Assessing Violent Recidivism in Sexual Offenders
Hollida Wakefield and Ralph Underwager*
ABSTRACT: Forensic and clinical psychologists have long been asked to make
predictions about violence, despite the fact that, in the past, such predictions
have been notoriously inaccurate. Several states now have sexual predator
laws which require predictions to be made concerning the likelihood of recidivism.
Since the U.S. Supreme Court in Kansas v. Hendricks (1997) upheld Kansas's
sexual predator laws, such requests are likely to increase in the future. Fortunately, there is now ongoing empirical research which has improved
psychologists' ability to predict violence in high risk groups. Several
schemes for predicting violence are in the process of research and development.
Learn How to Become - Prioritizing Your Mental Health at Work
Strategies and Tips for Improving Your Daily Well-Being
Workplace stress is not uncommon, left unchecked, it can lead to bigger problems, such as depression and anxiety. Provided is resource with online education and tips on how to make mental health a priority when on the job as well as information on how some organizations have made good mental health a part of their culture.
visit here to learn more >>
Programs are offered here as a community resource, and are not an endorsement of services or guarantee of treatment outcomes.
The Sexual Predator Laws
In the 1980s, several states began to pass sexual predator statutes that
required sexual offenders judged likely to reoffend to be civilly committed
until they were judged to be no longer at risk. The constitutionality of
these statutes was challenged in Kansas vs. Hendricks and on June 23, 1997
the U.S. Supreme court handed down a 5 to 4 decision confirming the constitutionality
of the statute. Currently there are sexual predator statutes in several
states, including Arizona, California, Illinois, Kansas, Minnesota, North
Dakota, New Jersey, Washington, and Wisconsin, and Delaware and Missouri
have prefiled bills along the same lines. Other states are likely to pass
such laws; 45 states and territories filed briefs supporting the Kansas law
(Gordon, 1998) and Michigan and New York have developed similar legislation
(Doren, 1998). The difficulties in predicting dangerousness have not stopped
courts from using such predictions in sexual offender commitment proceedings
and it is well-established that there is no constitutional barrier to using
such predictions in legal proceedings, including those that result in loss
of liberty (Janus & Meehl, 1997).
The sexual predator laws are applied to those who are about to be released
from prison following the completion of their sentences. To commit an individual,
the state must prove that (1) the person has committed sexual offenses in
the past, (2) the person currently suffers from a mental disorder, and (3)
the person is likely to commit a sex offense in the future. The civil commitment
follows a jury trial in which the jury finds that the person is likely to
engage in predatory acts of sexual violence in the future. Common to the
various sexual predator laws is the requirement that mental health professionals
assess the degree of sexual offense risk in thousands of individuals (Doren,
Predicting Recidivism and Base Rates
The prediction of future violence is difficult, complex, and controversial,
and psychologists and psychiatrists do not have a good track record in making
accurate predictions. But since John Monahan's (1981) influential book on
predicting violent behavior, there has been a great deal of research in
this area resulting in improvement in the ability of clinicians and researchers
to make these predictions (Monahan, 1996; Webster, Harris, Rice, Cormier,
& Quinsey, 1994).
The fundamental problem is that in the general population, violent behavior
is a low frequency event. Attempting to predict events in a population with
a low antecedent probability leads to an unacceptable level of false positives.
If the base rate for violence in a given population is very low, then the
most accurate prediction is always to predict that a given individual will
not be violent. Any assessments of individual cases will produce less accurate
results over the long run.
When statistical methods are applied to a population with a higher frequency
of violent behavior, i.e., the prison population or those with a history
of violence, more reasonable predictions can be made. Therefore, recent
research on high frequency violence populations indicates that the accuracy
of predicting future violent behavior can be improved over chance by the
use of actuarial methods.
