Cornell University Law School Social Science and Law

Predictions of Future Dangerousness

Predictions of Future Dangerousness: Civil Commitment of Sexually Violent Predators
It is a common knowledge that convicted criminals have high rates or recidivism when released from prison.  At one point does it become so likely that a criminal, particularly a violent one, will reoffend that he may be civilly committed to protect the community after serving his prison sentence?  Can social science provide a technical background to aid in drawing the line between ordinary criminals and ones with mental disorders qualifying them for commitment?  Can social science provide tools to aid in predicting which convicted criminals are likely to reoffend?

1.  The Law
A.  In 1990, following a pair of horrific and well-publicized crimes involving sexual violence perpetrated by offenders who had recently served time for similar crimes, the Washington State Legislature passed the Community Protection Act (CPA).  The CPA sought to redress what the public perceived as gaps in Washington’s criminal code which had allowed these dangerous offenders to roam free and offend again.  One provision of the CPA allowed prosecutors to seek involuntary civil commitment of certain sex offenders, who were termed Sexually Violent Predators (SVPs), at the end of their prison terms, with the goal of protecting the public from further crimes by these offenders.  Washington’s SVP law quickly served as a template for similar laws in several other states, including Kansas, where the legislature explained the reasoning behind passage of the new law thusly:

The legislature finds that a small but extremely dangerous group of sexually violent predators exist who do not have a mental disease or defect that renders them appropriate for involuntary treatment pursuant to the treatment act for mentally ill persons defined in [Kansas’s existing civil commitment statute] which is intended to provide short-term treatment to individuals with serious mental disorders and then return them to the community. In contrast to persons appropriate for civil commitment . . . sexually violent predators generally have antisocial personality features which are unamenable to existing mental illness treatment modalities and those features render them likely to engage in sexually violent behavior. The legislature further finds that sexually violent predators' likelihood of engaging in repeat acts of predatory sexual violence is high. The existing involuntary commitment procedure pursuant to the treatment act for mentally ill persons defined in [the existing statute] is inadequate to address the risk these sexually violent predators pose to society. The legislature further finds that the prognosis for rehabilitating sexually violent predators in a prison setting is poor, the treatment needs of this population are very long term and the treatment modalities for this population are very different [from] the traditional treatment modalities for people appropriate for commitment under the treatment act for mentally ill persons defined [in the current statute] therefore a civil commitment procedure for the long-term care and treatment of the sexually violent predator is found to be necessary by the legislature. K.S.A. 59-29a01.

In 1994, a Kansas district attorney filed a motion seeking civil commitment of a sixty year old repeat sexual offender named Leroy Hendricks a month before he was to be released from prison at the completion of his ten-year sentence for sexual abuse of his two stepchildren.  After a jury trial on the issue, Hendricks was found to be a Sexually Violent Predator and committed to a state run psychiatric hospital.  Hendricks appealed claiming that the K.S.A. 59-29a01,Kansas’s SVP law, was unconstitutional because it deprived him of his substantive due process rights, as well as subjecting him to Double Jeopardy and constituting an unconstitutional ex post facto law. 
http://video.msn.com/video.aspx?mkt=en-us&brand=msnbc&vid=942bd6d6-4793-4950-81e4-7289c254f561

B.  Without reaching Hendrick’s Double Jeopardy and ex post facto claims, the Kansas Supreme Court found that K.S.A. 59-29a01 unconstitutionally violated Hendricks’s substantive due process rights because it allowed commitment of individuals with a mental abnormality while previous substantive due process jurisprudence had only allowed mentally ill individuals to be committed. K.S.A. 59-29a01 read in relevant part:

(a) 'Sexually violent predator' means any person who has been convicted of or charged with a sexually violent offense and who suffers from a mental abnormality or personality disorder which makes the person likely to engage in the predatory acts of sexual violence.

(b) 'Mental abnormality' means a congenital or acquired condition affecting the emotional or volitional capacity which predisposes the person to commit sexually violent offenses in a degree constituting such person a menace to the health and safety of others.

After analyzing two United States Supreme Court cases involving civil commitment (Addington v. Texas, 441 U.S. 418 (1977), and Foucha v. Louisiana, 504 U.S. 71(1993)) and finding that they required a finding of mental illness before a civil commitment satisfied substantive due process, the Kansas Supreme Court invalidated the Kansas law.

