Assessment of social competence in preschoolers: Teachers versus peers.

Abstract

The predictive validity and the reliability of a teacher-based ranking method of assessing popularity were compared with those of the traditional picture sociometric approach. In each of two studies, both popularity measures were administered in several day-care classrooms. Behavioral observations of social interaction and teacher ratings of social competence were also collected. Very hi^h reliability estimates were obtained for the teacher popularity measure, with lower estimates obtained for the picture sociometric measure^ In both studies, when validated relative to the social competence ratings and behavioral observations, the teacher popularity measure was a more effective predictor. Stepwise multiple regression analyses were computed in which social competence measures were regressed on the popularity measures, with age treated as a covariate. With age effects partialed out, teacher popularity entered the equation prior to sociometric popularity and was a significant predictor of social competence. The sociometric popularity measure did not significantly improve this prediction. These findings support the use of teacher rankings of popularity as a measure of social competence in the preschool. Peer group social competence in preschoolers is currently a topic of considerable developmental and clinical importance. The methodology by which to adequately assess this construct is not yet clearly established. A common approach is to use indices of peer group popularity based on the sociometric technique (Oden & Asher, 1977). Until recently, the most frequently used technique for obtaining such popularity data with preschoolers was the picture sociometric method developed by McCandless and Marshall (1957). Children are shown photographs of their classmates and are asked in a variety of ways to indicate the pictures of their preferred playmates. In addition, rejection scores may be derived by asking the child to indicate those children with whom they do not like to play. Two factors are regularly cited in support of the sociometric approach to the assess