Immigrant political participation is associated with more positive majority immigration attitudes across European countries and Swiss cantons

Abstract

Immigrants are increasingly participating in politics, publicizing their political concerns and contributions. How does such political participation relate to national majorities’ immigration attitudes? Previous research suggested potential improvement of majority attitudes but also demonstrated the exacerbation of perceived threat. We investigated whether greater immigrant political participation is related to more positive or negative immigration attitudes among majority members. We implemented a cross-national and a cross-cantonal multilevel study drawing on the European Social Survey, the Second European Union Minorities and Discrimination Survey and the Swiss Migration-Mobility Survey with 43,632 participants in 26 European countries in Study 1 and with 1058 participants in 19 Swiss cantons in Study 2. Overall, higher levels of immigrant political participation were related to more positive attitudes. This association between participation and attitudes was stronger among left-wing than among right-wing nationals. Finally, we found no evidence that larger immigrant groups’ participation evoke threat and exacerbate attitudes. We conclude that immigrant political participation is associated with more positive majority attitudes.

Introduction

Close to five hundred undocumented immigrants engaged in a hunger strike for two months in Brussels, the capital of Europe in 2021 (Rankin, 2021). Their actions aimed to demonstrate the indispensability of immigrants’ labour, and to contest the criminalization of undocumented immigrants. Beyond specific protests and actions, immigrants (documented and undocumented) are increasingly participating in politics, sometimes successfully changing policies (Strijbis, 2015; Vintila & Martiniello, 2021). One key question is how such political participation is related to how national majorities perceive immigrants. Political participation is a tool to convey information about immigrants’ contributions to society and to create social change. In addition, political participation is also often seen as a crucial aspect of integration into society (Klarenbeek & Weide, 2020; Morales, 2011). Yet, national majority members might feel threatened by immigrants’ political participation if they see participation as questioning their dominant position in society (Bobo, 1999; Hindriks et al., 2015; Verkuyten, 2017). Indeed, sometimes even policy documents express ambivalence about immigrant participation: policy-makers might simultaneously demand active citizenship and label some forms of participation threatening (Klarenbeek & Weide, 2020). Despite the societal relevance of the phenomenon, little is known about the relation between immigrant political participation and national majority members’ immigration attitudes.

More broadly beyond immigrant political participation, there has been little attention on the impact of political participation on public opinion. Sociological and political science research has concentrated on the effects of political participation on policy outcomes and media coverage (see Amenta & Polletta, 2019 for a review). Only a handful of recent studies has investigated how protests shape public opinion (Dumas, 2018; Sawyer & Gampa, 2018). We contribute to this endeavor by investigating the link between immigrant political participation and public opinion. We provide a robust inquiry of these questions by replicating the analysis across 4 different datasets in 6 sets of analysis. We advance the state of the art by comparing the participation of immigrants of different origin, by examining different forms of participation beyond protests and by looking at immigrant participation and majority attitudes across national and regional contexts in Europe.

Theoretical background

Political participation of immigrants: improved immigration attitudes or exacerbated threat among national majority members?

From the national majority members’ perspective, the political participation of immigrants (and minorities) can be seen as a sign of their integration. In general, national majority members prefer that immigrants integrate or assimilate into society (Brown & Zagefka, 2011). In the political realm, such integration or assimilation would entail political participation in similar forms and subjects as the participation of the native population (Morales, 2011). Indeed, majority members in Europe find it somewhat important that immigrants participate for the success of immigrant integration (Drazanova et al., 2020). Based on this theorizing, higher immigrant participation within the conventional framework of participation would be related to improved immigration attitudes among national majority members. In line with this, two studies drawing on the same large-scale survey investigated how the five million strong “Día Sin Inmigrantes [Day without immigrants]” protests in 2006 related to immigration attitudes in the United States. Importantly, they focused not on the general public, but only on Latinxs (to use the gender-neutral term for Latinas and Latinos). The survey was fielded during the protests, providing a natural field experiment to compare responses before and after the protests took place across US cities. This natural field experiment showed that Latinxs exposed to immigrant protests endorsed more permissive immigration policy preferences (Branton et al., 2015; Carey et al., 2014). The protests were especially effective in changing the opinions of Latinxs born in the US who were not directly affected by the proposed anti-immigration measures themselves. In sum, political participation might be seen as a sign of immigrant integration and thus relate to more positive immigration attitudes among majority members.

