Candidate Ethnicity and Electoral Behaviour at the 2010 British General Election

Transcription

1 Work in Progress: Please do not quote or cite without permission Paper prepared for presentation at the ECPR General Conference, Reykjavik, August Candidate Ethnicity and Electoral Behaviour at the 2010 British General Election Stephen D. Fisher, Anthony Heath, David Sanders and Maria Sobolewska Universities of Oxford, Manchester and Oxford, Essex, and Manchester respectively. for correspondence: Abstract This paper considers the extent to which candidate ethnicity affected the party vote choice of both White British and ethnic minority voters at the 2010 British General Election. We find that white voters are less likely to support Muslim candidates, but do not appear to discriminate against ethnic minority candidates more generally. Ethnic minority voters were no more likely to vote for ethnic minority candidates or even candidates with whom they shared a Black or Asian identity. There is some limited evidence that Muslim voters were more likely to support a party if they had a Muslim candidate in their constituency. Introduction There is evidence of discrimination against non-white ethnic minorities in the labour market in Britain and elsewhere (Phalet and Heath, 2010, Heath et al 2008, Heath and Li 2008, Heath et al 2007). Ethnic minorities in Britain also experience disadvantages in education (Heath et al 2008, Heath and Brinbaum 2007), which affect second as well as first generation immigrants (Rothon et al 2009). Some of these disadvantages, which have been characterised as 'ethnic penalties', may in part be due to racism. The British Social Attitudes surveys have shown there is still a substantial proportion of the population who are willing to admit to being at least somewhat prejudiced against people from other races (Rothon and Heath 2003, Heath et al 2010). This context raises the question as to whether ethnic minorities are also at a disadvantage in the political area. This paper addresses the particular issue of whether ethnic minority candidates in UK parliamentary elections face an ethnic penalty from white voters. In addition to querying whether White British voters are more reluctant to vote for non-white candidates, we also ask whether ethnic minority voters are more inclined to support a candidate from the same ethnic group. Since the terrorist attacks on 11 th September 2001 in New York and on 7 th July 2005 in London there has been public debate and expressions of anxiety about the political integration of Muslims in Britain. There have also been claims that Islamophobia or intolerance of Muslims has become pervasive, to the point that bigotry against Muslims is socially acceptable (Kirkup 2011). So this paper also considers whether Muslim candidates in particular appear to suffer an electoral penalty at the polls and that hands of white voters, but also whether a shared Muslim identity encourages Muslim voters to be more likely to support Muslim candidates. 1

2 The next section outlines some important features of the context before we discuss previous research and possible mechanisms by which candidate ethnicity might affects voting. The data and methodology sections then precede the results and concluding discussion. Background on ethnicity in British electoral politics Britain s non-white ethnic minorities constituted over 8% of the population in the 2001 census. The five main minority groups as identified in the census question on ethnicity are Indian, Pakistani, Bangladeshi, Black Caribbean, and Black African. With the exception of some Black Africans from non-commonwealth countries, members of these groups are all entitled to vote and, if registered, they tend to vote as similar rates to the White British majority (Heath et al. 2011). They have also consistently been much more likely to vote Labour than whites (Heath et al. 2011, Saggar and Heath 1999), perhaps mainly because all of the legislation protecting or advancing the rights and opportunities of ethnic minorities has been passed by Labour governments. Regarding the descriptive representation of ethnic minorities, as Table 1 shows, there was a record number of ethnic minority candidates standing at the 2010 general election, and a record number elected. However, 26 MPs represents 4% of the total, less than half of the proportion of ethnic minorities in the population as a whole. Candidature is clearly an important issue here. Most of the increase in the number of ethnic minority MPs between 2005 and 2010 can be accounted for not by the increase in the number of such candidates, but by a new strategy by the Conservatives to ensure several were placed in safe seats (Sobolewska 2011). The raw success rate of EM candidates actually dropped from 2005, and at 20% it was well below the 33% rate for major party candidates on average. TABLE 1 ABOUT HERE Explanations for the levels of ethnic minority representation in parliament have been divided into demand and supply side. On the demand side there has been a perception that (mainly white) voters are less willing to support ethnic minority candidates leaving minority candidate success limited to a few areas with large ethnic minority populations. On the supply side, ethnic minorities are on average poorer and have lower levels of education and so fewer of them possess the kinds of civic resources normally thought to be necessary, and certainly associated with, participation in politics at the elite level. Also there are relatively few opportunities (for whites and minorities alike) to become a prospective parliamentary candidate in a safe or even winnable constituency for your party. So there is tough competition, which maybe even tougher for ethnic minorities if selectors in constituency parties believe that such candidates are likely to do less well in the polls than white candidates (Geddes, 1995). Previous research on the effects of candidate ethnicity This section first reviews what has been learnt about the electoral fortunes of minority candidates in Britain from analysis of constituency election results before considering international evidence and the mechanisms by which candidate ethnicity might influence voting behaviour. Until recent elections there have not been sufficient numbers of ethnic minority candidates 2

