Student evaluations of teacher behaviors as estimations of real-ideal discrepancies: A critique of teacher rating methods.

Abstract

Ratings imply comparative judgments between the values of the observer and his observations, for instance, a student's ratings of his teacher are estimates of the discrepancy between the student's ideals for the teacher's behavior and what he sees the teacher do. However, most methods for collecting teacher ratings make assumptions about ideals and about the discrepancies between ideals and observed behavior. In order to assess the relevance of direct measurement of ideals, students were asked to rate a teacher and to report their ideals for the teacher's behaviors. The judgments of ideal behavior varied across students and items; there were also interactions between ideals and observed responses. Thus, the authors recommend new approaches to teacher ratings.Students have ideas about ideal teacher behaviors. Students also see their teachers behave. One way to interpret students' evaluations of teachers is to say that each student estimates the discrepancy between his ideal for that behavior and what he sees his instructor do. When the instructor's behavior comes close to the student's ideal, the teacher receives a high rating; if the teacher is far from the student's ideal, the teacher receives a low rating.Most approaches to teacher evaluation do not assess the student's ideal directly. For example, students may be asked to rate the teacher's "flexibility" on a 5-point scale ranging from "outstanding" to "poor." In this approach, the student makes an explicit evaluation (i.e., estimates a discrepancy) based upon an implicit observation. The response of "poor" for a given attribute would be interpreted as the maximum discrepancy between the ideal and observed behavior for each student. However, the extent to which similar low ratings among a set of students truly reflects the constancy of the actual 'This paper is an expanded version of one presented at the annual meeting of the Midwestern Psychological Association in Chicago, 1969. 1 Requests for reprints should be sent to Leonard M. Lansky, Department of Psychology, McMicken Hall, University of Cincinnati, Cincinnati, Ohio 45221.psychological discrepancies depends upon the assumption that all students have the same ideal point along the attribute dimension.Coombs (1964) has discussed the problem of interpreting such evaluations. He has suggested that students stretch their maximum discrepancy distances to fit their evaluation scales. Figure 1 contains a hypothetical distribution of ideal and observed responses on a rating scale for "flexibility." If, for example, Student A has an ideal point at I A at the middle of the distribution of possible observations, the maximum discrepancy can be only about half as great as that of Student B, having his ideal point at IB, far to the left end of the observation distribution. A rating of "poor" by Student B would represent a much greater discrepancy between his ideal and his observation than a rating of "poor" by Student A, even though both students made the same rating response. According to this interpretation, individ...

The Use of Importance-Performance Analysis as an Evaluative Technique in Adult Education

“…Blai (1971) and Levinthal et al (1971) state that one problem with student evaluations is that they make assumptions regarding the student's ideals. Student ratings are actually a discrepancy between the student's ideals for the instructor's behavior and how the instructor actually behaves.…”

“…Student ratings are actually a discrepancy between the student's ideals for the instructor's behavior and how the instructor actually behaves. Knowing this, it is impossible to compare ratings between students because each student may have a different &dquo;ideal.&dquo; In an attempt to improve the accuracy of student evaluation Levinthal et al (1971) suggest using a two-scale rating format. One scale would look at student preferences, the other at student observations.…”

“…The previously mentioned research (Blai,1971;Levinthal et al,1971) suggests two ways to improve the current evaluation methods: (1) have the students/participants do the evaluating (as opposed to other methods) and (2) include a two-scale rating format based on preferences and observations. The model, Importance-Performance Analysis, meets both of these criteria.…”

On the Effectiveness of Students’ Evaluative Feedback to University Instructors

“…It has been suggested that such information is insufficient because it fails to identify the referent (standard) by which the evaluator arrives at his or her rating (Levinthal, Lansky, & Andrews, 1971;Sanders & Lynch, Note 9). The argument is that the interpretation of the ratings may take on a completely different meaning if it is done in the light of a perceived ideal rather than on the basis of actual performance alone.…”

“…Students differ in their conceptions of ideal teaching (Levinthal et al, 1971). They are almost equally likely to evaluate their instructor's performance in relation to their perceptions of an ideal instructor, other instructors in the department, other instructors in different departments, and what they perceive to be their instructor's potential (Gmelch,Note 1).…”