-in-a-
life-time event only repeated if conditions substantially changed. With the UK now facing a
chaotic exit from the EU, the conditions seem ripe for a renewed independence campaign.
Scots voted overwhelmingly to Remain within the EU, but the story is more complex. We
argue that those hoping for a second independence referendum saw Brexit as an
opportunity to create the conditions that would spur a second referendum. Using panel data
from the British Election Study, we examine whether Scottish voters voted tactically to leave
the EU. We argue that SNP voters were likely to interpret statements on the conditions for a
second independence referendum as an implicit signal to vote “Leave”. The results have
important implications for the role of referendums in representative democracy, strategic
voting, and the importance of intra-party division on individual vote choices.
Key words: Strategic Voting, Referenda, EU, Brexit, Scottish Independence
1
Introduction
Days after the surprise win for the UK campaign to leave the European Union (EU),
prominent Scottish Nationalists began calls for a second Scottish independence referendum.
Public opinion mirrored these appeals as support for independence surged to new highs (as
high as 59% in one survey).
1
This response may be unsurprising as each of the parties
represented in the Scottish Parliament had campaigned to remain in the EU and citizens
voted overwhelmingly to stay (62%). Yet, the story is more complex. The parties’ stated
positions masked major internal divisions over the parties’ ultimate policy goals and
strategy. For example, Nicola Sturgeon, leader of the Scottish Nationalist Party (SNP),
expressed strong support to remain in the EU, expounding on its numerous benefits for
Scotland. Prominent nationalists, such as former SNP leader Gordon Wilson, indicated
support for an alternate approach; vote to leave the EU to create the conditions for a new
independence referendum (Green 2016).
Faced with competing messages from prominent elites and opposing policy goals,
supporters of an independent Scotland faced a choice between a sincere vote to stay in the
EU and an instrumental vote for an uncertain second independence referendum. High levels
of aggregate support for the Remain camp may have masked the prevalence of
instrumental voting.
2
Indeed, areas such as Glasgow that voted in favour of Scottish
independence in 2014 faced lower turnout than more Unionist cities (despite Unionists
actually being more likely to support the leave campaign in pre-referendum polls).
3
Strategic
voting thus may be more prevalent than the aggregate indicators alone predict.
In this paper, we argue that many in support of an independent Scotland
demonstrated their support for independence by voting to leave the EU. Further, we posit,
based on the logic of strategic voting and elite signalling, that in the face of competing
1
See Curtis’s (2016) discussion of the strengths and weaknesses of these polls. Curtis’s discussion
implies that the real increase was somewhat smaller, but did increase to majority support for Scottish
independence.
2
We use a number of phrases particular to the campaigns throughout the paper. We use the terms
“Remain” and “Leave” to refer to the official electoral campaigns in support of remaining within the
EU or leaving the EU since these were the slogans and names endorsed by the campaigns themselves
in the run up to the election. We also use these terms to refer to the vote itself. We refer to the United
Kingdom leaving the EU as “Brexit” as politicians and the media often refer to the process.
3
For example, the district encompassing Glasgow city centre achieved a voter turnout of
approximately 56.3% in the EU referendum, whereas 73% of voters in the Edinburgh city centre
district (The Electoral Commission, 2016).
2
messages from non-partisan elites, those in support of Scottish independence, who also
supported the SNP, chose to go against the dominant message of their party and voted to
leave the EU. We add that the parallel argument on the conditions required for a second
independence referendum acted as an implicit cue for supporters to reinterpret the
leadership’s more explicit statements in support of remaining in the EU.
To test our argument, we use panel data from the third and eighth waves of the
British Election Study in Scotland. In particular, we connect respondents’ support for
independence in the 2014 referendum to their self-reported behaviour in the 2016 EU
referendum. In our analysis, we find that supporters of Scottish independence were more
likely to vote in favour of the UK remaining in the EU; however, they were also more likely
to pursue alternate voting strategies, depending on their identification with the SNP.
The theory and results from this analysis have important implications for the role of
referendums in representative democracy, the role that strategic voting can play, and the
role party politics, specifically intra-party division, has in influencing vote choices. Although
public discourse often celebrates referenda as a more “pure” form of democracy, the exact
question asked, as well as other political conditions, may create the opportunities for tactical
considerations. Forward looking, tactical voters may use their votes to set the future agenda,
rather than express their sincere preferences on future participation within the EU, which in
this specific case may be a second referendum on Scottish independence. These results
suggest that popularly elected leaders should think twice before using referenda to avoid or
meet their political aims on important issues, as the effect of elite cues may be limited.
Referenda and Voting Behaviour
Studies of referenda suggest that voters act in similar ways in referenda as they do in
other contexts. Much like studies of parliamentary elections, research has shown that
demographic indicators play a large role in referenda vote choices. A growing area of
research considers the relationship between referenda and election campaigns (see, for
example, Uleri and Gallagher 2016). Evidence from the parliamentary arena also suggests
that citizens might make voting decisions in referenda instrumentally.
3
Although referenda are a distinct subset of elections, voters’ choices likely reflect
their decision-making processes in other contexts. Numerous studies find evidence that
demographic considerations influence referenda outcomes. For example, scholars have
found evidence that national identity matters in votes on devolution and EU referenda in a
range of settings such as Wales (Jones and Scully 2012), the Czech Republic (Hanley 2007),
and the Baltic States (Mikkel and Pridham 2004).
Not all studies of direct democracy and voting behaviour focus solely on
demographic, economic or structural indicators. Past research on referenda in the EU, for
example, shows that support for the outcome depends on the popularity of the sitting
governmental parties (Franklin et al. 1995). Markowski and Tucker (2005) show further
evidence of individual level strategic behaviour in Poland. They find that many of those
against joining the EU sought to invalidate the 2003 referendum by not turning out to vote.
In particular, voters tried to deny the 50% mandatory threshold rather than turnout out to
vote against joining the EU.
Citizen information about the referenda likely also matters. De Vreese and Semetko
(2004), for example, show that the media played a role in the 2000 EU referendum in
Denmark. Citizens’ vote choice responded to exposure to public television campaign
advertisements. In a number of EU referenda, Hobolt (2005), moreover, finds that more
informed voters rely less on elite cues. In the Irish case, Marsh (2015) finds evidence that
party cues, incumbency and issues play an important role in vote choice.
Although referenda may seem to offer relatively simple, distinct alternatives at face
value, voters’ decision-making calculi may be more complex. Referenda can disrupt or limit
the usefulness of the normal set of cues and heuristics voters us to make informed decisions.