We must determine if Hendricks is denied substantive due process based not on his lack of character but, rather, on the merits of his challenge. Mental illness is defined in K.S.A. 59-2902(h) as meaning any person who: "(1) is suffering from a severe mental disorder to the extent that such person is in need of treatment; (2) lacks capacity to make an informed decision concerning treatment; and (3) is likely to cause harm to self or others." Here, neither the language of the Act nor the State's evidence supports a finding that "mental abnormality or personality disorder," as used in 59-29a02 (a), is a "mental illness" as defined in 59-2902(h). Absent such a finding, the Act does not satisfy the constitutional standard set out in Addington and Foucha.  Justice White, speaking for the majority of the United States Supreme Court in Foucha, clearly stated that to indefinitely confine as dangerous one who has a personality disorder or antisocial personality but is not mentally ill is constitutionally impermissible. Similarly, to indefinitely confine as dangerous one who has a mental abnormality is constitutionally impermissible. In re Care & Treatment of Hendricks, 259 Kan. 246, 264 (Kan. 1996) (internal citations omitted).

C.  After granting certiori the United States Supreme Court reversed the Kansas courts, holding that K.S.A. 59-29a01 did not offend the Constitution.

The challenged Act unambiguously requires a finding of dangerousness either to one's self or to others as a prerequisite to involuntary confinement. Commitment proceedings can be initiated only when a person "has been convicted of or charged with a sexually violent offense," and "suffers from a mental abnormality or personality disorder which makes the person likely to engage in the predatory acts of sexual violence." Kan. Stat. Ann. § 59-29a02 (a) (1994).  The statute thus requires proof of more than a mere predisposition to violence; rather, it requires evidence of past sexually violent behavior and a present mental condition that creates a likelihood of such conduct in the future if the person is not incapacitated.  As we have recognized, "previous instances of violent behavior are an important indicator of future violent tendencies." Heller v. Doe, 509 U.S. 312, 323, 125 L. Ed. 2d 257, 113 S. Ct. 2637 (1993); see also Schall v. Martin, 467 U.S. 253, 278, 81 L. Ed. 2d 207, 104 S. Ct. 2403 (1984) (explaining that "from a legal point of view there is nothing inherently unattainable about a prediction of future criminal conduct"). A finding of dangerousness, standing alone, is ordinarily not a sufficient ground upon which to justify indefinite involuntary commitment. We have sustained civil commitment statutes when they have coupled proof of dangerousness with the proof of some additional factor, such as a "mental illness" or "mental abnormality."  See, e.g., Heller, supra, 314-315 (Kentucky statute permitting commitment of "mentally retarded" or  “mentally ill" and dangerous individual); Allen v. Illinois, 478 U.S. 364, 366, 92 L. Ed. 2d 296, 106 S. Ct. 2988 (1986) (Illinois statute permitting commitment of "mentally ill" and dangerous individual); Minnesota ex rel. Pearson v. Probate Court of Ramsey Cty., 309 U.S. 270, 271-272, 84 L. Ed. 744, 60 S. Ct. 523 (1940) (Minnesota statute permitting commitment of dangerous individual with "psychopathic personality"). These added statutory requirements serve to limit involuntary civil confinement to those who suffer from a volitional impairment rendering them dangerous beyond their control. The Kansas Act is plainly of a kind with these other civil commitment statutes: It requires a finding of future dangerousness, and then links that finding to the existence of a "mental abnormality" or "personality disorder" that makes it difficult, if not impossible, for the person to control his dangerous behavior. Kan. Stat. Ann. § 59-29a02(b) (1994). The precommitment requirement of a "mental abnormality" or "personality disorder" is consistent with the requirements of these other statutes that we have upheld in that it narrows the class of persons eligible for confinement to those who are unable to control their dangerousness. Kansas v. Hendricks, 521 U.S. 346, 357-59 (1997).

