Introduction

Much research has focused on the relationship between specific factors and support for right-wing populist parties (RPP), such as perceived cultural and economic threats (Lucassen and Lubbers 2011), precarious working conditions (Gidron and Mijs 2019), and anti-immigrant sentiments (Hooghe and Dassonneville 2018). While some of these studies have focused on more traditional supply-side or demand-side explanations, more recent studies have focused on what Bos et al. (2017, p. 82) call “less rational” individual-level explanations (e.g., Jylhä et al. 2019).

In line with research on “less rational” explanations, this paper focuses on a topic largely left untouched: implicit racial bias. The study of implicit bias has become increasingly popular within political psychology and has been linked to both political preferences and discrimination (Greenwald et al. 2009a, b). Several studies (Arcuri et al. 2008; Payne et al. 2010) have similarly found a link between implicit attitudes and party support, but few have focused on right-wing populism (but see Bos et al. 2017, for an exception). To date, no large-scale study has, to the best of my knowledge, explored the relationship between implicit racial bias and support of RPP.

Implicit bias is relevant for the study of right-wing populism for two reasons. Implicit measures are less susceptible to social desirability bias, making them particularly suitable when measuring, e.g., prejudice or other types of socially sensitive attitudes (Gawronski et al. 2014). While previous studies acknowledge the issue of social desirability (e.g., Ackermann et al. 2018), it still permeates large parts of the literature. Social desirability bias is especially problematic if support of RPP has become more normalized (Stockemer 2017) while explicit prejudice remains stigmatized. Some voters might be open with their support of a right-wing populist party but unwilling to express prejudiced attitudes. As a result, studies might underestimate the link between attitudes such as racial bias and support for RPP.

Implicit bias furthermore reflects a partially independent construct. Implicit and explicit bias tend to correlate, but they can also diverge. It is possible to explicitly reject intolerant or prejudiced attitudes while still being affected by implicit prejudice. Experimental research (Schmuck and Matthes 2018) has shown how, e.g., right-wing populist ads can affect implicit but not explicit attitudes, and how these implicit attitudes can affect support of RPP.

Using data from France, Belgium, and the Netherlands, this paper explores whether implicit racial bias is associated with the support of RPP and whether the association is independent of explicit attitudes. As such, the paper makes two significant contributions. First, the article introduces and validates the use of implicit measures in the study of right-wing populism in three different contexts. While psychologists (e.g., Gawronski et al. 2014) have called for political scientists and sociologists to include implicit measures, such research is still largely absent. As one of the first articles on RPP to use implicit measures, the paper focuses on whether—and not why—implicit attitudes are associated with the support of RPP. Second, the article introduces the use of the Project Implicit database to a wider audience. While the database, containing over 2.3 million participants in 36 countries, has been public for several years, it has almost exclusively been used by social psychologists. This article will hopefully encourage social scientists within other disciplines, such as political science and sociology, to use the database to enrich our understanding of political behavior.

Implicit bias

Implicit bias is typically viewed as an automatic association between a concept and positive or negative sentiments (Greenwald and Krieger 2006).Footnote 1 It tends to operate beyond conscious control, although the degree of unconsciousness has been contested (Gawronski et al. 2006). Implicit bias was initially seen as being stable across time and context and acquired early in life through socialization, but subsequent research has demonstrated the malleability of implicit bias (Rudman et al. 2001; Payne et al. 2017). Brief exposure to, e.g., certain types of music (Rudman and Lee 2002), stereotypical ads (Arendt 2013), or national flags (Butz et al. 2007) can substantially change our implicit attitudes. These changes do, however, tend to be short-lived, and it has proved difficult to create long-term changes in implicit bias (Lai et al. 2016). Implicit bias is thus contextually and situationally malleable but typically reverts to a certain baseline over time. It has led some researchers to describe implicit bias as “permanent yet unstable” (Payne et al. 2017, p. 234).

Explicit and implicit bias tend to overlap, especially when it comes to less socially sensitive topics. They can, however, also diverge, and implicit bias tends to have better predictive validity for certain types of behaviors (e.g., automatic/habitual) and during certain situations (e.g., during time pressure, ambiguity, and stress). Similarly, a change in explicit bias is not always mirrored by a change in implicit bias. We typically update our explicit attitudes if we learn that a certain belief is incorrect, but our implicit attitudes will often remain unchanged and can still influence future actions (Arendt 2013).

