Statistical Effects of Class Size and Teaching Ability in Determining Teacher Ratings

Cohen (1983) considered a statistical problem associated with the effect of class size on teacher rating.Specifically, Cohen considered comparisons of mean scores [xbar]A and [xbar]B of two instructors teaching classes of different sizes.A heuristic proof was given that, under certain conditions and among the better lecturers, those who are evaluated by a smaller number of students have an advantage.We examine the effect of class size on teacher ratings for some special cases involving classes of modest size.We discover that class size can have a noticeable impact on teacher ratings but that the advantage lies sometimes in having a larger class and sometimes a smaller.Moreover, it appears that in practice the effect of class size on teacher ratings should be quite small.