A central focus of the public discourse about the European refugee crisis has been the large share of young men among the asylum seekers. Media outlets have written numerous stories depicting them as a problem, highlighting their aggressive behavior Footnote 1 and their need to be taught European gender norms. Footnote 2 Stories even have titles as blunt as “Abnormal number of young men a problem for Sweden.” Footnote 3 Politicians have been quick to politicize these young men. Geert Wilders, leader of the anti-immigrant Party for Freedom in the Netherlands, said “Masses of young men in their twenties…[are]…an invasion that threatens our prosperity, our security, our culture and identity.” Footnote 4 The head of the Alternative for Germany in Berlin said there are “…increasing problems with these so-called groups of young men.” Footnote 5
To what extent does this reflect the real fears of the European public? Perhaps these sentiments are an example of the hyperbole common among the media and politicians. It is equally plausible that young immigrant men are actually preferred because of their economic potential, which has been shown to positively affect attitudes toward asylum seekers (Bansak, Hainmueller, and Hangartner Reference Bansak, Hainmueller and Hangartner2016). In demographically ailing European societies, these young men can fill labor shortages and are likely to burden the welfare state less than women and older immigrants—countering an important driver of anti-immigrant attitudes (Fietkau and Hansen Reference Fietkau and Hansen2018).
However, other scholarly work suggests that we should see weak public support for young immigrant men. First, young immigrant men may be seen as posing a security threat. These concerns are known drivers of anti-immigration attitudes (Erisen and Kentmen-Cin Reference Erisen and Kentmen-Cin2017; Huysmans Reference Huysmans2006; Lahav and Courtemanche Reference Lahav and Courtemanche2012) and include fears of terrorism, sexual assault, theft, and other violence. Second, host communities may perceive that young men threaten their culture. Anxiety about the cultural impact of immigration extends to many domains, including language (Hopkins Reference Hopkins2015), the national identity (Sides and Citrin Reference Sides and Citrin2007), and norms and values (Sniderman, Hagendoorn, and Prior Reference Sniderman, Hagendoorn and Prior2004). Across these domains, a key finding is that citizens who are more anxious about immigration’s cultural impact are more opposed to immigration (Hainmueller and Hopkins Reference Hainmueller and Hopkins2014).
This study tests whether attitudes toward young immigrant men differ from attitudes toward other immigrants and also which of the three competing mechanisms described above (economic potential, security threat, and cultural threat) best explains attitudes toward young men. The evidence comes from a conjoint experiment conducted in Germany on attitudes toward immigrant groups with randomly varying shares of young men. To date, little research has been done on attitudes toward young immigrant men. Bansak, Hainmueller, and Hangartner (Reference Bansak, Hainmueller and Hangartner2016) include both age and gender in their conjoint experiments, and Fietkau and Hansen (Reference Fietkau and Hansen2018) vary the gender of two profiles of young immigrants in their survey experiments. Neither study focuses on attitudes toward young men specifically. Instead, they focus on immigrants’ economic and cultural characteristics more generally.
This study also contributes by focusing on groups of immigrants, in contrast to individuals. Although focusing on individuals is important (see Bansak, Hainmueller, and Hangartner Reference Bansak, Hainmueller and Hangartner2016; Turper et al. Reference Turper, Iyengar, Aarts and van Gerven2015), understanding how to best settle immigrants in communities in ways that maximize public acceptance and thereby foster integration success (Dancygier and Laitin Reference Dancygier and Laitin2014) also requires understanding attitudes toward the composition of groups of immigrants. After all, for logistic and other reasons, settlement is often done in groups. For example, 300 people were expected to move in within a few days of the opening of a refugee housing center in Berlin. Footnote 6
The main finding from the conjoint experiment is that respondents strongly prefer immigrant groups with few young men. Additionally, immigrant groups with many young men are significantly more likely to be perceived as security and cultural threats. There is no evidence that the number of young men in a group impacts evaluations of economic potential.
RESEARCH DESIGN
To study attitudes toward young immigrant men, I conducted a conjoint experiment in Germany. Two features make Germany an appealing case. First, Germany took in the largest absolute number of asylum seekers in 2015 and 2016, and on a per-capita basis, took in more asylum seekers than every other European country save Austria, Sweden, and Hungary. Footnote 7 Second, the share of young men among recent arrivals in Germany is typical of European countries. Specifically, 41% of asylum applicants in Germany were young men in 2016, whereas the median among all European countries was 42%. Footnote 8 These features enhance the generalizability of the inferences drawn from this study.
