Is it racism? The belief in cultural superiority across Europe Open Access

Some scholars maintain that modern racist thinking in Europe is expressed as the belief that European countries hold a superior culture. Theorists of ‘cultural racism’ (Blaut 1992; Wren 2001) maintain that such a belief in cultural superiority has substituted biological racism, or that racist beliefs are expressed as beliefs in both racial and cultural superiority. Critics of the theory point out that the concept of cultural racism is backed more by theoretical deliberations than by empirical evidence (Rattansi 2007; Siebers and Dennissen 2015). The present research seeks to address this shortcoming in the discussion of cultural racism. Although the present research argues that it is reasonable to make a conceptual distinction between a belief in cultural superiority and racism, the two can still be intertwined, which can only be demonstrated by empirical analyses.

The theory of cultural racism contends that a belief in a hierarchy of cultures revives old racism, albeit in new clothes (Blaut 1992). This theory maintains that both old-fashioned biological racism and the idea of some cultures being superior propagate a hierarchy of social groups, and that they are both racist. Blaut (1992) argued that the horrors of institutionalised racism under National Socialism, culminating in the Holocaust, brought biological racism into disgrace, and lead to new egalitarian norms in Europe. Consequently, old-fashioned racist ideas of a hierarchy of races became less acceptable. ‘Racist practice now needed a new theory’, Blaut argued (1992, p. 293). Discrimination of ethnic outgroups was no longer justified by references to biology and race, but instead had to refer to a hierarchy of cultures. Recent contributions to the theory of cultural racism have repeated the claim that a reference to a hierarchy of cultures is racist (Wren 2001; Poynting and Mason 2007; Balibar 2008; Meer and Modood 2009). Specifically, theorists of cultural racism contend that negative attitudes towards immigrant groups reflect cultural racism (e.g. Wren 2001; Ramos et al.2020).

So far little empirical research has backed up claims by the theory of cultural racism, but the theory has had considerable effect on current thinking on intergroup relations in Europe. For instance, when the European Social Survey set out to assess racism across Europe, it did not restrict its measurements to biological racism, it also included a measure of what the European Social Survey referred to as cultural racism¹: the belief that some cultures are ‘much better’ than others.

Critics of the notion of cultural racism have argued that it is based more on theoretical deliberations than on empirical evidence (Rattansi 2007; Siebers and Dennissen 2015). Critics also suggest that declaring a belief in cultural superiority as racist threatens the usefulness of the term racism, degrading it to a rhetoric vehicle (Rattansi 2007) and undermining the understanding of different forms of prejudices (Siebers and Dennissen 2015). Racism is only one of several forms of prejudices. People's tendency to think in terms of group boundaries is not limited to racism (Barth 1998); any perceived difference between groups will suffice to develop prejudice and even hostility towards the other group (Tajfel and Turner 1979; Diehl 1990).

Critics of the notion of cultural racism also included George M. Fredrickson, a historian and anti-racist activist. His historical analysis led him to conclude that devaluations of a specific religion should not be confused with racism – a distinction between the two was ‘crucial’ (Fredrickson 2015, pp. 6-7). Fredrickson suggested using the term ‘culturalism’ for religious intolerance and the tendency to condemn or persecute ‘others for what they believe, not for what they intrinsically are’ (p. 6).

The present research holds that in theory, a distinction between culturalism and racism seems reasonable. Biological racism claims to find biologically inherited dissimilarities between people of different geographic origins; culturalism makes no such claims of inherent differences. Culturalism and biological racism also differ in their relations to science. Biological racism is fundamentally a scientific theory and has been refuted by modern science (Rutherford 2020). In contrast, science cannot refute a belief in cultural superiority if this belief is based on values, such as the idea of human rights, the equality of genders (Okin 1999), democracy (Caldwell 2009), or economic development (Caldwell 2009).

