Hanson Full

8/13/2019 Hanson Full

1/15

Journalof Consulting and linicalPsychology1998,Vol. 66, No. 2,348-362 Copyright 1998bytheAmericanPsychologicalAssociation,Inc.0022-006X/98/S3.00

Predicting Relapse:AMeta-AnalysisofSexual OffenderRecidivismStudies

R. KarlHansonandMoniqueT.BussiereDepartment of theSolicitorGeneral of Canada

Evidence from 61 follow-up studies was examined to identify the factors most strongly related torecidivism among sexual offenders. Onaverage,thesexualoffense recidivism rate was low(13.4%;n =23,393). There were, however, subgroups ofoffenders whorecidivated at highrates. Sexualoffense recidivism was best predicted bymeasuresof sexual deviancy (e.g.,deviant sexual prefer-ences, prior sexual offenses) and, to alesserextent, by general criminological factors (e.g., age,total prior offenses). Those offenders who failed to complete treatment were at higher risk forreoffending than thosewhocompleted treatment.Thepredictorsofnonsexual violent recidivismandgeneral (any) recidivism weresimilar to thosepredictors found among nonsexual criminals (e.g.,prior violentoffenses, age, juveniledeliquency). Our results suggest thatapplied risk assessmentsofsexual offenders shouldconsiderseparately the offender's risk for sexual and nonsexual recidivism.

Assessing chronicity is crucial for clients whose sexual be-haviors have brought them into conflict with the law. Manyexceptional criminal justicepolicies,such as postsentence de-tention (e.g., Anderson&Masters, 1992),lifetime communitysupervision, and community notification, target those sexual of-fenderslikely toreoffend. Clinicians need to judge whether theclient's behaviors are truly atypical of the individual (as theclientwould like us to believe) or whether the client merits avirtuallypermanent label as a sexual offender.

Sexualassaultis aserious social problem,withhigh victim-ization rates among children (10% ofboys and 20% of girls;Peters,Wyatt,&Finkelhor,1986) and adult women(10-20%;Johnson & Sacco, 1995; Koss, 1993a).Given the large numberofvictims, it is not surprisingthat a significant portion(10-25%) of male community samples (e.g., university students,hospitalstaff) admittosexualoffending(Hanson &Scott, 1995;Lisak &Roth, 1988; Templeman& Stinnett, 1991).

One of the simplest and mostdefensibleapproaches torecidi-vismpredictionis toidentify astable patternofoffending. Be-haviorisinfluencedby avarietyof internalandexternal factorsthat can and do change over the life course. Nevertheless, itdoesnotrequireanyspecialexpertisetopredictwithconfidence

R. Karl Hanson and Monique T. Bussiere, Corrections Research. De-partment of the Solicitor General of Canada.

We thank Margaret Alexander, Lita Furby, Gordon Hall, RoxanneLieb, Robert Freeman-Longo, Robert Prentky, Mark Weinrott, andSharon Williams for help in locating articles for this review. The com-ments of Jim Bonta,BillMarshall, Andrew Harris, and RobertMcGrathonanearlier versionarealso appreciated.Aswell,wethank JeanProulx,JohnReddon, and David Thornton for access to their original datasets.

Theviews expressed are thoseof theauthorsand do notnecessarilyrepresentthe views of the Ministry of the Solicitor General of Canada.

Correspondence concerning thisarticleshouldbeaddressedto R.KarlHanson, CorrectionsResearch,Department of the Solicitor General ofCanada, 340 LaurierAvenueWest, Ottawa, Ontario, Canada KlAOPS.Electronic mail may be sent [email protected].

the continuationof any behaviorthai has occurred frequently,in many different contexts, and despite the best efforts tostopit.

Somesexualoffenders reportawell-established, chronic pat-tern of offending (e.g., Abel, Becker,Cunningham-Rathner,Rou-leau, &Murphy, 1987).More typically, however, sexual offend-ers deny recurrent deviant sexual interests or behavior (Ken-nedy & Grubin, 1992; Langevin,1988). In the absence of anestablished pattern, risk assessments need to rely onother,rele-vantinformation.Determining whatis ''relevant''requires theo-retical assumptions about thenatureof sexual offending.

One approach is to assume that sexual offending is similartoother criminalbehavior. Thetypicalcriminalcommitsavari-etyof offenses (e.g., theft, assault,drugs) with relativelylittlespecialization (M. R. Gottfredson & Hirschi,1990). Becausemanysexualoffendersalso engageinnonsexual criminalactivi-ties (Broadhurst & Mailer, 1992; Hanson, Scott, & Steffy,1995),the same factors that predict general recidivism amongnonsexualcriminals may also predict sexual recidivism amongsexual offenders.

Theextensive researchon thepredictionofrecidivismamongnonsexual criminals (Champion, 1994; D. M. Gottfredson &Tonry,1987)has identifiedareliableset ofboth static (histori-cal) and dynamic (changeable) risk factors (e.g., Bonta, 1996;Gendreau, Little, & Goggin, 1996). Specifically, the persistentcriminal tends to beyoung, have unstable employment, abusealcoholanddrugs,holdprocriminalattitudes,andassociatewithother criminals (Gendreau et al., 1996). Thesecharacteristicscan be considered to define a "criminal lifestyle," a conceptsimilar toDiagnosticandStatistical ManualofMental Disor-der s(4th edition) antisocialpersonality" (American Psychi-atric Association, 1994), Hare's psychopathy (Hare etal.,1990), and M. R.Gottfredson andHirschi's (1990) lackofselfcontrol.''

Someevidence, however, suggests that sexualoffending maybedifferent fromother typesofcrime.Although sexualoffend-ers frequently commit nonsexualcrimes, nonsexualcriminals

348


2/15

PREDICTING RELAPSE 349

rarelyrecidivate with sexualoffenses (Bonta & Hanson,1995b;Hanson et al.,1995).As well, many persistent sexual offendersare judged to be low risk by scales designed topredictgeneralcriminal recidivism (Bonta & Hanson, 1995b).

Rather than emphasizing generalcriminologicalrisk factors,sexual offender risk assessments may concentrate on sexualdeviance.Allsexualoffending is, bydefinition, socially deviant,but not all sexual offenders have deviant sexual interests orpreferences. Some date rapists, for example, may prefer consen-sualsexualactivitiesbutmisperceivetheirpartners'sexual inter-est(e.g.,'No'means'yes' )(Hanson&Scott, 1995; Mala-muth& Brown,1994).In contrast, the sexual lives of some boy-objectpedophilesmay be completely focused on their preferredvictim type (Freund &Watson,1991;Quinsey,1986).

