Student Perceptions of High Course Workloads are Not Associated with Poor Student Evaluations of Instructor Performance

Journal of Engineering Education, Jan 2007 by Dee, Kay C

All calculations of correlation coefficients, linear and quadratic regressions, and Mann-Whitney tests were conducted using SPSS for Windows® (SPSS Inc.). Determination of statistical outliers, quartiles and associated means, and comparisons of correlation coefficients and linear regression parameters were conducted using Excel© (Microsoft Corporation).

III. RESULTS

Figure 1 shows that Rose-Hulman student ratins of overall instructor performance were neither linearly nor quadratically related to student ratings of course workload required in relation to other courses of equal credit. Figure 2 presents the means and standard deviations of scores on all of the Rose-Hulman evaluation items, for courses in the highest and lowest quartiles of ratings on the item assessing overall instructor performances. Examining Figure 2 in conjunction with Table 1 reveals that courses in the lowest quartile (i.e., with the poorest ratings of overall instructor performance) received different (p

In contrast to the workload-related results, ratings on a number of other Rose-Hulman evaluation items were strongly correlated with ratings of overall instructor performance (Figure 3 and Table 2). For example, the strength of agreement with statements such as "The professor used teaching methods that helped me learn," 'The professor met the stated course objectives," and "The professor generally was well-prepared for class" was strongly associated with better ratings of overall instructor performance (Figure 3). Items most strongly correlated with overall instructor performance ratings (Table 2) tended to focus on student perceptions of the professor's teaching/presentation methods, preparation, sensitivity to students and interest in the subject, and overall learning experience and course quality. The items listed in Table 2 also yielded the strongest correlations when considering only humanities courses (with correlations ranging from 0.880 to 0.673) and when considering only mathematics and science courses (with correlations ranging from 0.892 to 0.703), with some minor re-ordering and in the case of humanities courses the substitution of "Grading was objective and impartial" for "The professor seemed genuinely interested in teaching this subject." Qualitatively similar results were obtained from the Tulane data (Table 3), in which items most strongly correlated with overall instructor performance ratings tended to focus on student perceptions of the professor's teaching/presentation methods, interest in teaching and students, and overall learning experience. Some items on the Tulane evaluation were phrased similarly to items on the Rose-Hulman form; a few of these "matched" items yielded differing correlations with overall instructor performance (Table 3).

Items most weakly associated with overall instructor performance ratings (Table 5) tended to focus on the number or percent of responses to evaluation items and student perceptions of course pace, workload, textbook, and coordination between laboratory exercises and course materials. The items listed in Table 5 also yielded the weakest correlations when considering only humanities courses (with correlations ranging from -0.003 to 0.381) and when considering only mathematics and science courses (with correlations ranging from -0.022 to 0.333), with some minor re-ordering. The correlations between the workload evaluation item and the overall instructor performance item were similar across engineering (0.07), mathematics and science (0.05), and humanities (0.08) courses.