The concept of cultural racism suggests a strong correlation between the belief in biological racism and the belief in cultural superiority. But the notion of culturalism as separate from racism also allows for a correlation. Even if a belief in cultural superiority is distinct from the belief in biological racism, it is still likely that some people maintain both beliefs. Moreover, there may easily be a grey area between culturalism and racism, further contributing to a correlation between the two.

A distinction between culturalism and racism as concepts does not prevent the two from being intertwined. Whereas authors such as Fredrickson and Rattansi argued against equalising culturalism with racism, they did not reject the claim that these two views can be linked. In Fredrickson's words, ‘there is substantial gray area between racism and culturalism’ (Fredrickson 2015, p. 7). A link between culturalism and racism is easily observed among groups hostile to immigration. Here, many people who uphold ideas of cultural superiority also maintain ideas of biological racism. Online-discussions in social media frequently demonstrate this link (Hughey and Daniels 2013; Richeson and Sommers 2017) and supporters of far-right political movements combine culturalist and racist arguments (Rydgren 2007; Cutts et al.2011; Brinkmann and Panreck 2019; Ahmed and Pisoiu 2020). Consequently, the link between culturalism and racism seems evident within subgroups in a society. But the claim of the theory of cultural racism is much more radical. It implies that a belief in cultural superiority is itself racist. The present research seeks to test empirically how the belief in cultural superiority relates to traditional racism within whole nations. Even if a conceptual distinction between culturalism and racism may be ‘crucial’ (Fredrickson 2015), racism and culturalism may still overlap. Specifically, expressed beliefs in cultural superiority might often indicate a racist orientation.

So far, few studies have tested the claims by the theory of cultural racism with methods that can either support or reject them (an exception is a study by Siebers and Dennissen 2015, who rejected the notion of cultural racism based on data from the Netherlands). A hallmark of a test of a theory or a hypothesis is that the test must be able to disconfirm a hypothesis, not merely illustrate it. Some research uses qualitative methods to illustrate what the author describes as cultural racism (Wren 2001), other researchers define beliefs in cultural superiority as representing cultural racism without using methods that might validate or question this interpretation (e.g. Ramos et al.2020). In another recent contribution, scholars focused on biological racism (Caller and Gorodzeisky 2021). Although Caller and Gorodzeisky (2021) restricted their analysis to biological racism, their study proves to be a valuable starting point for the present research.

Caller and Gorodzeisky (2021) suggested that a belief in a hierarchy between groups would not gain foothold in a society with strong social cohesion. Strong social cohesion, as described by Caller and Gorodzeisky (2021), comes from powerful and robust connections between individuals as well as groups within a society (see also Chan et al.2006). Research often refers to the Nordic countries in Europe as examples of countries with a strong social cohesion (Delhey and Newton 2005). The Nordic countries also have a strong welfare state, using taxes to redistribute wealth (see Esping-Andersen and Myles 2009). Caller and Gorodzeisky (2021) argued that social cohesion as well as welfare states’ redistribution of wealth would contribute to egalitarianism and therefore buffer against widespread belief in biological racism: it may seem difficult to reconcile living in a relatively egalitarian society with a belief in a nature-given hierarchy between social groups.

Compatible with their theory, Caller and Gorodzeisky (2021) found the least support to biological racism in Sweden, the Netherlands, and Norway, which the authors referred to as examples of countries with strong social cohesion (see also Holmberg and Rothstein 2017). In contrast, the samples from Estonia, Czechia, and Portugal expressed the strongest support to biological racism. Caller and Gorodzeisky (2021) suggested that the increased endorsement of biological racism in these countries was largely explained by weak social cohesion.

The arguments presented by Caller and Gorodzeisky (2021) might be extended to suggest that any ideas of a hierarchy involving groups of people would gain little support in the Nordic countries, even the idea of a hierarchy of cultures. It might seem reasonable to expect people in egalitarian countries rarely to express a belief in a hierarchy of cultures.