D.  Five years after Hendricks, the United States Supreme Court took up the constitutionality of K.S.A. 59-29a01 once more in Kansas v. Crane.  In Crane, the Court considered what showing of reduced volitional capacity was required before an SVP could be committed.  While the Kansas Supreme Court had interpreted Hendricks as requiring a showing of complete lack of volitional control, the state argued that no such showing should be required.  The Supreme Court chose an intermediate path, requiring that some such showing was required, but that it need not be absolute:

[In Hendricks], we did not give to the phrase “lack of control” a particularly narrow or technical meaning. And we recognize that in cases where lack of control is at issue, “inability to control behavior” will not be demonstrable with mathematical precision. It is enough to say that there must be proof of serious difficulty in controlling behavior. And this, when viewed in light of such features of the case as the nature of the psychiatric diagnosis, and the severity of the mental abnormality itself, must be sufficient to distinguish the dangerous sexual offender whose serious mental illness, abnormality, or disorder subjects him to civil commitment from the dangerous but typical recidivist convicted in an ordinary criminal case.  Kansas v. Crane, 534 U.S. 407, 413 (2002).

2. The Social Science
A.  The role of social science
In considering SVP laws, the courts have considered how much deference social scientific consensus and knowledge are due.  In its decision in the Hendricks, case, the Kansas Supreme Court criticized a Washington State Supreme Court ruling upholding Washington’s SVP Act:

With regard to the requirement of mental illness, the [Washington] court reasoned that the terms "mental abnormality," "mental disorder," and "mental illness" are largely synonymous. Thus, the reasoning continues, by defining a sexually violent predator as a person who suffers from a mental abnormality or personality disorder, the requirement is satisfied. At the heart of the court's reasoning is a selective and inconsistent use of the American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders (3d ed. rev. 1987). On the one hand, the court reasons that because "anti-social personality disorder" is defined in the DSM-III-R, it is a mental disorder, and on the other hand, that "mental abnormality" is a mental disorder, notwithstanding that it is not defined as such in the DSM-III-R. Simply stated, mental illness means whatever the Washington court says it means. [internal citations omitted].

The United States Supreme Court was much less deferential to social scientists in its Hendricks decision:

Hendricks nonetheless argues that our earlier cases dictate a finding of "mental illness" as a prerequisite for civil commitment, citing Foucha, and Addington. He then asserts that a "mental abnormality" is not equivalent to a "mental illness" because it is a term coined by the Kansas Legislature, rather than by the psychiatric community. Contrary to Hendricks' assertion, the term "mental illness" is devoid of any talismanic significance. Not only do "psychiatrists disagree widely and frequently on what constitutes mental illness," Ake v. Oklahoma, 470 U.S. 68, 81, 84 L. Ed. 2d 53, 105 S. Ct. 1087 (1985), but the Court itself has used a variety of expressions to describe the mental condition of those properly subject to civil confinement. See, e.g., Addington, 441 U.S. at 425-426 (using the terms "emotionally disturbed" and "mentally ill"); Jackson, 406 U.S. at 732, 737 (using the terms "incompetency" and "insanity"); cf. Foucha, 504 U.S. at 88 (O'CONNOR, J., concurring in part and concurring in judgment) (acknowledging State's authority to commit a person when there is "some medical justification for doing so"). Indeed, we have never required State legislatures to adopt any particular nomenclature in drafting civil commitment statutes. Rather, we have traditionally left to legislators the task of defining terms of a medical nature that have legal significance. Cf. Jones v. United States, 463 U.S. 354, 365, n. 13, 77 L. Ed. 2d 694, 103 S. Ct. 3043 (1983). As a consequence, the States have, over the years, developed numerous specialized terms to define mental health concepts. Often, those definitions do not fit precisely with the definitions employed by the medical community. The legal definitions of "insanity" and "competency," for example, vary substantially from their psychiatric counterparts. See, e.g., Gerard, The Usefulness of the Medical Model to the Legal System, 39 Rutgers L. Rev. 377, 391-394 (1987) (discussing differing purposes of legal system and the medical profession in recognizing mental illness). Legal definitions, however, which must "take into account such issues as individual responsibility . . . and competency," need not mirror those advanced by the medical profession. American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders xxiii, xxvii (4th ed. 1994).

B. Predictions of future dangerousness
How effective of a prediction of future dangerousness can social science make?  The answer to this question inevitably informs the factual debate over the appropriateness of SVP laws.  The more accurate a prediction that an SVP is likely to reoffend upon release, the more persuasive the factual argument for incapacitating that individual via civil commitment becomes.  Thus, social scientists have set about making estimates of future dangerousness based on the characteristics of an offender.
 