This includes sexual violence since the research on predicting violent recidivism
in general is relevant to predicting sex offender recidivism. In the procedures
for assessing violent recidivism, sex offenses have been included in the
category of violent recidivism. The sexual offense does not have to involve
actual physical violence. Webster, Douglas, Eaves and Hart (1997) state
that "all sexual assaults should be considered violent behaviour"
(p. 25). Boer, Wilson, Gauthier, and Hart (1997) define sexual violence as
"actual, attempted, or threatened sexual contact with a person who
is nonconsenting or unable to give consent" (p. 328). Webster et al.
(1994) include all sex offenses in their sample. Hanson and Bussière
(1996), however, conclude on the basis of their meta-analysis that sexual
recidivism is best predicted by a different set of factors, which includes
The base rate for sexual recidivism for certain offenders is high enough
that an actuarial prediction method can improve the accuracy of prediction
when the definition of recidivism is in keeping with the sexual offender
commitment laws. The recidivism rate, however, differs among various studies.
Hanson and Bussière (1998) report that only a minority (13.4%) of
their total sample of 23,393 subjects from their meta-analysis committed
a new offense within the average 4- to 5-year follow-up period. Even with
studies with thorough record searches and follow-up periods of 15 to 20
years, the recidivism rate never exceeded 40%. A universal finding in the
literature is that incest offenders have the lowest rates of reoffending.
In contrast, Doren (1998), in a review of the research, reports that the
true recidivism base rate over 25 years for extrafamilial sexual abusers
is 52% and for rapists is 39%. Doren, who is involved with the sexual predator
program at Mendota Mental Health Institute in Wisconsin, uses the recidivism
rates from Prentky, Lee, Knight, and Cerce (1997). This is an extremely
high risk sample. The Prentky, et al. sample consisted of 251 men who were
committed to the Massachusetts Treatment Center for Sexually Dangerous Persons
(MTC). Persons who were charged after being released from MTC and persons
who were residents at MTC but were previously discharged, reoffended and
were recommitted were included in the sample. Also, a charge, not a conviction,
was used as the index of reoffense.
In addition, the figures of 39% and 52% are estimates from the survival analysis;
the percentage of new offenses at the end of the study period (25 years)
was 26% for rapists and 32% for child molesters. Doren maintains that the
survival analysis provides a more accurate approximation of actual recidivism.
Clinical Versus Actuarial Predictions
There are two general approaches to risk assessment — actuarial and clinical.
In the clinical approach, the clinician makes a judgment about the person
based on his or her training, experience, and clinical impressions of the
person being assessed. Actuarial methods are the quantified, impartial, and
systematic use of factors from the file or history of the person. The essential
requirement is that the data are quantified, and statistical procedures based
on the laws of probability are employed. The prediction thus rests on base
rates, known probabilities, specific historical factors, and past conduct.
A clinical interview alone cannot be used to make an accurate prediction.
There is no more solidly established fact in the science of psychology than
the superiority of statistical, actuarial procedures over clinical judgments
in making decisions (Dawes, Faust, & Meehl, 1989, 1993). More than 50
years of research shows that in making decisions the information to be relied
upon first is the statistical and actuarial data that are available. Clinical
judgments, based only on personal opinions unsupported by empirical, quantified
data, must be regarded with considerable caution and relied upon over statistical
information only when there is a credible, compelling, and cogent basis
for doing so. The risk prediction schemes that have resulted in improvements
over chance in the ability of psychologists to make accurate predictions
about sexual recidivism have a strong actuarial base.
Risk Factors Associated With Recidivism
Risk factors associated with violent behavior, including sexual offender
recidivism, are both static and dynamic (for a good discussion of this see
Proulx, Pellerin, Paradis, McKibben, Aubut, & Ouimet, 1997). Static
factors that predict recidivism, such as age, offense history, childhood
family factors, cannot be changed. Dynamic factors, such as criminal attitudes
and progress in treatment, are potentially changeable. Research on an actuarial
and more accurate system for predicting the probability that an individual
will engage in violent behavior in the future mainly uses static factors,
although dynamic factors can modify the prediction.