Most researchers view implicit bias, similar to explicit bias, as an attitude or a trait. A growing number of researchers have, however, started to question whether implicit bias is a strictly individual-level phenomenon. Reviewing recent findings, Payne et al. (2017, p. 236) conclude that the “evidence is more consistent with [a] situational view.” Instead of reflecting individually held attitudes and stereotypes, implicit bias should be viewed as a social phenomenon, reflecting a certain context or situation. If we sample a group of people who live in a certain context (such as a nation or a city) where Minority X is associated with criminality, participants will on average implicitly associate Minority X with criminality regardless of their explicit attitudes. Scoring high on an implicit racial bias test might then primarily reflect and measure the context we are in.

Regardless of how implicit bias is conceptualized, most studies use the Implicit Association Test (IAT) to measure implicit bias. In the IAT, participants are tasked with pairing words (e.g., Black/White names) with positive or negative attributes (e.g., Good/Bad). The test is designed to give little pause for deliberation, making the IAT more difficult to control and manipulate than explicit measures. While participants might be unwilling to disclose explicit bias, the IAT will still tap into this construct. As a result, implicit tests have become especially popular when dealing with socially sensitive topics.

Implicit bias and right-wing populism

Few studies, especially in a European context, have focused on the link between implicit bias and voting preferences. In a Dutch study, Bos et al. (2017) found that implicit preferences of Partij voor de Vrijheid (PVV) were associated with a significant increase in the likelihood of voting for the party. Roccato and Zogmaister (2010) similarly found implicit measures to predict voting intentions above and beyond explicit measures in Italy. Several US-based studies (Payne et al. 2010; Arcuri et al. 2008; Lundberg and Keith Payne 2014) have found a similar link between implicit preferences and voting, especially for undecided voters.

Party preference or voting is mostly a deliberate decision. As a result, the link between implicit preferences and voting or party support is relatively weak (Gawronski et al. 2014; Bos et al. 2017). The association becomes even weaker when controlling for explicit attitudes, as people who implicitly prefer a party also tend to prefer the party explicitly. While it can still be important to look at implicit preferences, it might be more relevant to study implicit prejudice. Previous research (Payne et al. 2010; Greenwald et al. 2009a, b) has found a strong association between implicit racial bias and support for specific candidates, parties, and policy proposals. Compared to implicit preferences, the overlap between explicit and implicit prejudice is often weaker and the link between implicit prejudice and party support is usually more robust (Gawronski et al. 2014).

Despite the relevancy of implicit bias, few studies have explored the relationship between implicit bias and support of RPP. A noticeable exception is Schmuck and Matthes (2018). Using experimental methods, the authors found a link between exposure to right-wing populist ads and voting intentions for RPP. The effect was partially mediated by implicit bias, but implicit bias was also directly related to voting intention. In a similar study, Arendt et al. (2015) found that exposure to right-wing populist ads affected implicit but not explicit attitudes toward foreigners, even when participants explicitly negated the content of the ads.

These studies show the importance of implicit attitudes for party support, but also why the effects might occur in the first place. In line with research on implicit stereotypes (Rudman and Lee 2002; Mastro et al. 2014), exposure to stereotypical portrayals of a minority (e.g., in ads, media, etc.) affects implicit bias beyond conscious control. Changes in implicit bias could, in turn, make certain frames, such as that of the New Right in France, resonate more with the “lived experiences, attitudes and preconceptions of many people” (Rydgren 2005, p. 426). While a specific frame (e.g., anti-immigration) and a specific explicit attitude might remain static over time, support for parties advocating a certain frame could increase if the resonance between the frame and the attitude increases (Bonikowski 2019). Living in a “biased context” (Payne and Hannay 2021, p 931) could similarly increase the resonance, as our attitudes become validated by our surroundings.

In summary, implicit bias is expected to affect support for RPP in at least two ways. Implicit bias can affect party support by influencing explicit attitudes, preferences, and behavior (Gawronski et al. 2014; Schmuck and Matthes 2018), which could increase the resonance between certain frames and beliefs. Individuals with high implicit bias might then be especially likely to find right-wing populist frames appealing. Implicit racial bias can also reflect certain explicit attitudes that we are unwilling to disclose because of social desirability bias (Greenwald et al. 2009a, b). While implicit bias is independent of explicit bias, it is not unrelated. A person’s explicit bias will partially reflect the person’s implicit bias even if they are unwilling to disclose it.

