Nationalism and ethnic heterogeneity: The importance of local context for nationalist party vote choice

Electoral Studies 39 (2015) 129e141

Contents lists available at ScienceDirect

Electoral Studies journal homepage: www.elsevier.com/locate/electstud

Nationalism and ethnic heterogeneity: The importance of local context for nationalist party vote choice* Peter John Loewen*, Maxime Heroux-Legault, Carolina de Miguel Department of Political Science, University of Toronto, Canada

a r t i c l e i n f o

a b s t r a c t

Article history: Received 18 August 2014 Received in revised form 14 April 2015 Accepted 21 April 2015 Available online 29 May 2015

This article explores the individual-level correlates of nationalist party vote choice and the extent to which these correlates are conditioned by an individual's local context. We argue that the influence of individuals' policy positions on nationalism should vary in importance for predicting voting for nationalist parties in localities where voters feel threatened culturally or economically. To test this argument we use the case of support for the Bloc Quebecois in the Canadian province of Quebec and data from the 2011 Canadian Vote Compass. We show that voters' policy positions on nationalism become more important in predicting a vote for the Bloc Quebecois when the percentage of English speakers (our proxy for ethno-cultural threat) increases in their locality. By contrast, we find that the relationship between nationalism and support for the Bloc Quebecois is not conditioned by economic hardship in the place where an individual lives. To test the robustness of our findings, we reestimate our models using a different dataset from multiple elections e the Canadian Election Study as well as an additional modelling approach. Our findings contribute to the broader vote choice literature by examining the role that local context plays in individuals' choice of parties. Furthermore, our findings lend support to arguments made in the literature on the importance of an ethno-cultural calculus among voters voting for nationalist parties. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Nationalism Vote choice Canada Big data

1. Introduction As with many countries in the world, Canada witnessed a rise in support for nationalist parties1 in the second half of the 20th century, coinciding with the election of an official

separatist government in the French-speaking province of
be
cois would go on to hold Quebec in 1976. The Parti Que referenda on secession from Canada in 1980 and again in 1995.2 In the interim, Canada also saw the rise of a FrenchCanadian nationalist party within its federal Parliament. In

* This work relies heavily on the 2011 federal election edition of Vote Compass Canada, a voting engagement application. The original research team on this project included Peter Loewen, Yannick Dufresne, Gregory Eady, Jennifer Hove, and most especially Clifton van der Linden, who is the Founder and Executive Director of Vox Pop Labs. More details on the project can be found at www.votecompass.ca. Loewen is funded by an Early Researcher Award granted by the Ontario Ministry of Economic Development. * Corresponding author. E-mail address: [email protected] (P.J. Loewen). 1 We use the term “nationalist parties” to refer to a broad category of parties in which issues of nationalism are an important dimension of political competition. One of the goals of these parties is to defend the interests (cultural, linguistic, economic) of a geographically-defined nation or community. The comparative literature us es a variety of terms for these types of parties such as regional parties, ethno-regional parties, autonomist parties and separatist parties. The findings of this paper could be potentially extended to parties that defend the interests of a community or group that is not geographically defined such as the United Kingdom Independence Party. 2 They were out of office between 1985 and 1994.

http://dx.doi.org/10.1016/j.electstud.2015.04.005 0261-3794/© 2015 Elsevier Ltd. All rights reserved.

130

P.J. Loewen et al. / Electoral Studies 39 (2015) 129e141


be
cois was successful in having MPs 1993, the Bloc Que elected in two-thirds of Quebec's constituencies. As a result they held the status of the second largest party in Canada's House of Commons.
be
cois (PQ) and Bloc Que
be
cois The rise of the Parti Que (BQ) has motivated substantial research on the determinants of support for nationalist parties in Canada. This work, together with a broader comparative research agenda on nationalism has provided valuable insights into the individual-level determinants of vote choice for nationalist parties. For example, scholars have found that support for nationalist parties is driven by voters' policy positions on issues of nationalism as well as by voters' positions on socio-economic policy issues linked to the traditional left-right dimension. In addition, support for nationalist parties is often correlated with feelings of attachment to one's nation and various individual-level demographic factors. Surprisingly few scholars have moved beyond individual-level models to examine whether individual determinants of nationalist vote vary across regions in a given country, or across time within the same locale. This omission of context and the reliance on individualistic models is understandable for both theoretical and practical reasons. Rational choice theory argues that voting is an individual act (Downs, 1957, Enelow and Hinich, 1984) and as such the principal focus ought to be on individual-level variations. Some scholars focus on sociological determinants of voting, but even then such determinants are measured and conceptualized as characteristics of the individual voter (Lazarsfeld et al., 1944, Berelson et al., 1954). There are also practical considerations that justify this focus on individual factors since the modal election study has a limited number of respondents (generally a few thousand) spread over dozens or hundreds of geographic units. Accordingly, estimating contextual effects at a local level is to invite problems of low statistical power. Put differently, it is hard to uncover meaningful local variation in the determinants of voting when only a small number of respondents live in each locale.3 In this paper we link the individual to their social and political context. More specifically, we examine the modelling assumption that the determinants of support for nationalist parties have a constant effect across localities, and we explore both the individual-level correlates of nationalist party vote choice and the extent to which these correlates are conditioned by an individual's local context. To
be
cois do so we focus on the case of support for the Bloc Que in the province of Quebec. Our main argument is that policy positions on nationalism should become more important in
be
cois in localities where predicting vote for the Bloc Que voters feel threatened culturally or economically. To test this argument we use data from the 2011 Canadian Vote Compass, a uniquely large survey of voters.4

3 We do note, however, that related and very promising work in “small area estimation” (Park et al., 2004; Jiang and Lahiri, 2006; Ghitza and Gelman, 2013) is presenting an easily applied solution to this problem. 4 Restricting our analysis to the province of Quebec in Canada, we still have more than eighteen-thousand respondents.

We first estimate an individual-level model of vote choice, testing the comparative effects of nationalist preferences, left-wing policy preferences, and leader
be
cois across the evaluations on support for the Bloc Que province of Quebec. Our findings are largely consistent with findings in the Canadian and broader comparative literatures that two main cleavages (nationalism and ideology) underly support for nationalist parties. We then turn to explore how two local contextual factors (percentage of English speakers and percentage of unemployed) condition the effect of these individual-level determinants of nationalist vote. We estimate these contextual effects at the level of census divisions (roughly corresponding to cities and towns) using a two-step model recommended by Jusko and Shively (2005). In support of our argument, we find that voters' policy positions on nationalism become more important in pre
be
cois when the percentage dicting a vote for the Bloc Que of English speakers (our proxy for ethno-cultural threat) increases in their locality. By contrast, we find that the relationship between nationalism and support for nationalist parties is not conditioned by economic hardship in the place where an individual lives. To demonstrate the robustness of our findings, we then test our arguments using a different dataset from multiple elections e the Canadian Election Study e as well as an additional modelling approach. Our findings have implications for how we think about nationalist parties' electoral strategies, especially the tradeoff these parties face between catering to identitarian policies or to classical ideological policies. If nationalist parties want to maximize electoral support across geographic constituencies they may have to tailor their message differently in different localities. More specifically, the emphasis they will place on nationalist policy positions should not be the same everywhere. Furthermore, our findings lend support to arguments made in the literature on the importance of an ethno-cultural calculus among voters voting for nationalist parties. Finally, this paper seeks to contribute to the broader vote choice literature by examining the role that local context plays in individuals' choice of parties. We assume that voters employ different calculi depending on the environment in which they are situated, and we try to show this by explicitly modelling how individual-level determinants of vote choice vary by local conditions that should plausibly affect voters' calculus. In what follows, we review the Canadian and comparative literatures on individual level determinants of vote choice for nationalist parties and apply their insights to our
be
cois (Section 2). In turn, we explore case of the Bloc Que how local context matters to the study of vote choice for nationalist parties and derive some empirical expectations of the ways in which contextual variables exacerbate or attenuate the effects of individual-level predictors of nationalist vote (Section 3). We then review our data and modelling strategy (Section 4), and we follow this with our analyses and discussion of results using two different datasets and methodologies (Section 5). We end with a discussion of the implications of our findings and further avenues of research (Section 6).