In general, the factors most strongly related to violent and sexual recidivism
include having the characteristics of psychopathy as defined by a high PCL-R
score (i.e. Hare, 1991, 1996, in press; Hart & Hare, in press; Rice,
1997), a history of criminal behavior, and being young. Rice and Harris
(1997) report that the combination of psychopathy, measured by the PCL-R,
and sexual deviancy, based on phallometric test results, resulted in the
highest recidivism rate in their sample of sex offenders.
Hanson and Bussière (1996) report on the factors that predicted sexual
offender recidivism in their meta-analysis. This is an important study in
that, at the time of their study, there had been 87 different articles containing
61 different sets of data that dealt with factors that are related to sexual
offender recidivism. Not all of the studies report the same factors.
the meta-analysis allows the factors from the different studies to be considered
as a whole.
As noted above, Hanson and Bussière (1996) found that, overall, the
rate of sexual offense recidivism in the 61 studies they performed their
meta-analysis on was low. When recidivism was defined as any offense (not
just sexual reoffending), the overall recidivism rate was higher, 36.3%. They also report that sexual offenders classified as "mentally disordered
sexual offenders" under the sexual psychopath laws of several states
were only slightly more likely to reoffend than other sexual offender groups.
Their review also suggested that sexual recidivism is best predicted by
a different set of factors than those that predict general or nonsexual
violent recidivism. Although general criminological variables, such as age
and prior offenses, showed some relationship with sexual recidivism, they
report that the strongest predictors were variables related to sexual deviance.
No single factor was sufficiently related to recidivism to justify its use
in isolation, but the authors presented predictor factors that had a statistically
significant relationship to recidivism. They do not present an organized
scheme for considering and weighing the factors and making a prediction.
Risk factors for sexual recidivism identified by Hanson and Bussière's
(1996, 1998) meta-analysis are shown in Table 1 (listed in the order of
the strength of the correlations between the risk factor and recidivism).
Risk Factors for Sexual Recidivism
(Hanson & Bussière,
||Plethysmograph preference for children (r = .32)
||Scale 5 (Masculinity/Femininity) of the MMPI (r = .27)
||Severely disordered (r = .25)
||Deviant sexual preferences (pretreatment) (r = .22)
||Prior sexual offenses (r = .19)
||Any personality disorder (r = .16)
||Negative relationship with mother (r = .16)
||Scale 6 (Paranoia) of the MMPI (r = .16)
||Low motivation for treatment (r = .15)
||Victim stranger (r = .15)
||Antisocial personality disorder (r = .14)
||Plethysmograph preference for boys (r = .14)
||Victim female child (r = -.14)
||Prior offenses (any nonsexual) (r = .13)
||Anger problems (r = .13)
||Age (r = -.13)
||Early onset of sexual offending (r = .12)
||Prior offenses (r = .12)
||Victim related child (r = -.11)
||Single (never married) (r = .11)
||Diverse sex crimes (r = .10)
|(Note that 13,
16, and 19 are negative
Some factors that clinicians have assumed to be related to sexual offense
recidivism, such as denial of the sex offense, empathy for victims, a history
of being sexually abused as a child, and general psychological problems
were not found to predict sexual offense recidivism (Hanson & Bussière
Violence Prediction Methods
Hanson and Bussière's (1996, 1998) meta-analysis is not a violence
prediction method. They do not present an organized scheme for considering
and weighing the factors and making an actuarial prediction. A number of
the factors correlated with recidivism also overlap with each other and
are therefore likely highly intercorrelated. This means that simply counting
the factors likely produces an overinterpretation because more than one
factor will be counted on the same basis for an individual. For example,
a person with a diagnosis of an antisocial personality disorder will also
be counted for any personality disorder. This inflates the number of factors
counted. It is therefore a mistake to do a risk assessment by simply counting
the number of risk factors that the person may have.