The theory of cultural racism makes the opposite prediction. It argues that egalitarianism in European countries has increased the tendency to express beliefs in cultural superiority to legitimise ethnic discrimination. This should apply to the Nordic countries specifically, given that the Nordic countries are more egalitarian than most other countries (e.g. Esping-Andersen and Myles 2009). This view on the Nordic countries is illustrated by Wren (2001), who presented Denmark as a typical example of a nation inflicted with cultural racism.

The theory of cultural racism seems to provide less clear predictions for countries where biological racism has a stronger foothold, such as in several Eastern European countries (see Caller and Gorodzeisky 2021). The theory might suggest that people in such cultures are less motivated to endorse ‘cultural racism’ to legitimise racist views. Alternatively, they could endorse both biological racism and a belief in cultural superiority. The latter prediction is also a reasonable deduction from Caller and Gorodzeisky's (2021) arguments.

After testing how widespread the belief in cultural superiority is in various countries, the present research will compare this belief with endorsements of biological racism. Theorists of cultural racism claim that a belief in cultural superiority generally reflects racism. Other scholars disagree, maintaining that culturalism is separate from racism (Siebers and Dennissen 2015) or that culturalism only in some contexts is mixed with racism (Rattansi 2007; Fredrickson 2015). The present research uses these alternative suggestions to develop two competing models and then tests both with confirmatory factor analysis (CFA, e.g. Loehlin and Beaujean 2017). The two models will be referred to as ‘Cultural Racism’ and ‘Culturalism is Separate’.

In the model ‘Cultural Racism’, the item on cultural superiority is one of three indicators of a racist orientation; the remaining two indicators of a racist orientation ask for beliefs in biological racism. The Cultural Racism model assumes that the four remaining items will group in a separate factor, not referring to fundamental beliefs about a hierarchy of groups, but to immigration. The model calls this second factor Attitudes to immigration.

The alternative model, called ‘Culturalism is Separate’, introduces a clear distinction between views on cultural differences on the one hand and racism on the other. With this approach, the estimated Racism factor uses the two items on beliefs in biological racism and the demand for immigrants to be white as indicators. The remaining items all refer to cultural differences between groups, so this model estimates them as indicators of a factor labelled Culturalism. The analysis will compare the two models’ ability to explain nationally representative data from 21 countries, analysing each country separately. After an initial test of the original models, each of the two models may be adapted to fit the data in single countries, if necessary. The article will focus on overall findings, but details on country-specific results are available in the online supplemental material.

Analyses used data from the European Social Survey, Round 7, collected in 2014 and 2015 (ESS Round 7 2014). The data were collected with computer aided personal interviews in 20 European countries plus Israel. Within each country, the European Social Survey selected samples using strict probability methods, ensuring that a national sample was representative of the country's population aged 15 and over. Only respondents who reported being born in the country, holding citizenship in the country, and not belonging to an ethnic minority group were included in this research. The resulting total sample size was 33,517; country-specific sample sizes were between 1,062 (Switzerland) and 2,678 (Germany). Gender was approximately equally distributed; from 47% females (Denmark) to 62% females (Lithuania). Below, the text will refer to countries, which will be short for the nationally representative sample from the specific country. The raw data are available from the website for the European Social Survey (ESS Round 7 2014). The data had few missing responses, descriptive statistics are included in the online supplemental material.

The European Social Survey used two items to assess the belief in biological racism and one item to assess the belief that some cultures are superior to others. In their English versions, the two items on biological racism were: ‘Do you think some races or ethnic groups are born less intelligent than others?’ and ‘Do you think some races or ethnic groups are born harder working than others?’. The item assessing a belief in cultural superiority was: ‘Thinking about the world today, would you say that some cultures are much better than others or that all cultures are equal?’. All three items used yes/no answers (although specified for the item on cultural superiority as either ‘Some cultures are much better than others’ or ‘All cultures are equal’). In addition, the items included a Don't know option, which was coded as missing data (5.3% and 4.9% for the two items on biological racism; 6.9% for the item on whether some cultures are better than others). The two items assessing beliefs in biological racism correlated at .39, both correlated moderately with the item on cultural superiority (r_s = .22 and .24).