In the Scottish referendum, for example, the SNP acted as the primary advocate of an
independent Scotland, whereas the other major parties largely advocated remaining in the
UK. Conversely, the EU referendum placed all major parties in Scotland on the same footing
according to their official positions, advocating to remain in the EU. This policy convergence
created a context that amplified the role played by intra-party dissent among current and
former party elites. While no Scottish party leaders advocated the Leave position, the lack of
policy disagreement among the major parties likely created an opportunity for a wider
range of voices. Indeed, Scottish voters may have perceived goals for independence and a
4
vote to remain in the EU at odds. Whereas voters in England and Wales likely voted Leave
for sincere policy reasons; voters in Scotland voted Leave for tactical reasons.
Voter Coordination and Tactical Voting
Electoral rules induce strategic voting behaviour. Under single member districts in a
unidimensional setting, for example, voters may select a candidate other than their most
preferred, but who has a higher chance of winning the election to reduce the chance that an
even less preferred candidate wins (Downs 1957). Tactical voting occurs in a range of
systems, including those with open and closed list proportional representation and mixed
member rules (e.g. Gschwend 2007).
In the context of a referendum, a tactical vote would entail a choice to support the
less preferred outcome because the voter hopes that her vote will lead to a new status quo
supporting a third potential outcome. From the perspective of a tactical voter, a vote to leave
the EU, might have created the dissatisfaction within Scotland (where hypothetically over
60% of voters supported independence because of their desire to remain in the EU) to hold a
new, more successful referendum.
4
Although this perspective requires a number of
assumptions of what would happen following a Leave vote, past referenda on the EU
suggest that voters acted tactically or instrumentally in these settings (e.g. Markowski and
Tucker 2005).
Importantly, voters supporting Scottish independence would not have had to
develop this logic on their own. Although far from the official position of the party,
prominent former leaders of the SNP and even some media discussed the merits of a tactical
vote in the run up to the referendum (e.g. Green 2016). Voters who prioritized Scottish
independence had the signals and logic presented to them. A vote for Leave could be
construed as support for a second Scottish independence referendum. Following this logic,
those who voted for Scottish independence in 2014 would likely be those who would prefer
a second Scottish independence referendum. Therefore, we would expect that these
4
Indeed, public support for a second independence referendum soared in the weeks following the EU
vote (see Curtis 2016). The initial bump was, however, short lived. An August poll carried out by
YouGov found support for a second referendum declined to 46%
(https://yougov.co.uk/news/2016/09/01/).
5
individuals would be more inclined to vote in favour of Brexit to shift the status quo and
spur a new Scottish independence referendum. This leads us to our first hypothesis.
H1: Voters who supported Scottish independence are more likely to vote to leave the EU.
Partisan Identification and Mixed Signals
Strategic behaviour requires that voters not only hold information about their
current vote choices and how their fellow citizens will likely act, but also about the
implications of that vote choice (Meffert and Gschwend 2010). Lago (2008), for example,
shows that citizens in Spain use past election results as a heuristic to predict future coalition
participation. Voters might also use parties’ participation in coalitions to infer information
about the parties’ positions (e.g. Fortunato and Stevenson 2013; Spoon and Klüver N.d.).
Past electoral behaviour is less useful in referenda, however, as levels of information
available to voters depend on the salience and intensity of campaigns (Hobolt 2005).
Research has demonstrated that parties serve as a linkage between voters and
government (Lawson 1980, 2005; Dalton et al. 2011). One of the major functions of parties is
to mobilize voters and to serve as a heuristic or cue when it comes time to vote (Dalton and
Wattenberg 2000). Parties can, moreover, be thought of as brands, each with a distinct
product to offer (Downs 1957; Aldrich 1995; Cox 1997; Lupu 2013). By providing voters
with a heuristic or cue to follow, parties can reduce the cost or burden of voting to the
individual. In many cases, parties also provide direct information about how to vote
tactically under specific electoral rules (Gschwend 2007).
The mechanism that leads voters to deviate from their party’s official message
emerges from elite policy convergence, intra-party dissent, and conflicting priorities. This
perspective does not necessarily negate the expectation that elite statements inform voters
about how they should vote in elections. Indeed, a substantial body of research emphasizes
partisanship and party campaigns as tools for elite messaging and voter heuristics.
However, this research often treats parties’ messages as unified, although party leaders and
members often disagree. Scholars show, however, that parties are often divided on their
policy goals and strategies because in order to win elections they must accommodate and
attract voters with a diversity of interests. This is especially true in winner take-all systems.
6
From the perspective of an electorally motivated leader, a party would position itself
to maximize its potential votes (e.g. Adams 1999; Spoon 2011) which would be broad
enough to allow voters to project their own goals on the party (Shepsle 1972; Somer-Topcu
2015), and would emphasize the issues that encourage an image of competence (Petrocik
1996; Green and Jennings 2011a; Hobolt and de Vries 2012; Greene 2015). Importantly, these
strategies require the somewhat unrealistic assumption that party leaders, MPs and
members express a consistent message.
Contrary to this assumption, groups within parties hold and often express diverse
preferences. Recent research in several different contexts illustrates this. Ceron (2012; 2013;
2014), for example, shows that intra-party factions and MPs in Italian parties diverge in non-
election years, although they converge on the party leader’s position prior to an election. The
distribution of preferences within French and German parties, moreover, increases when
they are in government and the party expects to be punished by voters for the economy
(Greene and Haber 2014). More broadly, parties have reduced the requirements for voting in
leadership elections, but the leadership campaigns have become less competitive (Kenig
2009a, 2009b). Parties with parliamentary delegations from diverse backgrounds write more
diversified platforms (e.g. Greene and O’Brien 2016) and their MPs address a greater range
of issues in parliament (see, for example, Bäck et al. 2014). Evidence of disagreement within
Scottish parties, therefore, is not unique. Party leaders, MPs and members hold a
distribution of preferences over a diversity of issues and dimensions.
Further, parties’ messages are often the result of intra-party decision-making and
compromise. Spoon and Williams (N.d) find that parties’ responsiveness to public
Euroskepticism, for example, is conditioned by intra-party division. Parties’ manifestos are,
moreover, more likely to respond to shifts in voter preferences when party leaders are less
constrained by their activists (Schumacher et al. 2013). Importantly, public perceptions of
intra-party division hold real consequences for parties, as voters consider them less
competent and are less likely to vote for them (e.g. Greene and Haber 2015).
Despite the mixed messages from parties and elites, evidence suggests that many
voters use simple symbols such as partisan identification to make decisions about politics.