3 for systematic analysis, however there were various cases where non-white candidates appear to have done especially badly in the polls because of their race or ethnicity. The failure of the black Conservative candidate John Taylor to win what should have been the safe seat of Cheltenham in 1992 is perhaps the most famous example. More comprehensive analysis of the 1987 and 1992 elections by Lohé (1993) showed that minority candidates typically did worse. More recently, constituency level analysis from the 2010 election shows that ethnic minority candidates from the Conservative and Labour parties did worse than white candidates in areas with few non-white residents (Curtice et al. 2010). This echoed similar findings for the Conservative ethnic minority candidates in 2005, but there was no discernable difference in the performance of Labour candidates on the basis of ethnicity at that election (Curtice et al. 2005). There is also evidence that in areas with large ethnic minority populations, the particular ethnic group membership of the candidate can affect electoral performance. In 1997 despite an 11 percentage point fall nationally in the Conservative share of the vote, the party were up by 5 points in Bethnal Green and Bow and down by less than a point in Bradford West. Both seats have large Muslim populations and saw Conservative Muslim candidates stand against non-muslim Labour candidates (Curtice and Steed 1997). In Bethnal Green and Bow the Labour candidate, Oona King, was from a mixed Jewish and African American background, while in Bradford West Labour fielded a Hindu candidate. While these examples are limited they may be indicative and suggest a need for further investigation of the preferences of ethnic minority voters for candidates from particular ethnic groups. There are various mechanisms by which white voters may become less willing to vote for ethnic minority candidates. The most simplistic is racism and prejudice on the basis of colour, for which there is some evidence from the US, both from experiments (Terkildsen 1993) and observational studies (Carsey 2009). In addition, coming from a certain ethnic group may suggest to voters that the candidate is more likely to possess certain traits or ideological positions, either because these are in fact more common among the group or because of stereotypes. In US experimental studies, Sigelman et al. (1995) found that ethnic minority candidates were perceived to be more compassionate than Anglos, while McDermott (1998) found that Blacks were thought to be more liberal. There is also some evidence that Black candidates maybe be vulnerable to negative publicity that is not explicitly racial but plays on racial stereotypes (Berinsky et al. 2010). Mechanisms that could lead whites to be less likely to vote for minority candidates could also affect support for Muslim candidates. Discrimination may be stronger because of greater social distance and distrust, often labelled Islamophobia. In the US, Benson et al. (2011) found that 40% of survey respondents said they would be unwilling to vote for a Muslim candidate for President, and in their experimental study, Muslims, along with Atheists, were the most discriminated against groups. It is not clear that white British voters would discriminate in the same way. For instance, Ford (2008) found no significant difference between the levels of hostility to Asians and Blacks. The tendency for voters from ethnic minority groups to be more likely to vote for a candidate from their own ethnic group has been well documented in the US (Barreto 2007, Philpot and Walton 2007, Wolfinger 1965). There are a number of reasons for this behaviour, several of which would also be relevant in a UK context. Shared non-white identity and the implications this has for shared interest in anti-discrimination legislation and race-relations policy could generate affinity between non-white voters and candidates. Potentially shared experience of, 3

