Candidate selection and ethnic minority aspirants: Exploring the effect of party selectors’ biases in a PR system

Abstract

This study examines whether ethnic majority party selectors’ ethnic biases (e.g. beliefs about the political traits of ethnic minority aspirants) impede the selection of ethnic minority aspirants in a proportional representation system. To this end, a quasi-experiment was conducted among local party chairs in Flanders (Belgium). The participants were asked to evaluate both a hypothetical ethnic minority aspirant and an ethnic majority aspirant. When the participants were negatively biased against the ethnic minority aspirant, the selection chances of the ethnic minority aspirant vis-à-vis the ethnic majority aspirant declined. Additionally, the quasi-experimental data were linked to an analysis of the real-life candidate lists composed by the participants for the 2018 local elections in Flanders. The participants’ ethnic biases could not predict the percentage of ethnic minority candidates on the real-life candidate lists. Instead, the presence of ethnic minority voters, aspirants and co-selectors determined the diversity of the lists.

Keywords

ethnic minorities political representation candidate selection bias discrimination

Introduction

Whether defined as indigenous, immigrant-origin, language or religious groups, ethnic minorities fail to be proportionately represented in political assemblies across the world. Their under-representation is considered as one of the major deficits of modern democracies, as it not only limits the substantive representation and political participation of ethnic minorities, but also reduces the legitimacy of political institutions as a whole (Ruedin, 2013; Phillips, 1995).

Political scientists examining their under-representation have primarily focused on the success of ethnic minority candidates (Azabar et al., 2020; Celis and Erzeel, 2017; Janssen, 2020; Martin and Blinder, 2020; Portmann and Stojanović, 2019; Teney et al., 2010; Thrasher et al., 2017), whereas research focusing on the candidate selection stage is more scarce, albeit highly relevant (Shah, 2014; Van Der Zwan et al., 2019). Because political parties provide legitimacy and resources to their candidates, party candidates enjoy greater electoral prospects than independents (Brancati, 2008; Tolley, 2019). Furthermore, by deciding which candidates run for safe seats or are assigned eligible list positions, party selectors can further increase the eligibility of some of their candidates.

This article taps into the discussion about the role of party selectors by answering the question whether party selectors’ ethnic biases influence the selection of ethnic minority aspirants. Thus far, attention to how party selectors’ biased expectations influence the selection of ethnic minority aspirants has been limited to party selectors anticipating possible electoral discrimination against ethnic minority candidates (Durose et al., 2013; English, 2020; Norris and Lovenduski, 1995). This study offers a more comprehensive answer to this question as, in addition to party selectors’ electoral considerations, attention is paid to whether party selectors’ selection decisions are influenced by their own ethnic attitudes and biased expectations about the political traits of ethnic minority aspirants. These other types of biases have already been proven to affect voters’ support for ethnic minority candidates (Martin and Blinder, 2020; Portmann and Stojanović, 2019; Visalvanich, 2017).

Gaining insight into this topic is especially relevant in a European context, where the selection of candidates is generally controlled by the party elite or a small group of delegates (Chiru et al., 2020). In political systems characterized by inclusive selection methods, such as the US, the preferences of voters and selectors largely coincide, and thus findings about voters’ ethnic biases (Jacobsmeier, 2015; Jones, 2014; Weaver, 2012) can be expected to affect both the selection and election of ethnic minority candidates.

This study focuses on the local level in Flanders (Belgium). Not only is the Flemish local candidate selection process dominated be a small group of selectors (De Winter et al., 2013), the Flemish context also offers party selectors a unique combination of institutional incentives to select ethnic minority aspirants and the local level is the most accessible for ethnic minority aspirants (see the section The Flemish Local Political Context). Nevertheless, ethnic minority candidates remain strongly under-represented in Flemish local politics, which makes it highly likely that it is the personal preferences of party selectors which impede their presence. Therefore, the Flemish local context acts as a most-likely case.

To examine the effect of party selectors’ ethnic biases on the selection chances of ethnic minority aspirants, this article adopts a twofold approach. First, a quasi-experiment was conducted among Flemish local party chairs, who are the dominant party selectors at the Flemish local level (De Winter et al., 2013), to measure their ethnic biases and how these biases affect the selection chances of hypothetical aspirants. Second, it was examined whether the participants’ ethnic biases affected the real-life candidate lists the participating local party chairs and their parties composed for the 2018 Flemish local elections, which took place a few months prior to the quasi-experimental research. This approach reflects the reality of selection processes in PR systems, where parties look for the most promising candidates, whilst ensuring that their list reflects (part of) the diversity within the population (De Winter et al., 2013). Additionally, whilst the quasi-experiment increases the internal validity by allowing control over extraneous factors, examining how biases affect the real-life list composition enhances the external validity.

The results show that party selectors adopted a variety of ethnic biases and that these biases affected their support for individual ethnic minority aspirants. These biases, however, failed to influence the percentage of ethnic minority candidates on real-life candidate lists. Instead, the presence of ethnic minority voters, aspirants and co-selectors predicted the diversity on the lists.

Party selectors and ethnic biases

Before people become elected representatives, they have to pass four stages of political recruitment. To start, they have to be allowed to stand for office. If so, they subsequently need to apply for candidacy, get selected as candidates, and get elected as representatives (Ashe and Stewart, 2012; Norris and Lovenduski, 1995). Party selectors’ influence starts in the second stage. Aspirants applying for candidacy can be either self-recruited or recruited by party selectors, but aspirants from under-represented groups need more encouragement to stand for office than aspirants from dominant groups (Shah, 2014). Once applied, aspirants need the support of the selectorate. How open party selectors are to diverse aspirants is expected to depend on the (in)formal candidacy rules, the inclusiveness of the selectorate, the decentralization of the selection process and the selection method (Hazan and Rahat, 2010).

Focusing on the candidate selection stage and informal candidacy rules, this article examines whether party selectors’ ethnic biases hinder the selection of ethnic minority aspirants. Party selectors’ ethnic biases stem from people’s tendency to engage in ingroup favouritism and outgroup derogation, and refer to people’s affective, cognitive and behavioural responses to others based on their ethnicity (Fiske, 2002). Different, theoretically distinct, types of ethnic biases can be adopted by party selectors, and they can occur both separately or simultaneously.

