Measures of both preferred and actual classroom and school environment were administered to 1,675 secondary school students in New South Wales (Australia). Shortened versions of the My Class Inventory, Classroom Environment Scale, and Individualized Classroom Environment Questionnaire, as well as the Quality of School Life questionnaire were administered. Alpha reliability estimates indicated extremely high test reliability for the 19 scales in the four instruments. Maximum likelihood factor analysis revealed four factors from all the scales: (1) preference for peer conflict; (2) preference for individualization; (3) preference for teacher-managed structure; and (4) actual rather than preferred environment. Cluster analysis, using a non-hierarchical ISODATA approach, revealed a three-cluster solution: schools were broadly labelled as self-survival, indifferent, or cooperative and motivated. Multivariate analysis of variance indicated significant effects for school type, sex, and grade. English and mathematics achievement were more highly correlated with actual school environment than with preferences in environment. (GDC)