Grades, learning, and student evaluation of instruction

Abstract

In the first experiment, the author employed three different grading standards, identified as stringent, moderate, and lenient, in separate sections of the course Psychology of Learning. Other aspects of the course were the same for each section. The different grading standards resulted in substantially different grade distributions in the anticipated direction. Evaluations of both the instructor's performance and the couse decreased as the stringency of the grading criteria increased. Every item on the evaluation questionnaire was systematically influenced by the grading criteria. The amount learned, as measured by performance on the objective part of the final test, increased as the stringency of the grading criteria increased. In the second experiment, the subjects were the students in two sections of the course Introductory Psychology, taught by the author during the same term. The same textbook, course notes, and grading criteria were employed. The manipulated variable was test frequency, with the two sections receiving either weekly or biweekly tests. Students in the section receiving weekly tests scored 11.9% higher, on the average, over all tests. This resulted in a substantial difference in grade distributions. The ratings of the section receiving lower grades were substantially lower on every item of the rating form. Students appear to rate instructors on the basis of a global impression (“liking”) which they form, which is strongly influenced by the grade the student receives.