Comments on “Sex Differences in Cognitive Style”

Abstract

Given an uncertain task, children's response latencies will vary. Kagan (2) has described this variability as an aspect of cognitive style designated impulsivity-reflectivity: small latency values correspond to impulsiveness and larger values to reflectiveness. When studied cross-sectionally, grade school children exhibit a trend of increasing impulsivity, i.e., response latencies decrease with increasing age (2) . Lewis, Rausch. Goldberg, and Dodd (3) applied Kagan's impulsivity-reflectivity construct to study of relations among latencies and errors on a marching-figures task and psychometric intelligence (IQ) in a sample of preschool children (23 males and 25 females, 44 mo. old, 2nd with a slight high-SES bias). The purpose of their study was to test hypotheses of homogeneity of correlation among these measures berween preschool and school-age children and between preschool male and female samples (cf. 3 for detail). From Kagan's observations of school-age children, authors predicted heterogeneity of correlation across preschool male and female samples. Specifically, they proposed that the number of errors on a matching-figures task is more a function of response time for boys and more a function of IQ for girls (p. 563) .' Lewis, et al. (3) reported that only correlation between errors and latency was significantly different from zero for preschool males; for females only errors and IQ were significantly correlated. However, these results are not relevant to issue of sex differences since contrasts called for are between correlations for male and female samples, not independent contrasts with hypothetical (null) zero-order coefficients. The relevant contrasts are reported in a section in which these rs are compared with those reported by Kagan (cf. 3). Compar~sons of sample rs for males and females among preschool children and children in Grades l, 3, and 4 indicated that none of four (i.e., one per age group) hypotheses of homogeneity of correlation across sex between measures of error and latency could be rejected, that only one of three rs between errors and IQ could be rejected, and that none of three between latency and IQ could be rejected. In short, of 10 hypotheses which assert homogeneity of correlation across sex, only one (anent 4th grade Ss) could be rejected. Moreover, since 10 contrasts are nor independent, staristical significance of even this difference is questionable (cf. 1, p. 472). Thus, in terms of intercorrelations among number of errors, response latency, and psychometric intelligence, data reported by Lewis, et al. suggest that there are no sex differences in facets of cognitive style among preschool children and elementary school children in Grades 1, 3, and 4.