4 or at least understanding and sympathy with colour-based discrimination may also contribute to this. Non-white candidates might focus on non-white voters as core supporters and devote more effort to mobilizing them (e.g. Leighley 2001, Griffin and Keane 2006, Butler and Broockman 2011). Ethnic minority names on a ballot paper might act as useful heuristics for voters in low information environments (Barreto 2007). There may be a desire among ethnic minorities to see more non-white faces in Parliament. Or similarly, ethnic minority candidate may spark group loyalty, pride and interest (Tate 1991). While these are good reasons to expect ethnic minorities in Britain to vote for other ethnic minorities, most are more likely to motivate ethnic minorities to vote, not for any non-white candidate, but someone from their specific ethnic group. The different ethnic minority groups in Britain have separate identities and are far from an homogenous group with a strong shared non-white immigrant-origin identity. Sometimes their ethnic identities are actually hostile to each other. For example, the attitudes of Indians, Bangladeshis and Pakistanis to each other in part reflects the attitudes of people in those countries to each other. There have also been tensions, and some violent incidents, between Asian and Black groups, such as the 2005 Lozells riots. The Bradford West and Bethnal Green and Bow 1997 examples would seem to suggest that these differences are likely to be important. The final question that this paper addresses is the extent to which Muslim voters in particular are more likely to support Muslim candidates over others. Why might we see such behaviour? To some extent the Islamophobia discussed above could provoke solidarity among Muslims in Britain. But even without this, they share strong religious beliefs and identity which could lead Muslim voters to support Muslim candidates. Mosques and Islamic organizations may reinforce this. Laurence and Maxwell (2010) in their broad discussion of immigrant incorporation into European politics remarked specifically on the extent to which Pakistanis and Bangladeshis in Britain have successfully organized and mobilized their own communities to elect their own members. Since these two groups account for the large majority of Muslims in Britain, it should be noted that what may appear in the analysis as a Muslim effect could in fact be driven in part by voting on the basis of shared ethnicity within the two largest Muslim ethnic groups. Before proceeding further it is worth striking a note of caution regarding the extent to which we should expect findings from the US to translate to the British context. Not only are the histories and contemporary contexts of race and ethnicity within the two countries rather different, but there are important institutional differences too. While candidate traits and ideology are clearly important and legitimate concerns in US electoral behaviour because parties are weak and legislators frequently vote against their party, the high levels of legislative party cohesion in Britain s parliamentary system mean that policy outcomes depend almost entirely on which (usually single) party forms the government. Given the greater importance of parties relative to candidates in Britain, there is a more limited basis for candidate ethnicity to matter. Thus it is striking that Black and Erikson (2006) found no significant effect of candidate background on electoral behaviour in Canada, which also has a Westminster system of government. So if it is the case that white voters in Britain are systematically less willing to vote for ethnic minority candidates after controlling for factors affecting party preference, it suggests that whites are willing to compromise substantive representation in order to avoid voting for minority candidates. 4