In the conjoint experiment, respondents evaluated groups of immigrants for settlement in their community. Each group comprised 60 immigrants, randomly varying in terms of three attributes: origin countries, education levels, and shares of young men. Groups were presented and evaluated in pairs, and each respondent evaluated four pairs of groups.
The key attribute is Young Men, which identifies the share—0%, 25%, 50%, 75%, or 100%—of the 60 immigrants who are men under the age of 25. Allowing the share of young men to vary so substantially reflects demographics: as stated above, approximately 41% of asylum seekers in 2016 were young men. It is therefore plausible that some communities have received groups of immigrants with overwhelming majorities of young men. Footnote 9
The other two attributes—country of origin and education level—are included to create groups that respondents will perceive as more realistic and to reduce the risk that a lack of relevant information about the groups drives the results. Footnote 10 First, the countries of origin are Afghanistan, Albania, Eritrea, Nigeria, Serbia, and Syria. The number of immigrants per country was shown as an attribute and was randomly selected as 0, 10, 20, 30, or 60, subject to the constraint that the total number of immigrants is always 60. These countries were chosen to increase the realism of the immigrant groups: at least 5,000 asylum seekers from each of these countries entered Germany in 2015 (Bundesamt für Migration und Flüchtlinge 2015). Second, Education presented the share of the immigrants with a university degree and was randomly selected to be 0%, 10%, 20%, or 30%. These values were also chosen to reflect the reality of asylum seekers in Germany: estimates suggest that about 20% attended university (Rich Reference Rich2016).
For each pair of groups, respondents were first asked, “If you had to choose between them, which of the two groups would you prefer be settled in your neighborhood?” Responses are coded into a binary variable, Settlement Preference, which is one if a group is selected and zero otherwise. Because respondents were forced to choose one of the two groups, this item allows me to separate support for a specific group from overall support for immigration. Footnote 11 Appendix A contains additional information on the conjoint design.
Three additional questions were asked about each group. These items are designed to test the three mechanisms proposed for explaining attitudes toward young immigrant men: Economic Potential, Security Threat, and Cultural Threat. Each item asked respondents to indicate their agreement (on a seven-point scale from “strongly disagree” to “strongly agree”) with a statement about the group. For Economic Potential, the statement was “Few members of Group X will find jobs in my neighborhood.” For Security Threat, the statement was “Group X would be a safety concern for my neighborhood.” Finally, the statement for Cultural Threat was “Group X would adapt well to German culture.” Following Bansak, Hainmueller, and Hangartner (Reference Bansak, Hainmueller and Hangartner2016), I recode each scale into a binary variable. For Economic Potential and Cultural Threat any of the “disagree” responses, which represent perceptions of high economic potential or cultural threat, are coded as one. Similarly, the “agree” responses for Security Threat are coded as one.
RESULTS
The experiment was administered through Respondi to a sample of 2,130 Germans in December 2016. Footnote 12 Respondents were drawn from their approximately 100,000 panelists to be representative of the 18- to 75-year-old German population on age, gender, state, and education. To correct for differences between the sample and population on these covariates, all estimates use post-stratification weights. Footnote 13 More details about the sampling procedure and construction of survey weights are available in Appendix A. The units of analysis are immigrant groups, and the quantity of interest from the experiment is the effect of Young Men on Settlement Preference. This quantity is interpreted as the change in a group’s probability of being preferred for settlement caused by a change in the share of young men. These effects are estimated by weighted least squares regression, where the group attributes Young Men and Education enter as a series of dummy variables. Standard errors are clustered by respondent to account for autocorrelation induced by the forced-choice outcome (see Hainmueller, Hopkins, and Yamamoto Reference Hainmueller, Hopkins and Yamamoto2014). Figure 1 presents the estimated effects of Young Men. The corresponding regression table is available in Appendix B.