To start, party selectors’ ethnic attitudes are taken into account. This type of bias is connected to explicit prejudice, as it can be considered as a blatant form of anti-ethnic minority feelings (Pearson et al., 2009). Previous research has shown that voters’ antipathy towards ethnic minorities strongly diminishes their support for ethnic minority candidates, whereas voters with more positive views on immigration favour ethnic minority candidates over ethnic majority candidates (Martin and Blinder, 2020; Visalvanich, 2017). I expect to find similar effects among party selectors.

H1: When party selectors hold negative ethnic attitudes, the likelihood that an ethnic minority aspirant gets elected declines.

Additionally, party selectors’ perceptions about the political characteristics of ethnic minority aspirants are examined. Like voters (Campbell and Cowley, 2014; Jacobsmeier, 2015; Jones, 2014; Lerman and Sadin, 2016; Weaver, 2012), party selectors can use aspirants’ ethnicity as an indication of their political traits. The advantage of using informational heuristics is that decisions can be made when only limited or ambiguous information is available (Dovidio and Gaertner, 2000; McDermott, 1998). However, a disadvantage of heuristics is that they can lead to biased decisions based on faulty inferences (Lau and Redlawsk, 2001). Discrimination based on heuristics is linked to ‘aversive’ or ‘modern’ racism (Visalvanich, 2017). This type of discrimination is primarily prevalent among well-intentioned people who may still discriminate against ethnic minorities based on more subtle biases (Pearson et al., 2009).

The first political characteristic under scrutiny regards the perceived ideological distance between the party selector and the aspirant. As party memberships are often composed of ideologically congruent party members and ideological misfits (van Haute and Carty, 2011), party selectors are likely to be confronted with aspirants with diverging ideological profiles. The selectorate will try to seek out those aspirants whose ideological position is perceived as close to their own (Dodeigne et al., 2019) because when they do not, parties will appear internally divided, which will affect their electoral success (Greene and Haber, 2015). When party selectors are not yet acquainted with the aspirants, they will make inferences about their ideological position based on heuristics, including the aspirants’ ethnicity. Whereas people project their own ideological position on ingroup members, they are likely to engage in ideological stereotyping of outgroup members (Lerman and Sadin, 2016). For ethnic minorities, these stereotypes reflect the idea that they are more leftist than ethnic majority citizens, given their traditional support for leftist parties (Jacobsmeier, 2015; Jones, 2014). Therefore, the following is expected.

H2: When party selectors stereotype ethnic minority aspirants as more ideologically distant than ethnic majority aspirants, the likelihood that an ethnic minority aspirant gets elected declines.

Furthermore, party selectors are likely to prefer aspirants who seem competent and trustworthy (Bach and Veit, 2018; Greene and Haber, 2015; Ouyang et al., 2017). Ethnic minorities are, however, often stereotyped as lazy, incompetent, unhelpful and dishonest (Rooth, 2010; Stevens and Gorgoz, 2010), although some ethnic minority groups are stereotyped less negatively than others (Binggeli et al., 2014; Bye et al., 2014). Moving to the political context, voters equally tend to perceive ethnic minority candidates as less competent than majority candidates, but results on trustworthiness stereotypes have been mixed (Weaver, 2012; Visalvanich, 2017; Lerman and Sadin, 2016; Jones, 2014; Campbell and Cowley, 2014). I hypothesize that:

H3: When party selectors stereotype ethnic minority aspirants as less competent than ethnic majority aspirants, the likelihood that an ethnic minority aspirant gets elected declines.

H4: When party selectors stereotype ethnic minority aspirants as less trustworthy than ethnic majority aspirants, the likelihood that an ethnic minority aspirant gets elected declines.

Finally, party selectors could engage in strategic or imputed discrimination (Bateson, 2020; Norris and Lovenduski, 1995). This implies that party selectors refrain from selecting ethnic minority aspirants because they expect voters to refuse to vote for ethnic minority candidates. As vote-maximizing actors, parties try to select aspirants whom voters will support and to avoid those who could cost votes to the party. Concerns about voter discrimination against ethnic minority candidates voters (Besco, 2020; Portmann and Stojanović, 2019) could thus reduce party selectors’ support for ethnic minority aspirants (Durose et al., 2013; English, 2020).

H5: When party selectors perceive ethnic minority aspirants as more electorally appealing than ethnic majority aspirants, the likelihood that an ethnic minority aspirant gets elected declines.

Until now, the focus has lied on the probability that a single selector would select individual aspirants. However, it is likely that party selectors’ ethnic biases will have a smaller effect on the composition of entire candidate lists. First, when drawing their candidate lists, party selectors try to find a balance between selecting the most promising candidates and composing a list which reflects (part of) the diversity within the population (De Winter et al., 2013). Thus, in ethnically diverse districts, party selectors might select ethnic minority aspirants even though they are negatively biased against them. Second, because candidate selection processes are seldomly controlled by a single selector, the negative biases of one selector can be offset by the positive biases of another selector. Third, party selectors’ biases can only affect ethnic minority applicants' selection chances when a minimum supply of ethnic minority (potential) applicants is present.

H6: The influence of party selectors’ ethnic biases is smaller at the list level than at the aspirant level.

The Flemish local political context

To answer the research question, this study focuses on the local level in Flanders, the Dutch-speaking part of Belgium. Despite the growing size of the ethnic minority population and their increased politicization over the last several decades, ethnic minorities remain strongly under-represented at all political levels in Belgium. For instance, anno 2014, 7.3% of the members of the federal parliament and 25% of the members of the Brussels parliament had a non-EU background (Celis and Erzeel, 2017; Janssen et al., 2017), whereas people of non-EU descent constituted 13.6% and 42% of the Belgian and Brussels population, respectively (STATBEL, 2021). In the 2018 local elections in Flanders, ethnic minorities were under-represented as well. Ethnic minorities¹ made up 15.8% of the Flemish population in 2017 (Agentschap Binnenlands Bestuur and Statistiek Vlaanderen, 2019), but only constituted 4.8% of the candidates and 3.2% of the elected representatives⁹. On an aggregate level, ethnic minorities thus appear to be worse off at the local level than at higher political levels. However, their level of representation strongly differs between municipalities. Whereas 54% of the candidate lists did not include any ethnic minority candidate, several lists running in diverse cities included more than 20% ethnic minority candidates, many of whom also got elected.² Although comparative figures about the political representation of ethnic minorities are scarce, especially at the local level, a recent study of English (2018) shows that at the national level Belgium performs better than some other European countries such as France, Germany, Italy and Spain, but worse than Great-Britain and the Netherlands.

