This study involved an inservice education program designed to enhance teacher behavior in the use of Science - A Process Approach. The major method of gaining information on teacher behavior was direct observation and involved the development and utilization of a Classroom Observation Rating Form (CORF). Twenty-three elementary teachers comprised the sample. Three observers formed two teams to conduct a total of 92 classroom observations of science and mathematics lessons. Using data from these 92 observations, a reliability coefficient of .98 was found for the CORF. Results of data analyses showed that between the first and last science observations, there were some significant differences in the employment of strategies, as derived from CORF scores, which indicated a move toward more teacher-directed lessons. However, between the first observation in math and the last science observation, the net change in use of strategies was toward more student-centered lessons. There was no significant correlation found between years of teaching experience and change in the employment of teaching strategies. Primary teachers showed greater changes in teaching behavior than did the intermediate teachers. (Author/EB)