Classroom climate has been found to predict a significant portion of the variance in student achievement, independent of student background and intelligence quotient scores. This study sought to more clearly define classroom climate by determining to what extent climate measures teacher characteristics, student characteristics, and classroom characteristics such as curriculum and class size. After delineating the domains measured by classroom climate and establishing a climate construct, concrete and manipulable variables that covaried with the climate scale were identified. Data were obtained from 895 junior and senior high school classes. Students and teachers responded to questionnaires and interviews, and each class was observed on three separate occasions. Attention focused upon classroom climate variables of: (1) teacher concern, punitiveness, authoritarianism, favoritism, enthusiasm, and clarity; (2) student decision-making, peer attitudes, competitiveness, cliqueness, satisfaction, compliance, and apathy; (3) classroom physical appearance; and (4) instructional practices: knowledge of results, task difficulty, and organization. Evidence from the study indicated that the climate construct is affected by a wide range of variables that merge together in the classroom context, and the construct is most affected by the variables most proximate to the classroom. Two major findings support this evidence. First, climate scores are sensitive to variation from several domains. The background and beliefs of the students and teachers and the conditions within the classroom all affect the climate of the class. However, these variables accounted for, at most, 18 percent of the variance in the climate scores and were not considered a threat to the construct's validity. Second, teacher perceptions of classroom occurences were closely related to the climate scores. Five appendices provide additional information. (JD)