P.J. Loewen et al. / Electoral Studies 39 (2015) 129e141

2. Individual-level determinants of nationalist voting What underlies support for nationalist parties? Existing research suggests that support for nationalist parties is driven by voters' policy positions on nationalism, which can range from demands for cultural or linguistic recognition within a constitutional framework or demands for greater political or fiscal autonomy to demands for outright independence. The expectation is that the more recognition or autonomy voters desire for their region the more likely they are to vote for nationalist parties. The fact that voters' policy positions on nationalism explain voting for nationalist parties is not particularly surprising given that nationalist parties emerge in part to defend these nationalist positions and to push for policies of cultural recognition, autonomy and even political independence of their region.5 For example, in the case of Canada scholars have found that political support for sovereignty translates into
be
cois (Be
langer and electoral support for the Parti Que Nadeau, 2009). We expect a similar relationship with
be
cois. support for the Bloc Que Nationalist parties also compete for votes on the traditional left-right ideological dimension. In the abstract, nationalism and ideology are potentially orthogonal to each other. In practice, however, political entrepreneurs will often frame their nationalist demands in the language of left and €l and The
rien (2008), for instance, write: right politics. Noe “Still, in so far as they acquire a political expression, social identities almost always do so in the language of left and right.” They detail that nationalism has been wielded equally by tenants of right-wing and left-wing discourses, including fascists, as well as progressive proponents of decolonization and democratization. More directly relevant to the current research project, they explain that “sub-state identities that sought recognition and autonomy in multinational countries were always tied either to the left or to the right, depending on the class and political alliances that defined them.” On this topic, they agree with Guibernau (2013) and Erk (2010)
bec nationalism is social-democratic or that modern Que left-wing. Based on this research we expect left-wing voters
be
cois. This hyto be more likely to support the Bloc Que pothesis is strengthened by previous work by Canadian scholars who have found that ideology, quite apart from policy positions on nationalism, is likewise thought to increase the probability of supporting a nationalist party in Quebec, whether at the federal or provincial level (Blais
langer and Nadeau, et al., 2002, Gidengil et al., 2012, Be
be
cois we expect left2009). In the case of the Bloc Que

5 What is less obvious, and we will show in this paper, is that the importance of these nationalist positions is not constant across localities. 6 We do not have a previous expectation of which dimension e leftright or nationaliste will be more important in the case of support for
be
cois, but we posit that both factors will matter to explain the Bloc Que vote choice for this party. The extent to which voters place more or less importance on the nationalist dimension versus the ideological dimension in voting for nationalist parties has become a source of scholarly debate and has important consequences for the nature of polices that nationalist parties pursue: identitarian policies versus classical ideological programs. In the literature on Spain for example “research has consistently found that voters are more attentive to ideology than to nationalism (de la Calle, 2005; Balcells, 2007).” (Rivero, 2011).

131

wing voters to be more likely to vote for this party given its general policy orientation as a pro-labour, left-wing party and the background of its leader as a former trade union representative.6 In addition, support for nationalist parties has been associated with feelings of attachment to one's nation (nationalist sentiment), and a series of individuallevel demographic variables such as level of education, age, gender and religion. We expect that all these factors will be predictors of vote for the BQ as they are for many nationalist parties in other developed democracies (mostly in Western Europe). The goal of this paper however is to go beyond this individuallevel model of vote choice to explore whether these individual-level predictors of vote vary by local context. It is not surprising that nationalist policy positions (or left-right ideology) correlate with support for nationalist parties; but what is less obvious is that the importance of these nationalist positions might vary across local context, which poses particular challenges and opportunities for nationalist parties. In what follows we explore how local context might modify the effect of nationalist policy positions on vote for the BQ within the province of Quebec, and we derive some testable hypotheses. 3. The role of local context The question of support for nationalist or secessionist movements or parties has also been approached from a rational choice perspective that emphasizes individuals' calculus of voting based on contextual variables. Scholars such as Sorens (2005) and Sambanis and Milanovic (2011) posit that individuals are more likely to pursue secession or autonomy for their region if they expect that the new arrangement will improve their community's welfare in cultural, political and/or economic grounds. Therefore individuals tend be more supportive of secessionist options if they live in regions that are very different than the rest of the country. For example, Sambanis argues that “regions with higher levels of income, resource-rich regions, or ethno-culturally distinct regions should all generate some measure of demand for autonomy” (Sambanis and Milanovic, 2011, 6). In a similar vein, Sorens posits that economic and linguistic factors explain support for secessionist parties. He finds that high-income regions are more prone to request independence than poorer regions and that regions with a region-specific minority language are also more likely to opt for secession (Sorens, 2005).7 Scholars of Canadian politics have applied a similar approach to explain the attitudinal bases of support for sovereignty in Quebec. According to this literature, support for sovereignty is the product of a rational calculus on the part of Quebecers that has two main components: one linguistic and one economic. Quebecers are thought to ask themselves what gains or losses they would experience regarding the protection of French and the vitality of their economy in the case of Quebec independence. Findings

7 Sorens finds that the linguistic effect is decidedly smaller than the economic effect. Sorens also devotes substantial attention to other variables.

132

P.J. Loewen et al. / Electoral Studies 39 (2015) 129e141

suggest that Quebecers who expect gains in these two areas following independence are more likely to support independence than other voters (Mendelsohn, 2003, Blais et al., 1995, Blais and Nadeau, 1992, Nadeau and Fleury, 1995, Howe, 1998). Others give more attention to regional dif
ferences tied to economic concerns. For instance, Gagne and Langlois (2002) find that support for sovereignty in Quebec is markedly lower in the region of Outaouais than in the rest of regions within the province of Quebec. They surmise that the reason for this is that citizens of Outaouais are likelier to work for or have family working for the federal government, and thus would be worried about sovereignty for economic reasons.8 However, because of limited data, they cannot test this hypothesis (p. 115). In sum, both the comparative and Canadian literatures suggest the importance of an economic and ethno-cultural calculus in individuals' attitudes towards nationalist or secessionist options, a calculus that may depend on regional or national contextual factors. We build on these insights to argue that a similar economic and ethnocultural calculus should operate at the local level to affect individuals' support for nationalist parties. In particular, we argue that French-speaking voters in Quebec should place greater emphasis on nationalism when they perceive that their language/identity or their economic situation is at stake. If voters feel culturally or economically threatened in their locality then they should be more critical of the status quo and more inclined to think that an alternative scenario of autonomy (or even independence) for Quebec would improve their economic and cultural situation. If this is the case, voters should be more likely to prefer nationalist or secessionist policy options and to turn to parties that pursue these options. In order to test this argument we look at how the influence of nationalist policy positions on vote for the Bloc
be
cois varies according to two broad local conditions: Que the percentage of English speakers, which we use as a proxy of ethno-cultural threat, and average rates of unemployment, which we take as a proxy for general economic uncertainty or hardship. In particular we expect that the influence of nationalist policy positions on the vote for
be
cois will be greater in local contexts in which the Bloc Que there are more English speakers or in contexts with higher unemployment. Feelings of cultural threat or economic insecurity will be greater when individuals are exposed to more speakers of another language or to more dire economic circumstances respectively; and this day-to-day reminder of a conflict over cultural or economic resources should spill into politics, and make voters more likely to prioritize policy positions on nationalism when voting for
be
cois. It should likewise be a ready trope for the Bloc Que enterprising politicians who wish to prime feelings of nationalism and link them to vote choice.9 This leads to the following two hypotheses:

Hypothesis 1. The positive effect of nationalist policy po
be
cois should increase sitions on the vote for the Bloc Que as the percent of English speakers increases in a locality. Hypothesis 2. The positive effect of nationalist policy po
be
cois should increase sitions on the vote for the Bloc Que as the percent of unemployed people in a locality increases. In sum, we argue that the importance of nationalism in
be
cois is conditional on an invoting for the Bloc Que dividual’s cultural and economic environment. As explained in the previous section, it is not surprising that policy positions on nationalism have a positive (and sig
be
cois. Nationnificant effect) on the vote for the Bloc Que alist parties emerge in part to defend these nationalist positions and to push for policies of cultural recognition, political and economic autonomy and even political independence of their region. And we expect that these nationalist positions will be strong predictors of support for the BQ (among other factors). Our argument however is that the effect of nationalist policy positions on vote for nationalist parties is not the same everywhere; instead it varies depending on the economic and ethno-cultural local context of an individual. This challenges the assumption of many individual-level models of voting that the effect of nationalism on vote is constant across space (i.e. electoral districts or other geographic units). In the sections that follow, we outline an empirical strategy that allows us to estimate the contextual conditionality of the factors associated with voting for nationalist parties.

4. Data and modelling strategy Our paper employs data collected in the 2011 Canadian federal election by Vote Compass, a voting advice application hosted by the Canadian Broadcasting Corporation website.10 The application asks respondents thirty policy and attitudinal questions related to Canadian politics, in addition to a number of questions related to sociodemographics, voting behaviour, and evaluations of political parties and their leaders. The application received nearly two-million responses during the thirty-seven day campaign. The benefit of the Vote Compass data is that it allows us to estimate separate individual-level models within precise and low levels of geography. Whereas samples from standard election surveys range in the tens or hundreds of respondents per constituency, Vote Compass data number in the hundreds or thousands. To address the issue that data input and respondent effort cannot be as closely monitored in the Vote Compass data, we limit our set to those respondents who answered a complete schedule of sociodemographic questions. We make the assumption that while respondents may enter political positions on numerous occasions to see how their estimated political positions change, they are less likely to

8

The federal capital, Ottawa, is located close to the Outaouais region. As we discuss in our conclusion, it is difficult to know the extent to which observed variation in the importance of nationalism is a result of varying campaign messages, though we suspect that it does reflect this, at least in part. 9

10 This particular application, which is just one of a class of VAAs, has also been deployed in partnership with the Wall Street Journal and the Australian Broadcasting Corporation. As a matter of disclosure, Loewen has a passive financial interest in this application.

P.J. Loewen et al. / Electoral Studies 39 (2015) 129e141

enter demographic information multiple times. A respondent who is clicking or ‘speeding’ through the tool will skip non-mandatory questions. Once this filter is applied, and once we limit the data to French-speaking Canadians in Quebec, we are left with approximately 18,000 respondents. These respondents are nested within 98 census divisions.11 Our main modelling strategy is appropriate for the nature of our hypotheses and the multi-level character of our data e individuals nested within census divisions. This strategy follows closely that of Jusko and Shively (2005). In this two-step strategy an individual-level model is estimated in each geographic unit (which in this case is a census division).12 The estimation of separate models implies uniquely estimated variance and covariance structures. Coefficients of interest are then captured from each of these models and regressed against variables measured at the level of the geographic unit within a generalized

problem of respondent self-selection in two manners. First, we replicate our analyses using data from the campaign period survey of the 2011 Canadian Election Study (CES). These data are collected by random sampling over the phone. We also use the CES data to address concerns that results are election-specific by conducting some analyses using a pooled sample of several federal elections [2004, 2008 and 2011]. Second, we present results from a more conventional hierarchical linear modelling approach. All these results are presented in section 9 (appendix).

4.1. Individual-level model The first step in our procedure is to estimate a series of linear additive models of the probability to vote for the Bloc
be
cois. Formally, we estimate the following individualQue level model of vote choice:

YðVOTEÞij ¼ aj þ Nationalismij b1j þ Ideologyij b2j þ Covariatesij b3j þ εi j;

least-squares framework which weights coefficients according to the precision of their estimate (Jusko and Shively, 2005, Lewis and Linzer, 2005). The second level coefficients can then be inspected to estimate how much the influence of some individual-level variable varies by the levels of some contextual factor. The advantages of this approach are clearly articulated by Jusko and Shively (2005), and among them are straightforward interpretation of results, ease of visual inspection, and no loss in statistical efficiency in comparison to more standard HLM approaches (eg Snijders and Bosker, 1999). We recognize that the results that follow are estimated using self-selected respondents and a relatively new technology.13 We also recognize that our modelling strategy differs from arguably more conventional approaches. Accordingly, we wish to highlight that we address the

11 For purposes of comparison, we note that there are 75 federal constituencies (or ridings) in the province of Quebec. We use the federal constituency as the unit of analysis when using the CES data. The mean number of respondents is 189 per census division. The minimum is 11, while the maximum is 8942. The number of respondents at the 25th and 75th percentiles is 58 and 228, respectively. 12 We note an implicit assumption of this approach: the relevant level at which contextual effects occur is the local city or municipality, rather than a neighbourhood, for example. We return to a discussion of this in our conclusion. 13 As it happens, our Vote Compass data appear relatively representative prior to weighting. For example, data for our dependent variable (support
be
cois) from the VC and from CES closely mirrors prefor the Bloc Que election survey data. The percentage of BQ supporters in the Vote Compass is 37.05% and pre-election surveys situate support for the BQ around the high thirties during the months prior to election. [Cite 308 Blog]. Figures are similar in the CES data with 33.06% supporting the BQ in the Campaign Period Survey. We also note a recent paper by Fournier et al. (2013b) which compares vote choice models estimated from VAA data and conventionally-sampled data. While the levels of variables differ, sometimes substantially, the conclusions about the comparative importance of various variables do not.

133

(1)

where, for example, b1 indicates the effect of a respondent i's policy positions on nationalism on a nationalist vote choice in census division j. As the subscripts imply, the model is estimated separately in every census division, with separate error-covariance structures. For ease of interpretation and comparability across census divisions, we estimate an ordinary least squares regression at the first level.14 The dependent variable is whether the respondent
be
cois. It is a indicated a vote intention for the Bloc Que dummy variable. The key independent variable in our model is nationalism. Nationalism captures the voter's policy positions on nationalist issues and is an index based on the average response to three statements asked in the Vote Compass survey. Respondents were asked to indicate how much they agreed or disagreed with the following statements: “The federal government should have a say when it comes to decisions about culture in Quebec”, “Quebec should be formally recognized as a nation in the Constitution” and “Quebec should become an independent state.” The items exhibit reasonable internal consistency, returning a Cronbach's a ¼ .70. The scale runs from 1 to 5 with higher values indicating more support for autonomy, cultural recognition or independence. The scale has a mean of 3.86 and a standard deviation of 1.01. Nationalist offerings are often bundled together with left-wing policies. To parcel out nationalism, we also construct a measure of left-wing preferences, as in a classic economic left-right dimension. To generate this dimension, we compile an index composed of three categories of questions about the economy, fiscal redistribution and

14 This increases ease of interpretation because we can directly compare the size of coefficients across constituencies. Substantively, our results do not change if we use a logit estimation, but they prove more difficult to demonstrate.