The under-representation of ethnic minority candidates in Flemish municipalities is remarkable because of the highly favourable institutional context. On the one hand, the local level is the most accessible level for ethnic minority aspirants since ethnic organizations and networks have more impact on selection processes at the local level than at higher levels (Akhtar and Peace, 2018). On the other hand, the Flemish local context strongly facilitates the selection of ethnic minority candidates. To start, as opposed to single-member districts, the Flemish PR system grants parties the opportunity to balance their ticket (Dancygier, 2014). Its semi-open list system also stimulates parties to anticipate voters’ demand for certain types of candidates (Valdini, 2012). Furthermore, voting is compulsory for all Belgian citizens. Ethnic minority citizens who have acquired citizenship thus constitute a more important electorate in Belgium than in non-compulsory voting systems where ethnic minority turnout depends on contextual factors (Fraga, 2016). Finally, non-Belgian citizens enjoy active suffrage at the local level. Given that ethnic minority voters are strong supporters of ethnic minority candidates (Azabar et al., 2020; Teney et al., 2010), these elements should equally motivate party selectors to select ethnic minority aspirants. That ethnic minority candidates remain under-represented despite this favourable context makes it likely that ethnic majority party selectors, who constitute the vast majority of the selectors, discriminate against ethnic minority aspirants. Therefore, the Flemish local context acts as a most-likely case to test the aforementioned hypotheses.

The candidate selection process at the local level is dominated by the local party chair and the local executive committee. In most cases, the local party elite selects the candidates and drafts a candidate list, including a predefined list order in which the candidates will appear. This model list is afterwards presented to the local party members who can either approve or disapprove this model list (De Winter et al., 2013).

Data and methods

A twofold approach was used to test whether party selectors’ ethnic biases affect the selection of ethnic minority aspirants. First, their effect on the selection chances of individual aspirants was measured by means of a quasi-experiment among Flemish local party chairs. Second, to test their effect on the overall list composition, the quasi-experimental data were linked to an examination of real-life candidate lists composed by the party chairs who took part in the quasi-experiment.³

Quasi-experiment

1,379 party chairs, each presiding over a local party branch of one of the seven regional parties in Flanders, were invited to participate in an online quasi-experiment. The local party chairs were chosen as the research population because they constitute the dominant actors in the local candidate selection process (De Winter et al., 2013). Invitations were sent via email. The email addresses of the party chairs were collected via the party headquarters as part of the RepResent Local Chairs Survey 2018, except for those of the liberal party, Open Vld, whose email addresses were collected by the author by screening local party websites. 597 party chairs took part in the quasi-experiment in spring 2019, which results in a response rate of 43.30%. Supplemental Appendix 1 includes the response rate per party.

Each participant was presented two aspirants, one ethnic minority aspirant and one ethnic majority aspirant. The aspirants were presented in a randomized order. Their ethnicity was cued by granting them a common Moroccan (Mohammed El Amrani) or Flemish name (Victor Claes) (STATBEL, s.d.). A Moroccan name was chosen as the Maghrebi community is the most prominent non-Western minority group in Flanders whose political role is one of the most debated (Janssen, 2020) and because citizens of Turkish or Maghrebi descent are most severely affected by prejudice in general (Alanya et al., 2015). Furthermore, a Moroccan name only denotes a foreign origin and not a foreign nationality as passive suffrage in Flemish local elections is limited to Belgian and non-Belgian EU citizens. Both aspirants were given a male forename to control for aspirants’ gender. Hereafter, aspirants with a Moroccan and Flemish name are referred to as ‘Moroccan’ and ‘Flemish’ aspirants.

Both aspirants were randomly assigned a low or high socioeconomic status (SES), and one of two profiles with personal information and policy stances were randomly distributed between them (see Supplementary Material). Aspirants’ SES is used by party selectors to make inferences about their political qualities (Carnes, 2018), and aspirants’ ideological policy stances affect parties’ willingness to select aspirants (Dodeigne et al., 2019). To ensure that the two policy profiles were highly similar, a pre-test was conducted among university students in which they were asked to evaluate different profiles of policy stances. The profiles were composed of stances included in the 2018 election programmes of centrist, centre-leftist and centre-rightist parties of both small and large municipalities. Two profiles were not evaluated differently on any of the dimensions and were selected for the experiment.

The participants were asked to evaluate each aspirant on five 7-point Likert scales (see Supplementary Material). The scales measured the aspirants’ perceived ideological position (1 = very leftist and 7 = very rightist), competence (1 = very incompetent and 7 = very competent), trustworthiness (1 = very untrustworthy and 7 = very trustworthy), electoral appeal (1 = selecting this aspirant would be a big electoral disadvantage and 7 = (…) a big electoral advantage) and chance to get selected (1 = very small and 7 = very big).

Two manipulation checks were included as well. The first measures non-compliance. After each aspirant was presented, the participants were asked to answer a question about the content of each aspirant profile to test whether they had read it. The second aims to control for social desirability. After having evaluated both aspirants, the participants were asked to guess the aim of the experiment. As robustness checks, the analyses discussed below were replicated without the participants who did not answer both content-related questions correctly (N = 165) and who thought the objective had something to do with ethnicity or prejudice (N = 173). The results (not shown) provided identical patterns as the models discussed below.

In the last part of the quasi-experiment, the participants were asked to answer several questions about themselves and their party. Most important are the ideological self-placement of the participants and their ethnic attitudes. The former was measured on a 7-point scale (1 = very leftist and 7 = very rightist), and the latter were measured by asking the participants to evaluate eight statements about non-western ethnic minorities (1 = totally disagree and 7 = totally agree) (see Supplementary Material). Other characteristics are used as control variables (see the Variables section).

