“…They were instructed to mark 100% if virtually all students engaged in the behavior, 75% if most students engaged in the behavior, 50% if about half engaged in the behavior, 25% if some but not half engaged in the behavior, and 0% if virtually none of their students engaged in the behavior. The mean responses for each of the nonverbal behaviors served as the dependent variable in this study, Andersen, Andersen and Mayton (1985) report a reliability estimate of .93 using this method indicating good instrument reliability. The independent variable for all analyses was the grade level of the teacher, which served as the marker to view developmental changes in student nonverbal behaviors.…”