Effect of gender on ethnic homophily of Jewish and Arab pupils in Israel
Eyal Erlich, Yossi Shavit, Uri Shwed
The ethnic segregation of friendship networks is a persistent fact of social life in western societies and especially in conflict-ridden societies, such as Israel. Even in an ethnically diverse field, that provides opportunities for inter-ethnic friendships, social networks tend to be ethnically segregated. In Israel, Jews and Arabs are separated in most areas of life. Nevertheless, in recent years a growing number attend mixed Arab-Jewish schools.
The individual preference for same-ethnic friendships contributes to the persistent segregation of pupils’ friendship networks. In this paper, we focus on the effect of Arabs and Jewish pupils' gender as it affects the social tendency to form homophilic ethnic friendships. Gender is a key factor of peer relationships throughout children’s development. Gender differences appear at an early age, with girls tending to have more diverse social networks than boys. However, the tendency for gender homophily varies with age. In the early years of adolescence, gender boundaries begin to blur. The number of social relationships between boys and girls increases during the children's adolescence, towards the formation of romantic relationships. Romantic relationships are characterized by stricter ethnic homophily that reflects the desire of groups to empower their ethnic identity. Yet, ethnic homophily varies by gender. In conservative societies, and particularly in Arab society in Israel, women play a key role in reproduction and maintenance for the collective identity. They are often central in ethnic and national transformation, not only in the transmission of its cultural artefacts, but also as markers of the boundaries of collectivities. Women's commitments to ethnic collectivity, among these societies, may bring them to be more restricted than boys in choosing others to associate. Hence, we expect ethnic homophily to be higher among Arab girls than among all other sub-groups.
The study consist 93 classes of 4th to 11th graders, which included at least two Arab students and two Jewish students. The students were asked to complete a questionnaire that included questions about their relationship with each of the other students in their homeroom class. Having full classroom networks, we fit exponential random graph models. These models allow to assess ethnic homophily while controlling for background, context, and network properties; they also allow to include estimators of other boundary dimensions at the student level, i.e., to assess whether similarity in identity predicts friendship.
When assessing the tendency for ethnic homophily, Jewish girls, Jewish boys and Arab boys prefer to engage with members of same ethnic group. In primary school their tendency for ethnic homophily is stronger than in middle/secondary school. Arab girls, however, tend to have homophilic relationships in middle/secondary school but not in primary school. 
← Schedule