Current study

Few previous studies have focused on the association between implicit racial bias and support for RPP. The current study uses correlational data, which means it cannot test whether implicit bias has a causal effect on the support of RPP. It can, however, show if there is a significant relationship between implicit racial bias and support of RPP. In this study, I look specifically at the association between racial bias toward Black people and support of RPP. In light of existing research (e.g., Arendt et al. 2015), I expect implicit racial bias to have a positive association with support for RPP. As such, the paper starts off with the following hypothesis:

H1

Negative implicit racial bias predicts support for RPP.

As implicit attitudes are often correlated with explicit attitudes (Nosek et al. 2007), an effect could result from an association between explicit bias and support of RPP. Implicit racial bias would then not provide any incremental validity of its own. For certain types of attitudes, such as political preferences, the correlation is significant. For others, such as racial bias, the correlation is relatively small (Gawronski et al. 2014) and implicit racial bias has typically predicted outcomes above and beyond explicit bias. We thus expect:

H2

Negative implicit racial bias predicts support for RPP even when controlling for explicit attitudes.

Support for RPP has become increasingly normalized in Europe, and voters of, e.g., Front National (now known as Rassemblement National) are relatively open with their support in experiments and surveys (Stockemer 2017, but see also Harteveld et al. 2019). As such, voters may be open with their support of RPP while still unwilling to disclose socially sensitive attitudes. This dynamic can lead researchers to underestimate the link between racial bias and party support (e.g., Rydgren 2008). The association between implicit racial bias and support for RPP will, however, be relatively unaffected by social desirability. Implicit racial bias might then be an especially good predictor when it diverges from explicit attitudes, i.e., for individuals who report positive explicit bias and negative implicit bias. I thereby test the following hypothesis:

H3

The relationship between negative implicit racial bias and RPP support will be stronger for individuals with positive explicit racial bias.

The convergence of explicit and implicit attitudes can, however, also lead to an elaboration of attitudes, which increases the predictive validity of both explicit and implicit attitudes (Friese et al. 2012). Individuals with high values of both explicit and implicit bias might then be especially prone to supporting RPP. High levels of implicit racial bias might also, in line with the situational account of implicit attitudes (Payne et al. 2017), be indicative of a biased context rather than a biased individual. Living in a biased context can increase the acceptability of both expressing explicit racial bias (Paluck and Green 2009) and supporting RPP (Álvarez-Benjumea 2020). I thus test the following hypothesis:

H4

The relationship between negative implicit racial bias and RPP support will be stronger for individuals with negative explicit racial bias.

Some people have neither positive nor negative explicit racial bias. Not having strong explicit attitudes can sometimes increase the impact of implicit bias. Lundberg and Keith Payne (2014) showed how implicit attitudes primarily affected undecided voters, as these voters tended to be “swayed” by their implicit attitudes. Implicit attitudes had, in contrast, little effect on already decided voters. People with more elaborate bias (either positive or negative) might then be relatively unaffected by implicit racial bias, while the effect is more pronounced for people without a specific bias. This leads us to the paper’s final hypothesis:

H5

The relationship between negative implicit racial bias and RPP support will be stronger for individuals with no explicit bias.

Three different cases

The study focuses on data from three different countries: France, Belgium, and the Netherlands.Footnote 2 While the aim of this paper is to validate any potential findings in different contexts, contextual differences will likely influence the relationship between racial bias and support of RPP.

In Belgium and France, RPP made early electoral breakthroughs during the 80 s and 90 s. By the end of the 1990s, both Front National (FN) and Vlaams Blok (VB) had established themselves as influential (but isolated) political forces (Cammaerts 2018). Unlike FN, VB was primarily a regional party; while they received around 10% support in the 1999 federal election, the support came almost exclusively from the Flemish part of Belgium.Footnote 3 In contrast, RPP were largely unsuccessful in the Netherlands until the emergence of Lijst Pim Fortuyn (LPF) in 2002. Unlike VB and FN, LPF employed a mix of liberal and right-wing populist talking points (Kiess et al. 2016). While RPP in the Netherlands now enjoy similar levels of success as their Belgian and French counterparts (Cammaerts 2018), they still employ more liberal and civic frames (Halikiopoulou et al. 2013; Kiess et al. 2016). These party differences could affect the association between RPP support and implicit racial bias. Implicit racial bias might, e.g., be a better predictor of support for more extreme RPP, or RPP that use more ethnic and overtly xenophobic frames.

