School development in an inner city: An analysis of factors selected from Comer's program using latent variable structural equations modeling

Date of Completion

January 1992


Education, Administration|Education, General|Education, Elementary|Education, Educational Psychology




In this study, a structural equations model representing the causal relationships among the variables researched in James Comer's School Development Program (1981, 1986, 1988) was proposed, and a modified subset of the model was estimated and tested using structural equations modeling (Joreskog & Sorbom, 1989). Developed through synthesis of the writings of Comer and his colleagues, this model postulates that school climate has a direct influence on classroom climate, social behavior, student self-concept, and student achievement. The modified model suggests that in addition to the direct paths, the influence of school climate on student achievement and social behavior is mediated by classroom climate and self-concept. Classroom climate directly affects self-concept, and behavior directly affects achievement. Data used were obtained through the Yale Child Study Center. The responses of 127 students from eleven inner-city, elementary schools in Norfolk, Virginia, were used to test the proposed structural model.^ Initial evaluation of the measurement model showed lack of construct distinction between school and classroom climate. These two variables were collapsed into one: classroom climate, because analysis of the variance composition of the climate measures showed variation of climate by class within school, but little variation across schools. Estimation and testing of the just-identified model using LISREL VII, and $\chi\sp2$ difference tests, resulted in significant paths from classroom climate to self-concept, from self-concept to behavior, and from behavior to achievement. These results suggest that the effect of classroom climate on achievement is mediated by self-concept and behavior, and that the effect of self-concept on achievement is mediated by student behavior. This three-path model accounted for about 40% of the variance in self-concept, 62% of the variance in behavior, and 22% of the variance in achievement. A Tucker-Lewis index of.97 signified a good fit of this trimmed model to the data. The hypothesized model based on Comer's School Development Program was therefore supported in part by the data, strengthening the theory that climate and self-concept influence student achievement. ^