Developmental and individual differences in fluid intelligence: Evidence against the unidimensional hypothesis

The Brinley plot and partial correlation methods are used to investigate conjointly the theories that increasing speed of processing is responsible for changes in mental age (MA) and chat speed of processing gives rise to within-age differences in fluid intelligence (IQ). A group of 107 children ranging in age from 4 to 9 years completed three reaction time (RT) tasks (spatial, verbal and selective attention). Children were grouped first by chronological age then by fluid IQ, and Brinley plots of RT and RT difference scores (argued to give a purer measure of cognitive speed) were produced for each grouping. Two Brinley methods were compared: fixing the intercept of the line of best fit at the origin, and allowing it to vary. These produced quire different functions with correspondingly different implications for the speed theory. It is argued that previous findings of developmental speed increases (as measured by Brinley slope) may be an artifact of the fixed-intercept method. When intercept was allowed to vary, findings indicated that speed did not change developmentally. Furthermore, when absolute RT and RT difference score findings were compared, partial correlations indicated that RT was related primarily to age rather than MA or IQ, whereas RT difference scores were related directly to IQ, but not to age or MA. We argue that the results are best accommodated by the theory that speed underlies stable IQ differences but not increases in MA. Instead, we argue that MA change may be better accounted for by children's increasing attentional ability.