Analysis of group differences in processing speed: Brinley plots, Q-Q plots, and other conspiracies

Abstract

Researchers in a growing number of areas (including cognitive development, aging, and neuropsychology) use Brinley plots to compare the processing speed of different groups. Ratcliff, Spieler, and McKoon (2000) argued that a Brinley plot is a quantile-quantile (Q-Q) plot and that therefore Brinley plot regression slopes measure standard deviation ratios rather than relative speed of processing. We show that this argument is incorrect. Brinley plots, by definition, are not Q-Q plots; the former are based on unranked data and the latter are based on ranked data. Furthermore, the relationship between standard deviation ratios and slopes is a general property of regression lines and has no implications for the use of Brinley plot regression slopes as processing speed measures. We also show that the relative speed interpretation of Brinley plot slopes is strongly supported by converging evidence from a metaanalysis of visual search, mental rotation, and memory scanning in young and older adults. As to Ratcliff et al.'s hypothesis that age differences in response time are attributable to greater cautiousness on the part of the elderly, rather than true processing speed differences, this hypothesis has been extensively tested in previous studies and found wanting.