The opinions of experts in prediction in civil commitment hearings should help the courts, but over thirty years of commentary, judicial opinion, and scientific review argue that predictions of danger lack scientific rigor. The United States Supreme Court has commented regularly on the uncertainty of predictive science. The American Psychiatric Association has argued to the Court that "[t]he professional literature uniformly establishes that such predictions are fundamentally of very low reliability." Scientific studies indicate that some predictions do little better than chance or lay speculation, and even the best predictions leave substantial room for error about individual cases. The sharpest critique finds that mental health professionals perform no better than chance at predicting violence, and perhaps perform even worse.

Given these critiques, we should expect the rules of evidence, and specifically the reliability standards of Daubert v. Merrell Dow Pharmaceuticals, Inc., to require the exclusion of predictive expertise from the civil commitment process. Daubert overturned the seventy-year-old Frye v. United States standard for admissibility of expert opinion, replacing it with a standard focused (at least in part) on the scientific reliability of expert opinion. Civil commitment is a matter of state law, governed by state rules of evidence, and not all states have followed the Supreme Court's lead. But this should not matter: Given their notorious unreliability and deep division in the professional community, we might predict that no court would admit predictive opinions either under Daubert or under Frye. Yet no appellate court has ever ordered exclusion of expert psychiatric testimony about danger in a civil commitment case, either before or after Daubert.

This Article uses the example of predictive expertise both to develop a methodology for assessing Daubert's notion of "fit," and to find factors for applying that standard. The methodology appraises civil commitment for its characteristic features: its substantive standards; its burdens of proof; its patterns of proof for danger, including expert testimony; and the legal definition of danger. This methodology helps to identify four factors for assessing the fit of expertise in a given case, and to ratify the extraordinary fit between predictive opinions and civil commitments. The argument proceeds in two phases. Part I discusses how Daubert and its progeny have articulated a dual standard of reliability and fit. The first section concludes that, under Daubert, the "reliability" of an opinion means not just its scientific validity, but also its practical utility for resolving disputes. It also concludes that Daubert's test for "fit" asks not just about bare relevance; it also suggests a comparison of the expert's specialized inferences to the inferences required for fact-finding. The second section then reviews the available modern standards for validating predictive expertise, and concludes that the relevant scientific community has reached a state of guarded optimism about its reliability for judicial use. Finally, the third section examines the substance and process of civil commitment, assessing how predictive expertise fits the features of commitment, so as to articulate with specificity elements of Daubert's fit criterion.

Part II turns to evidentiary doctrine, analyzing those cases that have admitted expert predictive opinion in civil commitment proceedings,whether the jurisdiction uses tests from Frye, Daubert or no test at all. This Part confirms the widespread receptivity of courts to expert opinion evidence about future danger. It also critiques the existing rationales for those results, and proposes a rationale that fully explains the case law and accommodates developments in predictive science. This Article concludes that Daubert should permit the admission of expert predictions, but not because of their scientific reliability. Instead, the Article uses the example of civil commitments to develop the notion that Daubert's concern with evidentiary fit better explains the courts' receptivity to this form of predictive testimony.