134

P.J. Loewen et al. / Electoral Studies 39 (2015) 129e141

social spending.15 The index has a Cronbach's a ¼ .61. The index runs from one to five, with higher results indicating an ideologically left-leaning respondent. The mean of the index is 3.36, and its standard deviation is 0.56. As Fournier et al., (2013a) demonstrate, leadership was more important to vote decisions in Quebec in the 2011 election than in the previous election. Moreover, the importance of leader evaluations for vote choice in Canada is well-established Clarke (1984), Johnston et al., (1992), Clarke et al., (2009). Accordingly, we include a measure of the rank of a respondent's evaluation of Gilles Duceppe, the
be
cois. Finally, we include controls leader of the Bloc Que for gender, age, and educational attainment.16 4.2. Contextual-level model Following the estimate of n ¼ j different individual-level models, we then capture the coefficient on nationalism in each model and regress each separately on two aggregatelevel characteristics in each census division, namely the log of long-term average unemployment (M ¼ 1.99, sd ¼ 0.46) and the log of the percentage of English-speakers (M ¼ 0.65, sd ¼ 1.47). These measures are drawn from the most recently available census data. We do so in a feasible generalized least squares framework, with known variances Jusko and Shively (2005), Lewis and Linzer (2005). Estimates of the effect of policy positions on nationalism thus takes the form:

b1j ¼ aj þ EnglishSpeakersj b1j þ Unemploymentj b2j þ εj ; (2) The intercept in this model gives us the effect of nationalism on voting for a nationalist party when the share of English speakers is 0 and the long-term unemployment rate is also 0. We can then see how the effect of nationalism on nationalist voting increases or decreases as these two contextual factors change. 5. Results Table 1 presents regression results for an individual
be
cois. The model level model of voting for the Bloc Que
be
cois on nationregresses vote intention for the Bloc Que alism, ideology, evaluations of the Bloc leader, age, gender, education. The model is estimated in an OLS framework. Estimates for age, nationalism and leadership evaluations

15

The statements are: “When there is an economic problem, government spending usually makes it worse.”; “The federal budget deficit should be reduced, even if it leads to fewer public services.”; “Canada should seek closer economic relations with the USA.”; “How much of a role should the private sector have in health care?”; “The government should fund daycare instead of giving money directly to parents.”; “It should be easier to qualify for Employment Insurance.”; “Workers should contribute more to their government pension plan (CPP/RRQ) so that it can offer bigger pensions.”; “How much should wealthier people pay in taxes?”; “How much tax should corporations pay?”. 16 Unfortunately, the Vote Compass data does not include information on other potential control variables such as income with sufficient frequency to include in our estimations.

Table 1 Individual-level model. Variable

Coefficient

(Std. Err.)

Male Age Education Nationalism Leader rank Ideology (Left) Intercept

0.031** 0.001** 0.002 0.161** 0.148** 0.008 0.642**

(0.006) (0.000) (0.001) (0.004) (0.003) (0.005) (0.023)

N R2 F (6,18471)

18,478 0.474 4824.842

Significance levels: y: 10% *: 5% **: 1%. This table presents estimates of an individual-level model of vote for the
be
cois. The model is an OLS. The same model is estimated in each Bloc Que of 98 census divisions to generate first-level estimates for the effects of Nationalism and Ideology.

are clearly distinguishable from zero. Being young, a nationalist, and favouring Duceppe (the Bloc leader) over other leaders all increase the likelihood of voting for the Bloc. Interestingly, having a left ideology is not significant in predicting vote for the BQ, contrary to expectations in the literature.17 This is likely attributable to both the collinearity of this variable with evaluations of the leader of
be
cois18 and with the presence of a strong leftthe Bloc Que wing alternative in this election, namely the New Democratic Party.19 All in all, these results largely accord with the picture of nationalist voting outlined at the start of our paper, which conforms with much of the literature on individual-level determinants of voting for nationalist parties. The next step is to test our main argument, which is that these individual-level determinants vary across localities. More specifically, we test how the effect of nationalism varies depending on the social and economic context in which individuals live. To do so, we first estimate separate individual-level models in each unit of geography, in this case the 98 census divisions. The form of these 98 individual-level models is the same as that in Table 1 and in equation (2).20 We then capture the coefficient on nationalism in the 98 models and estimate a second regression where these coefficients are regressed on two variables observed at the second level of analysis: the natural log of the percentage of English speakers and the natural log of the long term unemployment rate in each census division.

17 Of course, left-wing policy views can still predict nationalist party voting in the 98 separate models we estimate at the constituency level. In 24% of these, we find a positive and significant effect for left-wing policy preferences. We find a positive but insignificant effect in another 35%. In 11% of cases we recover a significant negative effect. 18 Indeed, the variables are significantly correlated (r ¼ .38,p ¼ .00). Furthermore, the left-wing variable is significant and positive in a regression omitting leader evaluations. 19 In section 9 (appendix), we attempt to allay this concern by showing that our results are not anomalous to this election. Instead, they do remain significant in a model which pools data from three election studies. See Tables 7 and 8. 20 We do not present the 98 census-level models. They are available upon request.

P.J. Loewen et al. / Electoral Studies 39 (2015) 129e141

135

Fig. 1. This figure demonstrates the relationship between the presence of English speakers and the effects of nationalism. Each point represents the nationalism coefficient estimate in one census division. The size of the circle indicates the population of the census division. The red line is a linear fit of the data. The grey lines indicate the associated 95% confidence interval. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2. This figure demonstrates the relationship between the long term unemployment rate and the effects of nationalism. Each point represents the nationalism coefficient estimate in one census division. The size of the circle indicates the population of the census division. The red line is a linear fit of the data. The grey lines indicate the associated 95% confidence interval. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

This allows us to estimate how first level variables are conditioned by contextual factors at the second level. Figs. 1 and 2 plot the coefficients for nationalism estimated in the 98 census divisions. The size of the points corresponds to the population in each census division, which we include to show that the size of the census division is not a confounding variable that is driving the results. We also estimate a linear fit between the estimated coefficients and the level 2 variable of interest, as well as an associated 95% confidence interval. These figures suggest that the effects of nationalism increase with the share of English speakers in a census division (Fig. 1), but not with the rate of long term unemployment (Fig. 2). In other words, nationalism is a more important determinant of
be
cois in census divisions where there vote for the Bloc Que is a high proportion of English speakers. This finding is confirmed in Table 2. We find that the
be
cois is effect of nationalism on vote for the Bloc Que greater as the logged percentage of English speakers in the census division increases. Division of the coefficient by the constant tells us for each 1 point increase in the logged percentage of Anglophones in a district, the effect of nationalism increases by 13% (i.e. 0.02/0.154). Given that the log of the percentage of Anglophones in a district varies from 2.1 to 4.4 (as the percentage varies from 0.11 to 57), this is a respectable effect. Indeed, an analytical inspection tells us that the likelihood that an individual with average
be
cois in the levels of nationalism votes for the Bloc Que most French riding is 0.11, ceteris paribus.21 However, an individual with an average level of nationalism living in the most English census division has markedly higher likelihood of 0.24.22 In other words, the likelihood increases on the order of 120%.

Table 2
be
cois, FGLS. Variation of the effect of nationalism on vote for the Bloc Que

21 22

0.154 þ (0.02*2.15) ¼ 0.11. 0.154 þ (0.02*4.4) ¼ 0.24.

Variable

Coefficient

(Std. Err.)