FIGURE 1. The Effect of Young Men on Preferences Over Immigrant Groups
Note: Points are OLS Estimates with 95% Confidence Interval Bars Based on Clustered Standard Errors
Figure 1 reveals a negative relationship between a group’s share of young men and its probability of being preferred for settlement. The 0% young men group’s estimate of 0.087 indicates that they receive a premium of 8.67 percentage points over the baseline group (50% young men) in their likelihood of being preferred for settlement. For the 25% young men group, the premium is only 6.3 percentage points over the baseline. Footnote 14 In contrast, groups with more than 50% young men face a penalty to their selection probability. When young men make up three-fourths of the group, the penalty is 9.2 percentage points. For the group composed entirely of young men, the penalty is 18.1 percentage points.
These estimates translate into large differences in predicted preference rates. Averaging over the four categories of Education to calculate predictions, the group composed entirely of young men has a predicted preference rate of 34%. In contrast, the comparable rate for groups with no young men is 61%. The group with 50% young men—the closest in the experiment to the observed rate of 41% among recent asylum seekers in Germany—is predicted to be selected in approximately 52% of pairings.
Exploratory analyses reported in Appendix C show that these preferences are not meaningfully different for (1) male and female respondents, (2) young and old respondents, (3) high- and low-education respondents, and (4) East and West German respondents. Further, the results of a pilot study conducted among a sample of Mechanical Turk respondents in the United States show similar low levels of support for groups of young immigrant men. Footnote 15
WHAT EXPLAINS THESE ATTITUDES?
I now explore three explanations for the attitudes uncovered above: Economic Potential, Security Threat, and Cultural Threat. To do this, I fit a weighted OLS regression for each of these scales. As above, these models have standard errors clustered by respondent, and Young Men and Education each enter the regression as a series of dummy variables. Figure 2 presents the estimated effects for Young Men. The corresponding regression table is in Appendix B.
FIGURE 2. The Effect of Young Men on Group Evaluations
Note: Points are OLS Estimates with 95% Confidence Interval Bars Based on Clustered Standard Errors
First, I explore the possibility that respondents perceive young immigrant men as likely to find jobs and contribute to the German economy. Given that young men are in their prime earning years and are very likely to enter the work force, it is reasonable to expect respondents to see them as having high economic potential. The results in the first panel of Figure 2 do not support this. None of the estimated effects of Young Men on Economic Potential are significantly different from zero or from one another, and the estimates are close to zero.
To further explore the role of economic concerns, an exploratory analysis in Appendix D considers whether the results follow the logic of the labor market competition (LMC) theory: that fear of losing one’s job to an immigrant fuels anti-immigrant attitudes (see Hainmueller and Hopkins Reference Hainmueller and Hopkins2014). To do this, the effect of Young Men is estimated separately for respondents most likely to compete with young immigrant men—young male respondents with low education—and all other respondents. The results show no significant differences in the effect of Young Men between respondent groups, and hence, no evidence that LMC explains attitudes toward young immigrant men. This finding and the Economic Potential findings correspond to each other: young immigrant men are not seen as having higher economic potential than other immigrants and, therefore, are not a significant-enough competition to trigger the effects of labor market threat. This reiterates the point that for LMC to fuel anti-immigrant attitudes, immigrants must be perceived as serious economic threats (Dancygier and Donnelly Reference Dancygier and Donnelly2013).
The second panel of Figure 2 evaluates the effect of Young Men on Security Threat. The estimates show a positive effect: as the share of young men in an immigrant group rises, so too does the group’s probability of being rated as a high security threat. Relative to the 50% young men group, the 0% and 25% young men groups are 5.2 and 3.8 percentage points less likely to be perceived as a high security threat. In contrast, the 75% and 100% young men groups are 2.9 and 6.5 percentage points more likely to be seen as a security threat than the baseline. All four estimates significantly differ from the baseline. Further, the 75% and 100% estimates are also significantly different from each other, whereas the 0% and 25% estimates are not. As the baseline rate is about 42%, these estimates indicate that approximately half of all 100% young men groups are rated as a high security threat.
Finally, the third panel of Figure 2 assesses the impact of Young Men on perceptions of Cultural Threat. The estimates reveal that as the number of young men in a group increases, perceptions of cultural threat increase, i.e., evaluations of the group’s potential to adapt to German culture become more pessimistic. Specifically, a change in the share of young men from 50% to 0%, 25%, 75% or 100% changes the likelihood of a group being perceived as a cultural threat by −4.7, −2.9, 0.1, and 3.2 percentage points, respectively. All of these estimates save for 75% young men are significantly different from zero. Further, the 75% and 100% effects and 25% and 75% effects are significantly different from each other, whereas the 25% and 0% estimates are not.