Because self-reports are highly vulnerable to self-presenta-tion biases, the assessment of deviant sexual interest is bestsupplemented byother sourcesofinformation, suchas a sexualoffense historyandphallomelric assessment (i.e., direct moni-toring of penile response; see Launay,1994).In general,offend-ers with the most deviant sexual histories tend to show deviantor abnormal sexualinterestsonphallometric assessments(Bar-baree & Marshall, 1989; Freund & Watson, 1991; Quinsey,1984,1986).Specifically, deviant sexual interests are most prev-alent among those who 'Victimize strangers, use overt force,select boy victims, or select victims much younger (or mucholder) than themselves (Barbaree & Marshall, 1989; Freund &Watson, 1991; Quinsey, 1984,1986).

Afurther consideration in the assessment of sexual offendersconcerns symptoms of general psychological maladjustment.Sexual offenders rarely meet diagnostic criteriaformajor mentalillnesses, but they often show signs of low self-esteem, sub-stance abuse problems, andassertivenessdeficits (W. L. Mar-shall, 1996).Muchof thecurrent treatmentandtheory concern-ing sexualoffendingemphasizes poor coping strategies and neg-ative emotional states asprecursors to offending (Laws, 1989,1995; Pithers,Beal,Armstrong,& Petty,1989).There have beenfew attempts, however, to examineempiricallythe relevance ofthese psychological symptoms to sexual recidivism. Empiricallyexamining the assumedrelationshipbetween distress andsexualoffending is important because,fornonsexual recidivism, sub-jective distress haseithernorelationshipor anegative relation-shipwith recidivism(Gendreau etal.,1996).

Once detected, sexual offenders' motivation to change mayalso be related to recidivism. Those offenders who accept re-sponsibility, express remorse, and comply with treatment (goodclinical presentation) should be at lower risk than those whodeny any problems and actively resist change (poor clinicalpresentation).Motivationtochangeisdifficult toassess,how-ever, because thereareclear benefitsto 'appearing''willingtochange, and many sexual offenders have the socialskillsneces-sary togain the confidence of sympathetic clinicians.

In agreement withFurby,Weinrott, and Blackshaw (1989),we believe that group comparisons within follow-up studiesprovide "by far the best sources ofdata for theidentificationofrecidivismrisk factors (Furby et al., 1989, p. 27). The absoluterecidivism rates vary across studies as a result of differences infollow-up periods, definitions, andlocal criminal justice prac-tices. These factorsarecontrolled, however,whentherecidivist

nonrecidivist comparisons are made within a single follow-upstudy.

Previousnarrative reviews have examined a small numberofstudies and risk factors (Furby et al., 1989; Hall, 1990; Quinsey,Lalumiere,Rice,&Harris,1995).Their conclusions have beententative and, at times, contradictory. Rapists were consideredhigh risk by Furby et al. (1989) but not by Hall (1990) orQuinsey, Lalumiere, et al. (1995). Quinsey, Lalumiere,et al.(1995)did, however, report that general criminality (nonsexualoffenses) and sexual deviancy (prior sexoffences, phallometricassessments)predictedsexual recidivism. Noneof thepreviousreviews have considered general psychological adjustment orclinical presentation variables as predictors.

The present study provides a quantitative review of the sexualoffender recidivism literature. All the participants were sexualoffenders, but we examined three types of recidivism: sexual,nonsexual violent, and general (any). Sexual andnonsexualviolentrecidivismwere considered separately, because prelimi-naryevidence suggested that they may be predicted by differentsets of characteristics (Hanson etal., 1995; Marques, Nelson,West,&Day,1994).Quantitative summaries,ormeta-analyses,havebecomeastandard featureofresearchreviewsinpsychol-ogy and medicine (Rosenthal, 1995; Spitzer,1995).Meta-analy-ses have several advantages over the traditional narrative formsofreview: (a) Many studies can be considered simultaneously;(b) large,pooled samples yield high statistical power; (c) nu-merical estimates indicate the relative magnitude of effects; and(d) the generalizability of findingsacrossstudies can be tested.

It was expected that thebestpredictor of sexual offense recid-ivism would be ahistoryof sexual deviancy. On thebasisofpreviousreviews (Quinsey, Lalumiere,etal.,1995),it wasalsoexpected that criminal lifestyle variables would berelated toboth sexual and nonsexual recidivism. The relevance of the otherfactors, namely,psychological symptoms andclinicalpresenta-tion, wasless clear. Psychological symptoms havebeenunre-lated to recidivism among general criminal populations (Gen-dreau et al.,1996),but sexual offending may be a special case.The clinical presentation variables may also havelittlepredictivevalue giventhedifficulty identifying sincereremorseandgenu-inemotivationto change as well as the active debate concerningtheefficacy oftreatmentfor sexualoffenders (W. L. Marshall &Pithers, 1994; Quinsey, Harris, Rice,& Lalumiere,1993).

MethodSample

Computersearchesofboth PsycLIT and theNational CriminalJusticeReference System were conducted using the following key terms:sex(ual) offender, rape, rapist, child molester,pedophile, pedophilia,exhibitionist, exhibitionism, sexualassault,incest,voyeur,frotteur,inde-cent exposure, sexual deviant, paraphilia(c),predict,recidivism,recidi-vist,recidivate,reoffend, reoffense,relapse,andfailure.Reference listswere searched for additional articles. Finally, letters were sent to 32established sexual offenderresearchersrequestingoverlookedor as yetunpublishedarticlesordata.

Tobeincluded, studiesneededtofollowup asampleof sex offenders;report recidivism information for sexual offenses, nonsexual violentoffenses, or any reoffenses; and includesufficient statistical informationto calculate the relationship between arelevant offender characteristic


3/15

350 HANSON AND BUSSIEREandrecidivism. Further description of the selection criteria can befoundin Hanson andBussiere (1996) and in the complete coding manual,which isavailableon request from the authors.

