The dimensionality of student ratings of teaching: integration of uni- and multidimensional models

The debate on the unidimensionality versus multidimensionality of students' ratings of teaching performance, and the relationship of this dilemma to summative and formative uses of these instruments is reviewed, with special attention to the implications of their development and application in the Spanish university system. A comprehensive theoretical framework, stressing the relation between effective teaching theory and the expected construct of teaching performance in students' view, is advanced. The structure of teaching competencies is analysed by means of confirmatory factor analysis (CFA) and similarity structure analysis (SSA) in two ratings forms, of 13 and 34 items. The forms were applied to large student samples from two Spanish universities, rating the performance of individual teachers. CFA and SSA show a common latent structure in the two samples, and support the conceptual distinction between aspects of teaching competence. The results suggest that the structure of the students' ratings can be interpreted as multidimensional as much as unidimensional. The reason for this apparent paradox is that some specific dimensions are more central (versus peripheral) in the construct of teaching skills. This structure is consistent with the conceptual framework, and would explain the ambiguity between unidimensional and multidimensional solutions of CFA outcomes. The implications of these results for the orientation of future research on the dimensionality of students' ratings and for their application in formative and summative evaluation of teaching are discussed.