Relations among Student Effort, Perceived Class Difficulty Appropriateness, and Student Evaluations of Teaching: Is It Possible to "Buy" Better Evaluations through Lenient Grading?.

Abstract

Some professors argue that it is possible to buy higher course evaluations by grading more leniently and requiring less work. This study investigated the relations of student effort, appropriateness of class difficulty, and course evaluations with a sample of 463 undergraduates from a mid-sized, Midwestern university. Findings support the treatment of difficulty appropriateness and student as distinct concepts. Results refute the grading leniency hypothesis. Consistent with cognitive dissonance theory and the validity hypothesis, student was positively related to all dimensions of course evaluation, and these relationships could not be explained by effort's relation to expected grades. Student was also significantly related to a number of student, instructor, and course characteristics. ********** Although many administrators claim that teaching is the most important professor activity, frequently, the only tool used to assess teaching effectiveness is student ratings. Ratings are used to make personnel decisions, including promotion and tenure. Reviewers (c.f., Aleamoni, 1999; Wachtel, 1998) have concluded that ample evidence exists that, in general, supports the validity of these ratings. However, ratings do appear to be susceptible to influence by factors that may be unrelated to teaching effectiveness. One of the most widely studied and contentious of these factors is expected course grade. Some professors argue that it is possible to buy higher evaluations by grading more leniently and requiring less work. This grading leniency hypothesis predicts a positive relation between expected grades and evaluations of teaching and a negative relation between student and course evaluations. Many studies have found a positive relation, typically .10 to .30, between expected grades and course evaluations (Feldman, 1997). Aleamoni (1999) reported 24 studies finding no relation and 37 studies finding a positive relation, with a median correlation of .14. Wachtel (1998) concluded, based on his review of the literature, that there was a moderate positive relation between expected grades and evaluations. The inconsistency in the findings of individual studies may stem from using global measures rather than separate dimensions to assess teaching effectiveness (Aleamoni & Thomas, 1980). Course evaluations are multidimensional (Cashin & Downey, 1992; Marsh & Dunkin, 1997), and, as a result, the relation of expected grades to evaluations may vary depending on the dimension (Frey, 1978). In addition to using global evaluation measures, most studies have tested only half of the grading leniency hypothesis and have not measured student effort. Rather, it is merely assumed that an instructor who grades leniently also expects little of his or her students. For example, Needham (1978) postulated that students give higher evaluations to professors who give the best grade per unit of student effort (p. 39); however, he did not empirically test his model. Moritsch and Suter (1988) found a relation between and halo error (i.e., the failure to differentiate among rating items) but did not report the relation of to the actual evaluation. Only a few studies could be found in which and expected grades were both measured directly. Brodie (1998) found the highest evaluations went to professors with the highest grades for the least amount of student effort. Similarly, Greenwald and Gillmore (1997) found a negative relation between grades and workload, such that higher grades were received in courses with lighter workloads. They also found higher grades resulted in higher student evaluations. Taken together, these findings seem to support the grading leniency hypothesis. Contrary to the grading leniency hypothesis, cognitive dissonance theory suggests a positive relation between student and course evaluations. According to this theory, as expended increases, enjoyment of the task increases (Reeve, 2005). …