These results suggest that non-economic factors were important for respondent evaluations of immigrant groups. The perception that young immigrant men pose a security threat combined with the low levels of support for their settlement echoes Lahav and Courtemanche’s (Reference Lahav and Courtemanche2012) finding that citizens across the ideological spectrum display exclusionary attitudes when their security is threatened. In this light, young immigrant men are likely an example of the “situational triggers” that Sniderman, Hagendoorn, and Prior (Reference Sniderman, Hagendoorn and Prior2004) describe: a group of immigrants who, due to the way they provoke security and cultural threats, have the potential to tap latent exclusionary attitudes among a broad swath of society. More broadly, in the on-going debates about how economic, socio-cultural, and security concerns shape attitudes toward immigrants (Hainmueller and Hopkins Reference Hainmueller and Hopkins2014; Sides and Citrin Reference Sides and Citrin2007), these results contribute an additional example of the importance of socio-cultural and security concerns in underpinning attitudes.
DISCUSSION
This study’s primary findings are that in comparison to other immigrants, young men receive far less support and that they are more likely to be seen as cultural and security threats. It is unlikely that these results are unique to the German context. The share of young men among recent asylum applicants in Germany is very close to the median of European countries. Hence, respondents were unlikely to encounter abnormal shares of young immigrant men relative to what other Europeans experienced. In contrast, support for young immigrant men may be even weaker in countries with higher unemployment rates or more expansive welfare states than Germany, where economic considerations weigh more heavily on attitudes. Nevertheless, because Germany has accepted a large number of asylum seekers in recent years, in absolute terms and relative to its population, host communities’ interactions with new arrivals may proceed differently in Germany than in other countries. Empirically assessing the generalizability of the results of this study is an important task for future research.
One limitation of this study is that the Young Men treatment manipulates not only the number of young men in a group, but also the group’s heterogeneity and respondents’ certainty about the group’s composition. Both of these features are worth considering as alternative explanations. First, results interpreted as preferences over the number of young men may instead represent preferences over group heterogeneity. Such preferences would result from the combination of the belief that heterogenous groups are less likely to socially interact with locals (and more likely establish an isolated community) and of a preference for immigrant communities that only have limited interactions with the host community. However, preferences for heterogenous groups are unlikely to completely explain the results, as this explanation cannot account for the Cultural Threat findings. Second, because the experimental design did not provide information about the non-young male component of the groups, attitudes toward young men cannot be disentangled from attitudes toward the rest of each group. For example, respondents who preferred a 0% group over a 25% group may not have done so because of their preference for few young men, but rather because of their preference for the immigrants they assume comprise the non-young male share of each group, e.g., younger women or older men. Future studies that elicit attitudes over immigrant groups with completely specified compositions will therefore be vital for fully identifying the relationships between immigrant group characteristics and public attitudes.
There is not a clear policy program for addressing low levels of public support for groups of young immigrant men. One option is to tailor refugee dispersion quotas, which prevent the concentration of young men within localities. Another avenue is to admit fewer young men, an approach taken by the Canadian government. Footnote 16 This approach may have severe humanitarian costs, however, as threats facing young men are often just as dire as threats facing others. An alternative is for policymakers to prioritize cultural integration and public safety, targeting the sources of low support for young immigrant men. Encouragingly, 72% of the respondents to a survey of local elected officials in Germany indicated that language training and education of refugees are very important goals (vhw-Bundesverband für Wohnen und Stadtentwicklung 2016).
This work shows that public attitudes toward young immigrant men differ significantly from attitudes toward other immigrants and that perceptions of young men as cultural and security threats underpin these differences. Given the large number of young men who have entered Europe as asylum seekers in recent years, these results suggest that finding public support for the settlement of asylum seekers may be more challenging for European governments than originally thought.
SUPPLEMENTARY MATERIAL
To view supplementary material for this article, please visit https://doi.org/10.1017/S0003055418000710.
Replication materials can be found on Dataverse at: https://doi.org/10.7910/DVN/SQH6IF.
Comments
No Comments have been published for this article.