Informational shortcuts such as partisan labels can be useful in a range of settings (e.g. Lupia
1994; Lupia and McCubbins 1998). Yet, voters may rely on overly broad heuristics in
7
complex informational environments (Lau and Redlawsk 2001). Parties divided on an issue,
and signalling mixed messages, likely complicate the use of overly simple heuristics. Voters
may be hearing two messages: the official (explicit) message of the party and a more subtle,
even hidden (implicit) message from some party elites.
5
Evidence suggests that in the case of
mixed or conflicting signals, their effectiveness decreases. Indeed, in the case of relatively
basic cues such as partisanship and gender, voters are more likely to hold incorrect beliefs
about their party’s positions when they conflict (e.g. Pyeatt and Yanus 2016).
In addition, when parties present a divided or disunited message, research suggests
that parties’ cues compete with other signals as their message is weakened. Voters are less
likely to perceive parties as competent on issues or vote for those parties when they also
perceive them as divided (e.g. Greene and Haber 2015). Factors such as the economy greatly
influence voters’ perceptions of the parties themselves (e.g. Green and Jennings 2011a,
2011b). Moreover, voters are more likely to interpret parties’ signals through an ideological
lens. They are more likely to perceive parties in a positive light and are more likely to
respond to parties’ campaigns when they are ideologically close (e.g. van der Brug 2004;
Bélanger and Meguid 2008; Vegetti 2014). This discussion implies that voters faced with
competing messages over how to achieve distinct policies on issues likely side with, and act
on, the signals that most readily agree with their position on the most important of these
policy dimensions.
This discussion has implications for the Brexit vote. Despite each of the major parties’
official positions supporting Remain, disagreement on the EU referendum gave voters the
opportunity to question and subsequently deviate from the parties official position.
Furthermore, signals from prominent supporters of the SNP acted as a clear signal from
some SNP elites that a vote in favour of Brexit could lead to a second Scottish independence
vote, provided a strong impetus for SNP supporters to vote in favour of Brexit. This effect
was particularly strong among those who were most in favour of a second Scottish
independence referendum: those individuals who voted for Scottish independence in the
2014 referendum. Indeed, those in support of an independent Scotland could easily
5
Importantly, this outcome is different than when the party system is highly divided or polarized.
Levendusky (2010), for example, has shown that when elites are more polarized, cues from the party
can be clearer. Voters are better able to identify the positions of the parties and their policy views are
more consistent with their issue preferences.
8
reinterpret statements of support to Remain in the EU as the responsible and necessary
positions for the party in government, while justifying their decision to vote Leave with the
SNP’s threats directed towards Westminster of a second referendum outcome if the UK
voted to leave the EU. Following this logic, we expect that supporters of the SNP who also
favour a second Scottish independence referendum would be more inclined to tactically vote
in favour of Brexit in the hopes of altering the status quo to spur a second Scottish
independence referendum. This leads to our second hypothesis.
H2: Voters who supported Scottish independence and identify with the SNP are more likely
to vote to leave the EU.
In summary, we hypothesize that supporters of Scottish independence voted
instrumentally in the EU referendum, voting in favour of Brexit in order to change the status
quo and bring about a second Scottish independence referendum. Further, we posit that,
because of division among the official SNP position regarding Brexit and the signals some
SNP elites sent regarding the possibility of a second Scottish independence referendum if
Leave won, this effect should be particularly strong among those who favour Scottish
independence and support the SNP.
Data and Methods
To test the above hypotheses, we use panel data from the British Election Study
(BES), focusing only on those respondents who had the right to vote in the 2014 Scottish
Independence Referendum. The data used in this study comes from either Wave 3, which
was conducted immediately following the Scottish independence referendum in September
2014, and Wave 8, which was conducted immediately preceding the Brexit vote in 2016. The
BES panel study has the benefit of covering the same individuals’ vote choices for the
Scottish independence and the EU referenda.
The dependent variable used to test our hypotheses is the likelihood of an individual
to vote for the United Kingdom to leave the European Union in June of 2016. To
operationalize this variable, we rely on the BES Panel Study Wave 8 question, “If there was a
9
referendum on Britain’s membership of the European Union, how do you think you would
vote.” We coded those who responded that they would vote to “Leave the EU” as a 1, while
those who responded that they would vote to “Stay in the EU” were coded as a 0. We
exclude from the analysis those who reported that they would not vote, or did not know.
This results in a mean value of 0.37 and a standard deviation of 0.48 for the full sample of
Scottish respondents.
6
The main independent variable used in testing H1 is a dichotomous variable
indicating whether an individual voted for or against Scottish independence in the 2014
referendum. We operationalize this variable using BES Panel Study Wave 3 data based on
the question, “And how did you vote in the independence referendum?” Individuals were
coded as a 1 if they voted for Scottish independence, and 0 if they voted against Scottish
independence. We omit those who did not vote in the Scottish independence referendum or
did not know how they voted referendum. This variable has a mean of 0.47, with a standard
deviation of 0.50 in the full sample.
7
To test H2, our independent variable is an interaction between a respondent’s vote
choice in the Scottish independence referendum, measured as described above, and whether
an individual identifies with the SNP. Support for the SNP was operationalized as a dummy
variable and is based on the BES Wave 8 question, “Generally speaking, do you think of
yourself as Labour, Conservative, Liberal Democrat, or what?” Those who answered this
question with a response of SNP were coded with a value of “1.” All others were coded as
“0.” This variable has a mean of 0.14 with a standard deviation of 0.34. The interaction
variable (vote for Scottish independence*identification with the SNP) ranges from 0 to 1,
with a mean of 0.15 and a standard deviation of 0.36.
In addition to the primary independent variables, we include a number of control
variables. First, we include a set of traditional vote choice indicators as control variables:
6
In the BES Scottish sample, 63% supported Remain; whereas, 62% of Scottish voters actually voted to
Remain on June 23, 2016 indicating that the Wave 8 data provides a fairly representative measure of
Brexit referendum vote choice.
7
We used Wave 3 data as respondents were more likely to remember exactly how they voted in the
Scottish independence referendum when asked soon after casting a ballot rather than years later. In
the sampl, 47% of Scots recalled voting for independence, whereas 45% of Scots voted for
independence on September 18, 2014. This suggests that the Wave 3 data provides a fairly
representative measure of Scottish independence referendum vote choice.
10
age, education, gender, marital status, and personal income.