Examination of real-life candidate lists

Additionally, because the quasi-experiment was conducted a few months after the 2018 local elections in Flanders, the data of the participants could be linked to the composition of the real-life candidate lists they composed for these elections. I started with checking whether the participating party chairs ran with their local party in the 2018 local elections in Flanders.⁴ For those who did, the ethnicity of their candidates was coded as follows. As Belgium is primarily populated by Dutch-speaking and French-speaking citizens,⁵ all candidates with a non-Dutch and non-French name were examined. Information about the ethnicity of these candidates was sought on news, party and candidate websites and social media accounts. When concrete information was found, the candidates were coded as either ethnic minority or majority candidates. Those of Western-European or Northern-European origin were not considered as ethnic minority candidates because party selectors’ ethnic biases are most likely to affect visible ethnic minorities, who are perceived as ‘others’ (Garrett, 2017; Meuleman and Billiet, 2003). It is also the political representation of visible ethnic minorities which is the most politicized in Belgium (Janssen, 2020). Candidates about whom no concrete information was found were coded according to the geographical distribution of their name (http://forebears.co.uk).

Variables

The analysis was conducted at the level of the party chairs and not at the level of the aspirants who were evaluated. To do so, for each dimension on which the aspirants were evaluated, a new variable was computed which compared the score each participant had assigned to the Moroccan aspirant to the score which was assigned to the Flemish aspirant. This approach allows to make statements about how the personal biases of the party chairs affect their willingness to select Moroccan aspirants vis-à-vis Flemish aspirants and makes it possible to make predictions about whether the ethnic biases of the participating party chairs also affect entire candidate lists. This cannot be achieved when conducting the analysis at the level of the aspirants.

Comparisons between how each participant evaluated the Moroccan and Flemish aspirant were made as follows. The dependent variable measuring the selection chance of the Moroccan aspirant vis-à-vis Flemish aspirant was computed by subtracting for all participants the score they had assigned to the Flemish aspirant from the score they had assigned to the Moroccan aspirant. Negative values mean that the Moroccan aspirant was less likely to get selected than the Flemish aspirant, whereas positive values mean the reverse. To calculate the perceived ideological distance between the participants and the Moroccan aspirant vis-à-vis the ideological distance between the participants and the Flemish aspirant, the ideological scores assigned to the aspirants were first subtracted from the ideological self-placement of the participant. Thereafter, the absolute ideological distance between the participants and the Flemish aspirant was subtracted from the absolute ideological distance between the participants and the Moroccan aspirant. Positive values implied that the participants perceived the Moroccan aspirant as more ideologically distant than the Flemish aspirant. Finally, to measure the perceived competence, trustworthiness and electoral appeal of the Moroccan aspirant vis-à-vis the Flemish aspirant, the score given to the Flemish aspirant was subtracted from the score given to the Moroccan aspirant. Negative scores indicated that the participant perceived the Moroccan aspirant as less competent, trustworthy or electorally appealing. These four variables are used as predictors.

A potential drawback of this approach, however, is that these comparisons do not take into account differences between the aspirants’ SES and policy profile. To check whether this was the case, robustness checks were calculated based on a comparison between ‘adjusted’ scores granted to the Flemish and Moroccan aspirants. To calculate these adjusted scores, I started by computing two regression models per evaluation dimension estimating how the SES and policy profile of Flemish and Moroccan aspirants affected their scores. Next, aspirants with a high SES and policy profile B were chosen as the baseline category because high-SES aspirants were perceived as more competent and slightly more trustworthy than low-SES aspirants, and aspirants with profile B were perceived as more centrist than aspirants with profile A. Subsequently, the scores granted to aspirants with a different profile were adjusted to match the baseline category. For instance, for Flemish aspirants, having a high SES increased their competence score with 0.45, whereas being assigned policy profile B reduced the competence score with 0.03. Consequently, Flemish aspirants with a low SES and policy profile A were increased with 0.42. These adjusted scores were subsequently used to compute identical variables as those discussed above. As the models based on the adjusted scores provided the same results as the models based on the unadjusted scores, the latter are discussed below. The regression models used to calculate the adjusted scores and the models estimating the selection of ethnic minority aspirants based on these adjusted scores are included in the Supplementary Material.

In addition to party selectors’ perceptions of individual Moroccan and Flemish aspirants, also other variables were included in the analysis. First, for those party chairs who participated in the quasi-experiment and who participated with their local party in the 2018 local elections, two additional dependent variables were calculated: the percentage of ethnic minority candidates on their list and the relative list position assigned to ethnic minority candidates. The latter was calculated as follows: $t_{i} - p_{j} / t_{i} - 1$ for which $t_{i}$ stands for the total amount of candidates on list $i$ and $p_{j}$ stands for the position of candidate $j$ on list $i$ (Van der Zwan et al., 2019). Candidates’ relative list positions range from zero (last candidate) to one (first candidate). Second, as a predictor variable, party selectors’ ethnic attitudes were taken into account. The participants were asked to evaluate eight statements about non-western ethnic minorities. The evaluation of three of these statements (‘people of non-western origin who live in Flanders enrich Flanders’ cultural life’, ‘people of non-western origin who live in Flanders increase criminality rates’ and ‘people of non-western origin who live in Flanders turned Flanders into a better place to live’) were computed into a single measure (1 = very negative and 7 = very positive) (Cronbach’s Alpha = 0.847). Third, the regional party affiliation of the local party chairs was included as a control variable because leftist parties are more open towards (ethnic minority) group representation (Ceyhan, 2018). Because only eight participants belonged to the far left, they were excluded from the analyses which included party affiliation. Fourth, the size of the ethnic minority population within each district (Agentschap Binnenlands Bestuur and Statistiek Vlaanderen, 2018; Stad Antwerpen, 2020)⁶ was controlled for because parties select more ethnic minority aspirants and assign them to better list positions when they are running in more diverse districts (Geese and Schacht, 2018; Tolley, 2019; Shah, 2014). Fifth, a control variable was included measuring the proportion of ethnic minority party members. This variable acts both as an indication of the ability of the local parties to attract ethnic minority voters and as a proxy of the supply of ethnic minority aspirants (Ashe and Stewart, 2012). The participants of the quasi-experiment were asked to estimate how many people were a member of their party and how many of them had an ethnic minority background. Sixth, the presence of ethnic minority co-selectors was controlled for as they positively affect the selection of ethnic minority candidates (Tolley, 2019). Anticipating the under-representation of ethnic minority party chairs (only 7 of 597 participants had a minority background⁷), I asked the participants about the ethnic composition of their local executive committee. These committees are, besides the party chair, the most important actor in the candidate selection process (De Winter et al., 2013). When such a committee was present, the percentage of ethnic minority members was controlled for.

