Christopher Slobogin, Dangerousness and Expertise Redux

• Jurisdiction: United States,Federal
• Publication year: 2006
• Citation: Vol. 56 No. 2

EMORY LAW JOURNAL

Volume 56 2006 Number 2

ARTICLES

DANGEROUSNESS AND EXPERTISE REDUX

Christopher Slobogin^*

ABSTRACT

Civil commitment, confinement under sexual predator laws, and many capital and noncapital sentences depend upon proof of a propensity toward violence. This Article discusses the current state of prediction science, in particular the advantages and disadvantages of clinical and actuarial prediction, and then analyzes how the rules of evidence should be interpreted in deciding whether opinions about propensity should be admissible. It concludes that dangerousness predictions that are not based on empirically derived probability estimates should be excluded from the courtroom unless the defense decides otherwise. This conclusion is not bottomed on the usual concern courts and commentators raise about expert prediction testimony-that the associated false-positive rates are too high. In fact, because they produce better-than-chance results, both clinical and actuarial risk assessments are, with certain caveats, sufficiently probative to meet the expertise threshold, whether that threshold is defined by Frye or Daubert. Rather, the rationale for the conclusion reached in this Article is that prediction testimony should only be inadmissible when its prejudicial impact outweighs its (admittedly weak) probative value. Clinical prediction testimony should be excluded from the government's case-in-chief under this formula, because of its relative invulnerability to adversarial testing. But the government should be able to use actuarial prediction testimony, as empirical evidence suggests that this type of testimony is likely to be taken for what it is worth. Finally, because of judges and juries' demonstrably strong predilection toward findings of dangerousness in commitment and sentencing proceedings, the defense should always be able to use both clinical and actuarial testimony to rebut the state's case.

INTRODUCTION

Over two decades ago, I wrote an article entitled Dangerousness and Expertise that made a number of suggestions about how the rules of evidence should apply to expert testimony aimed at predicting future antisocial behavior.¹This Article addresses the same issue and reaches the same basic conclusion: Expert prediction testimony based on empirically derived probability assessments usually should be admissible, whether proffered by the government or the defense, but clinical prediction testimony, which is still the predominant method of proving dangerousness, should be excluded unless the defense uses it in its case-in-chief. Despite the similarity in end result between the two articles, a number of intervening developments require a new analysis of the way courts make determinations about future violence.

First, as society grows increasingly frustrated with the traditional punishment system's failure to protect its citizens, dangerousness has become increasingly popular as a criterion for depriving people of liberty. The most obvious example of this phenomenon is the "sex predator law" movement. Since 1990, about one third of the states have enacted laws that permit indeterminate post-sentence commitment of sex offenders considered to be

"predisposed" to violent behavior.²On authority of these statutes, well over

2,500 people have been committed as sexual predators and thousands of others have been subjected to sexual predator hearings.³Even more expansive "dangerous offender" laws, which permit post-sentence confinement of any offender perceived to have violent tendencies, have also been enacted.⁴

Meanwhile, dangerousness determinations remain important in many other areas of the law. While criminal sentencing since 1970 has generally moved toward fixed terms based on backward-looking culpability assessments,⁵recent Supreme Court decisions could push the states back toward indeterminate dispositions grounded on risk assessments,⁶and several states continue to make the risk of future violent crime an aggravating circumstance in death penalty cases.⁷Finally, the civil commitment laws of virtually every state permit involuntary hospitalization based on the likelihood of imminent harm to others,⁸and similar provisions authorize commitment of those found not guilty by reason of insanity.⁹

A second reason why a fresh look at methods of proving dangerousness is needed is that prediction methodology has significantly improved. Two decades ago, actuarial models for predicting violent behavior were in their infancy and the dominant method of evaluating dangerousness was essentially seat-of-the-pants, "clinical" speculation. Today the development of several sophisticated, empirically validated risk assessment instruments has made prediction much more of a science.¹⁰