Psychopathy Checklist-Revised (PCL-R)
Psychopathy is defined by Hare (1991):
Psychopathy can be differentiated from other personality disorders on the
basis of its characteristic pattern of interpersonal, affective and behavioral
symptoms. Interpersonally, psychopaths are grandiose, egocentric, manipulative,
dominant, forceful, and cold-hearted. Affectively, they display shallow
and labile emotions, are unable to form long-lasting bonds to people, principles,
or goals, and are lacking in empathy, anxiety, and genuine guilt and remorse.
Behaviorally, psychopaths are impulsive and sensation-seeking, and they
readily violate social norms. The most obvious expressions of these predispositions
involve criminality, substance abuse, and a failure to fulfill social obligations
and responsibilities (p. 3).
Although psychopathy is sometimes confused with antisocial personality disorder,
it must be differentiated from this diagnosis and the relationship between
the PCL-R diagnosis of psychopathy and the DSM diagnosis of APD is an asymmetric
one, at least in forensic populations. On average, about 90% of criminal
psychopaths meet the DSM criteria for APD, but only 20% to 30% of inmates
with APD also meet the PCL-R criteria for psychopathy (Hare, 1991, Hart
& Hare, in press).
The PCL-R has been validated on both male prison inmates and on male forensic
psychiatric patients. It shows very high levels of accurate prediction of
violence and recidivism and adds significantly to the best possible predictive
accuracy that is based solely on criminal history (Annon, undated; Hare,
1996, 1998; Harris, Rice, & Quinsey, 1993; Hart & Hare, in press;
Rice, 1997; Salekin, Rogers, & Sewell, 1996; Webster et al., 1994). The PCL:SV, a research and screening version that takes less time to complete
can be used as a screen for psychopathy in forensic populations. It can
also be used with civil populations (Hare, 1996, 1998).
The PCL-R requires collateral and file information in order to rate the
individual — ratings
cannot be made on the basis of interviews alone. Although the interview
is important for obtaining information about the individual's interpersonal
style, as well as gathering historical information, if an interview is impossible,
valid ratings can be obtained on the basis of collateral information alone
if there is sufficient high-quality information.
The authors recommend that evaluators attend a three-day training workshop
before using the PCL-R/PCL:SV (information on workshops can be obtained
from Robert Hare at (604) 822-3611 or on the webpage www.hare.org).
the PCL-R is used in several of the risk assessment schemes discussed below
it is important for professionals involved in sexual predator assessment
to be proficient in its use.
Rapid Risk Assessment for Sexual Offense Recidivism (RRASOR)
Hanson (1997) recently developed the Rapid Risk Assessment for Sexual Offense
Recidivism (RRASOR). The RRASOR, which is based on data from seven studies,
contains four items that are easily scored from administrative records:
prior sexual offenses, age less than 25, extrafamilial victims, and male
victims. The scale showed moderate predictive accuracy (r = .27) and Hanson
believes it is useful as a screening instrument in settings that require
routine assessments of sexual offender recidivism risk (see Table
The Rapid Risk Assessment
for Sexual Offense Recidivism
|Prior Sex Offenses
(Not including index offenses)
1 conviction; 1-2 charges
2-3 convictions; 3-5 charges
4 or more convictions; 6 or more charges
|Age At Release (Current Age)
More than 25
Less than 25
|Relationship to Victim
Violence Prediction Scheme
Webster, Harris, Rice, Cormier, and Quinsey (1994) developed the Violence
Prediction Scheme for assessing dangerousness in high risk men on the basis
of a sample of men committed to the maximum security division of a mental
health center in Canada. The scheme consists of two parts: an actuarial
component based on the Violence Risk Appraisal Guide (VRAG) and a 10-item
clinical scheme called the ASSESS-LIST (antecedent history, self-presentation,
social and psychosocial adjustment, expectations and plans, symptoms, supervision,
life factors, institutional management, sexual adjustment, and treatment
progress). Their sample included men committed for sexual offenses.