Table 1 includes an overview of the descriptives of all the variables included in the analysis. The effect of party selectors’ ethnic biases on the selection chances of individual aspirants was estimated by OLS regression models. Their effect on the real-life list composition was measured by fractional regression models because the dependent variables regarded proportion data. All models were checked for multicollinearity.

Table 1.

Descriptive statistics.

Variable		N	Mean	SD	Min	Max
Dependent	Perceived difference in selection chance	597	0.05	1.27	−6	4
	% ethnic minority candidates on list	468	0.04	0.06	0.00	0.36
	Relative list position of ethnic minority candidates	557	0.54	0.26	0	1
Predictor	Ethnic attitudes	584	4.55	1.30	1	7
	Perceived difference in ideological proximity	597	0	1.26	−4	4
	Perceived difference in competence	597	0.08	1.04	−6	4
	Perceived difference in trustworthiness	597	0.07	0.80	−3	3
	Perceived difference in electoral appeal	597	−0.08	1.18	−6	3
Control	Party affiliation	597	—	—	—	—
	Far left	8	—	—	—	—
	Greens	104	—	—	—	—
	Social-democrats	107	—	—	—	—
	Christian-democrats	131	—	—	—	—
	Liberals	80	—	—	—	—
	Regionalists	116	—	—	—	—
	Far right	51	—	—	—	—
	% ethnic minority population	581	10.14	8.78	1.00	51.00
	% ethnic minority party members	476	5.42	9.63	0.00	100.00
	% ethnic minority members in executive committee	572	4.18	9.48	0.00	100.00

Results

Figure 1 shows that a minority of the party chairs indeed held negative ethnic biases. 19% of the participating party chairs had negative ethnic attitudes (<4). Additionally, 34% of the party chairs perceived the Moroccan aspirant as more ideologically distant from themselves than the Flemish aspirant. The Moroccan aspirant was also evaluated as less competent by 21% of the participants, as less trustworthy by 16% of the participants and as less electorally appealing by 27% of the participants.

Figure 1.

Histograms showing the distribution of party selectors’ ethnic biases.

Table 2 shows that when party selectors were negatively biased against the Moroccan aspirant, the selection chances of the latter vis-à-vis the Flemish aspirant deteriorated. The only exception is party selectors’ ethnic attitudes, the effect of which was only statistically significant when the control variables were excluded from the model. H1 is therefore rejected. The effects of the other four ethnic bias variables are statistically significant in both models and are in the expected direction, thereby confirming H2−H5. The effects of these variables are shown in Figure 2. The main plots are based on the regression models including the control variables. Each plot is accompanied by a histogram showing the distribution of the scores granted by the participants. The y-axes show the difference in selection chances assigned to the Moroccan vis-à-vis Flemish aspirant. Recall that the participants were asked to indicate how likely they would be to select each aspirant on a 7-point scale (1 = very unlikely and 7 = very likely). A difference of −1 on the y-axis thus means that the participants positioned the Moroccan aspirant one step lower on the scale than the Flemish aspirant, whereas a difference of +1 means the reverse. All plots show that the party chairs who were not biased against either one of the aspirants were as likely to select the Moroccan aspirant as the Flemish aspirant. Participants who were negatively or positively biased against the Moroccan aspirant, however, were less or more likely to select the Moroccan aspirant than the Flemish aspirant. The upper left plot shows the effect of the perceived ideological proximity of the Moroccan vis-à-vis Flemish aspirant. Party chairs who perceived the Moroccan aspirant as much more ideologically distant from themselves than the Flemish aspirant (right end of x-axis) were less likely to select the former compared to the latter. The disadvantage of Moroccan aspirants amounted almost to 1 point. An opposite pattern emerges among the party chairs who perceived the Moroccan aspirant as much more ideologically close to themselves than the Flemish aspirant (left end of x-axis). The three other plots show a similar image. The participants who perceived the Moroccan aspirant as much less competent, trustworthy or electorally appealing than the Flemish aspirant (left ends of x-axes) were less likely to select the former vis-à-vis the latter. Among those participants who were the most negatively biased against the Moroccan aspirant, the disadvantage of the latter ranged from almost −1 (trustworthiness) to −2.5 (electoral appeal). Participants who perceived the Moroccan aspirant as much more competent, trustworthy and electorally appealing (right ends of x-axes) were more likely to select the Moroccan aspirant than the Flemish aspirant.

Table 2.

Linear regression models predicting the selection chance of the Moroccan hypothetical aspirant vis-à-vis the Flemish hypothetical aspirant.

	Model 1	Model 2
	B (SE)	B (SE)
Intercept	−0.21 (.15)	−0.03 (.24)
Ethnic attitudes	0.06^ (.03)	−0.01 (.05)
Perceived difference between Moroccan and Flemish aspirant in terms of …
Ideological proximity	−0.20*** (.03)	−0.17*** (.04)
Competence	0.22*** (.04)	0.21*** (.04)
Trustworthiness	0.26*** (.05)	0.29*** (.06)
Electoral appeal	0.45*** (.04)	0.40*** (.04)
Party (ref. Christian-democrats)
Greens	—	0.22^ (.13)
Social-democrats	—	0.30* (.13)
Liberals	—	0.24^ (.14)
Regionalists	—	0.15 (.13)
Far right	—	−0.46* (.20)
% ethnic minority population	—	0.00 (0.00)
% ethnic minority party members	—	0.00 (.01)
% ethnic minority members in executive committee	—	0.01 (.01)
Adjusted R²	0.52	0.55
Number of cases	584	445

Notes: Significant codes: p < 0.001 = ‘***’, p < 0.01 = ‘**’, p < 0.05 = ‘*’, p < 0.1 = ‘ ^ ’.

