P208 (02 473P208)
Social Networks and Intergroup Relations: New Questions and Challenges
Form of presentation:
Be the Smart Guy: The Role of Gender and Ethnicity in Ability Attribution Processes in the Classroom
Dorottya Kisfalusi | Institute for Sociology Centre for Social Sciences, Hungarian Academy of Sciences | Hungary
In this study we examine ability attribution processes among male and female Roma and non-Roma primary school students. Girls and members of certain minority groups are often evaluated as less competent than boys and majority students even with the same level of performance. Double standards theory (Foschi, 2000) provides an explanation for why this might happen: people use different standards for making inferences about others’ competence based on social status. Because of status generalization processes low status individuals such as women and members of minorities might be judged by a stricter standard than high status individuals. Social identity processes, however, might also play a role in ability attributions. Social identity theory (Tajfel, 1982; Tajfel & Turner, 1979) suggests that social groups try to establish a positive distinctiveness from other groups, and therefore, people evaluate in-group members more positively, than out-group members. We posit that students’ ability attributions are interdependent, therefore, they need to be addressed by proper social network methods.
We analyse cross-sectional dyadic peer nomination data from 21 primary school classes in Hungary (392 students from 16 schools, mean age=13 years) using exponential random graph models (ERGMs). Our findings are mostly in line with the predictions of social identity theory: controlling for grades, students are more likely to nominate their in-group peers as clever compared to classmates from the out-group, in terms of both gender and ethnicity. One exception has been found: boys are similarly likely to nominate both boys and girls as clever. In line with double standards theory we have found that non-Roma students are less likely to consider those peers as competent whom they perceive as Roma compared to those whom they perceive as non-Roma. However, similar association has not been found with regard to gender. We have also found that Roma students are less likely to consider those peers as clever whom they perceive as Roma, but who identify themselves as non-Roma, than those Roma peers who also identify with the Roma group. This phenomenon might also be driven by social identity processes: Roma students try to distance themselves from those peers who are perceived as Roma but who declare themselves having different ethnic identification.
The interplay between ethnic composition and out-group attitudes as possible explanation for ethnic homophily in schoolchildren's friendship networks
Mark Wittek | University of Cologne | Germany
A growing strand of literature applies network analysis to investigate ethnic homophily in social networks. As one possible cause of ethnic homophily scholars studied the relationship between the ethnic composition of classrooms and ethnically homogenous network structures in schoolchildren's friendship networks. This paper tries to contribute to this literature by analyzing the interaction between out-group attitudes and compositional aspects of the school context that could lead to ethnically segregated friendship networks. I will analyze whether individual and parental outgroup attitudes on the micro level are associated with the emergence of ethnically segregated structures in friendship networks on the classroom level. Furthermore, I will investigate whether the ethnic composition of classrooms moderates this link between the individual potential for ethnic homophily and actual homophilous networking behavior. Building on intergroup conflict theory as well as insights from social psychology I will argue that ethnic classroom composition can amplify the association between outgroup attitudes and ethnically homophilous friendship selection.
The first wave of the CILS4EU project (18716 students; 11700 parents; 958 classes; 4 European countries) will be used to carry out the empirical analysis. My analytical strategy is inspired by previous work concerned with ethnic homophily in schoolchildren's friendship networks (Smith et al. 2016) and follows a three step approach:
1. Estimate exponential random graph models for each school class to measure ethnic homophily above the opportunity structure and network endogenous processes. The ethnic homophily parameter is estimated as change statistic, which is interpretable as the increase in the conditional log-odds of a network by adding one same-ethnic friendship nomination given all other local network structures included in the model specification (like the overall number of edges, reciprocated ties, etc.)
2. Aggregate the results using an univariate meta-analysis.
3. Use metaregressions (with ERGM coefficients for ethnic homophily as the dependent variable) to investigate between-context variation in ethnic homophily.
The first aim of the study is to test if inter-ethnic friendship nominations send by students with (parental or individual) negative out-group attitudes are observed systematically less often than inter-ethnic friendship nominations send by students with positive out-group attitudes. An ERGM specification will be proposed that allows to differentiate between these two types of nominations for native and immigrant students separately. I will also discuss how this approach can be extended to study conditional cross-group and multiple cross-group nominations within the ERGM framework more generally.
In a second step, a set of nested metaregressions will be applied to estimate how well variation in ethnic homophily on context level is explained by the classroom-specific mean level of negative out-group attitudes, the ethnic composition of the classrooms and the interaction of these two factors.