These three items provided a brief measurement, but limitations in the measurements were compensated for by large sample sizes even within single countries. Four additional items were included in the analysis. One item requested views on whether immigrants would need to be white: ‘Qualification for immigration: be white’. Three items assessed aspects of culturalism: (1) ‘It is better for a country if almost everyone shares customs and traditions’; (2) ‘Would you say that [the specific country's] cultural life is generally undermined or enriched by people coming to live here from other countries?’; (3) A necessary qualification for immigration is to ‘come from a Christian background’. The four additional items had response scales from 0 to 10, except for the item on shared customs, which used a 5-point Likert scale (from Agree strongly to Disagree strongly).

The analysis used descriptive statistics of how often the idea of cultural superiority was endorsed in the various countries and it used CFA to investigate how the belief in cultural superiority related to biological racism. The supplemental material provides a brief description of CFA. Analyses with CFA incorporated sample weights provided by the European Social Survey. Data management and analyses used R 4.1 (R Core Team 2019); plotted maps were developed with the R package ‘rnaturalearth’ (South 2017). CFA was conducted with the R package ‘lavaan’ (Rosseel 2012). The online supplemental material was developed with R Markdown and the R packages ‘knitr’ (Xie 2021) and ‘kableExtra’ (Zhu 2021). All code for data management and analyses is available in the supplemental material.

Model fit in CFA was evaluated with the Chi-square, the Comparative Fit Index (CFI), the Root Mean Square Error of Approximation (RMSEA) and the 90% confidence interval for the RMSEA, as well as the Standardised Square Root Mean Residual (SRMR). Due to large sample sizes even within countries, models with a Chi-square based p-value below .05 were accepted. Consistent with recommendations in the literature (e.g. Mueller and Hancock 2010), the analysis aimed at a CFI of at least .95, an RMSEA not above .06, and SRMR not above 0.08.

CFA tested the two alternative factor models presented in the introduction to this paper. Three robustness checks were added. Firstly, one robustness check used the model with lower fit in the initial analysis and modified it until it had a good fit with the data, adding cross-loadings and residual covariances based on modification indices provided by lavaan. Would even the alternative model, once made to fit the data with added paths, support the initial findings?

A second robustness check added two binary indicators to the analysis. Two of the items in the CFA analyses – being white and having a Christian background as qualifications for immigrating – had a preponderance of zeros (that is, many respondents answered ‘Extremely unimportant’). For both these items, a binary variable was added to reflect the distinction between those who dismissed the idea of using whiteness or a Christian background as a criterion for being allowed to immigrate and those who considered this a relevant criterion. Data from Israel were treated somewhat differently. Here, respondents indicated whether a requirement should be that immigrants had a Jewish background, resulting in a preponderance of answers indicating complete agreement. Therefore, in Israel, the binary item on a specific religion distinguished between full agreement and all other responses. A third robustness check used only binary items.

Adding the two binary indicators in Robustness Check 2 resulted in a rather complex model; the aim of this robustness check was to verify that the originally better-fitting model (‘Cultural Racism’ or ‘Culturalism is Separate’) continued to have better fit, even with the added binary indicators. Due to the complexity of these models with added binary indicators, the requirements for RMSEA values were relaxed somewhat. Following MacCallum, Browne and Sugawara (1996) – who suggested 0.01, 0.05, and 0.08 for the RMSEA to indicate excellent, good, and mediocre fit, respectively – an RMSEA up to 0.08 was accepted in robustness checks when more complex models were tested against and modified to fit the data.

