STABILITY OF STUDENT EVALUATION OF INSTRUCTOR PERFORMANCE OVER TIME
The main hypothesis of this research was concerned with the stability over time of student ratings of instructor performance. The secondary hypothesis predicted that there would be exceptions to general stability of ratings in classes in which enrollment was small and in which unconventional instructional methods were used. The Student Instructional Report (SIR), an Educational Testing Service instrument originally validated at Northwestern University, was administered in 23 American University course sections in the fall of 1983 and in the spring of 1984 at the beginning and the end of each course. Student evaluations were matched, and paired t statistics were computed to determine if ratings were similar for the first administration (after four hours of instruction) compared to the second administration near the end of the course. The stability over time of early and late ratings was upheld. This was especially true for the organizational scale of the rating instrument, but for overall ratings, as well. Chi-square computations confirmed that smaller classes, those with 20 or fewer enrollees, and those classes with the greater variance from the conventional lecture-discussion methods were more likely to reflect rating instability--change in student opinion. Recommendations were that both practitioners and researchers take advantage of the stability of early student ratings to improve instruction.