Unlike the Hanson and Bussière (1996) analysis, the VRAG lists the
risk factors associated with recidivism and how to assess them, the weights
assigned to each, and provides a table of the relationship between the VRAG
scores and the probability of violent recidivism for 7-year and 10-year
follow-up intervals. There are difficulties in generalizing from this, since
the population made up of referrals in other jurisdictions is not identical
to the population used by Webster et al. (1994). Nevertheless, the VRAG represents
an advance over merely counting risk variables.
The variables included in the VRAG are the Hare PCL-R (psychopathy) score,
elementary school maladjustment, age at index offense, diagnosis of personality
disorder, separation from parents when the person was under age 16, failure
on prior conditional release, criminal history of nonviolent offenses, marital
status of never married, diagnosis of schizophrenia, victim injury in index
offense, history of alcohol abuse, and male victim in index offense. Weights
are assigned to these variables. All of them are related positively except
age, diagnosis of schizophrenia, and victim injury, which are negatively
After assessing and weighing the actuarial factors, the clinical factors
(ASSESS-LIST) are considered. These variables are those that they and others
have found to be related to violent behavior but have not reached the point
where they are added to the actuarial formula. Some of the factors discussed
by Hanson and Bussière (1996) above are included here. Webster et
al. (1994) recommend using these clinical factors in the overall prediction,
but stress using caution in altering the actuarial judgment.
In their 1998 book, however, the authors changed their advice to include
these clinical factors. They now state that there is "an extremely
high probability that clinically adjusted VRAG predictions are less accurate
than unadjusted scores" (Quinsey, Harris, Rice & Cormier, p. 163).
The VRAG shows excellent predictive accuracy (r = .47) for general violent
recidivism (Webster et al., 1994) and less for sexual recidivism (r = .20)
(Rice & Harris, 1997). The authors recommend using it both for general
and for sexual recidivism (Rice & Harris, 1997; Marnie Rice, personal
communication, 8/20/97 & 8/22/97). Some reviewers, however (i.e., Borum,
1996; Robert Hare, personal communication), have urged caution in generalizing
their findings to other populations. But Monahan (1995) states that "the
violence prediction scheme is so far superior to anything previously available
that not to seriously consider its use, at least on an experimental basis,
in other jurisdictions would be a difficult choice to justify" (p. 447).
Sex Offender Risk Appraisal Guide (SORAG)
Although the Webster et al. (1994) have recommended using the VRAG for assessing
risk of recidivism in sex offenders, they have recently developed a variation specifically intended for sex offenders, the Sex Offender Risk Appraisal
Guide (SORAG). This is described in Quinsey, et al., 1998. According to
Marnie Rice (personal communication, 3/16/98), the SORAG only does a little
better than the VRAG with sex offenders and has not been validated on as
many subjects. She states that there is not yet any big advantage to using
it, other than it was designed specifically for sex offenders and the scoring
of some of the variables fits with most people's intuition better than with
The HCR-20 (Webster, Douglas, Eaves, & Hart, 1997) arose out of an attempt
to provide a systematic way for clinicians to perform risk assessments in
a manageable procedure. (HCR-20 refers to the fact that there are 20 items
in three categories, Historical, Clinical, and Risk Management.) The original
HCR-20 was published in 1995 and the authors describe the current version
as a "work in progress" (p. vi.). Borum (1996) notes that the
HCR-20's primary value is as a checklist to prompt the examiner to cover
or consider the major areas of inquiry. Unlike the VRAG, the HCR-20 can
be best characterized as a structured guide to clinical assessment.
A similar instrument designed for use with sex offenders is the Sexual Violence
Recidivism-20 (SVR-20) (Boer, Wilson, Gauthier, & Hart, 1997). The authors
state that the SRV-R risk factors are not intended to be used as an actuarial
scale and they cannot recommend a decision-making algorithm. Instead, "evaluators
should consider the SVR-20 and any other case-specific factors deemed important,
and should integrate them in an unstructured or 'clinical' manner"
Minnesota Sex Offender Screening Tool (MnSOST)
The MnSOST (Minnesota Department of Corrections, 1997), which was developed
in response to the sexual predator laws in Minnesota, is designed to help
identify the most violent offenders and those offenders most likely to reoffend.
