A study of effects of teaching Project Physics (PP) on classroom social climates was made among a national random sample of classes. The population, divided into the experimental and control groups, included all classes with the same course, either PP or other physics courses, taught over a period from 1967 to 1970. Classroom climates were determined in 1967 and 1970 by factors selected from the revised "Learning Environment Inventory." Analyses showed that PP students shared more responsibility for class activity organization, were given more freedom in individual and small group study, and had a better access to materials. The work-pace and difficulty of the course were not sensibly different between the two groups. Teachers in PP became more sensitive to student's diversity. Moreover, effects of PP flexibility were found to be more related to available activities and materials than to teachers' understanding of the PP objectives. Physics instruction changes did not efficiently result in improvement of student's learning environment both in PP and other physics courses, unless the teacher was brought to a full understanding and supported to fulfill the course objectives. (Author/CC)