8
Second, we also incorporate
variables measuring EU efficacy and EU political knowledge. Our EU efficacy variable is
based on the BES Panel Wave 8 question, “How much do you agree or disagree with the
following statements? I have a pretty good understanding of the important issues at stake in
the EU referendum.” Respondents were able to choose from one of 5 responses, “Strongly
disagree,” “Disagree,” “Neither agree nor disagree,” “Agree,” or “Strongly agree.” Those
who answered “Don’t Know” were excluded from the analysis. The variable has a mean of
3.77 and a standard deviation of 0.89 in the full sample. Our EU political knowledge variable
is an additive variable derived from a battery of six (6) “true or false” questions about the
European Union in the BES Panel Wave 8. The variable has a mean of 1.42, with a standard
deviation of 1.91 in the full sample.
9
Our final control variable focuses on respondents’
perceptions of the clarity of the campaign material they received during the EU referendum
campaign. We create a logged scale of campaign clarity from the two measures of campaign
information included in Wave 8 of the BES panel. This variable is equal to the natural log of
the difference between the amount of campaign information respondents report having
received from the Leave and Remain campaigns.
10
This measure allows us to control for the
relative clarity of information provided by the separate campaigns in a single measure.
Higher values indicate that the respondent perceived the Leave campaign as having a
clearer campaign than the Remain campaign. On average, respondents reported the Remain
8
See the Appendix for which questions were used for each of these variables.
9
Respondents were asked to answer true or false to the following statements, “Each EU Member State
elects the same number of representatives to the European Parliament;” “Switzerland is a member of
the EU;” “Croatia is a member of the EU;” “The EU spends more on agriculture than any other policy
area;” “The European Court of Human Rights only has jurisdiction over EU members;” “The
European Union is made up of 15 member states.” If a respondent answered a question correctly s/he
was given 1 point for a maximum of 6 possible points.
10
The scale is based on the difference between responses to the two campaign information variables
on the question “To what extent do you agree or disagree with the following statements?” “The leave
campaign has provided clear information about why we would be better off leaving the European
Union.” “The remain campaign has provided clear information about why we would be better off
remaining in the European Union.” Respondents choose between a 5 point scale ranging from
“Strongly Disagree” to “Strongly Agree”. To estimate the scale we divide each of the campaign items
by 5 so that they range from 0 to 1 and then subtract the values. We then calculated the natural log of
the items (plus 1).
𝐶𝑙𝑎𝑟𝑖𝑡𝑦𝑆𝑐𝑎𝑙𝑒 = ln(1 + 𝐿𝑒𝑎𝑣𝑒𝐶𝑙𝑎𝑟𝑖𝑡𝑦
5+𝑅𝑒𝑚𝑎𝑖𝑛𝐶𝑙𝑎𝑟𝑖𝑡𝑦
5)
11
campaign as slightly clearer with a mean value of -0.04 and a standard deviation of 0.26 in
the full sample. For descriptive statistics for all variables included in our analysis, see Table
A.1 in the Appendix.
As our dependent variable is dichotomous, we use logistic regression (Long 1997).
Additionally, we use robust standard errors to estimate our models.
Analysis
In our first hypothesis (H1), we argue that those who voted in favour of Scottish
independence in 2014 were more inclined than those who voted against Scottish
independence to vote in favour of Brexit. Model 1 presents the results of a logistic regression
testing this hypothesis. In contrast to our expectation, an individual’s vote in the 2014
Scottish independence referendum is statistically significant and negative. This indicates
that if an individual voted in favour of Scottish independence in 2014, she was more inclined
to vote in favour of the United Kingdom remaining a part of the European Union.
[INSERT TABLE 1 ABOUT HERE]
We show the predicted differences in the likelihood of voting to Leave in Figure 1.
As the coefficient indicates, the likelihood of voting for Brexit is actually lower for those
respondents who reported voting for Scottish independence. Those supporting Scottish
independence have a lower probability of voting Leave of approximately 0.05 in comparison
to those not voting for Scottish independence.
11
[INSERT FIGURE 1 ABOUT HERE]
Model 2, adds a dichotomous variable indicating whether an individual supports the
SNP. When the SNP variable is included, the effect of the Scottish independence vote
variable remains negative, but becomes statistically insignificant. Additionally, the SNP
11
Despite the apparent overlap in the confidence intervals in the predicted likelihoods presented in
Figure 1, the effect is statistically different at the 95% level in Model 1.
12
dummy variable is negative but also statistically insignificant. This suggests that the main
finding of Model 1, that a vote for Scottish independence in 2014 is associated with a vote for
the UK to remain in the EU, is not robust. It further suggests that identifying with the SNP
has important effects that may condition the effect of voting for Scottish independence.
Model 3 presents the results of a test of H2 that SNP supporters who voted for
Scottish independence were more likely to vote in favour of the United Kingdom leaving the
European Union. The main independent variable in this model, the interaction between a
respondent’s identification with the SNP and her vote in the Scottish independence
referendum, is both positive (the expected direction) and statistically significant. This
indicates that those who identify with the SNP and who voted for Scottish independence
were more likely to vote for the UK to leave the European Union.
To explore the substantive impact of a vote for Scottish independence on an
individual’s vote choice in the Brexit referendum when she identifies with the SNP, we
computed a predicted effects plot based on Model 3 (holding independent variables at their
mean or median values for dichotomous variables). As in Figure 1, the y-axis of Figure 2
shows predicted likelihood of voting for the UK to leave the EU, the x-axis indicates whether
an individual identifies if they voted for Scottish independence or not. Those supporting the
SNP and voted against Scottish independence have a likelihood of voting for Brexit of
roughly 0.18. At the same time, those SNP supporters who voted for Scottish independence
have a likelihood of voting for Brexit of nearly 0.36. That is to say, among SNP supporters,
those who voted for Scottish independence were nearly about twice as likely to vote for
Brexit as those who voted against Scottish independence. Among non-SNP identifiers, a vote
against Scottish independence is associated with a likelihood of voting for Brexit of 0.37,
while a vote for Scottish independence is associated with a likelihood of voting for Brexit of
0.32. However, large overlap exists in the 95% confidence intervals of these groupings.
[INSERT FIGURE 2 ABOUT HERE]
Several of our control variables also confirm our expectations and reach standard
levels of statistical significance. In particular, we find that age, education level, and
perceived campaign clarity, to be statistically significant in all three models. The coefficient
13
for age is positive, indicating that older voters were more likely to vote in favour of the UK
leaving the EU. Education, conversely, has a negative relationship with voting for Brexit;
those with more education were less likely to vote in favour of the UK leaving the European
Union. The coefficient for the campaign clarity scale is positive, suggesting that those who
saw the Leave campaign as having a clearer message were more inclined to vote in favour of
the UK leaving the European Union, and vice versa. Since the average perception of
campaign clarity is negative, this suggests that the Remain camp actually benefited from
their campaign information in the aggregate. Additionally, efficacy was statistically
significant and positive in both Models 1 and 2, suggesting that those who felt that they had
a better grasp of the issues at stake in the Brexit referendum were more likely to vote in
favour of leaving the EU.