One approach to predicting future dangerousness is to allow an expert psychological evaluator to interview an SVP and predict his future dangerousness.  However, the efficacy of such predictions has been evaluated by studies that track released offenders and compare their actual rates of recidivism with the predictions made by evaluators.  When such studies were conducted, they arrived at an unexpected result: expert psychological evaluators were only slightly better predictors of future dangerousness than methods employing random chance.  See R. Karl Hanson & David Hanson, Improving Risk Assessments for Sex Offenders: A Comparison of Three Actuarial Scales, 24 Law and Human Behavior 119, 119 (2009).

Seeking to improve the accuracy of these predictions, social scientists and correctional officials sought an approach that would remove the subjectivity of expert evaluators.  By collecting demographic data on actual SVPs and looking for correlations between these demographic data and the actual recidivism rate of the SVPs, social scientists developed a number of tools for predicting future dangerousness.  Since these tools were based on demographic data, they were termed ‘actuarial’ predictors.  Most prominent among these tests is the STATIC-99, a ten-item survey, asking such questions as whether SVP has ever been convicted of a non-sexual violent offense or has lived with an intimate partner for more than two years. 

Actuarial predictors are current employed by more than 35 states, and have demonstrated predictive power as to the likelihood that an individual will engage in a repeat offense.  However, these predictions necessarily remain imperfect, and some questions remain unanswered.  For example, a 2009 study demonstrated that evaluators on opposing sides of a case employing the STATIC-99 reach statistically different results favorable to their respective sides.  See Daniel C. Murrie et al,  Rater (dis)Agreement on Risk Assessment Measures in Sexually Violent Predator Proceedings,  Evidence of Adversarial Allegiance in Forensic Evaluation?,  15 Psychology Public Policy and Law 19 (2009). 

C.  Scientific Criticisms of the Legal Approach
Some scientists, however, have leveled criticism at the involvement of social science in actuarial predictions of future dangerousness.  For example, one critic has pointed out the mismatch between the social political judgments called for in the courtroom, and the fact-based judgments employed by scientists:

The judgments made by professionals in SVP cases are largely political: they reflect societal value judgments, and they are vulnerable to the pressures of the political process. [Mental health] professionals judge whether the individual meets the legal criteria for a diagnosis based on mental condition and level of risk. These are questions that involve medical and psychological judgment and the application of legal thresholds. SVP experts are asked to calibrate their estimates against the legal thresholds for risk, which are expressed qualitatively using terms such as "likely" or "highly likely". Because these legal terms are so indeterminate, professionals must make political judgments that determine the balance between public safety and individual liberty. . . .
[M]mental health professionals decide whether individuals meet the legal mental disorder standard, which requires them to judge whether the disorder entails "serious difficulty in controlling behavior". This standard is apparently intended to invoke some notion of volitional dysfunction, a medically indeterminate concept of notorious opacity. To apply this standard professionals must take into account social and expert judgments.  Eric Janus, Sexually Violent Predator Laws: Psychiatry in Service to a Morally Dubious Enterprise, 364 Lancet 50 (2004).

However, social-science based predictions of future dangerousness are likely a fact of legal life in the United States for the foreseeable future.

While actuarial (statistical) prediction is not an ideal solution to the prediction of dangerousness, the approximate error rate of group predictions is known. Policy decisions about the cost of errors over predicting and under predicting dangerousness can be made in light of known probabilities.  The science of risk prediction is imperfect, however. . . actuarial methods are limited because offenders in any study group may vary on factors not measured.

Further Study:

1. Andrew Harris, Amy Phenix, R. Karl Hanson, & David Thornton.  Available at http://www.static99.org/pdfdocs/static-99-coding-rules_e.pdf.
2. Kansas v. Hendricks,  521 U.S. 346, 357-59 (1997).
3. In re Care & Treatment of Hendricks, 259 Kan. 246, 264 (Kan. 1996).
4. Kansas v. Crane, 534 U.S. 407, 413 (2002).
5. Eric Janus, Sexually Violent Predator Laws: Psychiatry in Service to a Morally Dubious Enterprise, 364 Lancet 50 (2004).
6. R. Karl Hanson & David Hanson, Improving Risk Assessments for Sex Offenders: A Comparison of Three Actuarial Scales, 24 Law and Human Behavior 119, 119 (2009).