5 Data Information on the ethnicity and religion of candidates comes from various sources, including the BBC, Pippa Norris (2010) and The Guardian. They have been cross-checked and validated by searches for candidate websites and other online information. Coding of ethnic group and religion is inevitably ambiguous in some cases. Candidates were coded as ethnic minorities if there was public information stating their background as such. These included candidates with mixed White and non-white minority backgrounds. There was insufficient information to adequately distinguish candidates from a Black African background from those of a Black Caribbean heritage, so these groups were combined, along with mixed Black heritage. Similarly, many candidates described themselves or were described as Asian, rather than more specifically Pakistani, Bangladeshi, Indian, etc. so for the purposes of analysis this category was used rather than attempting a more detailed coding. Since one of our aims is to assess the effects of co-ethnicity between candidates and votes and there are too few survey respondents with Asian backgrounds outside the Indian sub-continent, a classification of South Asian was generated, rather than a broader Asian category. The individual-level data on white voters come from the British Election Study (BES) internet panel for The cross-section face-to-face survey had insufficiently broad coverage of constituencies. Data on ethnic-minority voters come from the BES face-to-face post-election survey and the Ethnic Minority British Election Study (EMBES) 2010 survey, which is based on a face-to-face probability sample of the five main minority groups in Britain: Black African, Black Caribbean, Indian, Bangladeshi and Pakistani. The survey has at least 85% coverage of these populations, but it is disproportionately drawn from areas with high ethnic minority populations. Methodology The statistical models used to estimate the effects of candidate ethnicity are conditional logit models. These are similar to a multinomial logit model but allow party/candidate-and-votercombination specific (known as alternative specific) variables as predictors. This enables a single coefficient for the effects of candidate ethnicity, or co-ethnicity, to be estimated regardless of which party the candidate comes from. Another advantage of this more parsimonious specification is that it reduces the need to estimate the effects of different patterns of minority candidature. With three parties, there are eight different possible patterns of ethnic minority candidature (none, 3 x one, 3 x pairs, and all three with minority candidates). Each of these combinations occurred in at least one constituency and there are both white and ethnic minority respondents in each of the eight types but for some patterns of ethnic candidature there are too few respondents to sensible estimate a separate coefficient. For the white respondents we consider the effects of a party having an ethnic minority candidate and of having a Muslim candidate, operationalized with dummy variables. For the ethnic minority candidates we consider the effects of having an ethnic minority candidate, a co-ethnic candidate, a Muslim candidate and a co-muslim candidate. Note that the co-ethnic candidate variable takes the value 1 when the party has a candidate of the same ethnicity (Black or Asian) as the respondent, and 0 otherwise. Similarly the co-muslim candidate variable is 1 only when both the candidate and the respondent were Muslim, and is otherwise 0. The control variables in the models are Strength of party identification (a four-point scale 5

6 from 0 = not an identifier, through to 3 = very strongly identify with the party); Party like score (an eleven-point scale from 0 = strongly dislike to 10 = strongly like the party); Leader like score (an eleven-point scale from 0 = strongly dislike to 10 = strongly like the party leader); Party best on the most important issue (a binary variable equal to 1 if the party was given as the one thought best to solve what the respondent felt was the most important issue, and 0 otherwise); to capture campaign effects Contacted by party (a binary variable equal to 1 if the respondent was contacted by the party during the campaign, or 0 otherwise); and finally to capture Duvergerian strategic voting incentives, Distance from Contention (the distance from second place for third and lower placed parties, measured by the difference between the share of the vote for the party in question and that for the party that came second in the constituency at the 2010 election, or 0 if the party in question came first or second in the constituency). All of these are alternative-specific, so their values vary within respondents between parties. The controls were chosen because they are proximate to the vote choice. Potential additional control variables are likely to mediated by variables already in the models. So for instance, variables capturing social characteristics of the voter were not included in the models on the basis that these factors would influence voting through strength of party identification, preference, leader evaluations etc. (The only exception is that main effects of being a Muslim are included where co-muslim identity with a candidate effects are being estimated.) Additional models, not shown, suggest this, with only moderate idiosyncratic effects. Similarly, measures of party proximity on policy position issues and economic evaluations have not been included for reasons of data limitation and because they should operate on vote choice through party preference. Survey design weights were used as importance weights in the analysis. Results Table 2 shows the coefficients from three conditional logit models of vote choice for whites who voted for the Conservatives, Labour or Liberal Democrats. The control variables all operate as expected. Model (1) shows no significant effect of ethnic minority candidature, but Model (2) indicates that white voters are significantly less likely to vote for Muslim candidates. This finding is robust, in Model (3), after controlling for ethnic minority candidature. TABLE 2 ABOUT HERE. To more clearly understand the magnitude of the anti-muslim-candidate effect in Model (2) of Table 2, predicted probabilities from that model are presented for different scenarios in Table 3. The first row shows the predicted probabilities from the model of voting for each party for someone in a constituency with no Muslim candidates, who did not identify with any party and for whom all the other variables were at their overall (as opposed to alternative specific) means. Under these conditions the probability of voting Liberal Democrat is much higher than their actual vote share, not least because in practice they have fewer identifiers, are liked less and come third much more often than the two main parties. TABLE 3 ABOUT HERE The bottom three rows of Table 3 show how the baseline probabilities would change if each 6

