The study established the measurement invariance of a translated version of the Social Role Questionnaire to prove its usefulness as a measurement for gender-role attitudes of German adolescents and German-speaking adolescent refugees from Syria, Afghanistan, or Iraq (1). As scalar measurement invariance for both samples was achieved, a comparison was made between GRAs of adolescent refugees and German adolescents (2). Results show that only gender-linked attitudes were significantly higher for the adolescent refugees, and no significant differences in gender-transcendent GRAs between both groups were found, which only partly supports Hypothesis 2. Subsequently, the relationships of GRAs and the affective well-being was investigated (3). While neither GRAs predicted any facet of affective well-being, gender and age were found to predict some of them. In a final step, these relationships were examined regarding possible differences between the relationships of both groups (4). Such differences were not found.
Measuring gender role attitudes in adolescents with refugee experience and their german peers
To answer questions about differences and similarities in GRAs of adolescents with different cultural backgrounds, such as refugees from Middle Eastern countries and German natives, it was necessary to examine the usefulness of the SRQ (Baber and Tucker 2006) by demonstrating its measurement invariance. In line with Hypothesis 1, the results show that scalar measurement invariance can be assumed, which makes statements about differences and relationships possible. Therefore, group-differences in the latent means of the gender-linked and gender-transcendent scales were examined.
No significant differences in gender-transcendent attitudes between the two groups were found, but refugee adolescents reported more traditional gender-linked attitudes than their German peers. In terms of gender-transcendent attitudes, adolescents—regardless of their background—agreed on the extent of importance that should be attached to gender as a category for distinguishing social roles. Thus, the results only partly confirmed Hypothesis 2. Naz et al. (
2021) found the same pattern of differences between British and Pakistani young adults. They link their findings to a more traditional upbringing of the Pakistani participants based on the values and role expectations of their culture of origin in case of gender-linked attitudes. However, they also indicated similar acceptance of gender egalitarianism, as measured with the gender-transcendent scale, which Naz et al. (
2021) attributed to the contact with a more individualistic culture. These findings can be related to the findings of Arends-Tóth et al. (
2009), who found that attitudes toward gender norms and behaviors show varying degrees of agreement depending on the domains they address, such as marriage, family or employment, which are addressed in the scales of the SRQ (Baber and Tucker 2006). Gender role attitudes are often passed on from parents and their more traditional attitudes prevail to some extent (Kretschmer
2018). However, they are not static but dynamic and change depending on the situation and over the lifespan (Arends-Tóth et al., 2009; Fan and Marini
2000). Even though the adolescent refugees show more traditional gender-linked attitudes, it can be argued that both groups of adolescents are similarly open to disregarding gender as a useful category, which allows for a certain flexibility. Studies of GRAs of immigrants living in Germany (Idema and Phalet
2007; Kretschmer
2018), see also Phinney and Flores (
2002) for similar results in Hispanic immigrants in the US, found an association between less traditional GRAs and integration into the host society. A common proxy- measure of integration is use of language (Doucerain et al.
2016). As the current study relied solely on refugee adolescents who felt comfortable enough to fill out the SRQ in German, the sample analyzed in this study could be well integrated. This could reflect in their similar attitudes towards gender-transcendent gender-roles. Cross-cultural exchange might therefore foster less traditional attitudes in adolescent refugees in Germany. It is, however, not possible to rule out a possible influence of social desirability on the answers given to the SRQ. Even though Baber and Tucker (
2006) found no association between the SRQ and a social desirability scale they tested a very different sample. Further studies should therefore investigate the effect of social desirability, as well as the effect and direction of the relationship between integration and less traditional gender role attitudes based on longitudinal studies.
Hypotheses 3 and 4 explored possible implications of GRAs for the life of adolescents by examining associations with facets of affective well-being, such as well-being, sadness, anxiety, and anger. Although previous studies (Baird et al.
2019; Fragoso and Kashubeck
2000; Jaehn et al.
2020; King et al.
2019; Lengua and Stormshak
2000) suggested that gender differences in these outcomes may be due to differences in GRAs, the present results could not confirm this assumption. None of the outcome variables were predicted by gender-linked or gender-transcendent attitudes. To explain the lack of associations between GRAs and affective well-being, a look at studies with adult populations might be helpful.