Conclusion
In this paper, we have sought to understand the relationship between voting for
Scottish independence and Brexit. The prominent statement from the Scottish First Minister
that a new independence referendum would only occur following a material change” in
support for Scottish independence in public opinion
12
created the opportunity for a tactical
vote. Given the prominence of the EU in the Scottish independence referendum, those in
support of independence easily could have believed that Brexit would bring about these
conditions. Following this logic, we hypothesized that those who favour Scottish
independence from the United Kingdom would vote tactically by voting to Leave the EU.
We further hypothesized that this effect would be particularly strong among those who
favour Scottish independence and support the SNP, as they received mixed messages from
the party’s official position and former party elites. Ultimately, the SNP’s defining issue is
the goal of Scottish independence, and its supporters could easily understand this from any
of their statements.
Using panel data from the BES, we find that voters who favour Scottish
independence at the time of the independence referendum were generally more inclined to
12
Sturgeon’s commented on the need for a “material change”” during a debate between Scottish party
leaders held on April 6 by STV in the run up to the 2016 Holyrood elections (BBC, April 8, 2016). This
statement clarified the party’s earlier position requiring that polls consistently show over 60%
support for independence (The Scotsman, October 18, 2015).
14
vote against Brexit; however, those who favour Scottish independence and identify with the
SNP were more inclined to vote in favour of the UK leaving the European Union. These
findings support our second hypothesis and are consistent with a story in which SNP
supporters in favour of independence behaved as if they were voting tactically for a new
Scottish referendum.
Beyond the insight these findings provide regarding the case of Brexit, these results
also have implications for our understanding of voter behaviour more generally and the role
of party division in determining voter behaviour, more specifically. First, these results
demonstrate that, when provided with even small incentives, voters can easily diverge from
their true preferences and vote instrumentally to achieve a greater ultimate goal. As most
Scottish voters maintain a preference and affinity for the European Union, these results
suggest that many Scots chose to vote in favour of Brexit, ignoring their true preference
(staying in the EU), with the hope of achieving their long-term goal of independence
through a second Scottish referendum. This result adds further nuance to a theory of the
relationship between European integration and regional nationalism in Europe (e.g. Jolly
2015). Second, our results show that intra-party division plays a large role in influencing
vote choice. While parties’ campaign statements likely influence their supporters
preferences, intra-party division limits the extent of their ability. Messages from the party
leadership do not wholly determine their supporters’ votes. Furthermore, the salience of
major issues, such as Scottish independence, to voters likely determines their willingness to
deviate from the leadership’s priorities. Ultimately, referenda can also create the
opportunity for tactical voting depending on the potential outcome of the proposed policy
change.
15
BIBLIOGRAPHY
Adams, James. 1999. “Policy divergence in multicandidate probabilistic spatial voting.”
Public Choice 100(12): 103122.
Aldrich, John H. 1995. Why parties? The origin and transformation of political parties in America.
Chicago: Chicago University Press.
Bäck, Hanna, Marc Debus, and Jochen Müller. 2014. “Who Takes the Parliamentary Floor?
The Role of Gender in Speech-making in the Swedish Riksdag.” Political Research
Quarterly 67(3): 504-518.
BBC. 2015. “Election 2015: Sturgeon says only ‘material change’ could spark Scots
referendum.” April 8. http://www.bbc.co.uk/news/election-2015-scotland-32222806
(September 14, 2016).
Ceron, Andrea. 2012. “Bounded oligarchy: How and when factions constrain leaders in
party position-taking.” Electoral Studies 31(4): 689701.
Ceron, Andrea. 2013. “Brave rebels stay home: Assessing the effect of intra-party ideological
heterogeneity and party whip on roll-call votes.” Party Politics 21(2): 246-258.
Ceron, Andrea. 2014. “Inter-factional conflicts and government formation Do party leaders
sort out ideological heterogeneity?” Party Politics. DOI: 10.1177/1354068814563974.
Cox, Gary W. 1997. Making votes count: strategic coordination in the world’s electoral systems.
Cambridge University Press.
Curtis, John. 2016. “Indyref2 Has the tide turned since the Brexit vote?” BBC, June 27.
http://www.bbc.co.uk/news/uk-scotland-scotland-politics-36646789 (September 14,
2016).
Dalton, Russell J. and Martin P. Wattenberg. 2000. Parties without partisans: Political change
in advanced industrial democracies. New York: Oxford University Press.
Dalton, Russell J., David M. Farrell, and Ian McAllister. 2011. Political parties and democratic
linkage: How parties organize democracy. Oxford University Press.
Downs, Anthony. 1957. An economic theory of democracy. New York: Harper.
Fortunato, David, and Randolph T. Stevenson. 2013. “Perceptions of partisan ideologies: The
effect of coalition participation.” American Journal of Political Science 57(2): 459477.
Franklin, Mark N., Cees van der Eijk, and Michael Marsh. 1995. “Referendum outcomes and
trust in government: Public support for Europe in the wake of Maastricht.” West
European Politics 18(3): 101117.
16
Green, Christopher. 2016. “EU Referendum: Scottish Nationalists ‘may vote to leave to speed
up route to independence.’” The Independent (January 28).
Green, Jane, and Will Jennings. 2012a. “The dynamics of issue competence and vote for
parties in and out of power: An analysis of valence in Britain, 1979–1997.” European
Journal of Political Research 51(4): 469503.
Green, Jane, and Will Jennings. 2012b. “Valence as macro-competence: an analysis of mood
in party competence evaluations in Great Britain.” British Journal of Political Science
42(2): 311343.
Greene, Zachary. 2015. “Competing on the issues How experience in government and
economic conditions influence the scope of parties’ policy messages.” Party Politics.
DOI: 10.1177/1354068814567026.
Greene, Zachary, and Matthias Haber. 2015. “The consequences of appearing divided: An
analysis of party evaluations and vote choice.” Electoral Studies 37: 1527.
Greene, Zachary, and Diana Z. O’Brien. 2016. “Diverse parties, diverse agendas? Female
politicians and the parliamentary party’s role in platform formation.” European
Journal of Political Research 55(3): 435453.
Gschwend, Thomas. 2007. “Ticket-splitting and strategic voting under mixed electoral rules:
Evidence from Germany.” European Journal of Political Research 46(1): 123.
Hanley, Seán. 2004. “A Nation of Sceptics? The Czech EU Accession Referendum of 1314
June 2003.” West European Politics 27(4): 691715.
