Moving beyond button presses to enhance the reliability of congruency tasks

Developmental and individual differences in conflict processing are commonly assessed with button-press measures of response time and accuracy in congruency tasks such as the Eriksen flanker task. However, when the reliability of response time and accuracy congruency effects are reported, the measures frequently fall short of the reliability levels needed to draw firm conclusions. In Study 1, we re-analyzed data from nine hand-tracking experiments. We found that reaching behavior generated high congruency effect reliabilities across different age groups (children, young adults, and older adults) in arrow versions of the flanker task, with split-half reliabilities often above .85 in response time, initiation time, movement time, or movement curvature. Adult performance of the Stroop task revealed congruency effect reliabilities ranging from .76 in movement curvature to .89 in initiation time. In contrast, the Simon task revealed a good reliability estimate in movement curvature for children (.84) and acceptable estimates in response time (.72), movement time (.76), and curvature (.75) for adults. Study 2 tested the reliability of a release-and-press version of the arrow flanker task using a keyboard in one condition and a touchscreen in another condition (N = 51). The keyboard condition generated high congruency effect reliability estimates in response time (.91), initiation time (.94), and movement time (.88), while the touchscreen condition generated high-reliability estimates in response time (.85) and initiation time (.94), and a lower estimate for movement time (.78). These results indicate that the reliability of congruency tasks can be enhanced by moving beyond traditional button-press measures of performance.