Figure 2.

The significant effects of party selectors’ ethnic biases on the selection chance of the hypothetical Moroccan aspirant vis-à-vis the hypothetical Flemish aspirant. Notes: Marginal effects based on Model 2, Table 2. Other continuous predictors are held constant at their mean. For categorical predictors, the proportions of the categories are averaged. 95% confidence intervals are shown. Variables were calculated by subtracting the scores the participants granted to the Flemish aspirant from their score assigned to the Moroccan aspirant.

The only control variable which had a statistically significant effect in Table 2 is party affiliation. As opposed to the centrist Christian-democrats, greens, social-democrats and liberals were more likely to prefer the Moroccan aspirant vis-à-vis the Flemish aspirant. Far right parties, in turn, were less likely to nominate the former.

Subsequently, for those participating party chairs who ran with their local party in the 2018 local Flemish elections, information was gathered about the ethnic composition of the real-life candidate lists composed by the participants. A comparison between the party affiliation of the participants under examination in Table 3 (Model 2) and the party affiliation of all the individual party lists running in the 2018 local elections⁸ (N = 1,144) showed that this sample is fairly representative of the real-life elections. The degrees to which the different parties were under-represented or over-represented ranged from 0.78 (Far right) to 1.33 (Greens), with exception of the far left, which was strongly under-represented (0.57). For details see the Supplementary Material.

Table 3.

Fractional regression models predicting the percentage of ethnic minority candidates on the candidate lists composed by the participants of the quasi-experiment for the 2018 Flemish local elections.

	Model 1	Model 2
	B (SE)	B (SE)
Intercept	−4.54*** (.38)	−3.97*** (.35)
Ethnic attitudes	0.30*** (.07)	−0.04 (.07)
Perceived difference between Moroccan and Flemish aspirant in terms of …
Ideological proximity	−0.06 (.07)	0.07 (.06)
Competence	−0.09 (.08)	0.07 (.07)
Trustworthiness	−0.14 (.11)	0.02 (.09)
Electoral appeal	0.13^ (.08)	0.02 (.06)
Party (ref. Christian-democrats)
Greens	—	0.22 (.18)
Social-democrats	—	0.21 (.20)
Liberals	—	−0.02 (.20)
Regionalists	—	−0.49* (.20)
Far right	—	−1.05** (.35)
% ethnic minority population	—	0.05*** (.01)
% ethnic minority party members	—	0.02** (.01)
% ethnic minority members in executive committee	—	0.03*** (.01)
D²	0.08	0.47
Number of cases	458	353

Notes: Significant codes: p < 0.001 = ‘***’, p < 0.01 = ‘**’, p < 0.05 = ‘*’, p < 0.1 = ‘ ^ ’.

Table 3 shows that party selectors’ biases did not affect the percentage of ethnic minority candidates on real-life candidate lists. When only the bias variables were included (Model 1), party selectors’ ethnic attitudes and their perception of the electoral appeal of the Moroccan vis-à-vis Flemish aspirant had a significant positive effect. However, once the control variables were added (Model 2), the effects became statistically insignificant. The results are in line with H6 predicting that the effect of party selectors would be smaller at the list level than at the individual level.

Instead, many of the control variables exerted statistically significant effects (Table 3). To start, both regionalist and far right parties nominated significantly less ethnic minority candidates on their lists compared to Christian-democratic parties.⁹ Additionally, more diverse districts are associated with higher shares of ethnic minority candidates. To rule out the possibility that this effect was driven by outliers, an additional analysis was conducted which excluded all lists present in municipalities with an ethnic minority population of more than 22.5% (Q₃ + 1.5*(Q3 − Q1)). The results (not shown) confirmed the effect. Finally, the percentages of ethnic minority members in the party and the party’s executive committee positively affected the percentage of ethnic minority candidates. Due to the right skewed distribution of both variables, their effects were also tested using dummy variables (ethnic minority members present or not?). The effects of the dummy variables were statistically significant as well (results not shown).

Two additional fractional models showed that party selectors’ ethnic biases did not have an effect on the relative list position of ethnic minority candidates (results not shown). All control variables were statistically insignificant as well.

Finally, robustness checks measuring the effect of party selectors’ ethnic biases on the percentage and relative list position of solely candidates of Turkish and Maghrebi descent show no outspoken effect of the ethnic bias variables (results not shown). These checks were conducted because the ethnic minority aspirants presented in the quasi-experiment were assigned a Moroccan name. Party selectors’ ethnic attitudes had a significant positive effect on the percentage of Turkish and Maghrebi descent, but only when the control variables were omitted from the model. In the model with the control variables, both the competence and ideological proximity measure had a significant positive effect. None of the bias variables exerted a significant effect on the relative list position of Turkish and Maghrebi candidates.

Conclusion

This article discussed how party selectors’ ethnic biases contribute to the under-representation of ethnic minorities. The results showed that ethnic biases are present among Flemish party chairs, but that their effect on the selection of ethnic minority candidates is ambiguous. In general, party selectors’ ethnic biases diminished the participants’ support for individual ethnic minority vis-à-vis majority aspirants. However, party selectors’ ethnic biases did not influence the ethnic composition of the real-life candidate lists composed by the participants for the 2018 local elections.

Instead, the percentage of ethnic minority candidates on the real-life candidate lists was positively correlated with the share of the ethnic minority population, the degree of ethnic minority party members and the percentage of ethnic minority members in parties’ executive committee. The first two factors emphasize the importance of ethnic minority voters, although the second aspect also highlights the importance of the supply of ethnic minority aspirants, as party members are an important pool of potential candidates. The third factor, in turn, points to the relevance of ethnic minority co-selectors, as the local executive committee is, next to the party chair, the most dominant actor in the local selection process (De Winter et al., 2013). Not only are ethnic minority selectors more likely to select ethnic minority candidates, they can also be expected to be more likely to actively recruit ethnic minority aspirants (Tolley, 2019).

