Automatically Guilty? Measuring Associations between Evidence and Guilt using the DRM

Both real-life cases and laboratory research demonstrate that confession evidence is very convincing—even when it should not be. Could this be due to an automatic association between a confession and guilt? We tested this possibility using a Deese-Roediger-McDermott (DRM) paradigm 7-word list (Study 1) and context-rich story (Study 2). We hypothesized that participants would show more false recall for seeing “guilty” on a DRM list when the list included evidence that is closely associated with guilt than evidence weakly associated. Study 1 tested this hypothesis with a 7-word DRM list. Despite finding little support for our hypothesis, we did find patterns that suggest adaptations of these paradigms will help understand the phenomenon of wrongful conviction. We also discuss how this novel paradigm can be used to detect automatic associations, how it fits into the current knowledge of the DRM, and future research directions for guilt-automatic associations.