Various of those studies (e.g., Soltanpanah et al.
2018; Sweeting et al.
2014; van de Vijver
2007) show that it is not the GRAs themselves, but a possible discrepancy between the gender role lived and the GRAs represented, as well as a discrepancy between one’s own GRAs and the GRAs of the environment, which influence the well-being or mental health of men and women. These studies suggest that the match/mismatch with the general GRAs of the environment or the lived gender role are most relevant. Further studies should include measurements of such discrepancies to investigate a possible impact. Additionally, refugees themselves do not consider the differences in GRAs between themselves and German natives to be significant for their everyday lives (Forschungsbereich beim Sachverständigenrat deutscher Stiftungen für Integration und Migration [SVR-Forschungsbereich],
2019). This could further weaken the link between the GRAs and affective well-being and might offer an additional explanation for the absence of a moderation by group membership, as adolescent refugees might also not consider differences in GRAs to be important. Additionally, the internal consistencies of the affective well-being scales were partly low, restricting the meaningfulness of the findings. Nonetheless, with regard to gender differences in affective well-being, the results show similar findings to previous studies (Ravens-Sieberer et al.
2009; Torsheim et al.
2006). Girls reported more sadness, anxiety, or anger than boys. Anxiety and sadness were also positively predicted by age, which is also in line with findings by Ravens-Sieberer et al. (
2009). Surprisingly, when looking at well-being, no gender difference was found. Although girls seem to face more negative mental health outcomes, their well-being was equal to that of boys. Further research is needed to understand gender differences in affective well-being emerging during puberty and their possible relationships with GRAs. In (cross-cultural) research, further insights might be gained by including additional background variables such as socioeconomic background, religious affiliation, or GRAs of the immediate and broader environment (Kågesten et al.
2016).
Strengths and limitations
The investigation of measurement invariance of the German translation of the SRQ as a prerequisite for cross-cultural comparisons is a first important step to gain evidence-based insights into the differences and similarities of GRAs of adolescents living in Germany. Nonetheless, further differentiation of the samples, e.g., into sub-groups by gender, different countries of origin or ethnicities would have further improved the present study.
Due to the small number of participants, a further diversification was not possible but should be a priority in upcoming studies as the population of refugee adolescents in Germany is very heterogenous. This is especially true when considering that the final sample of refugee adolescents consisted of only those who felt comfortable to fill out the questionnaire in German. Their language skills might reflect a longer stay in Germany, arriving at a younger age or a general ability or will to orient toward the German culture. They therefore might be more similar to the German sample as expected in their understanding of the SRQ and GRA. The relatively small sample size was deemed appropriate for the present analyses by rule of thumb, but a larger sample would additionally allow further investigations of more complex relationships of gender role attitudes of adolescents in Germany, e.g., with gender, cultural background, country of origin, or interactions of these variables. Nonetheless, since studies of measurement invariance are more concerned with fitting the overall model and less with individual parameters, the power was still considered appropriate, and the sample of adolescent refugees is a unique feature of the study. Furthermore, the small amount of completed SRQs in Arabic, Farsi or any of the Kurdish languages made it impossible to test the MI of these translations from the German version. Here the linguistic differences are assumed to add to the measurement variance and testing MI would therefore be of great importance (Boer et al.
2018). So far, no comparisons between the results of these versions can be made.
Because the German translation of the SRQ used in this study was administered to an overall younger sample than it was constructed for, the investigation provided further insight into construct validity in addition to its reliability. The good model fit can be taken as a strong sign of construct validity (good model fit, acceptable values of internal consistency), but there is still little information on other forms of validity. Nonetheless, the findings of the current study suggest that the German version of the SRQ is a reliable and valid measurement for different groups of adolescents with different cultural backgrounds. It is therefore helpful in investigating GRAs in general (e.g., possible differences between those fleeing from other countries to Germany and its native population) and associations with other variables such as religious beliefs or career choices. Furthermore, cross-cultural research could also use the SRQ to examine whether and how GRAs change over time and if they converge across these groups during their development within the same country.