The implications of these results are threefold. First, party selectors’ practical and vote-seeking considerations can trump the effect of their personal biases. Second, party selectors’ biases are more likely to impede the selection of ethnic minority candidates when the selectorates are exclusive and homogenous, and especially when parties are running in non-diverse districts. Third, in single-member districts, where ticket balancing is not possible, party selectors’ ethnic biases can be assumed to result more easily in discrimination than in PR systems.

However, future research addressing the limitations of this study is required to get a more complete view on how party selectors’ ethnic biases affect the selection of ethnic minority aspirants. To start, the ethnic biases measured in this study were restricted to non-western minorities in general or minorities of Maghrebi descent in particular. Future research comparing how party selectors perceive aspirants of different ethnic minority backgrounds will enhance insights into which aspirants are most heavily disadvantaged. Additionally, this study only focused on a single political context and on the selection (chances) of ethnic minority aspirants. Examinations of other political contexts as well as studies focusing on how party selectors’ ethnic biases affect the recruitment of ethnic minority aspirants will help to understand when party selectors’ own biases are most influential.

Supplemental Material

sj-pdf-1-ppq-10.1177_13540688211041588 – Supplemental Material for Candidate selection and ethnic minority aspirants: Exploring the effect of party selectors’ biases in a PR system

Supplemental Material, sj-pdf-1-ppq-10.1177_13540688211041588 for Candidate selection and ethnic minority aspirants: Exploring the effect of party selectors’ biases in a PR system by Sigrid Van Trappen in Party Politics

Footnotes

Acknowledgements

I wish to thank the members of research group GASPAR for their useful comments on a previous version of this paper as well as Peter Thijssen for his constructive feedback on an earlier draft of this paper which was presented at the 2020 ‘Belgium: The State of the Federation’ Conference. Special thanks also goes to the anonymous reviewers, whose comments strongly improved this article.

Declaration of conflicting interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study is financially supported by Fonds Wetenschappelijk Onderzoek.

ORCID iD

Sigrid Van Trappen

Supplemental material

Supplemental material for this article is available online.

Notes

Author biography

Sigrid Van Trappen is a PhD researcher at the Department of Political Sciences of Ghent University. Her main research interest comprises the political representation of social groups. https://orcid.org/0000-0002-9010-916X

References

Agentschap Binnenlands Bestuur and Statistiek Vlaanderen (2018) Lokale Inburgerings- en Integratiemonitor editie 2018. Available at: https://integratiebeleid.vlaanderen.be/lokale-inburgerings-en-integratiemonitor-editie-2018.

Agentschap Binnenlands Bestuur and Statistiek Vlaanderen (2019) Lokale Inburgerings- en Integratiemonitor editie 2019. Available at: http://integratiebeleid.vlaanderen.be/lokale-inburgerings-en-integratiemonitor-editie-2019-0.

Akhtar

Peace

(2018) Ethnic minorities in British politics: Candidate selection and clan politics in the labour Party. Journal of Ethnic and Migration Studies 45(1): 1–17.

Alanya

Baysu

Swyngedouw

(2015) Identifying city differences in perceived group discrimination among second-generation Turks and Moroccans in Belgium. Journal of Ethnic and Migration Studies 41(7): 1088–1110.

Ashe

Stewart

(2012) Legislative recruitment. Party Politics 18: 687–707.

Azabar

Thijssen

Van Erkel

(2020) Is there such thing as a Muslim vote? Electoral Studies 66.

Bach

Veit

(2018) The determinants of promotion to high public office in Germany: Partisan loyalty, political craft, or managerial competencies? Journal of Public Administration Research and Theory 28: 254–269.

Bateson

(2020) Strategic discrimination. Perspectives on Politics 18: 1068–1087.

Besco

(2020) Friendly fire: Electoral discrimination and ethnic minority candidates. Party Politics 26: 215–226.

10.

Binggeli

Krings

Sczesny

(2014) Stereotype content associated with immigrant groups in Switzerland. Swiss Journal of Psychology 73: 123–133.

11.

Brancati

(2008) Winning alone: The electoral fate of independent candidates worldwide. The Journal of Politics 70(3): 648–662.

12.

Bye

Herrebrøden

Hjetland

, et al. (2014) Stereotypes of norwegian social groups. Scandinavian Journal of Psychology 55: 469–476.

13.

Campbell

Cowley

(2014) What voters want: Reactions to candidate characteristics in a survey experiment. Political Studies 62: 745–765.

14.

Carnes

(2018) The Cash Ceiling: Why Only the Rich Run for Office - and What We Can Do about It. Princeton, New Jersey: Princeton University Press.

15.

Celis

Erzeel

(2017) The complementarity advantage: Parties, representativeness and newcomers' access to power. Parliamentary Affairs 70: 43–61.

16.

Ceyhan

(2018) Why not ask the representatives? Parliamentary candidates' attitudes towards immigrant representation in Germany. Representation 54: 103–127.

17.

Chiru

De Winter

Vandeleene

(2020) Candidate selection. In: De Winter

Karlsen

Schmitt

(eds), Parliamentary Candidates between Voters and Parties. A comparative Perspective. Oxon: Routledge, pp. 54–77.

18.

Dancygier

(2014) Electoral rules or electoral leverage? Explaining Muslim representation in England. World Politics 66: 229–263.

19.

De Winter

Erzeel

Vandeleene

, et al. (2013) Nationale bemoeienis of lokale autonomie? Het lijstvormingsproces bij gemeenteraadsleden. In: Deschouwer

Verthé

Rihoux

(eds), Op zoek naar de kiezers: lokale partijafdelingen en de gemeenteraadsverkiezingen van oktober 2012. Brussel: Academic & Scientific Publishers, pp. 93–121.

20.

Dodeigne

Meulewaeter

Lesschaeve

(2019) Constituting the list amid time of personalisation of politics. The balance of congruent and popular candidates in Belgian political parties. In: Vandeleene

De Winter

Baudewyns

(eds), Candidates, Parties and Voters in the Belgian Partitocracy. Cham: Palgrave Macmillan.

21.

Dovidio

Gaertner

(2000) Aversive racism and selection decisions: 1989 and 1999. Psychological Science 11: 315–319.

22.

Durose

Richardson

Combs

, et al. (2013) 'Acceptable difference': Diversity, representation and pathways to UK politics. Parliamentary Affairs 66: 246–267.