Asof December 31,1995,our search yielded 87 usable documents(e.g., published articles,books, government reports, unpublished pro-gram evaluations, conference presentations). When the same data setwas reported in several articles, all the resultsfrom these articles wereconsidered to come from the same study. Consequently, the 87 documentsrepresent 61 different studies (countryoforigin: 30 United States, 16Canada, 10 United Kingdom, 2 Australia, 2 Denmark, 1 Norway; 45%unpublished; produced between 1943 to 1995, with median of 1989;meansamplesize of 475, median of 198, range of 12-5,000).

Most of thestudies examined mixed groupsofadultsexual offenders(.55 mixed offense types,6childmolcstersonly;52samplesofadults,6 adolescents only; 3 both adolescents and adults). The offenders camefrom institutions (48%),thecommunity (25%),orboth(27%). Nine-teenstudies focused exclusivelyoncorrectional samples, 1 1 on samplesfrom secure mental health facilities, and the remaining from othersources(private clinics, courts, mixture ofsources).Approximately onehalf of the samples (48%) were from sexual offender treatment pro-grams. When demographic information was presented, the offenderswere predominantly Caucasian (27 of 28 studies) and of lowersocioeco-nomic status (27 of 29 studies).The most common measures of recidivism werereconviction(84%),arrests (54%), self-reports (25%), andparoleviolations (16%). Multi-ple indexes of recidivism were used in 27 of 61 studies (44%). Themost common sourcesofrecidivism information were national criminaljusticerecords(41%),stateorprovincial records (41%), records fromtreatment programs(29%),and self-reports (25%). Other sources (e.g.,child protection records) were used in 25% of the studies. In 43% ofthe studies,multiplesources were used.In 15 studies,thesourceof therecidivism information was not reported. The reportedfollow-upperiodsranged from6monthsto 23years (median= 48months; mean =66months).

Quality ofStudiesAn important concern inmeta-analytic reviews is the quality of the

studiesreviewed.This issue was less of aconcernin the current review,however, because all studies used the best available design (i.e., thematched, longitudinal follow-up design;Furby et al.,1989). Neverthe-less, variationin theassessmentof thepredictor variablesand ofrecidi-vismcould affect theresults.Inmostcases,thepredictor variables weresufficiently explicit that there was little concern about reliability orvalidity (e.g., age, criminal history, victim type). There was, however,enoughvariabilityin therecidivism measures tojustify furtheranalysis.Consequently, the thoroughnessof the recidivism search wascoded foreachstudy usinga7-point scale ranging from (1)questionable methods(e.g., mail-in questionnaires only)/inadequate follow-up periods (10years). Eachstudy wasrated by the two authors,anddifferences were resolved bydiscussion. In a sample of 20 independent ratings,theraterreliabilitywas .72,using equation ICC(2,1) from Shrout andFleiss(1979).

Theratingsof theadequacyof therecidivism information ranged from3 to 7 (M =4.6, SD = 1.0), indicating overall acceptable levels ofdiligence in identifying recidivists.No studies used only self-report orwholly inadequate recidivism detection methods.

CodingProcedureEach document wascodedseparately by R. Karl Hanson and Monique

T.Bussierc usingacoding manual. When disagreement occurred, mostinvolved calculationerrors that were immediatelv corrected. In rare

casesofdifferencesofinterpretation, advicewassoughtfrom colleaguesfamiliar withforensic meta-analytic reviews.

Onlyone finding of each type of predictor variable was coded fromany one study (data set). Given several related variables, thevariablethatbest represented the category was selected first(e.g.,for the cate-goryof"any priorsexoffenses," allpriorsexoffenses" wasselectedbefore prior child molesting offenses ). Next, given conceptuallyequivalent findings, the selection was based on sample size andcom-pleteness of information. Finally, in those rare cases in which severaloptions remained, we simplyselectedthe median value. Further detailson thecodingand selectionprocedure areavailablein thecodingmanual.

Toillustrate the codingprocedure.Rice, Harris, andQuinsey(1990;Rice, Quinsey, & Harris, 1991; Quinsey,Rice,& Harris, 1995) reportedon therelationship between age andrecidivism in atleast three studiesfromthesame setting(MentalHealth Centre,Penetanguishene, Ontario),Onestudyexaminedrapists ( = 54;Riceet al.,1990),another exam-inedchild molesters (n =136; Rice et al.,1991),and a third examinedthe combined sample (n = 178; Quinsey, Rice, & Harris,1995).Weonlyused thefindings from Quinsey. Rice,andHarris(1995;Table 2)because it was based on the largest sample size and the longest follow-up period.

Even though each study could contribute only one finding per pre-dictor, studies frequently reported on more than one predictor variable.Consequently, it ispossible that the correlations within each study arethemselvescorrelated. Althoughweignoredthesepotential intercorrela-rions, the major consequenceof thisapproachwas tomake the testofdifferences between predictors conservative. Given thatthesample sizeswere generally large, thepotential loss of a small amountof statisticalpower was of little concern.

n ex ofPredictive AccuracyPredictive accuracy wascalculated usingrbecauseit isreadily under-

stood and thestatistical procedures for aggregating rs are well docu-mented (Hedges &Olkin, 1985; Rosenthal, 1991).Themagnitudeof acorrelation can be interpreted as an approximation of the percentagedifference inrecidivism rates between offenderswithorwithoutapartic-ular characteristic (Farrington&Loeber, 1989; Rosenthal,1991).If,forexample, theoverall recidivism ratewas 25% and blue eyes corre-lated .20withrecidivism, there would be a 20 percentage pointdifferenceintheratesbetween the groups (35% blue eyed vs. 15%non-blueeyed).Except with extreme distributions, this span of 20 percentage pointsshouldbecenteredaroundthe base rate (i.e.,2510percentage points).

Formulasforconverting study statistics(F, t, significancelevels) intorwere drawn from Rosenthal(1991).The correlations were calculatedfrom the most direct data available.If a studyreported both the rawfrequenciesand achi-square,forexample,thecorrelation wascalculatedfrom the raw frequencies. Nonsignificant findingswere assigned anr value of 0 (7.3% of findings). For five studies (Bonta & Hanson,1995a; Hanson,Steffy,& Gauthier, 1993b; Proulx,Pellerin, McKibben,Aubut, & Ouimet, 1995; Reddon, 1995; Thornton, 1995), the correla-tions were calculated directly from theoriginal raw data sets. Someofthe information from these unpublished data sets have been reportedpreviously (Bonta & Hanson, 1995b; Hanson et al., 1995; Hanson,Steffy, & Gauthier, 1992, 1993a;Proulx et al., 1997; Studer, Reddon,Roper,&Estrada, 1996).

