Partial information or facilitation? Different interpretations of results from speed-accuracy decomposition

Abstract

The speed-accuracy decomposition technique was developed by Meyer, Irwin, Osman, and Kounios (1988) to examine the time course of information processing. The technique allows for the estimation of the accuracy of guesses that are induced by the presentation of a response signal on a proportion of trials. Estimated guessing accuracy has been found to be above chance and to increase as time of guessing increases, suggesting that guesses are based on partial information that has accumulated prior to a response decision (sophisticated guesses). In this paper, a different interpretation of these data is presented. Results suggest that response signals may enhance the speed of regular processes, thereby violating the temporal-independence assumption that underlies the decomposition technique. As shown by Monte Carlo simulations, such facilitating effects of response signals can explain the results from the decomposition technique at least in part and possibly in full, even when guesses are actually at chance accuracy (pure guesses). The pure-guess model was supported by the results from an experiment designed to test between the alternative interpretations. These results point to the need for great caution in the attempt to infer the time course of information processing from guessing accuracies as estimated by the speed-accuracy decomposition technique.