Hobolt, Sara Binzer. 2005. “When Europe matters: The impact of political information on
voting behaviour in EU referendums.” Journal of Elections, Public Opinion and Parties
15(1): 85109.
Jolly, Seth. 2015. The European Union and the Rise of Regionalist Parties. Ann Arbor, MI: The
University of Michigan Press.
Jones, Richard Wyn, and Roger Scully. 2012. Wales Says Yes: Devolution and the 2011 Welsh
Referendum. University of Wales Press.
Kenig, Ofer. 2009a. “Classifying party leaders’ selection methods in parliamentary
democracies.” Journal of Elections, Public Opinion and Parties 19(4): 433447.
Kenig, Ofer. 2009b. “Democratization of party leadership selection: Do wider selectorates
produce more competitive contests?” Electoral Studies 28(2): 240247.
Lago, Ignacio. 2008. “Rational expectations or heuristics?” Strategic voting in proportional
representation systems.” Party Politics 14(1): 31-49.
17
Lau, Richard R., and David P. Redlawsk. 2001. “Advantages and disadvantages of cognitive
heuristics in political decision making.” American Journal of Political Science 45(4): 951
971.
Lawson, Kay. 1980. Political parties and linkage: A comparative perspective. New Haven: Yale
University Press.
Levendusky, Matthew. 2010. “Clearer cues, more consistent voters: A benefit to elite
polarization.” Political Behavior 32(1): 111-131.
Long, Scott. 1997. Regression models for categorical and limited dependent variables. Sage
Publications: Thousand Oaks, CA.
Lupia, Arthur. 1994. “Shortcuts versus encyclopedias: Information and voting behavior in
California insurance reform elections.” American Political Science Review 88(1): 6376.
Lupia, Arthur, and Mathew McCubbins. 1998. The democratic dilemma: Can citizens learn what
they need to know. Cambridge: Cambridge University Press.
Lupu, Noam. 2013. “Party brands and partisanship: Theory with evidence from a survey
experiment in Argentina.” American Journal of Political Science 57(1): 4964.
Markowski, Radoslaw, and Joshua A. Tucker. 2005. “Pocketbooks, politics, and parties: the
2003 Polish referendum on EU membership.” Electoral Studies 24(3): 409433.
Marsh, Michael. 2015. “Voting on Europe, again and again: Stability and change in the Irish
experience with EU referendums.” Electoral Studies 38: 170182.
Meffert, Michael F., and Thomas Gschwend. 2010. “Strategic coalition voting: Evidence from
Austria.” Electoral Studies 29(3): 339349.
Mikkel, Evald, and Geoffrey Pridham. 2004. “Clinching the ‘Return to Europe’: The
Referendums on EU Accession in Estonia and Latvia.” West European Politics 27(4):
716748.
Petrocik, John R. 1996. “Issue ownership in presidential elections, with a 1980 case study.”
American Journal of Political Science 40(3): 825850.
Pyeatt, Nicholas L. and Alixandra B. Yanus. 2016. “Sending mixed signals: The role of
partisanship in evaluations of political leaders.” Journal of Elections, Public Opinion
and Parties 26(4): 423434.
Schumacher, Gijs, Catherine E. de Vries, and Barbara Vis. 2013. “Why Do Parties Change
Position? Party Organization and Environmental Incentives.” The Journal of Politics
75(2): 464477.
18
Shepsle, Kenneth A. 1972. “The strategy of ambiguity: Uncertainty and electoral
competition.” American Political Science Review 66(2): 555568.
Somer-Topcu, Zeynep. 2015. “Everything to Everyone: The Electoral Consequences of the
Broad-Appeal Strategy in Europe.” American Journal of Political Science 59(4): 841854.
Spoon, Jae-Jae. 2011. Political survival of small parties in Europe. Ann Arbor: University of
Michigan Press.
Spoon, Jae-Jae and Heike Klüver. N.d. “Does anybody notice? How policy positions of
coalition parties are perceived by voters.” European Journal of Political Research.
Forthcoming.
Spoon, Jae-Jae, and Christopher J. Williams. N.d. “It takes two: How euroskeptic public
opinion and party divisions influence party positions.” Manuscript.
The Electoral Commission. 2016. “EU Referendum Results.”
http://www.electoralcommission.org.uk/find-information-by-subject/elections-and-
referendums/past-elections-and-referendums/eu-referendum/electorate-and-count-
information. (Accessed November 21, 2016).
The Scotsman. 2015. “SNP: 60% support needed before next independence referendum.
October 18. http://www.scotsman.com/news/politics/snp-60-support-needed-before-
next-independence-referendum-1-3920508 (September 14, 2016).
Uleri, Pier Vincenzo, and Michael Gallagher. 2016. The referendum experience in Europe.
Springer.
Van der Brug, Wouter. 2004. “Issue ownership and party choice.” Electoral Studies 23(2): 209
233.
Vegetti, Federico. 2014. “From political conflict to partisan evaluations: How citizens assess
party ideology and competence in polarized elections.” Electoral Studies 35: 230241.
Vreese, Claes H. De, and Holli A. Semetko. 2004. “News matters: Influences on the vote in
the Danish 2000 euro referendum campaign.” European Journal of Political Research
43(5): 699722.
de Vries, Catherine E., and Sara B. Hobolt. 2012. “When dimensions collide: The electoral
success of issue entrepreneurs.” European Union Politics 13(2): 246268.
19
TABLES AND FIGURES
Table 1. Vote Choice in the EU Referendum
Model 1
Model 2
Ind. Ref. Vote
-0.266**
-0.146
(0.114)
(0.142)
SNP Identification
.
-0.20
(0.155)
Ind. Ref. Vote*
.
.
SNP Identification
Age Group
0.148***
0.153***
(0.042)
(0.042)
Education
-0.066***
-0.067***
(0.012)
(0.012)
Gender
0.038
0.035
(0.126)
(0.126)
Marital Status
-0.039
-0.044
(0.122)
(0.123)
Personal Income
-0.032
-0.032
(0.022)
(0.022)
EU Efficacy
0.123*
0.124*
(0.070)
(0.070)
EU Political Knowledge
-0.003
-0.003
(0.015)
(0.015)
Campaign Info. Scale
5.985***
5.987***
(0.522)
(0.522)
Constant
-0.710
-0.700
(0.451)
(0.450)
AIC
1810.984
1811.445
BIC
1866.771
1872.810
Χ2
198.280
201.550
Log-Likelihood
-895.492
-894.722
Observations
1956
1956
Notes: *** p < 0.01, ** p < 0.05, * p < 0.10. We present the results of logistic regression
models in Table 1 predicting the likelihood that a respondent reports voting for
Brexit with Huber-White robust standard errors in parentheses.