Aggregation ofFindingsTwumethods were usedtoaggregatefindings.The first was themedian

r value across studies. Medianvalueshave been recommended formeta-analysis(Slavin, 1995) because theyarerelatively insensitivetooutliersand areeasyto calculateandinterpret.On theother hand, statisticsforestimating the variability of median values are not readily available.Median valuesdo not take into account factors thatmay influence the


4/15


5/15

352Table1Predictors of SexualRecidivism

HANSON AND BUSSIERE

VariableDemographic factors

AgeWith outlier

Single (never married)Married (currently)Employment instability

With outlierSocial class(low)

With outlierLow educationMinorityrace

Generalcriminality(nonsexual)Antisocial personality disorder

With outlierPrioroffenses(any/nonsexual)MMPI 4:Psychopathic DeviateAdmissions tocorrectionsJuveniledelinquencyPrior violent offensesPrior nonviolentoffensesSexual criminal historyPrior sex offenses

With outlierStrangervictim(vs.acquaintance)

With outlierFemale child victimEarly onset of sex offending

With outlierRelated child victim

With outlierMalechildvictimDiverse sexcrimes

With outlierExhibitionism

With outlierAny adult malevictimsVictims, children of both sexesRapistAge of child victim (young)Current sentence lengthDegree ofsexual contactAny child victimsForce orinjury to victims

With outlierSexual deviancy

Phallometricassessment (children)MMPI5: Masculinity-FemininityAnydeviant sexual preferencePhallometric assessment (boys)Deviant sexual attitudes

WithoutlierLegally classified as MDSOPhallometric assessment (rape)Clinical presentation andtreatment historyFailure tocomplete treatmentLengthoftreatmentEmpathyforvictimsDenialof sex offense

With outlierLow motivation for treatment

(self-report/clinical ratings)With outlier

Mdn

-.09-.08

.11-.08

.07

.07

.00

.01

.00

.00

.17

.16

.12

.10

.08

.06

.01-.02

.19

.20

.22

.26-.08

.14

.11-.12-.12

.06

.08

.08

. 1 1

.10

.10

.04

.06

.08

.05-.16-.05

.00

.02

.20

.17

.20

.15.09

.08

.03

.00

.18

.00

.03

.03

.03-.02

.02

r+

-.13-.10

.11-.09

.07

.39

.05

.20-.03

.00

.14

.09

.13

.10

.09

.07

.05

.00

.19

.29

.15

.38-.14

.12

.03-.11-.30

.11

.10

.11

.09

.03

.09

.09

.07

.05

.04-.03-.03

.01

.25

.32

.27

.22

.14

.09

.01

.07

.05

.17

.03.03

.02

.16

.01

.14

95% CI

-.11 .15-.08.12.07-.15-.05.13.00-.14.36-.42.00-.10.17-.23

-.07-.01-.04-.04.07-.21.05-.13.11-.15.00-.20.03-.15.02-.12.00-.10

-.08 -.08.17-.2I.27-31.06-.24.35-.41

-.16.12.05-.19-.03-.09

-.13.09-.32.28.09-.13.07-.13.08-.14.06-.12.01-.05.05-.13.07-.11.05-.09.01-.10.01-.08

-.10-.04-.05 .01-.04-.06

.22-.28

.29-.3S

.14-.40

.16-.30

.01-.27

.00-.18-.07-.09.01-.13

-.06-.16.10-.24

-.02-.08.00-.06-.05-.09.13-.19

-.08-.10.11-.17

Q51.62***

111.77***9.62

14.143.56

106.64***1.28

54.98***12.61*13.02*2.296.41

44.31***2.456.508.20

11.55*4.73

81.25***513.77***

8.2939.34***51.12***

1.0532.17***31.79

696.49***39.53**

1.8522.30***15.5749.59***9.07

52.51***122.39***

7.9314.98*25.17***76.75***7.26

172.68***36.79***4.731.11.17

1.4915.68**20.03***2.03

12.87*.75

3.292.29

19.21**.24

8.19*

Q

6,969(21)8,184(22)2,850(8)2,828(10)

762 (5)5,143 (6)1,622(6)6.003(7)2,304 (7)2.505(7)

811 (6)2,113(7)8,683(20)

393 (4)1,074 (4)1,486(7)1,421 (6)

685 (3)11,294(29)15,675(30)

465 (4)4,846(5)

10,198 (17)919 (4)

1,175(5)6,889(21)

11,270(22)10,294 (19)6,01 5)6,109(6)4.826(14)9,826(15)2.291 (5)7,598(9)

15 . 181 (25)1,828 (9)2.927(7)

828 (6) 3,683(24)1,564 (8)5,982 (9)4,853(7)

239 (3)570 (5)239 (3)439 (4)549 (5)

1,043 (3)320 (4)806 (6)

1,891 (4)4,670(3)

762 (6)5,143 (7)

435 (3)4,816(4)


6/15


7/15

5 HANSON AND BUSS1ERE

Table2PredictorsofNonsexual ViolentRecidivism

Variable Mdn 95% CIDemographic factors

AgeWith outlier

Minority raceSingle (never married)Married (currently)

General criminality(nonsexual)Juvenile delinquencyPrior violent offenses

With outlierAntisocialpersonality disorderMMPI 4: Psychopathic DeviatePrioroffenses (any/nonsexual)

WithoutlierSexual criminal history

RapistAnychild victimsAny adult male victimsRelated child victimYoung(vs.older)childMale child victimCurrentsentence length

WithoutlierPrior sex offensesFemale child victimVictims,children of both sexes

Sexual deviancyMMPI 5: Masculinity-FemininityPhallometricassessment: sexual

preference for rapeClinical presentation andtreatment history

Failure to complete treatmentPsychological maladjustment

AngerproblemsAlcohol abuse problemDepressionAnxietyGeneral psychological problem

Other psychological problemsLow intelligence

-.22-.16

.16.10-.07

.20

.21

.22

.18

.08

.12

.14

.22-.17-.11-.12-.08-.09-.02-.01

.00-.05

.01-.09

.17

.04

-.09.07

-.04-.03.00

.00

-.24-.22

.23.10-.10

.22

.21

.26

.19

.13

.11

.14

.23-.16-.13-.12-.11-.09-.02

.03

.02-.02-.02-.10

.03

.08-.09

.07-.04-.03.00

.07

-.27 .21-.25.19

.19-.27.05-.15-.15 05

.15-.29

.15-.27

.21-.31

.10-.28

.02-.24.08-.14

.11-.17

.20-.26-.20.12-.20.06-.17.07-.18.04-.15.03-.06-.02-.01 -.07-.01-.05-.07-.03-.09-.05-.21-.00-.09-.14-.01-.17-.22-.05-.04-.18-.14-.06-.13-.07-.09-.09.00-.14