Percentage of Anglophones (log) Percentage of unemployment (log) Intercept

0.021** 0.013 0.151**

(0.007) (0.022) (0.046)

N Log-likelihood c2ð:Þ

98 85.428 8.661

Significance levels: y: 10% *: 5% **: 1%. This table estimates a regression of the “nationalism” coefficient from OLS models (with the same variables as Table 1) estimated in 98 census divisions. The table demonstrates that the size of the “nationalism” coefficient varies with the natural logged percentage of Anglophones living within a census division. Coefficients do not vary with the natural logged percentage of long term unemployment in a census division. The model is a feasible generalized least squares with known variances (Jusko and Shively, 2005).

Scholars of Quebec politics may note an apparent contradiction here, namely that support for the Bloc
be
cois is traditionally higher in localities with fewer Que English speakers e such as the remote regions of Quebec e and lower in areas with more English speakers e such as Montreal. Our findings do not contradict this observation. Instead, we claim that given some individual-level of nationalism, a respondent will be more likely to vote for the nationalist party when they are surrounded by more Anglophones. This is reconcilable with the observation of greater overall support when there are fewer Anglophones as baseline levels of nationalism may be higher in those localities. In contrast to the significant conditional effect of percentage of English speakers, Table 2 shows that we do not find any conditional effects for the percentage of unemployed people in a census division. In other words, the influence of an individual's nationalism does not appear to depend on the underlying economic hardship in a locale. This does not accord with our second hypothesis, but it

136

P.J. Loewen et al. / Electoral Studies 39 (2015) 129e141

does correspond to a certain logic. Rational choice accounts suggest that individuals are more likely to support secession when they reckon that their economic situation will be better off after secession (Mendelsohn, 2003, Blais et al., 1995, Blais and Nadeau, 1992, Nadeau and Fleury, 1995, Howe, 1998). However, the place to find such an effect may be within individuals' economic circumstances, i.e. we should search for a direct effect. There may be, in other words, little about living in an economically depressed locale that primes nationalism. By contrast, there is a more logical link between living among Anglophones - that is, ‘others’ e and an amplified effect of nationalism. Living among others primes the importance of cultural difference, and provides a stronger link between this and nationalist vote choice. That said, we note that we are using just one contextual measure of economic hardship (unemployment) and this measure is a long-term average. It may well be the case that economic dynamics could condition the effect of nationalism on vote choice.23 Finally, as noted earlier, we analyze our Vote Compass data using a more conventional multilevel model. We estimate a model with the same individual level variables enationalism, ideology, leader evaluations, age, gender and educatione and the effects of the log of unemployment and the log of the percentage of Anglophones in the census division at the second level of analysis. To investigate the conditional effect of nationalism on the decision to vote for
be
cois, we add to the model an interactive the Bloc Que variable that is the multiplication of nationalism and the “percentage of Anglophones (log)” and an interactive variable that is the multiplication of nationalism and the “percentage of unemployment (log).” The analysis confirms the results we obtained using the two-step model. Results can be found in Table 6 and show that the interaction between nationalism and the “percentage of Anglophones (log)” is statistically significant.24

5.1. Reproducing the analysis using representative data There are two concerns that might arise from the use of Vote Compass data. First, given that respondents are self-

23 We note that we have also examined the effects of average and median household income. The effects of nationalism on vote choice do not vary with these variables either. 24 This paper does not look at the question of whether local contextual factors (such as the percentage of English speakers) predict support for
be
cois. This is an interesting, albeit separate question from the Bloc Que the one addressed in this paper. Our own preliminary analyses on this question suggests that the percentage of English speakers has either no effect (or a very small negative effect) on average support for the Bloc
be
cois. In other words, an increased presence of the “other” in a loQue cality does not increase the likelihood of voting for the BQ, whereas it does increase the likelihood that nationalism will be an important predictor of support for the BQ. This corroborates findings in the literature (Piroth et al., 2006; Lublin and Voss, 2002) that an increase in the share of non-francophones at the local level either has no effect (outside of Montreal) or a negative effect (in Montreal) on francophone's support for Quebec sovereignty. In terms of our second contextual variable our preliminary analysis shows that the percent of unemployment has a significant negative effect on support for the BQ, suggesting that the higher the percentage of unemployed people in a locality the smaller the likelihood
be
cois. that an individual will vote for the Bloc Que

selected, critics might wonder to what extent results hold over a representative sample of the general population of French-speaking Quebecers. Second, one might wonder about generalizability across elections. Are results specific to the 2011 election or do they hold for other election periods? In order to tackle these two questions we reproduce our analysis with a different dataset and a different modelling approach. First we use data gathered during the 2011 Canadian Election Study.25 The CES is Canada's national election study and is conducted by random sampling over the phone. Different questions were used in the Vote Compass and the CES surveys. For the CES replication the Quebec nationalism scale is the mean of five different questions, which ask respondents how much should be done for Quebec, whether they support sovereignty or not, how sovereignty will affect the state of the French language in Quebec and respondents' standards of living, and finally whether there is a reason to have a sovereigntist party in Ottawa or not. This scale has a Cronbach's a ¼ .70. To measure ideology on the left-right scale we use a set of variables related to spending in various left-leaning sectors, such as healthcare, welfare, education, environment and immigration and minorities. This scale has a Cronbach's a ¼ .52. We use ridings rather than census divisions as the geographical area of interest in these models, since the Canadian Election Study contains this information and not data on census division. This adds to the stringency of the robustness test. We will have more faith in our results if they hold across both census divisions and electoral ridings than if they hold for only one such regional division. It deserves to be noted that census divisions and ridings will typically be of inversely proportional sizes, and thus address a potential problem in the design. A possible concern would be that some census divisions are simply too big to accurately reflect neighbourly or immediate experiences of voters. Since census divisions are drawn around urban lines, whereas ridings are delimited by population, using both levels of analysis solves this problem. While the city of Montreal is a single census division, it is divided in many ridings, and so studying ridings gives us a greater degree of precision in this region. The opposite relationship holds for census divisions that are smaller than ridings, which typically occurs in rural parts of the province. Given the number of respondents in the 2011 CES survey, it is inefficient to run a separate model in every constituency. While we had, on average, more than 300 respondents per census division when using the Vote Compass data, we have to work with an average of 14 French-speaking respondents per riding when using the 2011 CES Campaign Period Survey. Accordingly, estimating a model in each constituency would make it difficult to obtain reliable estimates in the first step of the analysis. We thus opt for a slightly different approach. We estimate simultaneous regressions for the ridings in the highest and

25 In particular the data we use is from the campaign period survey. We use this data in order to have a better match with the VC data, which is a pre-electoral survey.

P.J. Loewen et al. / Electoral Studies 39 (2015) 129e141 Table 3 First level model depending on proportion of English speakers in the riding. Variable Lower 50% English Nationalism Ideology (Left) Leader rank Male Education Age Intercept Upper 50% English Nationalism Ideology (Left) Leader rank Male Education Age Intercept N Log-likelihood

Coefficient

(Std. Err.)

0.307** 0.194 0.225** 0.026 0.011 0.000** 0.396**

(0.109) (0.140) (0.022) (0.049) (0.012) (0.000) (0.133)

0.723** 0.131 0.145** 0.045 0.010 0.000y 0.416**

(0.112) (0.147) (0.023) (0.050) (0.012) (0.000) (0.138)

392 .

