Reliability and Convergence of Conflict Effects

Abstract

Abstract. Recent work in attentional control has suggested that conflict effects measured across different tasks are not reliable and by extension unrelated. The lack of correlation between these conflict effects is in juxtaposition not only to theoretical predictions of a domain-general attentional control mechanism but also to a large body of individual differences research that has used these tasks to show evidence for an attentional control construct and its relatedness to other psychological constructs. In an effort to address this, we fit hierarchical models to each task that modeled trial-to-trial variability in response times to assess the extent to which the parameter estimates for the conflict effect correlated across tasks. We compared this method of assessing shared variance to more traditional summed difference score estimates of the conflict effect by analyzing data from a large-scale individual differences experiment, in which N = 582 subjects completed a Stroop, Flanker, and Simon task. Across tasks, we found that while the reliability of the conflict was sufficiently high and the between-task conflict effect significantly correlated, the magnitude of the between-task correlation was low. We discuss the implications of these results as providing more support for a domain-specific than domain-general attentional control mechanism.