The evaluator rates the person on 21 items that are given different scores
and then ranks the inmate's dangerousness on a scale of 1 to 10. The MnSOST
is used to screen inmates for possible referral for civil commitment as
well as for determination of the person's community notification risk level.
The validity scale sample consisted of 256 Minnesota sex offenders released
from prison between 1988 and 1993. Sex offenders arrested for a subsequent
sex offense had higher mean scores than those who were not. The authors
state that with a cut off point of 47 and above, 41 of the 66 offenders
(62%) who had these scores were arrested for a subsequent sex offense. Since
the base rate for rearrest for the entire sample was 41%, the authors state
that the MnSOST provides a 50% improvement over chance in prediction of
There can be no question but that making the most accurate decision possible
serves the welfare and benefit of all involved. The aim to protect the society
from harm is advanced by accurate decisions both in precluding further criminal
acts and in avoiding the social and financial costs of unnecessary imprisonments.
The scientific evidence suggesting that a statistical, actuarial approach
improves accuracy of decisions made should be taken seriously. Given the
current climate of anxiety and fear about crime and the trend toward more
punitive and draconian punishment, including the factor of statistical,
actuarial methods in the decision process may serve both to increase the
accuracy of decisions and avoid potential well intentioned but possibly
erroneous rush to judgment. Given the high cost of imprisonment and the
ever more limited funds available, more accurate decision making will benefit
all by increasing the effectiveness of detention and deterrence at the lowest
Annon, J. (undated). Factors associated with the prediction of possible
violence. Unpublished manuscript.
Boer, D. P., Wilson, R. J., Gauthier, C. M., & Hart, S. D. (1997). Assessing
risk for sexual violence: Guidelines for clinical practice. In C. D. Webster
& M. A. Jackson (Eds.), Impulsivity: Theory, Assessment, and Treatment ()
(pp. 326-342). New York: Guilford.
Borum, R. (1996). Improving the clinical practice of violence risk assessment:
Technology, guidelines, and training. American
Psychologist, 51, 945-956.
Doren, D. M. (1998). Recidivism base rates, predictions of sex offender
recidivism, and the "sexual predator" commitment laws. Behavioral Sciences
Law, 16, 97-114.
Dawes, R. M., Faust, D., & Meehl, P. E. (1989). Clinical versus actuarial
judgment. Science, 243, 1668-1674.
Dawes, R. M., Faust, D., & Meehl, P. E. (1993). Statistical prediction
versus clinical prediction: Improving what works. In G. Keren, & C.
Lewis (Eds.), A Handbook for Data Analysis in the Behavioral Sciences: Methodological
(pp. 351-367). Hillsdale, NJ: Lawrence
Gordon, D. (1998, March). Keeping sex offenders off the streets. State
Hanson, R. K. (1997). The development of a brief actuarial risk scale for
sexual offense recidivism. Department of the Solicitor General of Canada,
Public Works and Government Services Canada, cat. No. JS4-1/1997-4E. (Available
in the publications section of Hanson's home page, http://www.sgc.gc.ca/ehome.htm)
Hanson, R. K., & Bussière, M. T. (1996). Predictors of sexual
offender recidivism: A meta-analysis. Cat. No. JS4-1/1996-4E, Public Works
and Government Services Canada. (Available in the publications section of
Hanson's home page, http://www.sgc.gc.ca/ehome.htm)
Hanson, R. K., & Bussière, M. T. (1998). Predicting relapse:
A meta-analysis of sexual offender recidivism studies. Journal of Consulting
and Clinical Psychology, 66, 348-362.
Hare, R. D. (1991). Manual for the Hare Psychopathy Checklist —
North Tonawanda, New York: Multi-Health Systems, Inc.
