Date of Award


Publication Type

Doctoral Thesis

Degree Name





Psychology, Experimental.




This study's purpose was to assess, first, whether one form of conceptual association posited by the systematic distortion hypothesis, the affective reaction that items evoke (evaluative salience) (Schweder, 1982), is implicated in the covariance structure of student ratings of diagnostically meaningful university instructor behaviour, and, second, whether evaluative salience threatens the construct and predictive validity of such ratings. End-of-term behaviour ratings were obtained from 490 students of 16 graduate student introductory psychology instructors using a modified version of Murray's (1987) 60-item diagnostic Teacher Behaviour Inventory. Inter-rater reliability for 58 items ranged from.50 to.86, with a mean of.61. 381 students from eight additional introductory classes rated the evaluative salience of each item using a 5-point scale, ranging from very bad to very good. Factor analysis of the behaviour ratings yielded four factors: General Instructional Effectiveness, Interaction, Enthusiasm, and Problematic Delivery. As predicted, removal of between-instructor variance had little effect on factor structure, indicating a high degree of systematicity in within-class disagreement and suggesting that a conceptual schema of "what-goes-with-what" may have been used. Also, as predicted, the factor structures of evaluative salience and partialled behaviour ratings were significantly correlated, thus implicating evaluative salience as a form of conceptual association. Criteria included student ratings of overall teacher effectiveness, course quality, attendance, and amount learned, as well as common exam performance. Partialling General Instructional Effectiveness, the factor with the highest mean evaluative salience, appreciably lowered all correlations. Whereas this suggested that students adjusted their ratings according to their general impression of instructor effectiveness, not all correlations were reduced to zero. With respect to predictive validity, class mean ratings on all factors correlated strongly with mean ratings of overall effectiveness and course quality, although none correlated with exam performance. Results are discussed with reference to previous Teacher Behaviour Inventory research, the study's limitations, and its implications for future research. (Abstract shortened by UMI.)Dept. of Psychology. Paper copy at Leddy Library: Theses & Major Papers - Basement, West Bldg. / Call Number: Thesis1992 .S748. Source: Dissertation Abstracts International, Volume: 53-12, Section: B, page: 6595. Thesis (Ph.D.)--University of Windsor (Canada), 1992.