In contrast with the above reasoning about immigrant integration, majority members might see politically engaged immigrants as competitors for power and influence and therefore threatening. In general, conflict and threat theories claim that majority group members respond with increased perceptions of threat and prejudice when they perceive minority members as menacing their position (Bobo, 1999; Stephan et al., 2016). The political participation of immigrants could thus be seen as threatening if majority members believe it endangers their cultural, economic or political status and position. In line with these arguments, two recent experimental studies found that national majority members’ attitudes worsened in response to immigrants’ political participation. More specifically, in these two studies Dutch majority members were shown vignettes about Muslim citizens with a migration background aiming for political power through engaging in party politics. As a result, the participating Dutch majority members felt more negative towards Muslim immigrants (Hindriks et al., 2015; Verkuyten, 2017). The reason for more negative attitudes was that national majority members in these experiments felt that the political power of the majority group was threatened. Thus the evidence on the effect of immigrant minority members’ political participation on attitudes is mixed: two studies found that immigrants’ political participation results in more positive immigration attitudes while two other papers documented more negative attitudes.

Arguing for a different direction of causality, it has also been suggested that immigrant minority members participate politically because national majority members immigration attitudes are already positive (Just & Anderson, 2014). Just and Anderson argue that more positive majority attitudes constitute a welcoming climate and thus create more opportunities for immigrant minority members to participate. Their argument is based on cross-sectional associations between country-level majority attitudes and immigrant political participation without testing explicitly the role of opportunity structures. Furthermore, ample social psychological and sociological evidence suggests the reverse: minority members (including immigrant minorities) participate more politically exactly in response to hostile environments, when they perceive more injustice (Goksu & Leerkes, 2022; van Zomeren et al., 2008). We thus theorize that it is primarily immigrant political participation that influences majority attitudes even though previous evidence is mixed on whether immigrant participation improves or exacerbates attitudes. Accordingly, we test this idea in 2 studies in which we examine the association between immigrant political participation and majority immigration attitudes. We also test the association of political participation and attitudes over indices of the political opportunity structure to substantiate our line of reasoning about the direction of effects.

Thus, the present paper aims to shed light on previous mixed findings and investigate whether higher immigrant participation will be related to more positive immigration attitudes (Hypothesis 1). Considering the mixed evidence and contrasting theoretical predictions on improving or worsening attitudes we do not put forward specific expectations on more positive or negative immigration attitudes as outcomes. We examine threat perceptions and support for permissive immigration policies as two facets of immigration attitudes. We focus on immigrant political participation to complement previous studies that investigated the effect of protests by historic ethnic-racial minority members on majority members’ attitudes (Sawyer & Gampa, 2018). We also extend the scholarship by examining political participation more broadly beyond protest activities.

Political participation of immigrants and national majority members’ political orientation

One reason for previous mixed results might be that left-oriented or right-oriented national majority members react differently to immigrants’ political participation. From a threat perspective, right-wing majority members may perceive politically active immigrants as especially threatening, to the extent that active immigrants are seen as competing for political power (Stephan et al., 2016). First, more right-oriented majority members feel more threatened by immigrants than more left-oriented majority members (Green & Sarrasin, 2018; Hodson & Dhont, 2015). Second, more right-oriented majority-members by and large prefer societal stability and less equal societies as opposed to social change and more equal societies (Jost et al., 2003). Immigrant political participation aims to create social change and more equal societies, thus right-leaners should respond to it unfavorably. Therefore, we expect that higher immigrant participation will be related to more positive immigration attitudes in particular among more left-oriented national majority members. In contrast, these relationships would be weaker or even reversed among more right-oriented national majority members (Hypothesis 2).

Political participation by larger groups of immigrants as perceived threat to national majorities’ political power

Following threat theories, a larger active minority group would be seen as especially threatening to majority members’ political power because larger groups are more likely to successfully challenge the majority group’s status and position (Stephan et al., 2016). Thus, another reason for the previous mixed findings could be that immigrants’ political participation is seen as less threatening by national majority members when the immigrant groups are smaller. To the best of our knowledge, this theorizing has not been empirically tested yet. Therefore, we investigate the interplay of immigrant political participation and immigrant presence. We expect that immigrant political participation would be related to more positive immigration attitudes when immigrant presence is relatively low. In contrast, it would be related to more negative immigration attitudes when immigrant presence is relatively high (Hypothesis 3).