A third reason why the topic addressed in this Article is worth revisiting is that the law of evidence has changed significantly. Until 1993, most courts evaluated the admissibility of expert testimony using one of two tests: The "balancing test" (derived from the balancing required by Federal Rules of Evidence 401, 402 and 403),¹¹which weighs the probative value of the evidence against the likelihood it will mislead or confuse the fact finder or unfairly prejudice one of the parties, and the Frye test, derived from Frye v. United States,¹²which looks at whether the basis of the testimony is "generally accepted" among professionals in the relevant field.¹³Today, however, the federal courts and more than half the states follow the test announced in Daubert v. Merrell Dow Pharmaceuticals,¹⁴which, at least on its face, is much tougher to meet. Daubert requires that expert testimony be validated or verified, preferably using scientific methodology that produces error rates rather than simply through assessment of whether the basis of the testimony is relevant or generally accepted.¹⁵

The reason Daubert, Frye, and other tests of expert admissibility are important in this context is that dangerousness assessments are often bottomed, if not wholly dependent, on the opinions of psychiatrists and psychologists. In some settings, particularly in connection with noncapital sentencing proceedings, dangerousness determinations may be based entirely on lay judgments, usually made by probation officers and judges. But in numerous other settings-in particular, capital sentencing, police power civil commitment proceedings, and sexual predator hearings-mental health professionals and other "experts" on dangerousness are almost always involved in assisting the decision makers who predict antisocial conduct.¹⁶

A number of commentators, writing both before and after Daubert raised the admissibility threshold, have asserted that this practice of relying on mental health professionals to assist in making legally relevant predictions should stop or be severely circumscribed.¹⁷The official organs of the mental health professions have long voiced similar views. Back in 1974, the American Psychiatric Association stated that "[p]sychiatric expertise in the prediction of

'dangerousness' is not established."¹⁸Four years later, the American Psychological Association came to much the same conclusion, asserting that "the validity of psychological predictions of violent behavior, at least in . . . sentencing and release situations . . . , is extremely poor, so poor that one could oppose their use on the strictly empirical grounds that psychologists are not professionally competent to make such judgments."¹⁹Today, despite improvements in prediction science, forensic professionals are still cautious about endorsing opinion testimony concerning violent propensities.²⁰

The courts, however, have paid no attention to such assertions. In 1983, despite amicus briefs incorporating statements like those quoted above, as well as John Monahan's 1981 conclusion, based on a meta-review of the studies, that two out of three long-term predictions of violence are wrong, the Supreme Court upheld the admissibility of prediction testimony from mental health professionals in death penalty proceedings.²¹That holding, in Barefoot v.

Estelle,²²still stands. If executions may be predicated on prediction testimony, presumably sex offenders and mentally ill people subject to commitment may be confined based on such opinion evidence as well.

Barefoot, however, only announced the constitutional minimum; evidentiary rules can demand more of experts, as Frye and Daubert do. Using a four-part analytical framework derived from traditional evidence law,²³I take an intermediate position, as I did in my first article on the subject, between complete exclusion and unquestioned acceptance of expert testimony on dangerousness. Traditional clinical testimony regarding a person's future behavior, while material, probative, and, on its face, helpful, is so prejudicial (and thus ultimately unhelpful) that in most settings it should be admissible only if the person first seeks to use clinical testimony to show that he or she is not dangerous. If the individual eschews such testimony, then the state should be limited to proving dangerousness using appropriately normed actuarial instruments or structured interview instruments that are tied to explicit probability estimates.

This Article develops this argument in eight parts. Part I describes the current state of prediction science. It explains the difference between the three primary prediction methodologies-clinical, actuarial, and structured professional judgment-and then evaluates their relative reliability. Part II examines the relevant evidentiary case law, which for the most part has not differentiated between prediction methodologies, and in any event has usually permitted any type of prediction testimony, regardless of its basis. Part III begins my own evidentiary analysis with an assessment of the materiality of prediction testimony, in particular the extent to which materiality is undermined by nomothetic (group-based) prediction or prediction based on...