Can carelessness be captured? Assessing careless responding in attitudes toward novel stimuli

Detecting careless responding has the potential to improve the quality of data obtained from research participants. In three samples (ns = 570, 602, 210), we used multiple indices of careless responding to predict the strength of implicit and explicit attitudes formed toward novel social groups as well as error rates on the Implicit Association Test (IAT; Greenwald, McGhee, & Schwartz, 1998). We tested each measure of careless responding on its own, simultaneously with other predictors, and with Bayesian analyses. In all three samples, there were strong and consistent effects such that more careful participants made fewer overall errors on the IAT; however, careless responding did not consistently predict implicit and explicit attitudes formed toward novel social groups. These results suggest that caution should be exercised when removing participants based on indices of careless responding.