H1. Muslim immigrants are more likely to experience EH than non-Muslim immigrants.
Results indicate that immigrants who identify as Muslim are more likely to experience EH compared to non-Muslim immigrants. Specifically, Muslims exhibit an estimated 7.4 percentage points (p.p.) higher probability of experiencing EH We also observe this ‘Muslim penalty’ when leaving out origin groups (i.e. 6.7 p.p., ..) and it remains significant when we include immigrants’ proficiency in the host-country’s language (L2) and schooling (.., M2), as well as labour market failure (.., M3). When we dissect the sample by generation, we observe that both 1G/1.5G and 2G Muslims experience a higher likelihood of EH. Notably, the Muslim penalty is significantly larger in 2G (+15.6 p.p.) compared to 1G/1.5G (+5.8 p.p.) (P < 0.001). These results lend support to H1.
H2. Muslim immigrants wearing traditional or religious clothing are more likely to experience EH than Muslim immigrants not wearing such clothing.
Findings reveal that wearing religious clothing is associated with 8.1 p.p. higher probability of EH (P < 0.01). This association holds true both for 1G/1.5G (Table 5, M2) and 2G Muslim populations (.., M3), with no statistical significant difference between the two generations (P = 0.4). As a second test of H2, we examined the relationship between wearing a headscarf and/or niqab and EH (.., M4). The results demonstrate that Muslim women who wear such clothing have a 9.6 p.p. higher probability of experiencing EH. This is a substantively large association, and also highly statistically significant (P < 0.01). Taken together, these two tests largely support H2.
H3. Sub-Saharan immigrants are more likely to experience EH than immigrants from other countries.
We also find origin group differences in EH. In line with H3, immigrants from Sub-Saharan Africa experience more EH than those from Asia (+10.1 p.p.). However, we do not find that immigrants from Sub-Saharan Africa are significantly more likely to experience EH than those from Northern Africa or Turkey. Furthermore, the results suggest that the impact of origin groups on EH is more pronounced in 2G. Formal tests for interaction effects reveal that the gap between those from Northern Africa and those from other origins is significantly stronger in 2G compared to 1G/1.5G (P < 0.001).
H4. 2G immigrants are more likely to experience EH than 1G immigrants.
H5. Immigrants who are more proficient in L2 are more likely to experience EH.
H6. Immigrants’ schooling increases the likelihood to experience EH.
Our prediction that 2G immigrants are more likely to experience EH than 1G/1.5G immigrants (H4) is supported by the findings. 2G immigrants exhibit a 9.3 p.p. higher probability of experiencing EH compared to 1G (.., M1, P < 0.01). We also note that those born abroad but who arrived before age 19 (1.5G) are more likely to experience EH (+3.4 p.p.) than those who were abroad and arrived at an older age in the host country (1G), confirming H4.
One of the reasons for expecting successive generations to experience more EH is their better proficiency in the host-country’s language (L2) and higher levels of education. These factors not only increase opportunities for interethnic contact (and thus EH) but also raise awareness of EH. Consistent with these expectations, we observe a positive association between L2 skills
and EH. Those with good L2 speaking skills have a 5 p.p. higher probability of EH (Table 4, M2), supporting H5. Also in line with our expectations (H6), we find a positive correlation between years of schooling and EH. An additional ten years of schooling is associated with a 6 p.p. higher probability of EH (P < 0.01).
This relationship between schooling and EH is particularly evident in 2G (P < 0.001), while it is not statistically significant in 1G/1.5G. Moreover, the difference between generations regarding the impact of schooling on EH is statistically significant (P < 0.01).
Overall, our findings suggest that both L2 skills and schooling are positively associated with EH, and these factors partially contribute to the higher likelihood of EH among 2G immigrants. Generational differences in EH diminish when we account for L2 skills and schooling (.. 4, M1 vs M2). In other words, the increased likelihood of EH among 1.5G and especially 2G immigrants is, in part, due to successive generations being more proficient in L2 and having higher levels of schooling.
H7. Immigrants experiencing greater labour market failure are more likely to experience EH.
Our final hypothesis emphasized the role of labour market failure (H7). Contrary to this hypothesis, we do not find a significant association between unemployment and EH (Table 4, M3). Having a paid job is not correlated with EH in both 1G/1.5G and 2G.
Even when we exclude perceived economic hardship (results not shown), no relationship emerges. Additionally, we assessed two other ‘objective’ measures of labour market failure. First, we considered a measure of what individuals cannot afford, even if they desire those items. We combined six yes/no indicators of the inability to afford specific expenses into a count variable (range 0–6). Second, we examined household income. However, neither measure is related to EH.
Despite the absence of associations with these factual indicators of labour market failure, we find that more subjective measures of labour market failure are predictive of EH: immigrants who perceive greater economic hardship are statistically significantly more likely to experience EH (P < 0.01). Immigrants who perceive great difficulty in making ends meet have an (5 × 0.017 × 100=) 8.5 p.p. higher probability of EH than those who find it very easy to make ends meet.
Additional evidence supporting H7 arises when examining perceived educational mismatch. Results reveal that immigrants who believe they are overeducated for their current job are more likely to report EH (.., M1). Those who perceive themselves as overqualified for their job exhibit an 8.1 p.p. higher probability of experiencing EH (P < 0.001). These findings hold when perceived economic hardship is added to the model (Table 6, M2), and perceived educational mismatch plays a role in both 1G/1.5G and 2G (.., M3 + M4, with no statistically significant difference).