22.06**45.76***11.21**9.2214.70**1.881.64

16.95**1.982.627.43

48.2641.09***44.99***

0.784.851.856.794.88

40.72***8.01

24.42***1.544.81

17.34***1.071.040.187.793.294.438.48*

3,376(7)3,530(8)1,981 (3)1,380(5)1,380 (5)

906(4)1,203 (5)1.585 (6)

494 (4)335 (3)

3,450(7)3,746(8)4,040(10)2,742(9)

898 (3)1,611 (6)

758 (4)1,245 (5)2,407 (5)2,788 (6)4,300(9) ,847 6)

869 (3)335 (3)290 (3)479 (3)231(3)340 (3)373 (4)383 (4)502 (5)695 (3)

Note. CI =confidence interval;n=aggregated sample size; MMPI=Inventory.aValuesinparentheses represent numberofstudies.*p


8/15

PREDICTING RELAPSE 355Table3Predictors of General Recidivism

VariableDemographic factors

AgeSingle (never married)Minority raceWith outlierMarried (currently)Education

General criminality (nonsexual)Juveniledelinquency

WithoutlierAdmissions to correctionsPrior offenses (any/nonsexual)Priorviolent offensesAntisocial personality disorderMMPI 4: Psychopathic Deviate

Sexualcriminal historyForceor injury to victimPriorsex offensesRelated child victimAny child victimsStranger victim(vs.acquaintance)Any adult male victimsRapistYoung(vs.older)childExhibitionistMalechild victimSexualintrusivenessVictims, children ofbothsexesFemale child victimSentence length

WithoutlierSexual deviancy

Phallometric assessment (children)Legally classified asMDSODeviantattitudestowardsex

With outlierMMPI 5: Masculinity-FemininityClinical presentation and treatment historyFailure tocomplete treatmentDenial of sexoffenseLow motivationfor treatmentPreviously failed treatment forsexual

offenseDevelopmental history

Negative relationshipwith motherSexually abused as childGeneral family problemsNegative relationship with father

Psychological maladjustmentAnypersonality disorderAlcohol use during offenseAlcohol abuse problemAnger problemsAnxietySeverely disorderedAny substance abuse problemGeneral psychological problemDepression

Other psychological problemsLow intelligence

With outlier

Mdn

-.18.14.08.10

-.11.00.20.20.22.25.18.23.09.11.12

-.16-.12

.08

.03

.05-.02

.05

.08-.08

.01-.08

.01

.04

.19

.04-.03-.04

.02

.19

.23

.13-.04

.14

.00

.06

.00

.22

.00

.13

.11

.08

.03-.03-.05

.00-.01

.00Note. Cl =confidence interval;n aggregated sample

r+

-.16.11.10.14

-.08.01

.28

.20

.25

.23

.20

.16

.10

.13

.12-.12-.08

.07

.07

.05-.03

.04

.03-.03

.02-.01

.00

.08

.11-.10

.06-.01

.04

.20

.12

.11-.07

.14

.10

.07

.02

.21

.12

.1

.09

.08

.03-.01-.01

.01

.00

.01size;

95%Cl

-.18.14.08-.14.06-.14.11-.17-.11.06

-.06-.08.22-34.15-.25.19-32.21-.25.14-.26.13-.19

-.03-.23.08-.18.10-.14

-.14.10-.11.05-.02-.16.03-.I I.03-.07

-.09-.03.01-.07.01-.06

-.12-.06-.04-.08-.04-.02-.06-.06

.03-.13-.02-.24-.16.04-.05-.17-.IO-.08-.09-.17

.13-.27

.02-.22

.05-.19-.12 .02

.03-.25

.01-.1 9

.00-.14-.09-.13.09-33.07-.17.08-.14

-.04-.22-.04-.20.00-.06

-.11-.10-.11-.09-.10-.12-.03-.03-.02-.04

Q

74.10***7.11

11.57*49.35***28.37***

1.9715.03*42.18***

2.2152.63***6.899.27.82.76

48.50***31.65**72.62***2.92

14.33**84.07***12.77*12.70*7.25

16.84***8.88*

53.04***1.13

24.48***5.93

1536***1.948.62*.39

5.3012.76**4.664.711.136.71*4.50.94

4.63***16.04***3.462.441.571.595.295.952.576.27

15.54**

n

8,250(14)5,038(9)2,919 (6)3,358(7)6,445(10)

914 (3)1,113 (5)1,574 (6)

834 (3)7,565(15)1,184 (5)3,977(7)

239 (3)1,304 (3)8,975(15)6,735(15)5,798(14)

465 (4)2,499(5)

14,753(19)1,056 (7)5,467(7)5,943(11)

438 (4)1,044 (4)6,052(11)1,034 (4)1,415 (5)

233 (3)1,053 (3)

338 (3)448 (4)239 (3)887 (7)408 (3)614(4)

1,380 (3)350 (3)473 (3)698 (4)335 (3)273 (3)

1,395 (3)3.552(5)

231 (3)284 (4)

3,332(3)422 (3)364(5)354 (5)

5,004(4)5,274 (5)

MMPI =Minnesota Multiphasic Personalityinventory;MttoU = mentally disorderedsexualoil"Values inparenthesesrepresentnumber ofstudies.*p


9/15

356 HANSON AND BUSSIERETable4Correlations with Recidivism for Each Category of Predictor

Typeof recidivism

PredictorCriminal lifestyleSexual deviancePsychological maladjustmentNegative clinical

presentationFailure tocomplete

treatment

Sexual.12 .02.19 .01.01 .03.00 .07.17 .07

Nonsexualviolence.16 .03.01 .03.02 .08

.08 .09

Any.21 .02.12 .02.02 .03.15 .07.20.07

Nore. Valuesrepresentaverage correlations 95%confidence interval.

sexual, recidivism.(Insufficient studies examined the relation-shipbetween negative clinical presentation andnonsexualvio-lentrecidivism.) Finally,failuretocompletetreatmentappearedto be a consistent risk marker for both sexual and generalrecidivism.

