Relation of Course, Instructor, and Student Characteristics to Dimensions of Student Ratings of Teaching Effectiveness
Heckert, Teresa M., Latier, Amanda, Ringwald, Amy, Silvey, Brenna, College Student Journal
This research investigated the relation of course, instructor, and student characteristics to student ratings of teaching effectiveness, both overall and within the dimensions of pedagogical skill, rapport with students, difficulty appropriateness, and course value/learning. Interest in the course content, expected grades, satisfaction with the time of day, and instructor sex were related significantly to all dimensions of teaching performance. Year in school and reason for taking a course were related to value ratings, and student sex and year in school were related to rapport ratings. Relations of background characteristics to teaching evaluations varied somewhat by dimension. Background characteristics were related most strongly to course value ratings and the overall evaluation, and least strongly to the pedagogical skill ratings.
Student ratings of instruction are the most widely used measure of college teaching effectiveness. Administrators use these ratings in making tenure and promotion decisions. Although some faculty members question the usefulness and accuracy of student perceptions, a substantial amount of research exists which supports the validity of such ratings (Aleamoni, 1999; Marsh & Dunkin, 1997; Wachtel, 1998). However, there is also quite a bit of research suggesting student ratings may be susceptible to influence by factors which may be unrelated to teaching effectiveness. For example, researchers have found that student and instructor sex may affect course evaluations (Centra & Gaubatz, 2000; Feldman, 1993). Furthermore, some research has found course characteristics, such as class size and level of the course, to be related to overall evaluations (Feldman, 1978). Other researchers (e.g., Greenwald & Gillmore, 1997; Marsh & Roche, 2000) have found a relation between expected grades and course evaluations. Although correlations are typically small, these relationships should not be ignored when interpreting an instructor's evaluations by his or her students. As Ludlow (1996) found, course characteristics may explain a substantial portion of the variability within a single professor's course evaluations over his or her teaching career.
Most of the research into the relations of background characteristics (i.e., course, instructor, and student) to course evaluations have utilized single overall or global evaluations of teaching. However, there is considerable evidence that student ratings are multidimensional (Cashin & Downey, 1992). Several researchers (Burdsal & Bardo, 1986; Frey, 1978; Marsh, 1983) have argued that the inconsistency in prior research on background characteristics may be resulting from a failure to separate evaluations into their appropriate dimensions and examine relations individually for each dimension. Marsh (1983) found background characteristics were related to those dimensions with which they were logically connected. For example, prior interest in the subject matter was strongly related to the dimension of learning/course value. He also found that no background characteristic was related to more than a few dimensions. Similarly, other researchers (e.g., Frey, 1978; Marsh & Dunkin, 1997) have found background characteristics to relate to only some dimensions. For example, Frey (1978) found that a rapport with students dimension was related to both class size and average class grade, whereas a pedagogical skill factor was relatively unaffected by these characteristics. Other researchers (Cramer & Alexitch, 2000; Cranton & Smith, 1990; Gigliotti & Buchtel, 1990) have also found that although background characteristics may be related to teaching ratings by students, the relations do seem to vary by dimension.
Although substantial research exists which points toward multidimensionality of student ratings, some researchers continue to argue that, for personnel decisions, it is not necessary to use multiple dimensions (Cashin & Downey, 1992). …