Reliability and Validity of the Static-2002 Among Adult Sexual Offenders With Reference to Treatment Status

• Author: Edward J. Peacock,Kevin T. Hansen,Howard E. Barbaree,Calvin M. Langton,Leigh Harkins
• DOI: 10.1177/0093854806296851
• Published date: 01 May 2007
• Date: 01 May 2007
• Subject Matter: Articles

616

RELIABILITY AND VALIDITY OF THE

STATIC-2002 AMONG ADULT SEXUAL

OFFENDERS WITH REFERENCE TO

TREATMENT STATUS

CALVIN M. LANGTON

University of Toronto and the Centre for Addiction and Mental Health, Toronto, Canada

University of Nottingham and Peaks Unit, Rampton Hospital, Nottinghamshire

Healthcare Trust, Nottinghamshire,UK

HOWARD E. BARBAREE

KEVIN T. HANSEN

University of Toronto and the Centre for Addiction and Mental Health, Toronto, Canada

LEIGH HARKINS

University of Birmingham, Birmingham, UK

EDWARD J. PEACOCK

Correctional Service of Canada, Ontario, Canada

Psychometric properties were reported for the Static-2002 using a sample of 464 sexual offenders followed for an average of

5.9 years after release. Correlations between theoretically relevant variables (using indices of lifestyle instability, sexual

deviance, psychopathy) and the Static-2002 content area subscales reflected the conceptual emphasis underlying the content

areas. An exploratory factor analysis revealed a five-factor solution generally congruent with the conceptual structure of the

Static-2002. Survival analyses indicated not all of the content areas had incremental validity in the prediction of either sexual

or any violent recidivism. Moderate-to-high levels of predictive accuracy for total score were achieved for treatment com-

pleters, dropouts, and refusers across four recidivism outcomes. Three risk categories were identified with significantly dif-

ferent rates of both sexual and any violent recidivism; failure rates and likelihood ratios for these risk categories over 5-, 7-,

and 10-year follow-up periods were reported.

Keywords: sexual offenders; recidivism prediction; actuarial risk assessment; Static-2002; Static-99; RRASOR; treatment

completion; treatment refusal; refusers; treatment dropout; dropouts; reliability; validity

The past two decades have seen marked progress in the assessment of risk and predic-

tion of criminal behavior within forensic mental health (Doren, 2002; Hanson, 2005;

Monahan et al., 2001). Research on static factors such as an offender’s developmental

history and prior criminal convictions has led to the promulgation of measures of sexual

offender recidivism risk that predict postrelease offenses with moderate levels of accuracy

(Hanson, Morton, & Harris, 2003; Harris, et al., 2003). These instruments represent what

Otto and Heilbrun (2002) have referred to as “forensically relevant instruments,” address-

ing as they do clinical issues or constructs in the context of an evaluation that are relevant

to legal proceedings (i.e., recidivism or violence risk). However, Otto and Heilbrun

included specific sexual offender risk assessment instruments among a number of forensi-

cally relevant instruments that they charge have been promulgated for use in applied con-

texts without adequate research and development. Of particular concern, Otto and Heilbrun

noted that comprehensive test manuals, as required by the Standards for Educational

CRIMINAL JUSTICE AND BEHAVIOR, Vol. 34,No. 5, May 2007 616-640

DOI: 10.1177/0093854806296851

© 2007 American Association for Correctional and Forensic Psychology

Langton et al. / RELIABILITY AND VALIDITY OF THE STATIC-2002 617

and Psychological Testing (American Educational Research Association, American

Psychological Association, & National Council on Measurement in Education, 1999), are

lacking for some risk assessment instruments and appropriate cross-validation research is

not completed for others. Given the potential for misuse, and the threat to the credibility of

forensic psychology as a professional discipline that is posed by the use of approaches that

have not been validated, it is unsurprising that Otto and Heilbrun called for immediate

redress. Although the absence of comprehensive test manuals remains an outstanding con-

cern for many of these instruments, issues of reliability and validity have been addressed

for some of them and cross-validation data collected (e.g., Barbaree, Seto, Langton, &

Peacock, 2001; Rice & Harris, 1997; Sjöstedt & Långström, 2001).

The study reported here contributes to the present knowledge base by investigating the

psychometric properties of a recently developed measure of sexual offense recidivism risk,

the Static-2002 (Hanson & Thornton, 2003). The measure was conceived as an improvement

over the Static-99 (Hanson & Thornton, 2000), an actuarial risk assessment instrument that

is used across jurisdictions and research contexts in North America as well as Europe and

Asia (R. K. Hanson, personal communication, June 11, 2003). Hanson and Thornton (2003)

identified four reasons for the revision. First, they wanted to increase the coherence and

introduce conceptual clarity to the measure. This should, in part, assist evaluators in deter-

mining whether variables external to the measure but present as features in a given case

require consideration in a comprehensive risk assessment. The second reason was to

improve the consistency of the scoring criteria. As Hanson and Thornton (2003) noted, the

Static-99 was the product of combining two earlier risk assessment scales, each with its own

scoring rules, complicating training and potentially affecting rater reliability in applied con-

texts. A third reason for a revision was the possibility of rare cases that had to be scored

counterintuitively; Hanson and Thornton (2003) provided the example of an offender whose

score must be lowered when he commits a new sexual offense following an index offense

involving nonsexual violence. The last reason given for undertaking a revision was an inter-

est in improving upon the predictive accuracy of the Static-99.

Development of the Static-2002 followed a similar procedure to that used in the develop-

ment and validation of the Rapid Risk Assessment for Sex Offense Recidivism (RRASOR;

Hanson, 1997) and the Static-99. Multiple data sets were used (10 with a total size of 4,596)

in the scale construction. With the emphasis on ease of data collection and the simplest

empirically based coding methods, Hanson and Thornton (2003) again identified variables

AUTHORS’ NOTE: The authors would like to thank their research assistants, Heidi Marcon, Duyen Luong,

Andrea Dalton, and Michele Adams, as well as the staff of the Warkworth Sexual Behavior Clinic and the

offenders. Sylvia Langton’s assistance in preparing the manuscript is also gratefully acknowledged as is the

coding advice provided by Karl Hanson and David Thornton for issues not addressed in the available coding

instructions for the Static-2002. Particular thanks are also due to Douglas Mossman for kindly providing the

formulae used to calculate the likelihood ratios and their confidence intervals. David Thornton, Dennis Doren,

Douglas Mossman, Jeff Abracen, Curt Bartol, Anne Bartol, and two anonymous reviewers are also thanked for

helpful comments on earlier versions of this article. The authors wish to acknowledge financial support for the

research described in the article from the Correctional Service of Canada, the Centre for Addiction and Mental

Health, and the Ontario Mental Health Foundation. All opinions expressed in the article are the sole opinions

of the authors and do not reflect opinions or policy of the Correctional Service of Canada or any other agency.

Correspondence concerning this article should be addressed to Calvin M. Langton, Peaks Unit, Rampton

Hospital, Retford, Nottinghamshire, DN22 0PD, England; e-mail: calvin.langton@utoronto.ca.