The empirical research is organized by the conceptual topics it addresses: perceptions and use of the insanity plea; characteristics of persons found not guilty by reason of insanity (NGRI), detention and subsequent release of persons found NGRI, postrelease recidivism rate and rehospitalization, and impact of the insanity standard. Review of the research suggests that it is relevant to some public policy issues; for example, lay persons may argue that the insanity defense should be abolished, because they believe offenders regularly invoke it and often win. Research documents the existence of these perceptions and their inaccuracies. The insanity defense is rarely invoked and less often successful. Some of this research also can speak to policy in ways that have not received wide recognition. Data on frequency of pleading, frequency of acquittal, and characteristics of those who plead and win should be used to address the following issues: the need for centralized court clinics to evaluate defendants for insanity; the type of staffing these clinics should have; and the size, type, and staffing of facilities needed to evaluate and treat acquittees. Findings across States are inconsistent, however. There is a need for cross-jurisdictional research that determines whether variations in findings are due to artifacts in prior work or to factors not yet thought to be relevant. The available research is not only of limited relevance to the current proposals on the insanity defense but also to the fundamental legal issues relating to the defense -- the language of the standard, the procedural implementation, and the disposition of those acquitted by reason of insanity. Ten notes and 37 references are provided.