The current studies

Our study is the first to investigate the relation between actual immigrant political participation and national majority members’ immigration attitudes. To test our hypotheses, we conducted two cross-sectional studies drawing on publicly available large-scale surveys, focusing on immigrants’ political participation at the country level and the regional level. In Study 1, we focused our inquiry on relations between immigrants’ political participation and national majority members’ immigration attitudes across different countries in Europe. In Study 2, we compared Swiss cantons. We chose Swiss cantons because they are subnational contexts with meaningful differences in attitudes towards immigrants (Green & Sarrasin, 2018). Furthermore, immigration levels in Switzerland are high compared to other European countries, therefore the country provides a useful test case for our predictions. We thus tested the relation between average immigrant political participation at the country or cantonal level and national majority members’ immigration attitudes (threat perceptions and immigration policy preferences) at the individual level. We focused on political participation in several forms, but did not take into account participation in national elections, because participation in national elections is less accessible to non-nationals in several European countries. In contrast, immigrants regularly participate in other forms, for example they would contact politicians, sign petitions, or participate in protests and boycotts (Vintila & Martiniello, 2021). In both Study 1 and Study 2, we investigated immigrant participation among immigrants of different origin, because attitudes towards different immigrant groups vary widely (Green & Sarrasin, 2018; Montreuil & Bourhis, 2001). We drew on data that sampled broader immigrant populations and also data that sampled specific immigrant or immigrant-origin groups, both from devalued and from more positively viewed immigrant groups. If we find comparable associations between immigrant political participation by these different immigrant groups and majority attitudes across countries and cantons, we can conclude with more certainty that our proposed key mechanisms hold across different social contexts.

Data and methods

Data

We identified suitable surveys through searching the GESIS archive and a systematic literature review (GESIS – Leibniz Institute for the Social Sciences, 2021). To construct our country and cantonal-level independent variable of immigrant political participation, we selected surveys that included items on political participation and drew on representative samples of immigrants across European countries and Swiss cantons. We matched the country-level or cantonal-level independent variables of immigrant political participation to the country or canton of the individual-level dependent variables of threat perceptions and policy preferences among national majority members. In the following sections the samples, measures and analytical strategy of the two studies are presented in parallel.

The data for both studies are available at: https://osf.io/spzy7/?view_only=233a535ae488448ebbde471f821609d0

Participants

Study 1

We used the European Social Survey (ESS) that included representative samples across European countries. We combined data from the ESS Round 8 (2016) and the ESS Round 9 (2018) to maximize the number of countries sampled (NSD - Norwegian Centre for Research Data, 2018). The ESS is a cross-national survey measuring the attitudes, beliefs and behaviour of diverse populations. Please see the technical report of each ESS survey for sampling information and complete list of questions (NSD - Norwegian Centre for Research Data, 2018).

We selected as national majority members those participants who were born in the country. The sample included altogether 43,632 participants in 26 countries (44.06% female, M_age = 50.37, SD = 18.83). Additional file 1: Table S1 in the Supplemental Materials shows the number of participants and descriptive statistics by country.¹

Study 2

Study 2 compared Swiss cantons using Swiss ESS data from Round 9 (2018) (NSD - Norwegian Centre for Research Data, 2018). We again selected as national majority members those participants who were born in the country. The final sample included 1058 respondents in 19 cantons (50.80% female, M_age = 49.99, SD = 19.67). Additional file 1: Table S2 in the Supplemental Materials shows the number of participants in each canton and descriptive statistics by canton.²

Measures

Individual-level dependent variables (Study 1 and 2)

The dependent variable threat perceptions was measured with three items, an example item is ‘Would you say that [country’s] cultural life is generally undermined or enriched by people coming to live here from other countries?’ (response scale ranging from 0 [cultural life undermined] to 10 [cultural life enriched]). Answers were reverse-coded so that higher values represent greater threat perceptions and averaged to create a 3-item scale. Threat perceptions were used in both Study 1 and Study 2 as a dependent variable. (Study 1: α = 0.88, αs by country from 0.753 to 0.945, M = 4.99, SD = 2.27) (Study 2: α = 0.787, as by canton from 0.566 to 0.887, M = 4.09, SD = 1.70) (see descriptive statistics and αs by country and canton in Additional file 1: Tables S1 and S2).

Policy preferences were measured with three items, an example item is ‘To what extent do you think [country] should allow people of the same race or ethnic group as most [country]’s people to come and live here?’ (response scale ranging from 1 [allow many to come and live here] to 4 [allow none]). Answers were averaged to create a 3-item composite scale with higher numbers indicating stronger anti-immigration policy preferences (Cronbach’s α = 0.876, αs by country from 0.672 to 0.960, M = 2.42, SD = 0.83). Policy preferences were only used as a dependent variable in Study 1, as there were no significant differences in policy preferences among national majority members and immigrants in Study 2 and cantonal-level variance was below 5% (see preliminary results in Supplemental Materials and see descriptive statistics and αs by country in Additional file 1: Table S1).