23.

English

(2018) Opinionated Opportunity: How Public Opinion Mediates The Representation of Immigrant Origin Groups in Britain and Europe. Manchester, UK: The University of Manchester.

24.

English

(2020) High rejection, low selection: How ‘punitive parties’ shape ethnic minority representation. Party Politics 1354068820973851.

25.

Farrer

Zingher

(2018) Explaining the nomination of ethnic minority candidates: How party-level factors and district-level factors interact. Journal of Elections, Public Opinion and Parties 28: 467–487.

26.

Fiske

(2002) What we know now about bias and intergroup conflict, the problem of the century. Current Directions in Psychological Science 11: 123–128.

27.

Fraga

(2016) Redistricting and the causal impact of race on voter turnout. The Journal of Politics 78: 19–34.

28.

Garrett

(2017) An easy concession or meaningful representation? Minority women in French politics. French Politics 15: 166–186.

29.

Geese

Schacht

(2018) The more concentrated, the better represented? The geographical concentration of immigrants and their descriptive representation in the German mixed-member system. International Political Science Review 40: 643–658.

30.

Greene

Haber

(2015) The consequences of appearing divided: An analysis of party evaluations and vote choice. Electoral Studies 37: 15–27.

31.

Hazan

Rahat

(2010) Democracy Within Parties : Candidate Selection Methods and Their Political Consequences. Oxford: Oxford University Press.

32.

Jacobsmeier

ML.

(2015) From black and white to left and right: Race, perceptions of candidates' ideologies, and voting behavior in U.S. house elections. Political Behavior 37: 595–621.

33.

Janssen

(2020) Shaping the (dis)advantage: the impact of partisan and demographic factors on ethnic minority candidates' success in preferential voting systems. Evidence from the Brussels case. Journal of Elections, Public Opinion and Parties: 1–24.

34.

Janssen

Dandoy

Erzeel

(2017) Het electorale succes van etnische minderheden in Brussel: De rol van kiezers en partijen. Res Publica 59: 389–412.

35.

Jones

(2014) Revisiting stereotypes of non-white politicians' ideological and partisan orientations. American Politics Research 42: 283–310.

36.

Lau

Redlawsk

(2001) Advantages and disadvantages of cognitive heuristics in political decision making. American Journal of Political Science 45: 951–971.

37.

Lerman

Sadin

(2016) Stereotyping or projection? How white and black voters estimate black candidates' ideology. Political Psychology 37: 147–163.

38.

Martin

Blinder

(2020) Biases at the ballot box: How multiple forms of voter discrimination impede the descriptive and substantive representation of ethnic minority groups. Political Behavior.

39.

McDermott

(1998) Race and gender cues in low-information elections. Political Research Quarterly 51: 895–918.

40.

Meuleman

Billiet

(2003) De houding van vlamingen tegenover “oude” en “nieuwe” migranten: diffuus of specifiek, Ministerie van de Vlaamse GemeenschapVlaanderen Gepeild! Brussel: pp. 137–176.

41.

Norris

Lovenduski

(1995) Political Recruitment: Gender, Race and Class in the British Parliament. Cambridge: Cambridge University Press.

42.

Ouyang

Haglund

Waterman

(2017) The missing element: Examining the loyalty-competence nexus in presidential appointments. Presidential Studies Quarterly 47: 62–91.

43.

Pearson

Dovidio

Gaertner

(2009) The nature of contemporary prejudice: Insights from aversive racism. Social and Personality Psychology Compass 3: 314–338.

44.

Phillips

(1995) The politics of Presence. Oxford: Clarendon Press.

45.

Portmann

Stojanović

(2019) Electoral discrimination against immigrant-origin candidates. Political Behavior 41: 105–134.

46.

Rooth

(2010) Automatic associations and discrimination in hiring: Real world evidence. Labour Economics 17: 523–534.

47.

Ruedin

(2013) Why Aren’t They There? The Political Representation of Women, Ethnic Groups and Issue Positions in Legislatures. Colchester: ECPR Press.

48.

Shah

(2014) It takes a black candidate. Political Research Quarterly 67: 266–279.

49.

Stad Antwerpen (2020) Stad in cijfers: Databank. Available at: https://stadincijfers.antwerpen.be/databank/.

50.

STATBEL (2021) Herkomst naar nationaliteitsgroep van herkomst per gemeente. Available at: https://statbel.fgov.be/nl/themas/bevolking/herkomst#figures.

51.

STATBEL (s.d.) Namen en voornamen. Available at: https://statbel.fgov.be/nl/themas/bevolking/namen-en-voornamen.

52.

Stevens

PAJ

Görgöz

(2010) Exploring the importance of institutional contexts for the development of ethnic stereotypes: A comparison of schools in Belgium and England. Ethnic and Racial Studies 33: 1350–1371.

53.

Teney

Jacobs

Rea

, et al. (2010) Ethnic voting in Brussels: Voting patterns among ethnic minorities in Brussels (Belgium) during the 2006 local elections. Acta Politica 45: 273–297.

54.

Thrasher

Borisyuk

Rallings

, et al. (2017) Candidate ethnic origins and voter preferences: Examining name discrimination in local elections in Britain. British Journal of Political Science 47: 413–435.

55.

Tolley

(2019) Who you know: Local party presidents and minority candidate emergence. Electoral Studies 58: 70–79.

56.

Valdini

(2012) A deterrent to diversity: The conditional effect of electoral rules on the nomination of women candidates. Electoral Studies 31: 740–749.

57.

Van der Zwan

Lubbers

Eisinga

(2019) The political representation of ethnic minorities in the Netherlands: ethnic minority candidates and the role of party characteristics. Acta Politica 54: 245–267.

58.

van Haute

Carty

(2011) Ideological misfits. A distinctive class of party members. Party Politics 18: 885–895.

59.

Visalvanich

(2017) When does race matter? Exploring white responses to minority congressional candidates. Politics, Groups, and Identities 5: 618–641.

60.

Weaver

(2012) The electoral consequences of skin color: The “Hidden” side of race in politics. Political Behavior 34: 159–192.

Supplementary Material

Please find the following supplemental material available below.

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

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

0.00 MB

0.19 MB