Despite these differences, RPP in France, Belgium, and the Netherlands share several important features. Similar to most RPP in Western Europe, these parties employ a “master frame” where certain cultures are portrayed as incompatible with national culture (Rydgren 2005; Cammaerts 2018). While this type of “cultural racism” is distinct from more classical biological racism, it still often implies—and reproduces—biological racism. Some, such as Barker (1981) and Van Der Valk (2003), have described it as “pseudo-biological culturalism”, where skin color becomes equated with a certain culture and mindset. By using more cultural frames, RPP can mobilize racist sentiments without necessarily being labeled as racist themselves (Rydgren 2005). It is thus not unlikely that RPP support in all three countries is, at least to an extent, affected by explicit and implicit racial sentiments.

There are, of course, many other differences between these three countries that affect the relationship between RPP and implicit racial bias. Black people are perceived differently in, e.g., France and the Netherlands (Essed 1991; Constant 2012), and the relevancy of race and racial bias is different in each country. The proportion of Black people similarly differs between the three countries which will, in turn, affect intergroup contact; 4.4% of people in France (Ministry for Europe and Foreign Affairs 2019), 2.8% in Belgium (NPdata 2022), and 1.5% in the Netherlands (CBS 2021) have a sub-Saharan background. These numbers do, however, say little of the frequency and nature of the interactions between the Black and non-Black populations. As shown in numerous studies, contact or exposure is a necessary but often insufficient condition to reduce, e.g., implicit racial bias (Rae et al. 2021). While it is beyond the scope of this paper to delve into these differences, they can—and likely will—affect the strength of the association between support of RPP and implicit racial bias.

Method

Participants

The data were collected from the Project Implicit database, which is publicly available at https://osf.io/kaqi5/. Project Implicit is a virtual laboratory founded by Greenwald, Banaji, and Nosek. For almost 20 years, volunteers have been able to take part in experiments related to implicit measures. After visiting the Project Implicit website, participants can complete one of several different implicit measures. This study consists of participants who completed the Race IAT. These participants also filled out a demographic survey and a short explicit questionnaire. I include all participants from France, Belgium, and the Netherlands who completed the test between the years 2008 and 2017, resulting in a sample of 41,803 participants. Participants who did not complete all parts of the test, including the Race IAT and the survey, are excluded from the sample. The final sample is overall younger, more educated, and more left leaning than the population in each respective country. Demographic details are presented in “Appendix A,” Table 3.

Measures

Implicit racial bias

All participants completed a Race IAT designed to measure implicit racial bias. Participants were tasked with pairing positive and negative words with pictures of Black and White people. Scoring was done in accordance with the updated scoring algorithm (Greenwald et al. 2003), excluding participants who completed the test too fast or too slow. Each participant was assigned an overall “d-score” based on their reaction speed, which ranges from − 2 (positive bias) to + 2 (negative bias), with 99.9% of participants having a score between − 1.5 and + 1.5.

Explicit racial bias

Explicit racial bias was measured by asking participants about their feelings toward Black people. Participants were tasked with evaluating Black people on a scale from 0 (Very cold/unfavorable) to 10 (Very warm/favorable). The scale was reverse coded for the final analysis, with higher values indicating more negative opinions of Black people.

Party support

Party support was measured slightly differently in the three countries. In Belgium, participants were asked about their party affiliation, while in the Netherlands, participants were asked about their favorite party. In France, participants were instead asked what party best represents their opinions. The measure was recoded into a dummy variable where participants scored a “1” if they reported support for a right-wing populist party and “0” if they did not.Footnote 4 Party data were not available for newer parties, and a list of all the parties can be found in “Appendix A,” Table 4. Following the classification in Huber and Ruth (2017), four parties were categorized as right-wing populist: Vlaams Belang (VB) and List Dedecker (LDD) in Belgium,Footnote 5Front National (FN) in France, and Partij voor de Vrijheid (PVV) in the Netherlands.

Control variables

Five control variables in total were used. Participants reported their current age, gender (male/female), race, and highest attained education. Age and gender were measured in the same way across all datasets. Education and race were recorded slightly differently in each country, reflecting national differences. As participants completed the survey between 2008 and 2017, we add year of completion as an additional control variable. A detailed description of these variables and the dataset can be found at https://osf.io/kaqi5/. A subset of participants also completed three additional measures: occupational background, Right-Wing Authoritarianism (RWA), and Social Dominance Orientation (SDO). A supplementary analysis including these controls can be found in “Appendix A,” Table 3, but the results remain largely unchanged.