Country and cantonal level independent variables (Study 1 and 2)

Study 1

To calculate the country-level independent variable of immigrant political participation, we used data from the European Social Survey (ESS) and the Second European Union Minorities and Discrimination Survey (EUMIDIS-II). First, for the ESS immigrant political participation measure, we selected as immigrants those participants who were born outside the country (See Additional file 1: Table S1 for sample sizes per country).³ Second, we used data from the Second European Union Minorities and Discrimination Survey (EUMIDIS-II) that sampled the most devalued immigrant and immigrant-origin minority groups in EU Member States in 2015–2016 (European Union Agency for Fundamental Rights (FRA), 2016). More specifically, the EU-MIDIS II sampled immigrants and descendants of immigrants from North Africa, Turkey and Sub-Saharan Africa in the countries in our study. Here we could only include 15 countries in the analysis where the EU-MIDIS II and the ESS Round 8 or 9 used for the dependent variables overlapped (See Additional file 1: Table S1 for sample size and mean political participation per country).

Country-level Immigrant political participation ESS was measured using eight items in the ESS, example items are: ‘There are different ways of trying to improve things in [country] or help prevent things from going wrong. During the last 12 months, have you done any of the following?’ ‘Have you worked in a political party or action group?’; ‘Have you worked in another organization or association?’; ‘Have you taken part in a lawful public demonstration?’. In a first step, we took the sum of affirmative responses to these items per immigrant respondent and in a second step we calculated a country-level mean of these sum scores (M = 0.98, SD = 0.54, ranging from 0.25 to 2.22).

Country-level Immigrant political participation EUMIDIS-II was computed from three items: “In the past 12 months, have you done any of the following?” “Have you ‘liked’ or ‘followed’ a political campaign on the internet?”, “signed a petition on paper or online?”, “taken part in a public demonstration?” As with the ESS political participation measure, we first took the sum of affirmative responses for each minority participant. Afterwards, the country-level mean of political participation was calculated (M = 0.27, SD = 0.16, ranging from 0.05 to 0.64).

Study 2

We calculated the cantonal-level independent variable of immigrant political participation drawing on the Swiss data from the European Social Survey and the Swiss Migration-Mobility Survey in Study 2 (D’Amato et al., 2018; NSD - Norwegian Centre for Research Data, 2018). First, we drew on the Swiss ESS. Second, we used data from the migration mobility survey (MMS). The migration-mobility survey (MMS) is the largest survey organized in Switzerland among the migrant population. It gathers information on the living conditions of residents of Switzerland who arrived within the last ten years. The MMS2016 focused on citizenships belonging to six languages (English, German, French, Italian, Spanish and Portuguese). This strategy covered 63% of the migrant population, but largely sampled European migrants, thus focused on immigrants who are often more positively viewed.

Canton-level Immigrant political participation ESS was the same as the Immigrant political participation ESS measure in Study 1 (M = 0.85, SD = 0.34, ranging from 0 to 1.8).

Canton-level Immigrant political participation MMS was measured using ten items from the MMS, example items are: “There are different ways of trying to improve things or help prevent things from going wrong. During the last 12 months, have you done any of the following in Switzerland?” “Worked in a political party or action group”; “Worked in another organization or association”; “Taken part in a lawful public demonstration”. We followed the same procedure as in Study 1, and calculated a cantonal-level mean of the sum of affirmative responses per immigrant respondent (M = 0.61, SD = 0.09, ranging from 0.13 to 0.85).

Individual level independent variables (Study 1 and 2)

Political orientation was measured with the item “In politics people sometimes talk of ‘left’ and ‘right’. Where would you place yourself on this scale, where 0 means the left and 10 means the right?” (Study 1 M = 5.12, S.D = 2.24; Study 2 M = 5.16, SD = 2.06).

Control variables (Study 1 and 2)

At the individual level, we controlled for gender, age and level of education (coded as standardized level of education across countries) (Ceobanu & Escandell, 2010). We also included percentage of immigrants at the country (Study 1) or cantonal level (Study 2) (Study 1 M = 12.69, SD = 6.67, Study 2 M = 29.52, SD = 7.74) (Eurostat, 2018; OFS - Swiss Federal Statistical Office, 2018).

Robustness checks (Study 1 and 2)

In both studies, we replicated the main effect of immigrant political participation controlling for the most common contextual determinants of immigration attitudes such as the GDP and unemployment rates (Green & Sarrasin, 2018). We also replicated the key results taking into account immigrants’ political rights in a country or a canton to account for political opportunity structures that could impact the levels of immigrant participation (Morales, 2011).

