Focusing on High-Conflict Trials to Better Measure Attention Control

Theories of working memory, intelligence, and executive functioning often propose a flexible, multi-purpose ability for controlling attention (Baddeley & Hitch, 1974; Shipstead et al., 2016). Yet, many commonly utilized tasks meant to measure attention control do not strongly correlate with each other. One explanation for this failure of convergent validity could be that, contrary to theory, there is no general “attention control” ability. An alternative explanation is that the tasks most often used are psychometrically flawed, such that no substantial cross-task correlation can be detected, even if the theory itself is correct. Recently, Moretti et al. (2025) found that a revised method of scoring common attention control tasks improved their reliability and convergent validity, as evidenced by stronger cross-task correlations and internal consistency estimates. In particular, they attempted to control for speed-accuracy trade-offs and congruency sequence effects. We reanalyzed archival data on Flanker and Stroop tasks (N ≈ 400) according to the specifications of Moretti et al. (2025). However, the revised scoring methods did not improve reliability or validity estimates.