Significance levels: y: 10% *: 5% **: 1%. This table presents a replication of the preceding results in Table 2 using data from the Canadian Election Study. The table presents results from two simultaneously estimated OLS regressions in electoral constituencies with the lowest and highest halves of Anglophones. The coefficients on the nationalism variable can be compared using a c2 test statistic. This suggests a significant difference in the size of these coefficients (c2 ¼ 7.11,p ¼ .01).

lowest half of the variable of interest, the percentage of English speakers or percentage of unemployment. We then compare the coefficients across these regressions and verify whether there are significant differences between them. Table 3 presents results on the conditioning effects of the share of English-speakers in a riding. Comparing the coefficients for nationalism across the two regressions, we find that the effect of nationalism is greater in constituencies with more Anglophones. A c2 test suggests a significantly greater effect (c2 ¼ 2.8,p ¼ .09). The effect is substantively large and similar to that estimated with the Vote Compass data. Indeed, the increase in the likelihood of voting for the Bloc in the most versus the least Anglophone constituencies is 55% greater, net all other effects.26 Table 4 presents results for the conditioning effects of unemployment. As with the Vote Compass data, we find that the effect of nationalism does not vary according to the extent of unemployment in a constituency (c2 ¼ 0.08,p ¼ .77). Finally, to address concerns that results from the Vote Compass data and from the CES data are specific to the 2011 election, we rerun the analysis with a pooled sample of 2006, 2008 and 2011 election data from CES. The results are presented in Tables 7 and 8 in section 9 (appendix). The pooled analysis confirms that the results hold for percentage of English speakers (and not for percentage of unemployment). In sum, our main findings regarding the importance of contextual effects hold regardless of the dataset (Vote Compass data or CES data) and methodology used for the

26

i.e. 0.454/0.293 ¼ 1.55.

137

Table 4 First level model depending on unemployment rate in the riding. Variable Lower 50% unemployment Nationalism Ideology (Left) Leader rank Male Education Age Intercept Upper 50% unemployment Nationalism Ideology (Left) Leader rank Male Education Age Intercept N Log-likelihood

Coefficient

(Std. Err.)

0.573** 0.275y 0.172** 0.027 0.003 0.001 0.429**

(0.101) (0.152) (0.022) (0.047) (0.011) (0.001) (0.156)

0.411** 0.067 0.209** 0.041 0.014 0.000 0.473**

(0.120) (0.137) (0.025) (0.052) (0.012) (0.000) (0.134)

392 .

Significance levels: y: 10% *: 5% **: 1%. This table presents a replication of the preceding results in Table 2 using data from the Canadian Election Study. The table presents results from two simultaneously estimated OLS regressions in electoral constituencies with the lowest third and highest third proportion of long term unemployment. The coefficients on the nationalism variable can be compared using a c2 test statistic. This suggests no significant difference in the size of these coefficients (c2 ¼ 1.05,p ¼ .31).

analysis. We find that nationalism matters more for French voters when they live amidst a higher proportion of Anglophones. We also find that the effect of nationalism does not vary depending on local economic conditions. Given the many differences between the two datasets and modelling approaches, including the different geographical areas and different measures, this replication suggests that the results hold across a variety of conditions. These results also hold over two other tests of robustness. Specifically, in section 9 we demonstrate that our two-step results are replicated when we use a different second stage estimator. Second, we demonstrate that using a more conventional hierarchical linear modelling approach with the Vote Compass data returns results that reflect our main findings. 6. Conclusion Our findings are important to scholars of nationalism and nationalist political parties. At the individual-level, we find that respondents are more likely to support a nationalist party when they hold more nationalist policy positions. This is unsurprising. However, our contribution is in showing that the importance of these nationalist positions in predicting vote for nationalist parties is not constant. We find that the effect of nationalist positions is conditional on an individual's social context. We show that nationalism plays a stronger role in explaining vote for a nationalist party if voters (in our case French Quebecers) live among citizens of another ethno-cultural group (in our case, English Quebecers). Interestingly, this effect does not vary according to the economic hardship in a voter's locale (which we capture through long-term unemployment). These findings should be of interest to scholars of

138

P.J. Loewen et al. / Electoral Studies 39 (2015) 129e141

nationalism who have made similar claims about the importance of an economic and ethno-cultural calculus in individuals' attitudes towards nationalist or secessionist options. Our findings seem to suggest that some of the expectations of these scholars hold at a local level of analysis. More broadly our findings might speak to some of the controversies regarding intergroup contact theory and inter-ethnic conflict. The literature is unresolved regarding how contact should affect such conflict. Some think that more contact means more mutual understanding and greater tolerance between groups (Kalin and Berry, 1982, Varshney, 2003). Others argue the opposite, that is, a majority will feel threatened when a minority becomes too strong and will vote to limit the powers of the minority (Tolbert and Grummel, 2003). Negative contact theory states that intergroup contact leads to increased prejudice (Barlow et al., 2012, Paolini et al., 2010). Others adopt a more nuanced view, and claim that while individual-level contact breeds tolerance and mutual understanding, more contact yields more conflict in the aggregate (Pettigrew et al., 2006). Our findings are in line with the second theory, since we find that when French Quebecers live among more members of the linguistic minority (English speakers), they are more likely to let nationalist positions determine their vote. Unfortunately, with the data at hand, we cannot explicitly test the mechanism that threat mediates the relationship between an individual's ethnocultural context and her voting behaviour; yet the results are in line with the idea that contact breeds national conflict rather than mutual understanding. It also points to a mechanism by which this might occur: namely by increasing the importance of nationalist preferences over ideological concerns in determining vote choice. These findings have implications for how we think about nationalist parties' electoral strategies, especially the trade-off these parties face between catering to identitarian policies and catering to classical ideological policies. Our results suggest that for nationalist parties to maximize electoral support across geographic constituencies they must offer their message differently in different localities. More specifically, the emphasis they place on nationalist policy positions should not be the same everywhere. At some level the geographic variation allows nationalist parties to navigate both dimensions. It also allows us to better understand why nationalist parties also take pains to articulate other policy positions: nationalism will play better in some places than in others, and something else should fill out a voter's decision calculus. Indeed, it is highly likely that our results reflect just such a tailoring and tar
be
cois. Sorting out both the causal geting by the Bloc Que ordering of this heterogeneity and the effective limits it places on various electoral strategies is beyond the scope of our paper. Nonetheless, it does provide a promising direction for future research, particularly in countries where parties are confronted with competing incentives to make nationalist and left-wing appeals. There are limitations to our findings, which suggest several extensions moving forward. First, we have specified a rather broad locale, namely towns and municipalities. It may be the case that the relevant context is the much more

localized and selected areas in which individuals transact their daily lives. Future research should consider and estimate models at different levels of geography, including neighbourhoods and towns. It may actually be the case that the effects of economic hardship do condition determinants of nationalist vote, but we have failed to capture the level at which this occurs. Furthermore, future models might want to consider controlling for additional individual-level and contextual-level factors. Second, our results are obtained in an election in which the party in question was in a marked decline. Indeed, the BQ found themselves outpaced by a left-wing but essentially non-nationalist rival (NDP). It may well be the case that the factors that condition support for a nationalist party change when the party is on the ascent rather than the descent, or depending on who is a direct competitor. Although we find support that the direction of the relationship holds in a pooled analysis of several elections, the role of election-specific factors should be further explored. Third, our results rely largely on a new and unconventional source of data. However, we are able to replicate these results using a more conventional dataset. We think this is thus a small limitation. Third, we have assumed that the factor conditioning the effect of nationalism is the presence of the “other.” Of course, this factor could be correlated with another unmeasured variable. Such a problem plagues most observational research, so we point to our theoretical justification for attributing explanation to this variable. In line with this, future work should find ways to better test the mechanism that an individuals' economic and ethnocultural context generate feelings of economic and ethnocultural threat which affect their voting behaviour. Finally, it would be good to consider different methods for estimating causal effects, in particular small-area estimation approaches (Park et al., 2004, Jiang and Lahiri, 2006, Ghitza and Gelman, 2013). Individuals live among other individuals. In doing so, they engage in contact and they share political and nonpolitical experiences. While we have well-established and tested models of vote choice, we know less about how they vary from one place to the next. Our principal contribution, then, is to show that the effects of well-known variables can vary, sometimes substantially, depending on context. Acknowledging and estimating this can help us both understand voters and evaluate theories of their behaviour. Appendix. Robustness checks and additional results We present three sets of results in this Appendix. The first shows that our second stage results are consistent with a model estimated using an ordinary least squares rather than a generalized setup. As Table 5 shows, the results are very similar to those in Table 2. Second, we conduct a robustness test by replicating our results using a more conventional multilevel model. We estimate a model with the same individual level variables enationalism, ideology, leader evaluations, age, gender and educatione and the effects of the log of unemployment and the log of the percentage of Anglophones in the census division at the second level of analysis. To investigate the conditional

