A reconsideration of the Cooper/Kline critique of the factor structure of the motivation analysis test

Abstract

Criticism of Cattell's Motivation Analysis Test (MAT) partly on the basis of a quasi-higher-order factor analysis of the MAT subscales on a sample of 109 subjects is untenable, given the methodological weakness inherent in the procedural application of the method by Cooper and Kline (1982). Cattell's (1982) response to the Cooper/Kline critique did not address these methodological issues directly, but instead concentrated on the controversy pertaining to intra-scale item-homogeneity. The debate over the structure of the MAT requires consideration of the factor analytic issues per se. The present reconsideration of the Cooper/Kline critique attempts to do this, and to provide a fuller perspective on the issues raised by Cooper and Kline. A series of articles recently published in the British Journal of Educational Psychology concerning the internal structure of the Motivation Analysis Test (MAT) has left the reader with a less than adequate account of reliability and validity of the instrument. Cooper/Kline (1982) criticized both the low item-homogeneities and also the factor structure of the MAT. Cattell (1982) responded to the Cooper/Kline critique and addressed the issue of intra-scale item-homogeneity. However the question of the factor structure of the MAT remained unresolved, and subsequently, Kline and Cooper (1982) reiterated their belief that the factor structures of the MAT were not stable across different samples. As Cattell did not address this issue in his response, it is germane to reconsider the evidence provided in the paper by Cooper and Kline. In particular, the present discussion focuses on (1) the validity of the MAT, (2) the number of subjects required, (3) the number of factors extracted, and (4) the question of simple structure. Finally, conclusions regarding the adequacy of the Cooper/Kline critique of the MAT are presented.