Hare, R. D. (1996). Psychopathy: A clinical construct whose time has come. Criminal Justice and
Behavior, 23, 25-54.
Hare, R. D. (1998). Psychopaths and their nature: Implications for the mental
health and criminal justice systems. In E. Simonson & T. Millon (Eds.),
Psychopathy: Concept, Etiology, Epidemiology, and Treatment. New York: Guilford.
Harris, G. T., Rice, M. E., & Quinsey, V. L. (1993). Violent recidivism
of mentally disordered offenders: The development of a statistical prediction
instrument. Criminal Justice and
Behavior, 20, 315-335.
Hart, S. D., & Hare, R. D. (in press). Psychopathy: Assessment and association
with criminal conduct. In D. Stoff, J. Breiling, & J. Maser (Ed.), Handbook
of Antisocial Behavior (), New York:
Janus, E. S., & Meehl, P. E. (1997). Assessing the legal standard for
predictions of dangerousness in sex offender commitment proceedings. Psychology,
Public Policy, and Law, 3, 33-64.
Kansas v. Hendricks (1997). 95 1649 and 95 9075. Supreme Court of the United
States. June 23, 1997.
Minnesota Department of Corrections (1997, February).
Minnesota Sex Offender
Screening Tool (MnSOST). Author.
Monahan, J. (1981). Predicting Violent Behavior: An Assessment of Clinical
Techniques ()(). Beverly Hills, CA: Sage.
Monahan, J. (1996). Violence prediction: the past twenty and the next twenty
years. Criminal Justice and
Behavior, 23, 107-120.
Monahan, J. (1995). Review of The Violence Prediction Scheme: Assessing
Dangerousness in High-risk Men, by C. D. Webster, G. T. Harris, M. E. Rice,
C. Cormier, & V. L. Quinsey. Criminal Justice and
Behavior, 22, 446-447.
Prentky, R. A., Lee, A. F. S., Knight, R. A., & Cerce, D. (1997). Recidivism
rates among child molesters and rapists: A methodological analysis.
Law and Human Behavior, 21, 635-659.
Proulx, J., Pellerin, B., Paradis, Y., McKibben, A., Aubut, J., & Ouimet,
M. (1997). Static and dynamic predictors of recidivism in sexual aggressors.
Sexual Abuse: A Journal of Research and Treatment, 9(1), 7-27.
Quinsey, V. L., Harris, G. T., Rice, M. E., & Cormier, C. A. (1998).
Violent Offenders: Appraising and Managing Risk (). Washington, DC: American Psychological Association.
Rice, M. E. (1997). Violent offender research and implications for the criminal
justice system. American
Psychologist, 52, 414-423.
Rice, M. E., & Harris, G. T. (1997). Cross-validation and extension
of the violence risk appraisal guide for child molesters and rapists.
Law and Human Behavior, 21, 231-241.
Salekin, R. T., Rogers, R., & Sewell, K. W. (1996). A review and meta-analysis
of the psychopathy checklist and psychopathy checklist-revised: Predictive
validity of dangerousness. Clinical
Psychology: Science and Practice, 3,
Webster, C. D., Douglas, K. S., Eaves, D., & Hart, S. D. (1997). HCR-20:
Assessing risk for violence, Version 2. Burnaby, British Columbia: Simon
Webster, C. D., Harris, G. T., Rice, M., Cormier, C., Quinsey, V. L. (1994).
The Violence Prediction Scheme: Assessing Dangerousness in High Risk Men ().
Toronto, Canada: Centre
of Criminology, the University
* Hollida Wakefield and Ralph Underwager are psychologists at the
Institute for Psychological Therapies,
13200 Cannon City Boulevard,
Northfield, MN 55057-4405.
A version of this paper was first presented at the 14th Annual Symposium
of the American College of Forensic Psychology, San Francisco, California,
May 3, 1998.
ISSUES IN CHILD ABUSE
ACCUSATIONS Volume 10, 92-98