Clique Formation among Primary and Secondary School Children

Abstract

Studies of friendship among primary and secondary school children have been mainly carried out in streamed schools and have shown, essentially, that children tend to make friends with others of similar attitudes, attainments, and backgrounds to themselves. The literature on this subject is massive and I shall restrict my attention to a few of the more telling studies. One of the earliest studies of children's groups in primary schools described how boys in a class of ten-year-olds formed two stable, peacefully co-existing cliques based, apparently, on the propinquity of their homes, religious denomination, and IQ; but not social class.1 Blyth did not investigate the pupils' attitudes and behaviour, nor did he say which groups were preferred by the teacher; in fact, the teachers appeared to have little awareness of the existence of the groups. In a later study which compared friendship in streamed and unstreamed schools it was found that girls in unstreamed classes tended to choose friends with a similar IQ to their own, but the tendency was less true for boys.2 In streamed classes the IQ range was, predictably, too narrow to allow children to form friendships with children greatly dissimilar in intelligence to themselves. Similarly, in streamed schools, the narrow social class range coupled with the social cleavage between the 'A' and the 'B' streams, meant that in these schools friendships were almost always between children of like social class. In the unstreamed classes, however, there was a slight tendency for children to group along class lines. An American study discovered no appreciable differences in the selection of friends in streamed and unstreamed schools.3 In both types of school there was a tendency for children to select friends of similar intelligence. In the most recent study it was shown that primary school children seemed, on the whole, to choose each other as friends when they were of similar ability and social class.4 The data in this N.F.E.R. survey, however, were not