The participants were undergraduate students at Princeton University in New Jersey. In the first experiment, case vignettes provided only legally relevant information about defendants' degrees of impairment in cognition or in behavioral control. Results revealed that participants' judgments of criminal liability or not guilty by reason of insanity reflected an exculpatory standard of substantial impairment in both cognition and control. In the second experiment, case vignettes provided realistic information about defendants' psychiatric diagnoses; the participants had to infer levels of impairment in cognition and control. Results revealed that participants made highly idiosyncratic inferences based on diagnostic categories; however, once made these inferences predicted judgments of not guilty by reason of insanity. Findings indicated that ordinary persons are not implicitly opposed to the two broad standards used in many courts for determining insanity. However, where reports of the case refer to psychiatric diagnoses and other extralegal indicators of the defendant's state of mind, community members will come to their own idiosyncratic conclusions about the defendant's degrees of cognitive and control dysfunction, resulting in controversies over the validity of the verdict and the application of the insanity defense itself. Tables, footnotes, and 35 references (Author abstract modified)