Norway was an interesting case. Here, few in the sample endorsed biological racism, but more than two thirds (67%) endorsed the belief in cultural superiority. The theory of cultural racism would claim that this result for Norway implied that the Norwegian sample was among the most racist ones included in the present research. Only in Portugal did the sample in a similar manner endorse one of the three items; in Portugal, 68% endorsed the belief that ‘some races or ethnic groups are born harder working’. The belief in cultural superiority was also frequently endorsed in Denmark (61%), but less often in other countries with strong social cohesion, such as Sweden (42%), Finland (40%) and the Netherlands (43%). The endorsement of the belief in cultural superiority was more rare in several countries often referred to as countries with low social cohesion, such as Hungary (29%) and Lithuania (32%), but it was more frequent in Estonia and Slovenia (both 38%). Czechia had the most frequent (57%) endorsement of the idea of cultural superiority among Eastern European countries.

Country scores for the two beliefs in biological racism were strongly correlated, r  = .59 [95% confidence interval = .20, .82]. But neither of these country scores for biological racism correlated with country scores for the belief in cultural superiority: r = .01 [-.44, .45]; r = -.02 [-.46, .43]. At the individual level, the two items on biological racism correlated at .39 [.38, .40], as estimated with Spearman correlation, r_s. Both these items on biological racism had only moderate individual-level correlations with the belief in cultural superiority: r_s = .24 [.22, .25] and r_s = .22 [.21, .23].

The present research tested two competing CFA models: one based on the notion of cultural racism, the other on the notion of culturalism being separate from racism. In 19 of the 21 countries, the model ‘Cultural Racism’ had clearly better fit (e.g. with a difference in RMSEA values from .01 in the Netherlands to .12 in Hungary). In eight countries, the model ‘Cultural Racism’ even gave sufficient fit without further modifications (Finland, Hungary, Ireland, Israel, Lithuania, Poland, Portugal, and Slovenia). In the remaining countries, adding a residual covariance between two items (immigrants are qualified if white and immigrants are qualified if having a Christian background) resulted in good fit², see Table S2 in the supplemental material for details). In Austria, the model also needed a cross-loading between the item on racially determined intelligence to Attitude towards immigration.

Only in two countries – Norway and Sweden – was the model ‘Cultural Racism’ not obviously the best fitting model. In Sweden, the model ‘Cultural Racism’ still had marginally better fit; in Norway, the model ‘Culturalism is Separate’ had marginally better fit. However, both models had too low fit in Norway and in Sweden to be accepted without modifications. In both Norway and Sweden, the ‘Cultural Racism’ model fit the data if a residual covariance was added between the items immigrants are qualified if white and immigrants are qualified if having a Christian background.

The model ‘Cultural Racism’, as shown in Figure 4, used the item ‘some cultures are much better’ as an indicator of racism. The factor loading of this item varied substantially when used as an indicator of racism; its lowest factor loading was uncovered in Norway (0.27), but the factor loading for this item was low also in Israel (0.32). However, in the other countries, the factor loading was at least 0.40, with 0.78 in Lithuania as its highest value.

The partially low factor loading for the item on cultural superiority suggested that this item was not consistently a good indicator of racism. Specifically, the low factor loading in Norway indicated that the item did not reflect racism in this country, compatible with the frequent endorsement of this particular item in Norway. But the results still supported other claims by the theory of cultural racism by showing a strong correlation between the two estimated factors (see for instance the correlation of .73 in Norway). A strong correlation between the two factors substantiated that those who expressed cultural concerns when confronted with immigration also tended to express views in line with biological racism.

In a few countries (Denmark, France, the Netherlands, Norway, Sweden), the ‘Culturalism is Separate’ model had good fit even when the model included no cross-loadings (specifically, the item on cultural superiority loaded only on the factor Culturalism). But this solution, maintaining two separate factors with no-crossloadings, showed that expressing a culturalist view (as estimated by the factor Culturalism) implied a high probability of also maintaining a racist belief: the correlations between the factors Culturalism and Racism in these countries was between .78 and .85, implying that most of the variance in culturalism (61% to 72%) overlapped with the variance of racism.

The second robustness check added binary indicators for items with a preponderance of zeros (see the section Analytical Strategy for details). Again, these analyses were performed on each country separately. Even when the two binary indicators were added to both models, the model ‘Cultural Racism’ had substantially better fit than the model ‘Culturalism is Separate’; detailed results are available in the supplemental material.

