Familiar strangers? : a sociometric investigation of gender cleavage during middle childhood

The term gender cleavage refers to imbalances in same- and cross-gender peer interactions and sociometric peer evaluations, favouring a same-gender (in-group) imbalance. The mechanisms underlying gender cleavage during childhood and adolescence are unclear, with several competing models encapsulating group-level and individual-level processes. In comparison to observational studies, sociometric investigations to date have contributed less substantially to an understanding of the roles of gender, age and maturational factors in gender cleavage, and to theoretical accounts of gender cleavage. On these bases, this thesis examined interlocking theoretical and methodological issues, focusing on middle childhood. Studies 1, 2 and 3 examined method-related and naturally-occurring variations in gender cleavage effects, using sociometric responses &Om a group preference record and three-choice nomination tasks collected by individual interviews with 308 boys and girls in Grades 2 to 6. Results revealed significant gender cleavage effects in both positive and negative peer evaluations from each sociometric method. Nevertheless, the cleavage effect was much less marked in negative peer evaluations, providing evidence that same-gender preference is not necessarily accompanied by cross-gender disliking. In Study 1, the strength of gender cleavage expressed in positive peer evaluations was similar in sociometric measures that encapsulated differing levels of intimacy in peer relationships. Gender cleavage expressed through negative peer evaluations was more pronounced for measures manifesting less intimate peer relationships. In Study 2, sociometric criteria reflecting child-regulated playground interactions produced more extreme expressions of gender cleavage than criteria reflecting adult-regulated situations such as classrooms. In Study 3, significant and strong gender cleavage effects were found in positive peer evaluations for male and female evaluators, and across all grade levels regardless of the sociometric method used. Negative peer evaluations exhibited more complex variations in gender cleavage effects, according to age, gender group and sociometric method. In-group preference evident in same-gender roster-based evaluations was more pronounced in girls than in boys. Out-group (cross-gender) evaluations became more positive and less negative with age in both gender groups. Study 4 investigated the relative applicability of three contrasting theories as models of gender cleavage during middle childhood. Gender stereotyping of peers' behaviour was a stronger predictor of gender cleavage than was behavioural compatibility. This finding confirmed the greater relevance of social-cognitive models than those based on behavioural dissimilarity as explanations for the maintenance of gender cleavage during middle childhood. However, gender cleavage was more strongly and consistently predicted by class-based gender segregation than by either gender stereotyping or behavioural compatibility. Hence, group processes overall appear more important than individual-level processes in the maintenance of gender cleavage during middle childhood. Gender cleavage was found to vary significantly ac-cording to both substantive and method-related factors, and consequently should not be viewed as a static aspect of peer relationships during middle childhood. Instead, gender cleavage should be seen as a dynamic and ever-changing social-developmental phenomenon, that can potentially be modified in school settings by interventions at a group level that involve everyday school practices and policies.