An observational instrument was developed to measure social competence in kindergarten classrooms. Social competence was specified in terms of four broad behavioral categories: obedience, cooperation, interaction, and autonomy. Each category was further particularized in terms of observable behaviors. Initial 5-minute observations using the instrument were conducted in kindergarten classes; a coding sheet listing broad and specific behavioral categories was employed. Revisions of the original coding sheet included the addition of the category "aggression" and related behavioral subcategories, reduction of the time unit to 2 minutes, and provision of a way to indicate behavioral appropriateness or inappropriateness. The behavior of subjects engaged in different activities was to be observed on six occasions and coded on the new sheet. Tests of interrater reliability and a pilot study were conducted. The latter involved 52 children from three kindergarten classes in a lower-middle-class Sydney suburb. Data, collected at the beginning and end of the school year, were reduced so that scores for each subject could be computed on each competency: obedience, cooperation, positive interaction, autonomy, and negative interaction. Discrete scores were then converted to continuous dichotomies. Results indicated a statistically significant change in autonomy over the year. Statistically nonsignificant changes in the other four competencies were also evident. (The scale and guidelines for using the instrument are appended.) (RH)