Reply to the note by Bruner and Tajfel.

Category width as a measure of discrimination1

“…Given these two conceptual emphases, and the lack of clarity m defining category width m terms of either one, arguments in the literature about niceties of measurement may seem somewhat displaced (Bruner & Tajfel, 1965, Gardner & Schoen, 1965. Various considerations seem to have colored both conceptions unduly For example, inasmuch as most studies report that females have narrower category widths than males, it is suggested that the narrow categorizer is more "conservative " However, as Crandall (1965) observed, this appellation may not apply to women's categonzmg tendencies when other, more interpersonal, domains of expenence are judged Similarly, the fact that the CWS refiects preferences for extreme responses m a relatively bemgn choice situation m which no actual stimulus discrimmation is called for would seem to weaken its status as a measure either of discrimmation or of risk takmg.…”

Breadth of categorization and the generalization of expectancies.

“…• Burner and Tajfel (1961) suggest that narrow categorizers are more sensitive to changes in the stimulus environment and show greater alteration in their categorizing breadth in response to change than do broad categorizers. However, agreement response set may have influenced their findings (Gardner & Schoen, 1965). A variety of personality variables have been related to category width.…”

A rejoinder.

“…The reply by Gardner and Schoen (1965) has not only done nothing to justify their original inadequate reporting of previous experiments (this cannot be altered by delayed afterthoughts); it has also managed to create even more confusion. Some of the omissions pointed to in our previous note (e.g., control of regression, lack of reference to Pettigrew's results) are again gingerly bypassed; new references are quoted with great gusto, whether relevant or not; empirical points are made which have nothing to do with our previous comments about the limping logic (not the empirical validity) of their original argument concerning uni-and multidimensional tests.…”