Stanford's Environment for Teaching study involved analysis of data collected in 1973 and 1975 by questionnaire and interview from 188 principals and over 200 teachers from a stratified sample of 46 elementary schools drawn from 34 San Francisco Bay Area school districts. Using aggregate indices, the study examined data relating school complexity (both organizational and instructional), morale, satisfaction with decision-making, and strategies for control, influence, coordination, compliance, and support between teachers and principals. The original data and some additional data on academic achievement in the classrooms of the teachers interviewed have been reanalyzed using a new method of data reduction that isolates variables more precisely, takes into consideration whether principals or teachers provided the information analyzed, and correlates the results with school complexity. While replications of the original analyses were successful, new and interesting patterns were revealed. The major finding of the new study is that patterns of association vary radically with the domain (for example, morale) and source of information, a finding that was obscured by aggregation in the original analyses. The reanalysis raises the question, "Which structural properties of the school are loosely coupled and which are not?", and offers tentative hypotheses. (Author/PGD)