Pearson correlation coefficients between first and second year classroom process data from the Texas Teacher Effectiveness Project are presented. These data are from 19 second and third grade teachers whose classrooms were observed during two consecutive school years. The stability of classroom behaviors varied greatly across the two years. Factors which affected stability included: frequency of the behaviors, reliability of the instruments, teacher vs. student control of the behaviors, and the amount of structure afforded by the contexts within which behaviors occurred. Possible explanations for the observed differences in stability of these process variables (both high and low inference) are offered, and implications of these data for classroom research are discussed. (Author/MV)