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

Course-averaged evaluation scores for all classes offered at RoseHulman Institute of Technology during the 2004-2005 academic year were obtained from the Rose-Hulman Office of Institutional Research, Planning, and Assessment. These evaluations were administered electronically via a Web-based form. Students had one week to complete evaluations-the final week of classes, prior to the final exam period each academic quarter. Table 1 shows the set of numerically-rated evaluation items with the applicable response scales; spaces were provided for students to type additional or explanatory comments as well. If fewer than five students submitted evaluations for a course, no data from that course were used for analyses. Evaluations of military science (i.e., Reserve Officers' Training Corps) courses and courses designated as graduate-level only (i.e., graduate seminars) were not used. The resulting data set consisted of information from 490 engineering courses, 390 mathematics and science courses, and 165 humanities courses. The response rates for the engineering, mathematics and science, and humanities course evaluations used in the present study were 78 ± 15%, 78 ± 15%, and 82 ± 13% (means ± standard deviations), respectively. Therefore, the data used in the present study are likely a fair to good representation of the perceptions of students in these classes.

A second data set consisted of course-averaged evaluation scores for all classes offered through the Tulane University School of Engineering from the fall of 1997 to the fall of 2002. These evaluations consisted of a 17-item "bubble sheet" paper questionnaire, which asked students to indicate whole numbers ranging from 1 to 5 to signify their level of agreement with a given statement. The evaluations were administered in each class by School of Engineering staff at the end of each semester, prior to the final exam period. Course instructors left their classrooms while evaluations were administered. Evaluation response rates were not calculated or tracked as part of the administrative process, but since these evaluations were administered during a normal class period with anticipated normal attendance, the data used in the present study are likely to be a fair to good representation of the perceptions of students in these classes. Further information on the Tulane evaluation form items and procedures can be found in reference [8]. If fewer than five students submitted evaluations for a course, no data from that course were used for analyses. The resulting data set consisted of information from 823 courses offered through the School of Engineering.

B. Data Analysis

Simple correlational analyses were chosen for the present study since they provide easily-understandable verbal information (numerical coefficients, significance values) and visual information (scatterplots, trendlines) accessible to a broad audience. The Pearson correlation coefficient is a common way to characterize the association between two variables; this parametric technique carries a number of assumptions about the variance and distribution of the data to be examined [9]. The nonparametric Spearman's rho correlation is based on relative ranks of data rather than on the observed numerical values of data, and does not depend on stringent assumptions about the shape of the population from which the observations were drawn [9].