CombinedRisk ScalesCombinations of variables should predict recidivismbetter

than any individual risk factors examined alone. Tbdate, riskscales for sexualoffendershave not received extensive examina-tion by researchers, but the available results can, nevertheless,provide some guidance.

There are several methods of combining variables. Clinicaljudgescanweigh information gained through interviews, formaltesting, and offense history. Alternately, statistical algorithmscan select optimal weightsto maximize the prediction oftheknown recidivism results (e.g., multipleregression).Suchstatistical methods will always provide the largest correlationsbecause they are designed to select optimal weights for thatsample.Athird methodis to useobjective riskscales,inwhichweightsareassignedinadvance basedoneither theoryorprevi-ousstatistical analyses.

As can be seen in Table 5, the predictive accuracy of clinicalrisk assessmentswasunimpressiveforsexual(.10),nonsexual

violent (.06), and general recidivism (.14). In contrast, thestatistical risk prediction scales(e.g.,stepwise regression) typi-callyproduced correlations substantially larger thanthosefoundforsingle variables(.46for sexual recidivism, .46 for nonsexualviolent recidivism,and .42 forgeneral recidivism).

The items selected by the statistical procedures, however,varied considerablyacrossstudies.Each scale included betweenthree and nine items; no single item was common to all sixstudies (Abel, Mittelman, Becker, Rathner, & Rouleau, 1988;Barbaree & Marshall, 1988;Hanson et al., 1993b; Prentky,Knight,&Lee,1997; Quinsey, Rice, &Harris, 1995; Smith&Monastersky, 1986).Themost common items werepriorsexualoffenses (used in four studies), deviant sexual preferences(three studies),maritalstatus(threestudies), anddiverse sexualcrimes and malechildvictims (both used in two studies). Thedifferences between these studiescan beattributedto thevaria-tionsinsamples, to thedifferent variables examined,and to therandom fluctuationstowhich stepwise methods areparticu-larlyvulnerable (Pedhazur, 1982).

An insufficient number of studies used objective risk scalestojustifyquantitative analysisofthesescales. Thesestudiesarediscussedbriefly,because this research is particularly importantforappliedrisk assessments.

Wewere able tolocateonly one study (Epperson, Kaul,&Huot,1995)inwhicha riskinstrumentwasspecifically designedfor sexual recidivism and subsequently cross-validated on anentirelynewsample(r =.27,withanartificial recidivism baserateof50%).The21items covered sexualandnonsexualcrimi-nal history, substance abuse, and employment. Many of theindividual items did not withstand cross-replication, however,and thescale iscurrently being revised.

Objective risk scales designed for general recidivism havepredicted nonsexual recidivism among sexual offendersbuthavenot predicted sexual recidivism. Bonta and Hanson (1995a,1995b) found (hat the Statistical Information on Recidivism(SIR) scale correlated .41 with general recidivism, .34 withnonsexual violent recidivism but only .09 with sexual recidi-vism.The SIR scale was developed on a sample ofCanadianfederal offenders and included items related to criminal history,age, and marital status (Bonta, Harman, Hann, & Cormier,1996).Similarly,theCommunity Risk/Needs scale usedby the

Table 5CombinedRisk Scales

VariableSexual recidivism

Clinical assessmentStatistical

Nonsexual violent recidivismClinical assessmentStatistical

General recidivismClinical assessmentStatistical

Mdn

.04

.44

.06

.45

.11

.42

r_

.10

.46

.06

.46

.14

.42

95%CI

.05-.15

.40-.52-.02-.14.37-.S4.08-.20.36-.51

Q

16.1015.06**1.10.65

17.81**.87

n

1,453 (10)684 (6)544(3)343 (3)

1.067 (8)453 (5)

Note. CI = confidence interval;n= totalaggregated samplesize.aValues in parentheses represent number of studies.**p


10/15


Correctional ServiceofCanada (CSC) predicted generalparolefailureamong sexual offenders(r = .23,n=809)onlyslightlylesswell than among nonsexualcriminals(r = .33, n =253;Motiuk &Brown, 1993; Motiuk & Porporino, 1989).Sexualoffense recidivism was not specifically examined in the CSCCommunity Risk/Needs studies.

The Violence Risk Appraisal Guide (VRAG)was developedto predict violentrecidivismamong patients at amaximumsecu-ritypsychiatric hospital and has been replicated on a subsampleofsexual offenders (Webster, Harris, Rice, Cormier, & Quinsey,1994). The 12 items of the VRAGaddresssuch factors as per-sonalitydisorders, early school maladjustment, age, marital sta-tus,criminal history, schizophrenia,andvictim injury (the lasttwo items were negatively weighted, meaning the presence ofthese factors reduced risk scores). In a replication sample of159 sexual offenders, Rice and Harris (1997) found that theVRAGcorrelated .47 with violent recidivism (sexual and non-sexual)butonly.20with sexualrecidivism.

Influence ofRecidivism MethodsAdditionalanalyses were conducted to identify possible ef-

fects ofdifferences inrecidivism methods (e.g., convictionsvs.other outcomecriteria).To reduce unnecessaryerrorvariance,these supplemental analyses were conducted on two predictorvariables for which there were many studies and the overalleffects were uncontroversial: namely,(a)priorsexual offensespredictingsexualoffense recidivismand (b) anyprior offensespredicting general (any) recidivism.

The findings(correlations)based on convictions were equiva-lent to those findings based on other recidivism measures forprior sexual offenses,x2(l ,N = 11,139,k =28)= .36,p >.50,and forprior (any)offenses,x*(l,N =7,565, k= 15) =1.93, p > .10.Aswell, the thoroughnessoftherecidivismsearchhad no influence on the magnitudeof the findings; prior sexoffenses,K\lN =15,675,k= 29) = 1.19,p > .25; prior(any) offenses, X2(l,N=6,192,k=14) =.04,p > .50.