7 party were alone in fielding a Muslim candidate. They show that the chances of voting for the party would decline by between 7 and 12 percentage points. The differences in the effects for the different parties are in part a function of the baseline probabilities and nothing should be inferred from them since the model has just one parameter for the effect of Muslim candidature on the log odds scale which is the same for all parties. Similarly the estimated effects on the probability scale vary with different values of the control variables, so the figures in Table 3 are intended to be indicative. Nonetheless, this simulation does suggest that the estimated effect of Muslim candidature in Table 2 is substantial. Turning to the electoral behaviour of ethnic minorities, Table 4 shows a similar set of conditional logit models to those in Table 2, but estimated with ethnic minorities who voted for one of three main parties only. Model (1) shows that ethnic minority voters were not significantly more likely to vote for ethnic minority candidates. Nor were they more likely to vote for co-ethnic candidates (Model (2)). Models (3) and (4) show some signs that Muslim voters are more likely to vote for Muslim candidates, but the parameter is not statistical significant. However, it should be noted that if we took the tiny and statistically insignificant main effect of Muslim candidature out of the model then we would see a significant effect of co-muslim candidature. Nonetheless, it is safer to conclude that there are only modest indications that Muslim voters prefer to vote for Muslim candidates. Conclusions To summarize the main results of this paper are that white British voters are less likely to vote for Muslim candidates, but there is insufficient evidence that minority candidates in general face an ethnic penalty at the ballot box. While there are slight, and statistically insignificant, indications that Muslim voters are more likely to vote for Muslim candidates than others, there is no sign that ethnic minority voters prefer to vote for ethnic minority candidates in general or even co-ethnic candidates. One possible explanation for the relative unwillingness of whites to vote for Muslims is that the Muslim candidates in 2010 were not as good as other candidates on average. If anything the reverse seems more likely given the more formidable difficulties ethnic minority candidates face in seeking prospective parliamentary candidatures from major parties. The absence of a co-ethnicity effect may be largely due to the unfortunately crude nature of the measure of co-ethnicity, based on an almost binary classification of minorities into Black (comprising Black African and Black Caribbean) and Asian (comprising Indian, Pakistani and Bangladeshi), plus a few others. Given this difficulty of adequately controlling for coethnicity, it may well be that the indications of a co-muslim candidature effect are actually driven in part by a co-ethnicity effect. Most Muslims in Britain are of Pakistani or Bangladeshi origin and it may be shared Pakistani or shared Bangladeshi heritage with candidates rather than shared religion that matters here. Perhaps the most striking feature of the results for the analysis of ethnic minority behaviour is the generally null findings. This suggests that there is relatively little evidence for demand among ethnic minorities in Britain for better descriptive or symbolic representation. Also while some research on the US suggests that there are benefits to be gained for parties from fielding ethnic minority candidates in generating support from minority voters, the null findings here suggest that that any such benefits appear to be uncertain at best. Moreover, in the case of Muslim candidates, any gains from Muslim voters seem to be outweighed by the 7

