Cluster Analysis in Sociometric Research: A Pattern-Oriented Approach to Identifying Temporally Stable Peer Status Groups of Girls
Abstract
A modern clustering technique was applied to age-10 and age-13 sociometric data with the purpose of identifying longitudinally stable peer status clusters. The study included 445 girls from a Swedish longitudinal study. The identified temporally stable clusters of rejected, popular, and average girls were essentially larger than corresponding groups identified according to the frequently used Coie, Dodge, and Coppotelli (CDC) classification method. It was further shown that all stably rejected and all (but one) stably popular CDC girls were members in corresponding clusters. Although no stable clusters of neglected or controversial girls were identified, some temporally volatile clusters were found, which were interpreted as largely accidental groupings. Rejected and popular clusters showed similar associations to some childhood adjustment correlates as the CDC groups and also have been documented by earlier sociometric research. The conclusion was that the clustering technique seems to be a promising classification method in longitudinal sociometric research.