Results

The analysis is conducted using a hierarchic binary logistic regression. For ease of interpretation and to combat issues of rescaling (Mood 2010), log odds are transformed into probabilities and presented as average marginal effects. Logit models with log odds are presented in “Appendix B,” Table 7, but the results remain unchanged.

Participants had overall moderately strong implicit racial bias in all countries, with an average d-score of 0.38 (SD = 0.41) in France, 0.33 (SD = 0.42) in Belgium, and 0.32 (SD = 0.42) in the Netherlands. Explicit racial bias was more positive, with an average score of 3.74 (SD = 2.18) in France, 4.03 (SD = 1.78) in Belgium, and 3.96 (SD = 1.7) in the Netherlands. Participants in all countries thus had, on average, implicitly negative but explicitly positive attitudes toward Black people. Explicit and implicit attitudes are weakly correlated in all three countries, ranging from 0.14 in France to 0.15 in the Netherlands and Belgium.

Model 1 in Tables 1 and 2 presents the results of the first analysis. Implicit racial bias is significantly associated with a higher probability to support RPP in all three countries. For every unit increase in implicit racial bias, the probability of supporting a right-wing populist party increases by an average of 3.98 percentage points in France, 2.21 percentage points in Belgium, and 2.8 percentage points in the Netherlands.

Table 1 Average marginal effects for logistic regressions on probability to support a right-wing populist party in France and Belgium
Full size table
Table 2 Average marginal effects for logistic regressions on probability to support a right-wing populist party in the Netherlands
Full size table

In Model 2, explicit attitudes is added to the regression. The coefficient for implicit racial bias decreases in all countries. The effect is most noticeable in France and Belgium, where the probability to support RPP is decreased by 1.3 and 1.1 percentage points, respectively. Explicit attitudes account for part of the effect, but the implicit measure still proves to be a significant predictor of support. In Model 3, survey year, age, gender, race, and educational background are added as additional controls. The coefficient decreases slightly in all three countries but the results remain significant. We thus find support for both Hypotheses 1 and 2.

Hypotheses 3–5 predicted that the effect of implicit attitudes should be different for individuals with positive, negative, and no explicit bias. To test Hypotheses 3–5, explicit racial bias was recoded into a categorical variable with three values: no bias, negative bias, and positive bias. Following the recommendations in Mize (2019), the interaction effect is analyzed using predictive margins and presented in Fig. 1. A table with all predictive margins can be found in “Appendix B.”

Fig. 1
figure 1

Predictive margins of implicit and explicit bias on support of RPP

Full size image

In all three countries, individuals with negative explicit bias have a higher predicted probability of supporting a right-wing populist party, and the probability increases with higher levels of implicit bias. A person in, e.g., France with no implicit bias (a value of 0) but with negative explicit bias has a 0.169 predicted probability of supporting Front National. A person with strong negative implicit bias (a value of 1) and negative implicit bias has a 0.27 predicted probability of supporting Front National. Hypothesis 4, which predicted that the effect of implicit bias on support would be higher for individuals with negative explicit bias, is thus supported.

Individuals with positive explicit attitudes have the lowest predicted probability of supporting a right-wing populist party and are relatively unaffected by implicit bias in all three countries. Dutch participants with no bias have a relatively low predicted probability of supporting a right-wing populist party, but the probability is significantly increased at higher levels of implicit bias. We thus find no support for hypothesis 3 and only limited support for hypothesis 5.

Discussion

The paper aimed to highlight the role of implicit attitudes in explaining support for RPP. While previous research has focused on various individual-level “irrational” factors, few studies have looked at the impact of implicit attitudes. The study found a significant effect of implicit racial bias on the probability of supporting RPP in all three countries, although the effect is relatively modest. In line with other research on implicit attitudes and voter support (e.g., Bos et al. 2017), explicit racial bias partially mediated the effect.

Implicit racial bias was more weakly associated with support of RPP than explicit racial bias (see Table 2). The finding is consistent with studies showing that implicit attitudes primarily affect automatic or unconscious behavior (Gawronski et al. 2014; Glaser and Finn 2013). Implicit racial bias was, however, not inconsequential, especially in France. Importantly, implicit racial bias proved to be an especially good predictor for people with negative explicit bias toward Black people in all three countries. People with both negative explicit and implicit biases thus seem especially prone to support RPP, having a predicted probability of, e.g., 0.27 in France. While implicit racial bias alone might be a modest predictor of RPP support, implicit racial bias in combination with explicit racial bias is an especially potent combination.