P.J. Loewen et al. / Electoral Studies 39 (2015) 129e141

effect of nationalism on the decision to vote for the Bloc
be
cois we add to the model an interactive variable that Que is the multiplication of nationalism and the “percentage of Anglophones (log)” and an interactive variable that is the multiplication of nationalism and the “percentage of unemployment (log).” The analysis confirms the results we obtained using the two-step model. Results can be found in Table 6. Third, we show that our main findings from the CES replication are maintained if we include data from the 2006 and 2008 Canadian Election Studies (Tables 7 and 8).

Table 5
be
cois, OLS. Variation of the effect of nationalism on vote for the Bloc Que Variable

Coefficient

(Std. Err.)

Percentage of Anglophones (log) Percentage of unemployment (log) Intercept

0.020** 0.014 0.154**

(0.007) (0.022) (0.046)

N R2 F (2,95)

98 0.08 4.15

Significance levels: y: 10% *: 5% **: 1%. This table presents an OLS estimate of the results presented in Table 2. It regresses the nationalism coefficient from OLS models (with the same variables as Table 1) estimated in 98 census divisions on our key contextual predictors. The table demonstrates that the size of the nationalism coefficient varies with the natural logged percentage of Anglophones living within a census division. However, the same cannot be said of the natural logged percentage of long term unemployment in a census division.

Table 6 Multilevel model. Variable

Coefficient

(Std. Err.)

Nationalism Ideology (Left) Male Age Education Leader rank Percentage of Anglophones (log) Percentage of unemployment Nationalism  Anglogphones (LogPct) Nationalism  Unemployment (LogPct) Intercept

0.147** 0.002 0.032** 0.001** 0.003y 0.150** 0.054** 0.000 0.011** 0.003 0.607**

(0.019) (0.006) (0.007) (0.000) (0.002) (0.003) (0.011) (0.042) (0.003) (0.010) (0.079)

Random Effects S.D.(Nationalism) S.D.(Intercept) S.D. (Residual)

0.005** 3.08e-08** 0.36**

0.002 6.91e-08 0.002

N Log-likelihood c2ð10Þ

13,153 5195.274 11,422.069

Significance levels: y: 10% *: 5% **: 1%. This table shows the results of our analysis when replicated using a multilevel model. The “Nationalism  Anglophones (LogPct)” variable demonstrates that the influence of “nationalism” on the decision to vote
be
cois is increased in the presence of more Anglophones. for the Bloc Que

139

Table 7 First level model depending on proportion of English speakers in the riding (Pooled elections). Variable Lower 50% English Nationalism Ideology (Left) Leader rank Male Education Age Election 2006 Election 2008 Intercept Intercept Upper 50% English Nationalism Ideology (Left) Leader rank Male Education Age Election 2006 Election 2008 Intercept Intercept N Log-likelihood

Coefficient

(Std. Err.)

0.421** 0.053 0.206** 0.011 0.012 0.000** 0.156** 0.103** 0.556** 2.176**

(0.094) (0.093) (0.017) (0.033) (0.008) (0.000) (0.043) (0.039) (0.090) (0.067)

0.686** 0.030 0.164** 0.026 0.005 0.000 0.147** 0.066 0.571** 2.177**

(0.090) (0.106) (0.016) (0.034) (0.008) (0.000) (0.043) (0.043) (0.099) (0.069)

801 .

Significance levels: y: 10% *: 5% **: 1%. This table presents a replication of the preceding results in Table 2 using data from the Canadian Election Study. The table presents results from two simultaneously estimated OLS regressions in electoral constituencies with the lowest and highest halves of Anglophones. The coefficients on the nationalism variable can be compared using a c2 test statistic. This suggests a significant difference in the size of these coefficients (c2 ¼ 4.15,p ¼ .04).

140

P.J. Loewen et al. / Electoral Studies 39 (2015) 129e141

Table 8 First level model depending on unemployment rate in the riding (Pooled elections). Variable

Coefficient

(Std. Err.)

Lower 50% unemployment Nationalism Ideology (Left) Leader rank Male Education Age Election 2006 Election 2008 Intercept

0.587** 0.009 0.183** 0.007 0.006 0.001 0.160** 0.130** 0.629**

(0.086) (0.104) (0.016) (0.032) (0.008) (0.001) (0.040) (0.038) (0.107)

Intercept

2.212**

(0.065)

Upper 50% unemployment Nationalism Ideology (Left) Leader rank Male Education Age Election 2006 Election 2008 Intercept

0.502** 0.033 0.189** 0.011 0.010 0.000 0.140** 0.029 0.527**

(0.101) (0.094) (0.018) (0.036) (0.009) (0.000) (0.047) (0.044) (0.102)

Intercept

2.121**

(0.072)

N Log-likelihood

801 .

Significance levels: y: 10% *: 5% **: 1%. This table presents a replication of the preceding results in Table 2 using data from the Canadian Election Study. The table presents results from two simultaneously estimated OLS regressions in electoral constituencies with the lowest third and highest third proportion of long term unemployment. The coefficients on the nationalism variable can be compared using a c2 test statistic. This suggests no significant difference in the size of these coefficients (c2 ¼ 0.41,p ¼ .52).

Table 10 Summary statistics: Second-level model, Vote Compass data. Variable

Mean Std. Description Dev.

Nationalism 0.139 0.105 Coefficient of Nationalism in the coefficient first-level regression. 0.654 1.414 Log of the percentage of Percentage of anglophones in the census division. Anglophones (log) 1.999 0.461 Log of the percentage of Percentage of unemployment in the census unemployment division. (log) N 98

Table 11 Summary statistics for the 2011 CES. Variable

Mean

Std. Dev.

Vote for the Bloc

0.331

0.471

Nationalism

0.526

0.279

Ideology (Left)

0.699

0.173

Leader Rank

3.786

1.294

Male

0.436

0.496

Education

6.725

2.168

Age

69.522 977.879

Table 9 Summary statistics: First-level model, Vote Compass data Variable

Mean Std. Dev.

Vote for the Bloc

0.433

Male

0.632

Age

67.501

Education

8.043

Nationalism

3.855

Leader rank

3.617

Ideology (Left)

3.36

N

N

Description

620 Dummy variable indicating whether the respondent plans to vote for the Bloc. 1108 Nationalism index based on answers to five questions. 1133 Ideology index based on answers to five questions. High values indicate left-wing preferences. 636 The position of Gilles Duceppe in relation to other leaders. A higher score indicates Duceppe is preferred. 1133 Dummy variable indicating whether the respondent is male. 1128 Ordinal variable indicating the respondent's education level. 1133 Continuous variable indicating the respondent's age.

