Methodological and conceptual issues arising from a study of student perceptions of secondary school class climate are discussed. Through a pilot study, 113 Likert-type items were collected, consisting of teacher-focused dimensions, classroom-focused dimensions, and instructional practices. Factor and cluster analyses were used at the item level. Three correlation matrices were examined: students as individual units, students as units within classes, and classes as units. Two levels of schooling, junior and senior high school, were also computed. Efforts were focused on finding similarities between correlation matrices, and 18 class climate series appeared to be useful for further study: student competitiveness; student decision making; peer esteem; student cliqueness; classroom dissonance; task difficulty; student apathy; organization; student compliance; classroom physical appearance; student satisfaction; knowledge of results; teacher clarity; teacher favoritism; teacher authoritarianism; teacher punitiveness; teacher enthusiasm; and teacher concern. The choice of unit items for analysis was critical to the outcome of the study, and differences will be further explored at the scale level. (FG)