The finding is in line with research on attitude elaboration, where a convergence of implicit and explicit attitudes increases the predictive validity of both measures. The finding can also be reconciled with the situationist account of implicit bias. If implicit bias reflects a person’s context or surroundings, high implicit bias could indicate a normalization of certain attitudes. This could, in turn, make right-wing populist messages “resonate” more with the individual, strengthening the link between explicit bias and RPP support. Living in a biased context could similarly lower the effect of social desirability, as people in biased contexts might feel freer to express both explicit racial bias and support of RPP.

Participants with positive racial bias remained relatively unaffected by implicit bias, even at very high levels of implicit bias. If implicit bias does reflect a biased context, this context might only affect behavior when it aligns with explicit attitudes and beliefs. Implicit bias can similarly only increase the resonance between RPP frames and explicit racial bias if the bias exists in the first place.

Participants with no explicit racial bias became more likely to support RPP with increasing levels of implicit racial bias, but primarily in the Netherlands. While some of these participants might have been “swayed” by their implicit racial bias, the effect was noticeably smaller than for participants with explicit racial bias.

In summary, this paper has found a link between implicit racial bias and support of RPP and shown how implicit racial bias can moderate the relationship between explicit racial bias and RPP support. The paper has thus demonstrated the usefulness of implicit measures to understand why and when people support RPP. As implicit racial bias primarily affected voters with explicit racial bias, using implicit measures might be especially fruitful when trying to understand why explicit bias turns into support for RPP. While this study relies on cross-sectional data, future studies might find it useful to analyze longitudinal or experimental data. Because implicit racial bias can affect information processing (Matthes and Schmuck 2019) and feelings of resonance, it might reasonably have a larger effect over time. Exposure to, e.g., stereotypical portrayals can cause certain implicit “gut-feelings,” which over time become more elaborate explicit attitudes (Arendt and Northup 2015). Using longitudinal or experimental data would also be critical to establishing causality. While it is likely that implicit racial bias affects the support of RPP, the reverse could also be true; supporting RPP might enhance and normalize both explicit and implicit racial bias (Rydgren 2003). The current study, which relies on non-experimental survey data, primarily highlights the relationship between support of RPP and implicit bias and not the direction of the relationship.

The principal purpose of this study was to study whether implicit racial bias is related to RPP support and not why it is related. The purpose was also to explore whether implicit racial bias was related to RPP support in different contexts, but not why the association might vary in different contexts. As this study shows, context does seem to matter: not only was the relationship between RPP and implicit racial bias different in all three countries, but the interaction between implicit and explicit racial also varied between contexts. Future studies should try to disentangle when and why these effects occur. This could be done by combining Project Implicit data with more experimental approaches, but also by combining geographical data from Project Implicit with, e.g., regional indicators of prejudice and discrimination (see Payne et al. 2019 for a similar approach). Future studies could also take several different attitudes into account, as different prejudiced attitudes tend to correlate (Akrami et al 2010).

A couple of limitations should be noted, both in relation to the paper and to the Project Implicit database. While the paper controls for demographic variables, the participants are self-recruited via the Project Implicit website. Participants are likely more informed about prejudice in general and not necessarily representative of the population in France, Belgium, or the Netherlands (see “Appendix A”). Supporters of RPP were furthermore underrepresented compared to other parties. The number of people who supported RPP was thus relatively small, especially in comparison with the overall sample.

Support for RPP, which served as the dependent variable, was also measured slightly differently in all three countries. These different operationalizations make cross-country comparisons less meaningful, which could be problematic for studies using a more comparative approach. Data were also not available for newer RPP in the Netherlands and Belgium.

Finally, explicit racial bias was measured using a 1-item feeling thermometer. Feeling thermometers are common to include in IAT studies as they—similar to the IAT—measure positive and negative sentiments toward a specific group. It is, however, possible that other types of racial beliefs could influence the association between implicit racial bias and support of RPP. While this study shows that implicit racial bias is independent of explicit racial bias, it might be more or less independent of, e.g., specific racial stereotypes.

In conclusion, the study has shown the importance of racial implicit bias and will hopefully encourage researchers to incorporate implicit measures when studying right-wing populism. By introducing the Project Implicit database to a larger audience, this paper will hopefully also stimulate future research interested in large-scale, cross-national datasets. While this study focused on racial implicit bias and RPP support, the database contains a variety of different measures, variables, and participants, which have—until now—been left unexplored outside of psychology.