Tables 1 and 2 presents descriptive statistics and correlations for our key variables in Study 1 and Study 2 respectively.

Table 1 Descriptive statistics and correlations for Study 1

Table 2 Descriptive statistics and correlations for Study 2

Analytical strategy

A series of two-level regressions were implemented in both studies to test the hypothetical effects of immigrant political participation (Hypothesis 1—level 2 hypothesis) and the conditional effects of immigrant political participation depending on individual political orientation (Hypothesis 2—cross-level interaction) and the conditional effects of immigrant political participation depending on immigrant presence (Hypothesis 3—level 2 interaction). Multi-level analysis was required because of the nested data structure of participants (individual level) within countries or cantons surveyed (country or canton level). In Study 1, a set of models predicted national majority members’ individual threat perceptions and the other set predicted national majority members’ individual immigration policy preferences. In Study 2, the same model-building steps were employed to test hypothetical effects on national majority members’ individual threat perceptions only.

The main analyses in both studies were run in consecutive steps starting from the null model (“Null model”); secondly including individual-level control variables—i.e. age, gender and education in one step (“Level 1 control” models); thirdly including the country-level control immigrant presence (“Level 2 control” models) and afterwards including country-level or canton-level predictor immigrant political participation to test Hypothesis 1 (“Immigrant political participation” models). Next, we tested the interaction between political orientation and immigrant political participation to test Hypothesis 2 (“Political orientation interaction” models). Finally, we added the interaction between immigrant presence and immigrant political participation to test Hypothesis 3 (“Immigrant presence interaction” models). When decomposing the significant interactions, we defined left-wing and right-wing as plus or minus one standard deviation from the mean of political orientation in each analysis (Hypothesis 2) and similarly high and low immigrant political participation and immigrant presence as plus or minus one standard deviation from the mean political participation or presence in each analysis (Hypothesis 3). We included data from all participants in the analysis who had valid responses on the dependent and control variables, we did not impute missing values for any of these variables. In Study 1, we used country-level design weights for all multilevel analyses to account for the unequal probabilities of participants being sampled (Heck & Thomas, 2020). The analysis was conducted using Mplus 8 (Muthén & Muthén, 2017). We only present in detail the analysis drawing on ESS data for the independent variables in the main text to avoid redundancy in both Study 1 and 2. We summarize the analysis drawing on the EU-MIDIS II (Study 1) and the MMS (Study 2) data in the main text and describe it in detail in the Supplementary Online Materials (SOM).

Results

Study 1

Threat perceptions

Table 3 presents the results for national majority members’ individual threat perceptions. Older people, less educated people and people more oriented towards the right end of the political spectrum perceived higher levels of threat, while gender had no effect on threat perceptions. Higher immigrant presence was related to lower threat perceptions.

Table 3 Study 1. Stepwise models predicting national majority members’ threat perceptions at the individual level as dependent variable using ESS Round 8 or 9 from immigrant political participation at the country level as independent variable using ESS Round 8 and 9

More importantly, when testing Hypothesis 1 we found that in countries where immigrants were more active politically, national majority members perceived lower levels of threat. Furthermore, the model including immigrant political presence and immigrant political participation explained 48% of the country-level variance in threat perceptions, while the model with only immigrant presence explained just 10% of the country-level variance, thus adding immigrant political participation greatly increased the explanatory power of the model.

Testing Hypothesis 2, we found that the cross-level interaction between immigrants’ political participation and national majority members’ political orientation was significant. To interpret the interplay between immigrants’ political participation and national majority members’ political orientation, we plotted this interaction. Decomposing the interaction showed that both right-oriented national majority members and left-oriented national majority members perceived lower threat when immigrants’ level of political participation was high as opposed to low (see Fig. 1). Meanwhile, the relationship was stronger for left-wing participants supporting Hypothesis 2. Confirming the stronger effect among left-wing participants, the line is steeper when we compare threat perceptions at low and high participation for left-oriented participants as opposed to right-oriented participants.

Fig. 1

Study 1. National majority members’ political orientation and immigrants’ political participation predicting majority members’ threat perceptions. Note: †p < 0.10, *p < 0.05, **p < 0.01, ***p < 0.001 (2-tailed) in all the figures

We did not find support for Hypothesis 3, as the interaction between immigrant political participation and immigrant presence was not significant.