DiscussionWhat factors are related to sexual,nonsexual violent, and

general (any)recidivism among sexual offenders? Three majorcategories of predictor variables were examined: criminal life-style, sexual deviance, andpsychological maladjustment. Theoffenders' clinical presentation and treatment compliance werealso considered as potential risk factors. Overall, the predictorsofnonsexual recidivism(violentornonviolent)were very simi-lar tothosefoundin theresearchongeneral(mostly nonsexual)offender populations (e.g., Gendreau etal., 1996). Sexual of-fenders who recidivated with nonsexual crimes tended to beyoungand unmarried and to have a history of antisocial behavioras juveniles and as adults. In contrast, the strongestpredictorsof sexual recidivism were factors related to sexual deviance.Criminal lifestyle variables did predict sexual recidivism, butthe best predictors were factors such as deviant sexualinterests,prior sexualoffenses, anddeviant victim choices (boys, strang-ers).Withthe exception of personality disorders, psychologicalmaladjustment had little or no relationship with any type ofrecidivism.A negative clinical presentation(e.g., low remorse,

denial, low victim empathy) was unrelated to sexual recidivismbut showed a small relationship with generalrecidivism.Failuretocomplete treatment, however, was asignificantpredictorofboth sexual and nonsexual recidivism.

The correctional literature tendsto minimize differences be-tween types of offenders (e.g., M. R. Gottfredson & Hirschi,1990),but the current results suggestthat sexualoffendersmaydiffer fromother criminals (seealsoHanson et al., 1995). Fornonsexual offending, sexual and nonsexual criminals seemmuch the same, but separate processes appear to contribute tosexual offending. In particular, not all criminals would be ex-pected to have deviant sexual interests (e.g., sexual interest inboys). Consequently, risk assessments should consider sepa-ratelytheprobabilityofsexualandnonsexual recidivism. Giventhatthepredictorsofnonsexualcriminalitywerealmost identi-cal to thosefound for general offenders (Champion, 1994; Gen-dreau etal.,1996; D. M. Gottfredson &Tonry, 1987), standardcriminalrisk assessment methods (e.g., Bonta, 1996; Gendreauet al., 1996) should predict equally general recidivism amongsexual offenders. The assessmentof sexual recidivismrisk, incontrast, needs to consider factors specially related to sexualoffending (e.g., sexual deviance, victim type).

Although the risk factors in each domain can be relativelyindependent, they mayalsointeract. Wefoundthat general crim-inality was significantly related to sexual recidivism, but thedirect relationshipwasweak (.10- .14).Twoindividual studies,however, found that the combination ofdeviant sexual prefer-encesandpsychopathy substantially increasedthe risk forsexualreoffending (Gretton,McBride,& Hare, 1995; Rice & Harris,1997).

The present findings contradict the popular view that sexualoffenders inevitablyreoffend. Onlya minority of thetotal sam-ple (13.4% of23,393)were known to have committed a newsexualoffense within the average 4- to5-yearfollow-up periodexamined in this study. This recidivism rate should be consid-eredanunderestimate becausemanyoffensesremain undetected(Bonta & Hanson, 1994). Nevertheless, even in studies withthorough records searches andlong follow-up periods (1520years), the recidivism rates almost neverexceeded40%. Lowrates of recidivism can, nevertheless, be worrisome, given theserious effects of sexual victimization (Hanson, 1990; Koss,1993b).

Inthis review, measures ofsubjectivedistress had no relation-ship to any type of recidivism; the averagecorrelationswerenearzerowith nosignificantvariability. Subjective distress is atransient state, and no measure of highlychangeable stateswould beexpecttopredict sexualoffense recidivism years later.

Previousresearch, however,hassuggested that negative emo-tionalstatesmaytriggerasexualoffensecycle.When recidivistshave been asked about thefactorsthat contributed to their newoffense, they frequently identify subjective distress (Pithers etal., 1989; Pithers, Kashima, Cumming, Beal, &Buell,1988).Similarly, McKibben, Proulx, and Lusignan (1994; Proulx,McKibben,&Lusignan,1996)foundthat when sexual offenderswere upset, they were likely to report deviant sexual fantasies(based onrepeatedassessments). Thesesignificantwithin-sub-ject correlations contrast with the nonsignificant between-sub-ject correlations between mood and recidivismfound for thesame population (Proulx et al., 1995). The extent to which


11/15

358 HANSON AND BUSSIEREsexualoffenders aredistresseddoesnotpredict recidivism,butsexualoffenders appear to react deviantly when distressed. Spe-cifically, sexual offenders oftenreport takingsolace insexualthoughtsandbehaviorwhenconfronted withstressful lifeevents(Cortoni, Heil, & Marshall, 1996).

One of our most important findings is that offenders whofailed to complete treatment were at increased risk for bothsexual and general recidivism. Reduced riskcould be due totreatmenteffectiveness; alternately, high-risk offenders may bethose most likely toquit,orbeterminated,from treatment. Ingeneral, psychotherapy dropouts tend to be young and unedu-cated and to have antisocial personality characteristics (Wierz-bicki & Pekarik, 1993).Attrition from treatment canalsobeinterpreted as a behavioral (vs. purely verbal) indicator of moti-vation tochange.

Somewhat surprisingly, a negative clinical presentation (i.e.,verbal expressions of denial or low motivation for treatment)was related to general recidivism but not sexual recidivism. Oneexplanation is that such negative clinical presentation may beshaped heavily by a hostile, belligerent style, a style more con-nected to a general criminal lifestyle than sexual deviance.Treatment effectiveness was not examined directly in thisreview because several narrative reviews (W. L. Marshall, Jones,Ward,Johnston, & Barbaree, 1991; W. L. Marshall &Fitters,1994; Quinsey etal., 1993) and at least twometa-analyses(Al-exander,1995;Hall,1995b)havebeen conducted. Overall, opin-ionremains dividedas to theeffectivenessoftreatment, partiallybecause of the difficulty of conducting research in this field(Hanson, 1997). The results of the current review, however,suggest that treatment programs can contribute tocommunitysafety through their ability tomonitor risk.Even if wecannotbesure that treatmentwillbeeffective, thereisreliable evidencethat those offenders who attend and cooperate with treatmentprograms are less likely to reoffend than those who rejectintervention.

The detailed results presented in Table 1 should be of consid-erable interest to clinicians conducting applied assessments ofsexualoffense recidivismrisk.Several cautions, however, needto beconsidered. First, among the large number of variablesexamined,some are likely to appear statistically significantbe-cause of chance only. These random findings aremost likelywhen only a limited numberofstudies examineda particularvariable (i.e., three to five). Incontrast,the findingsbasedonlargenumberofstudies(e.g.,10 ormore)areunlikelytochangeevenwiththe addition of a few new studies.