20
Figure 1. Likelihood of Voting In Favour of Brexit for Full Sample (Model 1)
13
Figure 2. Likelihood of Voting In Favour of Brexit for SNP Identifiers (Model 3)
13
Predicted likelihoods in Figure 1 and Figure 2 are estimated from 1000 draws of the variance-
covariance matrix based on the results presented in Model 1 and 3, respectively. Independent
variables are held at their mean values for continuous variables and at the median values for
dichotomous variables. The predicted probabilities are the median predicted value with 90% (lighter
lines) and 95% (darker lines) confidence intervals.
.15 .2 .25 .3
Pr(Leave Vote)
No Vote Independence
Vote in the Scottish Independence Referendum
.05 .1 .15 .2 .25
Pr(Leave Vote)
No Vote Independence
Vote in the Scottish Independence Referendum
21
APPENDIX
Operationalization of demographic control variables
Age is based on the BES Panel Wave 8 question, “What is your age?” This variable is ordinal
with a “1” denoting a respondent who is under the age of 18, a “2” is a respondent who is
between the ages of 18 and 25, a “3” is a respondent who is between the ages of 26 and 35, a
“4” is a respondent who is between the ages of 36 and 45, a “5” is a respondent who is
between the ages of 46 and 55, a “6” is a respondent who is between the ages of 56 and 65,
and a “7” is a respondent who is 66 or older.
Education is based on the BES Panel Wave 8 question, “What is the highest educational or
work-related qualification you have?” This variable is ordinal ranging from 1 (least
educated) to 18 (most educated).
Gender is based on the BES Panel Wave 8 question, “Are you male or female?” This is a
dichotomous variable, with males coded as a “1” and females coded as a “2.”
Marital status is derived from the BES Panel Wave 8. This variable is a dichotomous
variable, with a “1” denoting any individual who is either “Married,” “Living as married,”
or in a “Civil Partnership.”
Personal income is based on the question, “What is your gross household income?” This is
an ordinal variable ranging from “1” to “14,” with “1” being a gross household income of
less than £5,000 per year. The values then increase in increments of £5,000 per year until
reaching the 11th category, which is a gross household income of between £50,000 per year
and £59,999 per year. The 12th category is a gross household income of between £60,000 per
year and £69,999 per year, with the 13th category is a gross household income between
£70,000 per year and £99,999 per year, and the 14th category is a gross household income
over £100,000 per year.
22
Table A.1. Descriptive Statistics
Variables
N
Mean
Std. Dev.
Min
Max.
EU Ref. Vote
3467
0.371
0.483
0
1
Ind. Ref. Vote
5047
0.465
0.499
0
1
SNP Identification
8286
0.136
0.342
0
1
Ind. Ref. Vote* SNP Identification
5047
0.150
0.357
0
1
Age Group
7744
4.783
1.631
1
7
Education
8021
2.858
4.926
1
18
Gender
8286
1.476
0.499
1
2
Marital Status
8286
0.626
0.484
0
1
Personal Income
6436
4.990
3.023
1
14
EU Efficacy
3681
3.768
0.894
1
5
EU Political Knowledge
8286
1.419
1.909
0
6
Campaign Info. Scale
8283
-0.038
0.256
-1.609
0.693
23
Table A2: Vote Choice in EU Referendum (Sample Restricted by Party Support)
Model A1
(Tory)
Model A2
(Labour)
Model A3
(LibDem)
Model A4
(Green)
Model A5
(No Party)
Ind. Ref. Vote
0.699
-0.008
-0.563
-0.927
-0.331
(0.542)
(0.328)
(0.700)
(1.890)
(0.317)
Age Group
0.118
0.104
0.285
0.850
0.320***
(0.104)
(0.118)
(0.233)
(0.650)
(0.112)
Education
-0.047
-0.034
-0.111*
-0.140
-0.142***
(0.036)
(0.030)
(0.063)
(0.227)
(0.039)
Gender
0.286
-0.092
0.423
5.893*
-0.237
(0.328)
(0.336)
(0.676)
(3.161)
(0.306)
Marital Status
-0.238
0.132
-1.424**
2.427**
-0.451
(0.321)
(0.332)
(0.616)
(1.216)
(0.317)
Personal Income
0.012
-0.083
0.083
-0.275
-0.045
(0.044)
(0.074)
(0.100)
(0.254)
(0.055)
EU Efficacy
0.469***
0.159
0.806*
-1.029
0.002
(0.177)
(0.178)
(0.420)
(0.686)
(0.188)
EU Political
0.013
-0.013
0.078
0.186
-0.010
Knowledge
(0.034)
(0.035)
(0.078)
(0.137)
(0.046)
Campaign Info.
Scale
5.511***
6.285***
7.064***
7.311
7.089***
(1.060)
(1.622)
(1.773)
(4.594)
(1.456)
Constant
-2.102
-1.212
-4.599*
-14.422
0.759
(1.289)
(1.148)
(2.662)
(12.338)
(1.025)
AIC
350.336
313.306
95.067
41.764
293.385
BIC
388.596
353.095
123.508
61.654
331.254
Χ2
51.98
24.36
30.11
12.76
53.61
Log-Likelihood
-165.168
-146.653
-37.553
-10.882
-136.693
Observations
339
395
127
54
326
Notes: *** p < 0.01, ** p < 0.05, * p < 0.10. We present the results of logistic regression models in
Table A2 predicting the likelihood that a respondent reports voting for Brexit with Huber-White
robust standard errors in parentheses when the sample is restricted to only those respondents
supporting a specific party. The subset of party supporters is included in parentheses under the
model number. The results of Model A4 are the results for respondents who supported the Green
Party. These estimates must be interpreted particularly carefully as the number of observations for
this subsample is only 54.