We repeated the main analyses drawing on the EU-MIDIS II for our independent variable of political participation. The results for Hypothesis 1 replicated: higher participation was related to lower threat perceptions (H1). We did not test Hypothesis 2 because a sample of 15 countries does not suffice for testing cross-level interaction. When testing Hypothesis 3, we found a significant interaction between political participation and immigrant presence. In contrast with our expectations, high immigrant participation was related to lower levels of threat than low immigrant participation at both low and high immigrant presence and the effects were stronger at high immigrant presence. For details, please see the narrative description of results, Additional file 1: Table S3 and Fig. S1 in SOM.

Policy preferences

Table 4 presents the results for national majority members’ immigration policy preferences. Older, less educated, more right-wing oriented people and men had more anti-immigration policy preferences. When immigrant presence was higher, national majority members’ policy preferences were less against immigration.

Table 4 Study 1. Stepwise models predicting national majority members’ immigration policy preferences at the individual level as dependent variables using ESS Round 8 or 9 from immigrant political participation at the country level as independent variable using ESS Round 8 and 9

When testing Hypothesis 1, we found that in countries with higher immigrants' political participation, national majority members were less supportive of anti-immigration policies.In addition, the model with immigrant presence only explained 18% of the country-level variance in policy preferences while the model including immigrant presence and immigrant participation explained 67% of the variance. These results thus suggest that immigrant political participation explains a large part of the variation in policy preferences.

Next, we tested Hypothesis 2. National majority members’ political orientation and immigrants’ political participation interacted significantly to predict immigration policy preferences. Decomposing the interaction showed that the results supported Hypothesis 2: the effect of immigrant political participation on policy preferences was stronger for left-oriented than for right-wing oriented national majority members (Fig. 2). More specifically, high levels of immigrant political participation were related to less anti-immigrant policy preferences among both right-oriented national majority members and left-oriented national majority members. However, confirming the stronger effect for left-leaning individuals, the line is steeper when we compare low and high participation for left-wing participants as opposed to right-wing participants.

Fig. 2

Study 1. National majority members’ political orientation and immigrants’ political participation predicting majority members' policy preferences

Finally, we did not find support for Hypothesis 3, the interaction between immigrant presence and immigrant political participation was not significant.

We replicated the main analysis using the EU-MIDIS II immigrant political participation measure. The main effect of immigrant political participation was not significant, thus we did not find support for Hypothesis 1. Due to the low number of countries sampled we did not test Hypothesis 2. We found a significant interaction of immigrant presence and immigrant political participation when testing Hypothesis 3. Similar to the findings on threat perceptions, the pattern again contradicted our expectations as it was exactly at high immigrant presence that high (compared to low) immigrant political participation was related to less negative policy preferences. In contrast, at low immigrant presence there was no significant difference in policy preferences at high and low political participation. See narrative description of the results, Additional file 1: Fig. S2 and Table S4 in SOM for details.

Robustness checks

We replicated the main effect of immigrant political participation on majority immigration attitudes controlling for GDP, unemployment rates and immigrants’ political rights in a country. Please see the SOM for detailed narrative and models in Additional file 1: Tables S5 and S6.

Study 2

Threat perceptions

Table 5 presents the results for the models predicting national majority members’ threat perceptions. Younger, more highly educated and male participants reported lower levels of threat. Furthermore, more right-oriented participants reported higher threat. Higher immigrant presence was related to lower levels of threat.

Table 5 Study 2. Stepwise models predicting national majority members’ threat perceptions at the individual level as dependent variables using ESS Round 9 from immigrant political participation at the cantonal level as independent variable using ESS 9

When testing Hypothesis 1, we found that higher immigrant political participation was related to lower perceived immigrant threat among national majority members. Furthermore, the model including immigrant presence and immigrant political participation explained 94% of the variance (ESS data) at the cantonal level compared to 86% of the variance explained by the model with only immigrant presence.

Second, we did not find evidence in the direction of Hypothesis 2, the interaction between national majority members’ political orientation and immigrants’ political participation was not significant. We did not find support for Hypothesis 3 either, the interaction between immigrant presence and immigrant political participation was not significant.

We replicated these results drawing on the MMS for immigrant political participation. The findings were similar across the two sets of analysis, except that the main effect of immigrant political participation was only marginally significant (H1) using the MMS measure. Please see the SOM for detailed narrative and models in Additional file 1: Table S7.

Robustness checks

We replicated the main effect of immigrant political participation on majority immigration attitudes over GDP, unemployment rates and an index of immigrants’ political rights in a canton. Please see the SOM for detailed narrative and models in Additional file 1: Table S8.