The theory section will discuss the mechanisms that could be responsible for the association between outgroup attitudes, classroom composition and ethnic homophily on the network-level. By testing these theoretical considerations empirically the analysis will help to evaluate whether the interplay of outgroup attitudes and ethnic classroom composition could provide a possible explanation for ethnic homophily in schoolchildren's friendship networks.
The simultaneous development of ethnic segregation and health disparities among primary school children: a dynamic social network approach
Jochem Tolsma | Radboud University of Nijmegen | Netherlands
Ethnic health disparities are already observable among young children. Health-based selection of friends may therefore have as unintended by-effect ethnically segregated friendship networks. At the same time, friends may influence each other’s health behavior. Ethnic segregation in friendship networks may thus also reinforce ethnic health disparities over time. In the present contribution, we aim to investigate to what extent ethnically segregated friendship networks overlap with health segregated friendship networks in ethnically diverse Dutch primary school classes. We hypothesize that this overlap grows over time (age range 9-12) due to both selection and influence processes. Because ethnic identities will become more salient when children reach adolescence, the importance of health-based selection of friends (compared to ethnicity-based selection of friends) is expected to decline over time. Furthermore, we expect that influence process with respect to health behavior are stronger within ethnic groups than between ethnic groups. We will test our expectations based on data retrieved from the ERC funded ‘MyMovez’ project (4 waves; multiple ethnically diverse classes; for more information see: http://mymovez.socsci.ru.nl/). As our indicators of health and health behavior we will use BMI, physical activity (measured through a wearable lab) and eating habits. Friendships are measured by a nomination procedure.
Adolescents' Ethnic Self-Identification and the Formation of Interethnic Friendships
Adolescents in many Western countries attend multiethnic schools. This increased opportunity to engage in interethnic contact has long been recognized as being beneficial for the development of friendships between native- and immigrant-origin youth. Fostering interethnic friendships in turn is desirable because they can reduce prejudice and improve intergroup attitudes. However, the opportunity to engage in interethnic contact itself hardly guarantees the formation of interethnic friendships. In fact, even in ethnically mixed schools friendships tend to be more often formed between peers of the same ethnic group.
A crucial limitation of most existing studies is that they capture ethnicity by “objective” demographic characteristics such as the country of birth of students or their parents. But a rising number of immigrant-origin youth was born and raised in the host country. For these students, it is questionable to what extend they identify themselves with the country of origin of their parents, with the host country, or with both of them.
Against this background of contested ethnic identities, we ask how ethnic self-categorization affects friendship formation among native- and immigrant-origin peers. For immigrant-origin youth, we distinguish three ways in which they can categorize themselves: as members of the host country (host country identifiers), of their country of origin (heritage country identifiers), or of both countries (dual identifiers). In short, we suggest that these identifications affect both how immigrant-origin youth’ choose their friends and how they are perceived, and thus chosen as friends, by others. Based on various theoretical approaches, we derive respective hypotheses regarding youths’ friendship preferences.
We empirically test our hypotheses using data from the project “Friendship and Identity in School“ (FIS). FIS is a longitudinal German study of friendship networks of more than 2,000 students in schools with high shares of immigrant students. Our analysis relies on three waves of data, which are separated by nine months each. We use stochastic actor-oriented models (SAOM) for network dynamics, which can be regarded as agent-based simulation models that allow to examine how individual actor's preferences interact to create the network dynamics we observe empirically while controlling for relational mechanisms and the opportunity structure.
We find that native German youth tend to befriend immigrant-origin peers with host country or dual rather than heritage country identification. Immigrant-origin youth with heritage country identification, however, were more likely to befriend native peers than immigrant-origin peers from other ethnic groups. Heritage country identification thus is disadvantageous in two ways, as it not only makes immigrant-origin youth less attractive as friends to native peers, but also reduces the preference for befriending immigrant-origin peers from other ethnic groups.
In sum, our findings speak against the pessimistic notion that native-origin adolescents may reject immigrant-origin adolescents as friends irrespective of their ethnic self-identification. Likewise, immigrant-origin adolescents do not always prefer immigrant-origin adolescents over native ones. Instead, our study provides evidence that how immigrant-origin youth identify themselves affects not only to what extent they seek intra- or intergroup friendships, but also how attractive they are as friends to both immigrant-origin and native peers.