24
Table A3: Vote Choice in EU Referendum (Interactions for Parties other than SNP)
Model A6
(Tory)
Model A7
(Labour)
Model A8
(LibDem)
Model A9
(Green)
Model A10
(No Party)
Ind. Ref. Vote
-0.045
-0.444***
-0.318***
-0.237**
-0.225*
(0.129)
(0.126)
(0.117)
(0.115)
(0.126)
Party ID
0.817***
(
-0.799***
(
-0.737***
(
-0.370
(
-0.180
(
(0.175)
(0.187)
(0.305)
(0.709)
(0.190)
Ind. Ref Vote*
0.569
0.447
0.115
-0.487
-0.213
Party ID
(0.569)
(0.357)
(0.711)
(0.877)
(0.298)
Age Group
0.123***
0.144***
0.157***
0.144***
0.153***
(0.042)
(0.042)
(0.042)
(0.042)
(0.043)
Education
-0.072***
-0.071***
-0.063***
-0.064***
-0.065***
(0.012)
(0.013)
(0.012)
(0.012)
(0.012)
Gender
0.074
0.037
0.054
0.045
0.034
(0.127)
(0.127)
(0.127)
(0.126)
(0.126)
Marital Status
-0.072
-0.053
-0.047
-0.046
-0.038
(0.124)
(0.124)
(0.123)
(0.123)
(0.123)
Personal Income
-0.040*
-0.035
-0.031
-0.033
-0.032
(0.022)
(0.022)
(0.022)
(0.022)
(0.022)
EU Efficacy
0.114
0.138*
0.114*
0.128*
0.127*
(0.071)
(0.071)
(0.070)
(0.071)
(0.070)
EU Political
-0.004
-0.001
-0.002
-0.002
-0.002
Knowledge
(0.016)
(0.015)
(0.015)
(0.015)
(0.015)
Campaign Info.
Scale
5.911***
5.880***
5.950***
5.974***
5.976***
(0.523)
(0.525)
(0.521)
(0.523)
(0.522)
Constant
-0.715
-0.450
-0.713
-0.718
-0.783
(0.453)
(0.459)
(0.452)
(0.453)
(0.461)
AIC
1786.006
1785.419
1807.706
1811.293
1814.157
BIC
1852.950
1862.363
1874.650
1878.237
1881.101
Χ2
215.77
216.02
208.95
199.37
199.54
Log-Likelihood
-881.003
-885.710
-891.853
-893.646
-895.079
Observations
1,956
1,956
1,956
1,956
1,956
Notes: *** p < 0.01, ** p < 0.05, * p < 0.10. We present the results of logistic regression models in
Table A3 predicting the likelihood that a respondent reports voting for Brexit with Huber-White
robust standard errors in parentheses. Each model’s main independent variable is an interaction
between an individual’s vote in the Scottish independence referendum and the party with which
she identifies. The party in each model is noted in parentheses under the model number.
ResearchGate has not been able to resolve any citations for this publication.
Book
The referendum has become established as part of the decision-making process in many European countries. Experts from each country survey the historical experience and current debates in Austria, Denmark, Finland, France, Ireland, Italy, the Netherlands, Norway, Russia and eastern Europe, Sweden, Switzerland and the United Kingdom. The impact of the referendum experience upon European politics is assessed, and the merits of the use of the referendum are evaluated, with a discussion on the implications for political parties, party systems, and representative government.
Article
Do parties respond to voters’ preferences on European integration in elections to the European Parliament (EP)? Following recent research that shows political party responsiveness to Euroskeptic attitudes during EP elections is conditioned by party characteristics, this article seeks to understand how party unity on European integration affects party responsiveness to Euroskepticism. We argue that when Euroskeptic attitudes among voters are high and the parties are divided in their position on European integration, parties will be more responsive to voters and take a more Euroskeptic position. To test our theoretical expectations, we use data from the Chapel Hill Expert Survey, the Euromanifestos Project, and European Election Study from 1989-2009 for over 120 parties across 20 European Union Member States. Our findings have important implications for understanding the nature of democratic representation in the European Union.
Article
Do parties respond to voters’ preferences on European integration in elections to the European Parliament (EP)? Following recent research that shows political party responsiveness to Euroskeptic attitudes during EP elections is conditioned by party characteristics, this article seeks to understand how party unity on European integration affects party responsiveness to Euroskepticism. We argue that when Euroskeptic attitudes among voters are high and the parties are divided in their position on European integration, parties will be more responsive to voters and take a more Euroskeptic position. To test our theoretical expectations, we use data from the Chapel Hill Expert Survey, the Euromanifestos Project, and European Election Study from 1989-2009 for over 120 parties across 20 European Union Member States. Our findings have important implications for understanding the nature of democratic representation in the European Union.
Article
Does governing in coalitions affect how coalition parties’ policy positions are perceived by voters? In this article, we seek to understand the relationship between parties’ participation in coalition governments and their perception by voters. Policy positions are an important instrument through which parties compete for the support of voters. However, it is unclear to what extent voters can correctly perceive the positions of parties when they govern together with other coalition partners. We argue that because of the blurred lines of responsibility in multiparty cabinets, it is difficult for voters to correctly perceive the positions of coalition parties. What is more, we expect that the internal functioning of coalition cabinets affects the extent to which coalition parties struggle to get their message out to voters. We hypothesize that intra-cabinet conflict is negatively related to misperception. To test our theoretical expectations, we combine data on the left-right policy positions of political parties from the Comparative Manifestos Project with data on how these positions are perceived by voters gathered from the Comparative Study of Electoral Systems from 1996-2011. Our findings shed light on the relationship between party competition and coalition governments and its implications for political representation.
Article
Partisanship and gender are powerful heuristic cues used by citizens to understand their elected officials’ ideology. When these cues send complementary signals – a Democratic woman or a Republican man – we expect they will aid citizens in evaluating their leaders’ political ideology. However, when partisanship and gender send conflicting signals, we expect citizens will be more likely to misperceive their leaders’ beliefs. We test this proposition using ideological evaluations of incumbent US senators collected in the 2010 and 2012 Cooperative Congressional Election Studies. The findings support our hypotheses, illustrating voters’ reliance on both partisan and gender cues. Our results suggest potential consequences for not only Republican women, but also Democratic men.
Article
Parties’ parliamentary delegations contain a multitude of interests. While scholars suspect that this variation affects party behavior, most work on parties’ policy statements treats parties as unitary actors. This reflects the absence of strong expectations concerning when (and how) the parliamentary caucus matters for platform construction, as well as the difficulties inherent in testing such claims. Drawing on the literature on women’s descriptive representation, we argue that the makeup of the parliamentary party likely has important consequences for issue entrepreneurship, the scope of issues represented on the manifesto, and even the left-right position of election platforms. With the most comprehensive party-level study of women’s representation ever conducted, we test our three diversity hypotheses using data on the gender makeup of parties’ parliamentary delegations and the content of their manifestos for 110 parties in 20 democracies between 1952 and 2011. We show that as the percentage of women in the parliamentary party increases, parties address a greater diversity of issues in their election campaigns. Women’s presence is also associated with more left-leaning manifestos, even when controlling for parties’ prior ideological positions. Together, these findings illustrate a previously overlooked consequence of descriptive representation and provide a framework for understanding when and why the parliamentary party influences manifesto formation. They show that diversity—or lack thereof—has important consequences for parties’ policy statements, and thus the overall quality of representation.