A third robustness check used binary indicators only. Non-binary items were recoded to be binary; the idea was to test whether the superiority of the ‘Cultural Racism’ model might rest on the mix of binary items assessing the Racism factor and non-binary items assessing the factor Attitudes to immigration. Even with this modification, the superiority of the ‘Cultural Racism’ model was confirmed. The alternative model, ‘Culturalism is Separate’, failed to converge in 20 countries (in each case resulting in a non-positive latent variable covariance matrix). In contrast, the ‘Cultural Racism’ model converged, although resulting in a negative error variance in five of the 21 countries. Even if one were to disregard the non-convergence of the ‘Culturalism is Separate’ model, the ‘Cultural Racism’ model was superior, having much better fit estimates than the competing model (see the supplemental material for details).

Endorsements of the belief in cultural superiority varied substantially across countries, from 67% in Norway to 27% in France. These country scores were not linked to country differences in social cohesion in a similar manner as beliefs in biological racism are (see Caller and Gorodzeisky 2021). For instance, the nationally representative samples from Nordic countries – typical examples of countries with strong social cohesion (e.g. Delhey and Newton 2005) – expressed either moderate (40% and 42% in Finland and Sweden, respectively) or strong (61% and 67% in Denmark and Norway) support to the idea of cultural superiority. Conversely, in Eastern European countries with lower social cohesion, endorsements of the idea of some cultures being superior varied from 29% in Hungary to 57% in Czechia.

Approvements of the belief in cultural superiority varied across countries, but did they reflect racism? The current research held that in theory, culturalism and racism are most accurately seen as separate constructs, but empirical analyses may still find that they overlap. Overall, tests with representative samples from 21 countries analysed separately gave notible support to claims put forward by the theory of cultural racism. The link between culturalism and biological racism was evident either by factor analysis failing to distinguish between culturalism and racism (crossloadings had to be added), or by a very strong correlation between culturalism and racism. A strong correlation between culturalism and racism suggested that these two factors may constitute subdimensions in a general outgroup orientation rather than being separate beliefs.

The current research indicates that empirically, it is difficult to distinguish culturalism from racism in any of the 21 countries included in the analysis. The belief in cultural superiority tended to be an indicator of traditional racism. If this item was not an indicator of racism, then cultural concerns when faced with immigration correlated strongly with racism. Thereby, the current research supports Fredrickson's notion of a grey area between culturalism and racism. Yet this grey area was present not only among subgroups in a society as identified by earlier research, but in whole nations across Europe (and in Israel).

One caveat, however, is that the belief in cultural superiority appeared not to have the same implication across countries. For instance, the large differences between neighbouring countries Norway and Sweden in their endorsements of the idea of cultural superiority hint at differences in the implied meaning of expressed support to the idea of cultural superiority. Varying meanings of this item are even more evident in differences across countries in the item's ability to assess the Racism factor: the factor loading varied from 0.78 in Lithuania to 0.27 in Norway. These very different factor loadings across countries are a clear indication that the belief in cultural superiority (and the estimated factor Racism) did not have the same meaning across countries. Consequently, it seems not justified to generally equalise the belief that some cultures are superior with racism. For instance, declaring two thirds of the Norwegian sample to be racist because they endorsed the idea of cultural superiority seems not a valid conclusion based on the present data. Still, even with this qualification, the results provided clear support to claims put forward by the theory of cultural racism. For instance, the weak factor loading in Norway for the item on cultural superiority specifically did not prevent a very strong correlation between cultural concerns and racism.

The current research has several strengths: it was able to replicate results across 21 countries using large, representative samples from each country and thereby avoided potentially arbitrary findings based on a single sample. Also, with the full code provided in the supplemental material, it should be easy to reproduce the current findings and to test them with new samples. However, there are also limitations in the current data, which may be addressed in future research.