Description

0.495 31,392 Dummy variable indicating whether the respondent plans to vote for the Bloc. 0.482 35,308 Dummy variable indicating whether the respondent is male. 16.885 35,301 Continuous variable indicating the respondent's age. 1.833 35,338 Ordinal variable indicating the respondent's education level. 1.008 34,635 Nationalism index based on answers to three questions. 1.456 22,237 The position of Gilles Duceppe in relation to other leaders. A higher score indicates Duceppe is preferred. 0.563 32,357 Ideology index based on answers to nine questions. High values indicate left-wing preferences.

Table 12 Summary statistics: Summary statistics for the combined CES model. Variable

Mean

Vote for the Bloc

0.439

Nationalism

0.546

Ideology (Left)

0.724

Leader rank

3.873

Male

0.466

Education

6.72

Age Election 2006 Election 2008

6.245 0.207 0.333

Std. Dev.

N

Description

0.496 1393 Dummy variable indicating whether the respondent plans to vote for the Bloc. 0.262 2429 Nationalism index based on answers to five questions. 0.175 2086 Ideology index based on answers to five questions. High values indicate left-wing preferences. 1.346 1336 The position of Gilles Duceppe in relation to other leaders. A higher score indicates Duceppe is preferred. 0.499 2462 Dummy variable indicating whether the respondent is male. 2.217 2446 Ordinal variable indicating the respondent's education level. 667.674 2449 Continuous variable indicating the respondent's age. 0.405 2462 Dummy Variable indicating whether the data is from 2006. 0.471 2462 Dummy Variable indicating whether the data is from 2008.

P.J. Loewen et al. / Electoral Studies 39 (2015) 129e141

References Barlow, F.K., Paolini, S., et al., 2012. The contact caveat: negative contact predicts increased prejudice more than positive contact predicts reduced prejudice. Personal. Soc. Psychol. Bull. 38 (12), 1629e1643.
Nadeau, R., 2009. Le comportement e
langer, E.,
lectoral des Que
be
cois. Be Cambridge Univ Press. Berelson, B.R., Lazarsfeld, P.F., McPhee, W.N., 1954. Voting: a Study of Opinion Formation in a Presidential Campaign. University of Chicago Press. Blais, A., Nadeau, R., 1992. To Be or Not to Be Sovereignist: Quebeckers' Perennial Dilemma. Canadian Public Policy/Analyse de Politiques, pp. 89e103.
conomiques et linguistiBlais, A., Martin, P., Nadeau, R., 1995. Attentes e  la souverainete
du que
bec: une analyse prospective et ques et appui a comparative. Can. J. Polit. Sci. 28 (04), 637e657. Blais, A., Gidengil, E., Nadeau, R., Nevitte, N., 2002. Anatomy of a Liberal Victory. Peterborough, Broadview. Clarke, H.D., 1984. Absent Mandate: the Politics of Discontent in Canada. Gage Distribution Company. Clarke, H.D., Kornberg, A., Scotto, T.J., 2009. Making Political Choices: Canada and the United States. Cambridge University Press. Downs, A., 1957. An Economic Theory of Democracy. Enelow, J.M., Hinich, M.J., 1984. The Spatial Theory of Voting: an Introduction. Cambridge University Press. Erk, J., 2010. Is nationalism left or right? critical junctures in quebecois nationalism. Nations Natl. 16 (3), 423e441.
2013a. Riding the
langer, E., Fournier, P., Cutler, F., Soroka, S., Stolle, D., Be orange wave: leadership, values, issues, and the 2011 canadian election. Can. J. Polit. Sci. 46 (04), 863e897.
lineau, F., Harell, A., 2013b. Les syste mes d’aide au vote Fournier, P., Ge ^tre utiles pour l’e
tude des e
lections ?. Working paper peuvent-ils e
de Montre
al. Universite
, G., Langlois, S., 2002. Les raisons fortes: nature et signification de Gagne 
du Que
bec. PUM. l’appui a la souverainete Ghitza, Y., Gelman, A., 2013. Deep interactions with mrp: election turnout and voting patterns among small electoral subgroups. Am. J. Polit. Sci. 57 (3), 762e776. Gidengil, E., Nevitte, N., Blais, A., Everitt, J., Fournier, P., 2012. Dominance & Decline: Making Sense of Recent Canadian Elections. University of Toronto PressHigher Education. Guibernau, M., 2013. Nations without States: Political Communities in a Global Age. John Wiley & Sons. Howe, P., 1998. Rationality and sovereignty support in Quebec. Can. J. Polit. Science/Revue Can. de Sci. politique 31 (1). Jiang, J., Lahiri, P., 2006. Mixed model prediction and small area estimation. Test 15 (1), 1e96.

141

Johnston, R., Blais, A., Brady, H.E., Jean, C., 1992. Letting the People Decide: Dynamics of a Canadian Election. Cambridge University Press. Jusko, K.L., Shively, W.P., 2005. Applying a two-step strategy to the analysis of cross-national public opinion data. Polit. Anal. 13 (4), 327e344. Kalin, R., Berry, J.W., 1982. The social ecology of ethnic attitudes in canada. Can. J. Behav. Science/Revue Can. des Sci. du comportement 14 (2), 97. Lazarsfeld, P.F., Berelson, B., Gaudet, H., 1944. The People’s Choice; How the Voter Makes up His Mind in a Presidential Campaign. Lewis, J.B., Linzer, D.A., 2005. Estimating regression models in which the dependent variable is based on estimates. Polit. Anal. 13 (4), 345e364. Lublin, D., Voss, S., 2002. Context and francophone support for the sovereignty of quebec: an ecological analysis. Can. J. Polit. Sci. 1 (35), 75e101. Mendelsohn, M., 2003. Rational choice and socio-psychological explanation for opinion on quebec sovereignty. Can. J. Polit. Sci. 36 (3), 511e538.
s et appui a  la Nadeau, R., Fleury, C.J., 1995. gains linguistiques anticipe
du que
bec. Can. J. Polit. Science/Revue Can. de Sci. souverainete politique 28 (1). €l, A., The
rien, J.-P., 2008. Left and Right in Global Politics. Cambridge Noe Univ Press. Paolini, S., Harwood, J., Rubin, M., 2010. Negative intergroup contact makes group memberships salient: explaining why intergroup conflict endures. Personal. Soc. Psychol. Bull. 36 (12), 1723e1738. Park, D.K., Gelman, A., Bafumi, J., 2004. Bayesian multilevel estimation with poststratification: state-level estimates from national polls. Polit. Anal. 12 (4), 375e385. Pettigrew, T.F., Tropp, L.R., et al., 2006. A meta-analytic test of intergroup contact theory. J. Personal. Soc. Psychol. 90 (5), 751.
, P., Lublin, D., 2006. Francophone bilingualism, inter-group Piroth, S., Serre contact, and opposition to sovereignty among quebec francophones. Natl. Ethn. Polit. 12, 1e24. Rivero, G., 2011. Integrality and separability in multidimensional voting models: ideology and nationalism in Spanish regional elections, Center for Advanced Study in the Social Sciences, Fundacion Juan March 265. Sambanis, N., Milanovic, B., 2011. Explaining the Demand for Sovereignty. Snijders, T., Bosker, R., 1999. Multilevel Analysis: an Introduction to Basic and Advanced Multilevel Modeling. Sorens, J., 2005. The cross-sectional determinants of secessionism in advanced democracies. Comp. Polit. Stud. 38 (3), 304e326. Tolbert, C.J., Grummel, J.A., 2003. Revisiting the racial threat hypothesis: White voter support for california's proposition 209. State Polit. Policy Q. 183e202. Varshney, A., 2003. Ethnic Conflict and Civic Life: Hindus and Muslims in India. Yale University Press.