Introduction

Friendships serve important functions for both children and adults. Across the life-span, the dynamics of friendship are complex (Hartup and Stevens, 1997; 1999) and, as yet, poorly understood. Two observations are well established: an absence of friends characterizes those who seek clinical intervention for personal problems, and good adjustment is associated with having a social network of friends. It is less certain whether friends directly contribute to our well-being, or whether well-adjusted persons find it easier to make and keep friends. Regardless, friendships provide children and adolescents with experiences that facilitate growth of social competence, providing functions critical to development of self-concept and serving as resources of emotional support, buffering against life stress (Price, 1996). Indeed, recent evidence suggests that only for children without mutual best friends does victimization predict bad behavioral outcomes; peer friendships appear to be crucial in preventing escalating cycles of peer abuse (Hodges, Malone and Perry, 1997; Hodges, Boivin, Vitaro and Bukowski, 1999). Friendship relations are implicated in children's social and emotional growth, influencing children's school adaptations, adjustment to major life transitions, and the acquisition of social skills and values (Bukowski, Newcomb, and Hartup, 1996). The developmental significance of friendships includes who one's friends are, as well as the stability and quality of those friendship relationships across time. And, because the identity of one's friends accounts for more outcome variance than whether one has friends (Hartup and Stevens, 1997), it is important to study the qualitative features of friendship selection.