Setting-related characteristics of socially withdrawn behavior : an observational analysis

The summary label socially withdrawn refers to a low incidence of positive social interaction between a child and his/her peers and implies that such a child will behave consistently across different situations. The purpose of this observational study was to empirically determine the degree to which individual, socially withdrawn children would exhibit correlated percentages of positive social interaction with peers in two separate play settings.

A preschool-age boy and girl from a Headstart classroom identified by their teacher as socially withdrawn were selected as the primary subjects. For comparison purposes, another boy and girl were selected from the same classroom who were identified by the teacher as socially active. These four subjects were the targets of direct, repeated observation of their naturally occurring behaviors during indoor and outdoor free play. Also, a picture sociometric assessment procedure was administered to each child in the classroom at the beginning and end of the study to index subjects' social standing. Data were analyzed graphically and statistically.

Generally, withdrawn subjects exhibited extremely low percentages of positive peer interaction while socially active subjects displayed relatively higher percentages in both play settings. However, considerable between session variability was evident in subjects' peer interaction percentages. At the level of individual behavior categories only the most withdrawn child exhibited correlated (consistent) percentages of peer interaction across play settings, although in terms of behavior patterns, both withdrawn and one socially active subject displayed across-setting correlations. Sociometric measures were not sufficiently stable for further analysis.

The results of this study were equivocal in so far as establishing the across-setting consistency of withdrawn and socially active children's social behaviors. Rather, they more closely supported Bern and Allen's (1974) "differential consistency" hypothesis, i.e., that only some people are behaviorally consistent some time across some situations. That only one subject displayed across-setting correlations in her peer interaction tentatively supports the view that only some people are behaviorally consistent. Future research should attempt to replicate this finding with a larger sample and to search for situations and conditions functionally related to across-setting behavioral consistency.