Latent state-trait and latent growth curve modeling of inhibitory control.

Abstract

The reliability of inhibitory control task performance as well as the existence of an underlying unitary inhibitory construct have been questioned. The present study is the first to use a trait and state decomposition approach to formally quantify the reliability of inhibitory control and to examine its hierarchical structure. N = 150 participants carried out antisaccade, Eriksen flanker, go/nogo, Simon, stop-signal, and Stroop tasks on three occasions. By applying latent state-trait modeling and latent growth-curve modeling, reliability was estimated and divided into the amount of variance explained by trait effects and trait changes (consistency) and the amount of variance explained by situational effects and effects of Situation × Person interaction (occasion specificity). Mean reaction times for all tasks revealed excellent reliabilities (.89-.99). Importantly, on average, 82% of variance was accounted for by consistency while specificity was rather small. Although primary inhibitory variables revealed lower reliabilities (.51-.85), the majority of explained variance was again trait determined. Trait changes were observed for most variables and were strongest when comparing the first occasion to later ones. In addition, in some variables, those improvements were particularly high in initially underperforming subjects. An analysis of the construct of inhibition on trait level showed that communality between tasks was low. We conclude that most variables in inhibitory control tasks are mainly affected by stable trait effects, but there is only little evidence of a common, underlying inhibitory control construct at trait level. (PsycInfo Database Record (c) 2023 APA, all rights reserved).