8 costs associated with loss of support from white voters. Acknowledgements The authors would like to thank Per Block for assistance with the coding of candidate ethnicity. Previous versions of this paper were presented in Manchester and Oxford and we are grateful for helpful comments from participants. References Barreto, Matt A ISí Se Puede! Latino Candidates and the Mobilization of Latino Voters. American Political Science Review 101(03): 425. Benson, Brett V., Jennifer L Merolla, and John G Geer Two steps forward, one step back? Bias in the 2008 presidential election. Electoral Studies. Berinsky, Adam J et al Sex and Race: Are Black Candidates More Likely to be Disadvantaged by Sex Scandals?. Political Behavior 33(2): Black, Jerome H, and Lynda Erickson Ethno-racial Origins of Candidates and Electoral Performance: Evidence from Canada. Party Politics 12: Butler, Daniel M, and David E Broockman Do Politicians Racially Discriminate Against Constituents? A Field Experiment on State Legislators. American Journal of Political Science 55(3): Carsey, Thomas M The Contextual Effects of Race on White Voter Behavior: The 1989 New York City Mayoral Election. The Journal of Politics 57(01): Curtice, John, Stephen Fisher, and Robert Ford Appendix 2: An Analysis of the Results. In Dennis Kavanagh and Philip Cowley, The British General Election of Basingstoke: Palgrave: MacMillan, p Curtice, John, Stephen Fisher, and Michael Steed Appendix 2: The Results Analysed. In Dennis Kavanagh and David Butler, The British General Election of Basingstoke: Palgrave Macmillan, p Curtice, John, and Michael Steed Appendix 2 The Results Analysed. In David Butler and Dennis Kavanagh, The British General Election of Basingstoke: Macmillan Press, p Ford, Robert Is racial prejudice declining in Britain?. The British Journal of Sociology 59(4): Geddes, A The `Logic' of Positive Action?: Ethnic Minority Representation in Britain after the 1992 General Election. Party Politics 1(2): Griffin, JD, and M Keane Descriptive representation and the composition of African American turnout. American Journal of Political Science 50(4): Heath, Anthony F et al Ethnic Heterogeneity in the Social Bases of Voting at the 2010 British General Election. Journal of Elections, Public Opinion & Parties 21(2): Heath et al. Unequal chances: ethnic minorities in Western labour markets. (2007) pp. 715 Heath and Rothon. The second generation in Western Europe: Education, unemployment, and occupational attainment. Sociology (2008) Kirkup, James Tory chief Baroness Warsi attacks bigotry against Muslims. The Telegraph. 8

9 Laurence, Jonathan, and Rahsaan Maxwell The Two-Sided Pressure for Diversification of Political Parties in Western Europe. In Council of European Studies Conference, Montreal, Canada. Lomé, MJ Ethnic minority candidates in general elections. The Political Quarterly. Mcdermott, M L Race and Gender Cues in Low-Information Elections. Political Research Quarterly 51(4): Norris, Pippa (2010) British General Election Constituency Results Release Pantoja, AD, and GM Segura Does Ethnicity Matter? Descriptive Representation in Legislatures and Political Alienation Among Latinos*. Social Science Quarterly 84(2): Phalet and Heath. From ethnic boundaries to ethnic penalties: Urban economies and the Turkish second generation. American Behavioral Scientist (2010) Philpot, TS, and H Walton One of our own: Black female candidates and the voters who support them. American Journal of Political Science. Rallings and Thrasher (2007) British Electoral Facts Rothon, Catherine, and Anthony Heath Trends in racial prejudice. In British Social Attitudes. The 20th Report, Alison Park et al.eds. London: Sage, p Rothon et al. The Educational Attainments of the" Second Generation": A Comparative Study of Britain, Canada, and the United States. The Teachers College Record (2009) Saggar, Shamit, and Anthony Heath Race: Towards a Multicultural Electorate?. In Critical Elections: British Parties and Voters in Long-Term Perspective, Geoffrey Evans and Pippa Norris eds. London: Sage, p Sigelman, CK, L Sigelman, and BJ Walkosz Black candidates, white voters: Understanding racial bias in political perceptions. American Journal of Political Science Sobolewska, Maria Party strategies and the descriptive representation of ethnic minorities: the 2010 British general election. Centre for European Studies Conference, Barcelona. Tate, Katherine Black Political Participation in the 1984 and 1988 Presidential Elections. The American Political Science Review 85(4): Terkildsen, N When white voters evaluate black candidates: The processing implications of candidate skin color, prejudice, and self-monitoring. American Journal of Political Science. Wolfinger, Raymond E The Development and Persistence of Ethnic Voting. 9