The data used were collected in 2014 and 2015, ahead of Europe's immigration crisis in 2015/2016. Although the current analyses are not obviously affected by the timing of measurements, renewed measurements would be helpful as several European countries now have received more refugees or immigrants from outside Europe. For instance, will the results for Sweden (low endorsements of both biological racism and the belief in cultural superiority) have changed after the country allowed many refugees during the refugee crisis? Renewed data collections would help clarify whether the current findings can be replicated across time, in addition to the current replications across countries.

Also, replication in the current research does not include replication with different measurements and the analysis was restricted to one item on the belief in cultural superiority. The large sample sizes even within countries compensated for the use of a single item to assess beliefs in cultural superiority. Furthermore, the factor analysis built on various items as indicators of both culturalism and racism. But the factor analysis also detected that the belief in cultural superiority was unlikely to have the same meaning across contexts. The assumption that any belief in cultural superiority implies racism (that is, cultural racism) may have contributed to the European Social Survey assessing such beliefs with only one item, preventing a detailed investigation into different implications of beliefs in cultural superiority. Such analyses would be possible if new surveys include more detailed measurements, specifically of the belief in cultural superiority.

The results suggest that an expressed belief that some cultures are superior is unlikely to have one and only one meaning across contexts. In some contexts (or countries), the expressed belief in cultural superiority might be little more than a revised racist claim of immanent differences between people due to their background. In other contexts, the expressed belief that some cultures are superior might reflect a far more benevolent view. For instance, seeing some cultures as oppressive for women and sexual and religious minorities might lead not only to the belief that less oppressive cultures are superior. It might also lead to a willingness to accept refugees from oppressive cultures. Therefore, a detailed measurement of the belief in cultural superiority could consider Blum's (2020) classification of different forms of culturalism, some apparently qualifying for the characteristic of cultural racism, others not. Research with more detailed measurements could distinguish between applying the belief in cultural superiority to cultures and traditions as abstract entities, or to the dominant people in such cultures, or to individuals (that is, becoming essentialist like traditional racism is, e.g. Haslam and Whelan 2008). Extended assessments might also include measurements of unconscious discrimination, for instance as assessed by the implicit racial bias test (Banaji and Greenwald 2016). Research could then examine whether people maintaining culturalist views also have a tendency to discriminate based on ‘racial’ characteristics – even if they hold no traditional racist beliefs.

More detailed measurements could also help investigate a suggestion by the theory of cultural racism, which appears to be supported in the present research: the idea of a culturalist beliefs forming a subdimension within racism. The factor analysis in the present research showed that either was the belief in cultural superiority a substantial indicator of racism, or culturalism and racism had a very strong correlation. Does the latter imply that biological racism and culturalist views form two dimensions within a higher-order racist orientation? Testing this suggestion is possible with higher-order factor analysis, but this form of factor analysis would require many more items, and the estimation of at least four separate first-order factors instead of the two estimated in the current research.

The current research explicitly distinguished between culturalism and racism in theory and left it to the empirical tests to uncover links. Even with this approach somewhat critical to the theory of cultural racism, this research has given strong evidence not only to Fredrickson's (2015) notion of a grey area between culturalism and racism, but even to claims put forward by the theory of cultural racism. Factor analysis indicated that in most of the European countries included in this research, the belief in cultural superiority was more closely linked to racism than to cultural concerns. Furthermore, if a distinction between a belief in cultural superiority and racism was at all possible, culturalist ideas of the need to maintain a homogeneous culture were closely linked to racism. Biological racism is alive and kicking in Europe, and the problem may be more far-reaching than uncovered by measurements of biological racism alone. The present research suggests that even references to cultural superiority or the expressed need for cultural homogeneity will tend to reflect a racist orientation.

No potential conflict of interest was reported by the author(s).

Christopher Bratt is associate professor at the Inland Norway University of Applied Sciences (Norway) and honorary research fellow at the University of Kent (United Kingdom). His main research interest is in the field of intergroup relations; his work also includes research on topics such as health services, mental health, and consumers' contribution to environmental protecton.