Law in Contemporary Society

A Bridge Too Far: Use & Misuse of Predictions of Future Dangerousness

-- By RyanMcDevitt - 14 Feb 2008


Expert predictions of future dangerousness are used to sentence criminal defendants more severely, civilly commit sexually violent offenders, and even detain juveniles for crimes they have not yet committed. If the science behind violence risk assessment (VRA) or the courts' understanding of it are questionable, then its widespread use would mean that these people's civil liberties are being curtailed drastically and unfairly. What if psychologists can accurately predict future dangerousness with certain instruments, but one prediction method shows equivocal results that one expert interprets as indicating a high risk of recidivism and another interprets to predict innocuousness, and a judge must decide the defendant's fate? The danger of misapplication is just as severe as that of bad science. Three questions about VRA evidence are thus raised: how is it used, how good is the science, and how well do courts understand it?

How Is VRA Used?

Capital Cases

Interestingly, a challenge to several states' application of the death penalty led to the increased use of VRA in capital cases. As a result of a line of cases including Branch v. Texas, several states overhauled their jury instructions in death penalty sentencing to create three factors that the state has to prove to get a death sentence: intentionality, disproportionality (if relevant), and a probabilistic prediction of the defendant's continuing threat to society. Thus, expert testimony on future dangerousness is a practical necessity in capital cases in these states.

Civil Commitment

The most common use of predictions of future dangerousness, though, is in paternalistic measures aimed at protecting society, such as civil commitment and as probable cause for pretrial detention. Sixteen states' civil commitment statutes are modeled on Kansas' Sexually Violent Predator Act, which allows the civil commitment of 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 predatory acts of sexual violence. (K.S.A. 59-29(A)(01)) In practice, many states civilly commit convicted sex offenders deemed dangerous upon their release from prison. Ideally, they are treated and released, but some spend the remainder of their lives in state custody.

How Good Is the Science?

Kansas Sexually Violent Predator Act (Example)

The Kansas Act provides a four-factor test for determining the mental abnormality prong: i) history of sexual offenses, ii) mental abnormality, iii) volitional impairment, iv) as a result of mental abnormality, the individual is likely to engage in...sexual violence. There are problems associated with the application of VRA to each element.

i) History of Sexual Offenses

The first is a question of social policy: a history of sexual offenses seems like a reliable factor, since investigators can simply look at an offender's criminal record, but what about a first-time offender who could recidivate? If the purpose of civil commitment here is to prevent future danger to society, is it justifiable to curtail the offender's rights if he has no known history of sexual offenses?

  • How about "reoffend," "relapse," or even "repeat"? Did you really need to disinter a particularly ugly verb that no one has used in a published English sentence since 1677?

ii) Mental Abnormality

The diagnosis of "mental abnormality" is problematic because diagnoses of sexual disorders are often unreliable, with different psychologists agreeing on diagnoses at a rate little better than chance. (Marshall, cited in Miller)

iii) Volitional Impairment

Next, diagnosis of volitional impairment: the APA and ABA recommended that the volitional criterion be removed from insanity standards because it is "virtually impossible for a mental health professional to measure." (LaFond? , cited in Miller)

iv) Recidivism

Similar difficulty in assessment applies to the recidivism prong; part of the problem is that so many different tests exist to assess it that conflicting results are likely and it is difficult to expect a finder of fact to determine which instrument is best.


Overprediction is the norm. (Marquart, 452; Hart, 696) This is perhaps unsurprising, as so few offenders are truly dangerous that it is almost statistically impossible to underpredict. Evaluators miss some--false negatives--but the base rate of dangerousness is so low that any prediction with less than 100% certainty is likely to overpredict violence.

How Well Do the Courts Understand VRA?

Gatowski Study

One study asked 400 state court judges for their preference between the two dominant standards for admissibility of expert testimony and then evaluated their understanding of and ability to apply the Daubert standard. About half the judges were from states using the Daubert standard; the other half were from states relying either on the previous Frye standard or state rules representing some mixture of the two. The study found that although 91% of the judges preferred the Daubert standard and similarly high percentages found the scientific constructs of falsifiability and error rate to be useful measures of reliability, only 6% and 14%, respectively, clearly understood their meaning. (Gatowski, 445-447)


Chief Justice Rehnquist's concerns about Daubert --" imposes on [judges] either the obligation or the authority to become amateur scientists in order to perform that role"--have been realized.

  • This isn't an essay designed to carry a reader along through the development of an idea. It's a shopping list of arguments made elsewhere.

The Future

Menzies Argument

After two moderately successful attempts at building a better mousetrap (a multidimensional predictive instrument), Menzies came to the conclusion that predictions of dangerousness are too important to the legal system and society to be abandoned altogether, even if the data supporting some of the methodology is equivocal. (Menzies) He advocated that mental health professionals be limited to an advisory rather than determinant role in legal proceedings, because the step up to the greater role risks overreaching the existing science.

  • Excuse me? Why should the reader give a flying anything at all about someone called Menzies? As mentioned above, this isn't an essay anymore, just a cranial blood clot.


Limiting mental health professionals' predictions of future dangerousness to the sentencing phase of criminal trials because the evidence is likely to be more prejudicial than probative would help: obviously, that a person is dangerous generally--particularly when as a result of mental illness--does not necessarily mean they are guilty of the crime for which they have been charged. The increased role of judges as the gatekeepers of admissibility since the Daubert ruling means that improved, clearer legal standards may also be needed to help the judges avoid misapplication of the evidence rules or the standards on which experts are asked to testify.

  • What was there in this thesis that required you to adopt such crabbed means of communication? It isn't as though anything very innovative or doubtful were getting said.

Authorities/Works Cited

Gatowski, Dobbin, Richardson, Ginsburg, Merlino, & Dahir, “Asking the Gatekeepers,” Law & Human Behavior, Vol. 25, No. 5, (October 2001), pp. 433-457

Hart, Webster & Menzies, “A Note on Portraying the Accuracy of Violence Predictions,” Law & Human Behavior, Vol. 17, No. 6, (December 1993), pp. 695-700

Marquart, Ekland-Olson, & Sorensen, “Gazing into the Crystal Ball: Can Jurors Accurately Predict Dangerousness in Capital Cases?” Law & Society Review, Vol. 23, No. 3, (1989), pp. 449-468

Menzies, Webster, & Sepejak, “The Dimensions of Dangerousness…” Law & Human Behavior, Vol. 9, No. 1, (March 1985), pp. 49-70

Menzies et al., “The Dimensions of Dangerousness Revisited,” Law & Human Behavior, Vol. 18, No. 1. (Feb., 1994), pp. 1-28

Miller, Amenta, & Conroy, “Sexual Predator Evaluations,” Law & Human Behavior, Vol. 29, No. 1, (February 2005), pp. 29-54

Daubert v. Merrell Dow Pharmaceutical 509 U.S. 579 (1993)

Frye v. United States 293 F. 1013 (DC Cir. 1923)

Kansas Sexually Violent Predator Act, K.S.A. 59-29a01 et seq. (1994)

  • Why do this sort of thing in a Wiki? In many cases here there were available live links, and you should have made them.

# * Set ALLOWTOPICVIEW = TWikiAdminGroup, RyanMcDevitt


Webs Webs

r6 - 12 Jan 2009 - 23:07:36 - IanSullivan
This site is powered by the TWiki collaboration platform.
All material on this collaboration platform is the property of the contributing authors.
All material marked as authored by Eben Moglen is available under the license terms CC-BY-SA version 4.
Syndicate this site RSSATOM