10 Table 1 Numbers of Ethnic Minority Candidates and MPs Conservative Labour Liberal Democrat Total Cands MPs Cands MPs Cands MPs Cands MPs Source: Rallings and Thrasher (2007), with authors own figures for

11 Table 2 Conditional logit models of major party vote choice for whites Model (1) Model (2) Model (3) Lab Candidate *** *** *** (0.06) (0.06) (0.06) Lib Dem Candidate 0.50 *** 0.50 *** 0.50 *** (0.06) (0.06) (0.06) Strength of party id 0.52 *** 0.52 *** 0.52 *** (0.03) (0.03) (0.03) Party like score 0.58 *** 0.58 *** 0.58 *** (0.02) (0.02) (0.02) Leader like score 0.19 *** 0.19 *** 0.19 *** (0.02) (0.02) (0.02) Pty best Most Impt Issue 0.92 *** 0.92 *** 0.92 *** (0.06) (0.06) (0.06) Contacted by party 0.91 *** 0.91 *** 0.91 *** (0.09) (0.09) (0.09) Dist from contention *** *** *** (0.00) (0.00) (0.00) EM Candidate (0.11) (0.14) Muslim Candidate ** * (0.17) (0.21) Respondents BIC Source: BES 2010 Campaign Internet Panel Study, post-wave. White respondents only. Standard errors in parentheses. * p < 0.05, ** p < 0.01, *** p <

12 Table 3 Predicted changes in voting probabilities of whites in response to Muslim candidature Labour Lib Dem Conservative Baseline voting probabilities: Change in probability if: Lab Muslim cand, Others not LD Muslim cand, Others not Con Muslim cand, Others not Note: Simulations from model in Table 2. See text for more details. 12

13 Table 4 Conditional logit models of major party vote choice for ethnic minorities Model (1) Model (2) Model (3) Model (4) Lab Candidate 0.30 * 0.30 * 0.29 * 0.28 (0.14) (0.14) (0.14) (0.14) Lib Dem Candidate (0.17) (0.17) (0.17) (0.17) Lab Cand * Muslim (0.21) (0.21) (0.22) (0.22) LD Cand * Muslim (0.25) (0.25) (0.25) (0.25) Strength of party id 1.15 *** 1.15 *** 1.16 *** 1.17 *** (0.07) (0.07) (0.07) (0.07) Party like score 0.34 *** 0.34 *** 0.34 *** 0.34 *** (0.04) (0.04) (0.04) (0.04) Leader like score 0.19 *** 0.19 *** 0.18 *** 0.18 *** (0.03) (0.03) (0.04) (0.04) Pty best Most Impt Issue 0.60 *** 0.60 *** 0.61 *** 0.61 *** (0.13) (0.13) (0.13) (0.13) Contacted by party 0.54 * 0.54 * 0.50 * 0.48 * (0.21) (0.21) (0.21) (0.21) Dist from contention *** *** *** *** (0.01) (0.01) (0.01) (0.01) EM Candidate (0.15) (0.20) (0.23) Co-ethnic Candidate (0.26) (0.29) Muslim Candidate (0.28) (0.33) Co-Muslim Candidate (0.39) (0.42) BIC Source: EMBES and BES 2010 Post-election surveys. Ethnic Minority respondents only. N=1849. Standard errors in parentheses. * p < 0.05, ** p < 0.01, *** p <