Psychometric proprieties of a color-shape version of the switch task

Abstract

The impact of concussions on an individual's cognitive functioning has become a growing health concern over the past several years; however, the search for sensitive tests persists. The task-switching paradigm is known to be sensitive to various medical conditions, including concussion. Accordingly, we developed 2 versions of the color-shape switch task. Three different costs are computed from the raw scores: global switch cost, which is thought to be a measure of global cognitive control; local switch cost, which is believed to be a measure of cognitive flexibility; and working memory cost. The aim of this study was to evaluate psychometric characteristics of these costs. An ANOVA revealed a main effect of sex on local latency switch cost, with females exhibiting longer latencies than males, p = 0.05. No main effect of sex was observed on any other switch costs. Moreover, no main effect of experimenter or version of the task was observed. Local switch cost was significantly correlated with trails 4 and 5 of the Comprehensive Trail Making Test (rs > 0.21, ps < 0.04). No other significant correlation between costs and established neuropsychological tests was observed, indicating low convergent validity. The intraclass correlation coefficient estimates ranged from 0.23 to 0.77, suggesting low-to-moderate 1-week test-retest reliability. Results indicated a low switch costs; convergent validity. Moreover, results show that the traditionally computed switch costs are less reliable than the primary outcomes (i.e., reaction time and accuracy). Researchers and clinicians should rely on primary variables from the task-switching paradigm rather than computing the traditional switch costs to increase the psychometric properties of the tasks which is critical to advances in theoretical models of executive functions and evaluations of clinical populations.