Discussion

Immigrants face fundamental inequalities in expressing their political concerns. Still, they increasingly participate politically in many ways (Ataç et al., 2016; Strijbis, 2015; Vintila & Martiniello, 2021). In the current research, we showed that such immigrant political participation relates to more positive immigration attitudes among national majority members. From a psychological perspective, more active immigrants might be perceived in a more positive light if political participation is viewed as a sign of their integration. However, immigrant political participation might also be seen as a threat to national majority members’ status and position. Previous empirical studies provided mixed results on the impact of immigrant political participation (Branton et al., 2015; Carey et al., 2014; Hindriks et al., 2015; Verkuyten, 2017). We attempted to shed light on these mixed findings and we advanced the state of art by examining the actual political participation of immigrant minority members drawing on large scale surveys and by investigating political participation more broadly beyond protests. We implemented a cross-national study across 26 countries and a cross-cantonal study across 19 Swiss cantons with altogether 43,632 participants. Overall, higher levels of immigrant political participation were consistently related to more positive immigration attitudes (Hypothesis 1). Furthermore, immigrant political participation was more positively associated with the immigration attitudes of left-wing national majority members than right-wing national majority members in the cross-national analyses in Study 1 though not in the cross-cantonal analyses in Study 2 (Hypothesis 2). Finally, we did not find consistent evidence that participation would be especially threatening when immigrant groups are larger (H3) and relate to more negative immigration attitudes.

Overall our findings suggest that immigrant political participation is associated with more positive majority attitudes without necessarily inducing feelings of threat. First, we consistently found positive associations between immigrant political participation and majority attitudes. Second, immigrant political participation was positively associated with immigration attitudes over structural factors and political opportunity structures that could provoke feelings of threat: we replicated the main effect of immigrant political participation over the mere presence of immigrant groups, country-level and cantonal-level measures of unemployment, wealth and immigrant political rights (Green & Sarrasin, 2018). Third, based on threat theories, we expected threat effects especially in the presence of relatively large immigrant groups, because larger immigrant group might be seen particularly capable of contesting the status and position of the majority group, but these expectations were not confirmed (Bobo, 1999; Stephan et al., 2016). We did not find in either of the six set of analysis that immigrant political participation would be seen as more threatening when immigrant groups are larger. Moreover, in two sets of analysis we found that immigrant political participation is related to more positive attitudes exactly when immigrant presence is high. Interestingly, we only found these patterns when examining the political participation of the most devalued immigrant groups across Europe who were sampled in the EU-MIDIS survey i.e. immigrants and descendants of immigrants from North Africa, Turkey and Sub-Saharan Africa. The findings are thus in line with previous large-scale survey studies from the United States showing that a Day without Immigrant protest improved minority Latinx Americans’ immigration attitudes (Branton et al., 2015; Carey et al., 2014). In contrast, our findings contradict previous experimental research from the Netherlands showing that immigrant political participation provoked threat reactions among majority nationals, especially when citizens with a migration background were aiming for political power in party politics as a separate group (Hindriks et al., 2015; Verkuyten, 2017).

We speculate that the differences from the above-mentioned studies in our results might be due to differences in the aim or the form of immigrant political participation, in the perceptions of the immigrant groups concerned as well as in the political opportunity structures of immigrant political participation. First, the US protests voiced grave societal concerns but did not challenge majority political power directly. In contrast, the Dutch studies presented scenarios of immigrants vying for political positions. Though the data we employed do not describe the subject of immigrant participation, it could have consisted of claims that were less threatening for national majority members. Second, we excluded participation in national elections, as most immigrants do not have such voting rights (Vintila & Martiniello, 2021). The two Dutch studies, in contrast, investigated immigrant participation in party politics which could be seen as especially threatening, because that is where (arguably) real power lies in Western democracies. Third, the Dutch studies presented scenarios about Muslim immigrants. Contrary to the Dutch experiments, our research looked at the effective political participation of all residents of immigrant-origin, also replicating our results with the most devalued immigrant groups (using the EUMIDIS survey in Study 1) but without focusing on Muslim immigrants. As societal master narratives often pit European and Muslim values against each other, Muslim immigrants could be seen as especially threatening by majority Dutch nationals. Fourth, the Netherlands constitutes a special context for the participation of immigrants in party politics. The political opportunity structure facilitates the parliamentary entrance of small parties rooted in immigrant communities, unlike in most European countries. Therefore the Dutch results might be due to this special political context. Our research went beyond these previous studies by investigating the actual political participation of immigrants, by examining the participation of different immigrant groups, by looking at participation beyond protest and by drawing on high-powered large-scale surveys. Nevertheless, further research should disentangle the impact of different aims and forms of political participation by different immigrant groups in different contexts.

