Four decades of forensic research have left unanswered a fundamental issue regarding the best conceptualization of competency to stand trial vis-à-vis the Dusky standard. The current study investigated three competing models (discrete abilities, domains, and cognitive complexity) on combined data (N = 411) from six forensic and correctional samples. Using the Evaluation of Competency to Stand Trial-Revised (ECST-R), items representative of the Dusky prongs were used to test the three models via maximum-likelihood confirmatory factor analyses (CFA). Of the three, only the discrete abilities model evidenced a good fit, indicating that competency to stand trial should consider separately each defendant's factual understanding of the proceedings, rational understanding of the proceedings, and ability to consult with counsel. ECST-R competency scales, based on the current CFA, have excellent alphas (.83 to .89) and interrater reliabilities (.97 to .98).