Abstract
Behavioral science has greatly benefited from a simple but powerful dependent measure—response time (RT). This measure provides a window into cognitive processes, but it is not as simple as commonly assumed—meaningful information can be gained by breaking RT into subcomponents. Prior research has investigated the positive skew of RT distributions, leading to several models including Ex-Gaussian (Hohle, 1965) and Ex-Wald (Schwarz, 2001). A common theme of the models is that RT is comprised of multiple underlying components with different distributions, mainly a motor initiation and decision process. Here, we explored whether the decision process is the more meaningful cognitive measure that can be better leveraged as a dependent variable when distinguished from the motor process. Participants completed an object-sorting task (from the mobile application Airport Scanner; Kedlin Co., www.airportscannergame.com) in which they classified a series of objects by first touching and then swiping the objects to the top or bottom of their mobile device touch screen. The "touch and swipe" response method allowed RT to be separated into two subcomponents: 1) the time from stimulus appearance to initial contact with the screen—motor initiation, and 2) the time from initial contact to ultimate classification of the item—decision process. These two components of the overall RT only weakly correlated with each other across participants, supporting their status as unique subcomponents. Further, they were differentially related to trial accuracy and individual difference measures, and had different learning curves. Overall, the decision component appears to be the more interesting RT subcomponent, and it has strong relationships to cognitive effects—sometimes much stronger than overall RT. While overall RT is useful, these analyses suggest we could be missing, or occluding, additional effects, which can be recovered by separating RT into subcomponents to distinguish decision from motor initiation processes.
Meeting abstract presented at VSS 2018