Finally, our results point in the direction of stronger positive associations between political participation and attitudes among more left-leaning national majority members. These results align with robust findings on political orientation. Previous studies show that more right-leaning individuals feel threatened by social change and increasing equality, changes that immigrant political participation aim for (Jost et al., 2003). However, our results contrast those of the Hindriks et al. (2015), who did not find differences by political orientation. As our studies drew on considerably larger samples, the different results might be due to power necessary to detect such differences. Still, further research is necessary to clarify the role of political orientation.

While our main results on the association between immigrant political participation and majority nationals’ immigration attitudes were very consistent, we also found some differences in the different analyses, possibly because we investigated the participation of different immigrant groups. Most importantly, we replicated the main effect of immigrant political participation on national majority members’ attitudes in four out of six sets of analysis across countries and cantons. On the one hand, we found significant effects of immigrant political participation in Study 1, drawing on the ESS and the EUMIDIS-II for immigrant samples. Interestingly, we found no significant main effect in Study 1 on policy preferences (though we did on threat perceptions) when we drew on the EU-MIDIS survey for immigrant political participation, but participation of larger immigrant groups was related to more positive attitudes (both threat perceptions and policy preferences) in this analysis. Both the ESS and the EU-MIDIS included responses from immigrant or immigrant-origin participants from various countries of origin, also from outside the EU. Furthermore, the EU-MIDIS sampled participants from the most devalued and most visible immigrant groups. On the other hand, we only found a marginally significant effect when we drew on the Swiss Migration Mobility survey to capture cantonal immigrant political participation in Study 2. In contrast with the ESS and EU-MIDIS II immigrant samples, in Study 2, the Swiss MMS data set sampled immigrants mostly from the EU. Thus the MMS immigrant sample included participants who are further from the stereotypical immigrant pictured by European national majority members compared to the ESS and EU-MIDIS. We therefore speculate that what matters for national majority members’ immigration attitudes is the political participation of exactly those immigrants who embody the immigrant stereotype.

Our study contributed to the literature in several ways, most importantly we showed that higher immigrant political participation was related to improved majority attitudes without exacerbating perceived threat. But we were also limited by some of our methodological and data choices. First and foremost, our data is cross-sectional, not allowing for causal claims about immigrant political participation. Indeed, a study proposed just the opposite direction of effects, so that in countries with more positive immigration attitudes, immigrants would participate more. More specifically, Just and Anderson (Just & Anderson, 2014) found that in countries with more positive attitudes towards immigrants, immigrant participation is higher in un-institutionalized participation such as signing a petition, or participation a lawful demonstration, but they find no effect on institutionalized political participation such as contacting a politician. Just and Anderson explained their results by conceptualizing more positive attitudinal climate as part of the political opportunity structure of immigrant participation. For this reason, we replicated the effects of political participation on majority attitudes over indices of political rights for immigrants, a direct measure of opportunity structure. This replication suggests that the broader political opportunity structure in itself does not explain the relation between immigrant participation and majority attitudes. However, longitudinal and experimental studies would be necessary to clearly establish the direction of causality.

As further limitations, we had no information on the subject of immigrants’ political participation, on the participation of the most disadvantaged immigrant groups and on the specific effects of different forms of participation. First, some studies suggest that immigrants often participate politically on immigration-related issues, but these studies do not cover the whole European context (Vintila & Martiniello, 2021). Data on immigrant-led movements in different countries would allow us to overcome such limitations, but currently data is only available for a limited number of countries or cities, not allowing for multilevel analysis (e.g. Morales, 2011). Therefore, future large-scale studies should include questions on the subject of participation and cross-country studies on immigrant political participation should be implemented in a sufficient number of countries to allow for comparative multilevel approaches. Second, the immigrant samples in our surveys do not include the most disadvantaged immigrants, though for example undocumented immigrants have been mounting highly visible actions in many European cities, such as the 2-month-long hunger strike in Brussels in 2021. Future surveys should aim to reach also these groups. Finally, we decided to group together all forms of nonelectoral political participation reported in the surveys, but some forms of participation e.g. protests are more visible than others e.g. signing petitions. These different levels of visibility might impact majority perceptions differently, thus future research should investigate them separately.

At the same time, our sophisticated analyses disentangling effects at national or regional (i.e., cantonal) and individual levels and our use of measures of actual immigrant participation and large-scale surveys lend credence to our findings. Immigrant political participation not only serves as a tool for immigrants to voice their political concerns but, beyond the direct political goals of the participation, can also improve national majority members’ attitudes towards immigrants and immigration and thereby foster social cohesion.