Internal consistency and test-retest reliability of an affective task-switching paradigm

Affective flexibility refers to the flexible adaptation of behavior or thought given emotionally relevant stimuli, tasks, or contexts, and has been associated with the efficiency of emotion regulation and dealing with stress and adversity. Experimentally, individual differences in affective flexibility have been measured as behavioral costs (response times, errors rates) of switching between affective and neutral tasks. However, behavioral task measures can only be treated as trait-like characteristics if they have sufficient psychometric quality. We report an analysis of the test-retest reliability (interval two weeks) as well as internal consistencies of behavioral switch costs measured in an affective task switching paradigm. This paradigm elicits strong response time switch costs for both tasks, but higher when switching to the emotion than to the gender task. These ‘asymmetric switch costs’ suggest dominance of the emotional task rule. Reliability analyses indicated excellent internal consistency estimates (Spearman-Brown corrected r = .92 for both switch directions) and good test-retest reliabilities (ICC(2,1) of .78 and .82, respectively) for response time-based switch costs. Effect sizes and reliability estimates were substantially lower for switch costs calculated from error rates, which is consistent with previous literature discussing the psychometric properties of task-based cognitive measures. Reliability measures were lower but still acceptable for valence-specific response time-based switch costs, potentially due to lower trial numbers per cell when increasing granularity of the analysis. In conclusion, our results indicate that response time-based affective switch costs are well-suited as individual differences measure, and thus may be a valuable proxy for assessing affective flexibility.