The predictive accuracy of most of the variables was alsosmall(.10-.20range), and no variable wassufficiently relatedto justify its use in isolation. It was also unclear how bestto combine the variables because theirintercorrelations wereunknown andwouldbeexpected to be rather highforcertainvariables (e.g.,youngandunmarried). Consequently,we do notrecommend simplysummingtheitems, using either unit weightsorweights inferred fromthetables,tocreateriskscales.Thedevelopment of avalidated, actuarial riskscale for sexual of-fenserecidivism remains an important research goal. Neverthe-less, our results could be used to identify the factors worthconsidering in risk assessments. Although the average clinicalrisk assessment showed little accuracy, the most accurate clinicalrisk assessments required clinicianstoconsider a standard list

of risk factors before making theirjudgments (e.g., Smith&Monastersky, 1986,r= .29).

Anotherlimitationwasthat almostall thepredictorsofsexualoffense recidivism were historicalorextremely stable variables.Consequently,such variables cannotbeused to assess treatmentoutcome or monitor risk to thecommunity. Historical factorscannot improve, and it is difficult to change deviant sexualpreferences (Rice ctal., 1991) or antisocial or psychopathicpersonality disorder (Hare etal.,1990).The most changeable(dynamic)riskfactorwas treatment attendance.Toidentify dy-namic risk factors,further research is required using alternateresearch designs (such as timeseries).

Thelowrateof sexualoffense recidivism presents a specialchallenge to those interested in identifying risk factors. Oftenmany years (5-10) of follow-upare required toaccumulatesufficient cases forstatistical analyses. Consequently, research-erseither havetorelyonpreexisting data sets (with potentiallyoutdated measures) or set up new data sets that yield resultsmany years later. Today's clinicians can contributeto futureresearch by carefully assessing andrecording the factors thatareconsidered importantfor riskassessmentbuthaveyet to beadequately researched. Included inthis listis the use of sex asa coping mechanism (Cortoni & Marshall,1995),associationswith other sexual offenders (Hanson & Scott,1996),attitudestolerant ofsexual crimes (Bumby,1996),heterosocial percep-tion deficits (Hanson &Scott, 1995; Malamuth & Brown,1994),andunfulfilled intimacy needs (Frisbie, 1969; Seidman,Marshall, Hudson, & Robertson, 1994). As well, there is a needto examine the developmental precursors of sexualoffending,ashasbeen well documentedforgeneral criminality (Andrews&Bonta, 1994; Loeber & Dishion, 1983; Loeber & Stouthamer-Loeber,1987).It isonly throughthecollectiveeffortofpastandfutureresearchers thatwe canimproveourabilitytodistinguishbetween those sexualoffenders likelytoreoffend andthosewhohavestoppedforgood.

ReferencesReferences marked with an asterisk indicate studiesincludedin the meta-

analysis.Abel, G. G., Becker, J.V., Cunningham-Rathner,N.,Rouleau, J.L., &

Murphy,W. D.(1987). Self-reported sex crimes of non-incarceratedparaphiliacs. Journal ofInterpersonal Violence, 2, 3-25.

*Abel, G.G.,Mittelman,M,Becker, J.V.,Rathner, J., & Rouleau, J.(1988). Predicting childmolesters' response to treatment. In R. A.Prentky& V. L. Quinsey(Eds.),Human sexual aggression: Currentperspectives (pp. 223-234). New York: New \brk Academy ofScience.Alexander, M. (1995, September). Sex offender treatment: Does itwork? Paper presented at the NOTA/ATSA 1st Joint InternationalConference, Cambridge, United Kingdom.

American Psychiatric Association.(1994).Diagnostic and statisticalmanual ofmental disorders (4thed.).Washington, DC:Author.

Anderson, N.W.,& Masters, K.W.(Eds).(1992). Predators andpoli-tics:A symposium on Washington's SexuallyViolentPredatorsStat-ute [Special Issue], University ofPuget Sound LawReview, 15(3).

Andrews, D. A., & Bonta, J.(1994).The psychology of criminal con-duct.Cincinnati,OH: Anderson.

*Barbaree, H. E., & Marshall, W. L.(1988). Deviant sexual arousal,offense history and demographic variables as predictors of reoffense


12/15


among child molesters.Behavioral Sciencesand theLaw,6. 267-280.

Barbaree, H. E., &Marshall, W.L.(1989).Erectile responsesamongheterosexual child molesters, father-daughter incest offenders, andmatched non-offenders: Five distinctage preference profiles. Cana-dian JournalofBehavioural Science, 21,70-82.

*Beck, A. J., & Shipley, B. E.(1989).Recidivism ofprisoners releasedin1983,Washington, DC: Bureau of Justice Statistics.*Bluglass, R.(1980). Indecent exposure in the West Midlands. In D.West (Ed.).Sex offenders in the criminaljusticesystem (pp. 171-180). Cambridge, England: Cambridge Institute ofCriminology.

Bonta,J.(1996).Risk-needs assessmentandtreatment.In A. T.Harland(Ed.),Choosing correctional options that-work(pp. 18-32).Thou-sand Oaks, CA:Sage.

Bonta, J., & Hanson, R. K. (1994). Gauging the risk for violence:Measurement, impact and strategiesfor change (User Report No.1994-09). Ottawa, Ontario, Canada: Department of the SolicitorGeneral of Canada.

*Bonta,J.,& Hanson, R. K. (1995a). [ 10-year recidivism of Canadianfederal offenders]. Unpublishedrawdata.

Bonta,J.,& Hanson, R. K. (1995b,August).Violentrecidivismofmenreleasedfrom prison.Paperpresentedat the103rdAnnualConventionoftheAmerican Psychological Association, New brk.

Bonta, J., Harman, W.G.,Hann, R.G.:& Cormier, R. B. (1996). Thepredictionof recidivism among federally sentenced offenders:A re-validation of the SIR scale. Canadian Journal of Criminology,38,61-79.

*Broadhurst,R.G.,&Mailer,R. A.(1991).Sex offending andrecidi-vism(Tech. Res. Rep. No. 3). Nedlands, Western Australia: UniversityofWestern Australia,Crime ResearchCentre.

*Broadhurst, R. G., & Mailer, R. A. (1992). The recidivism of sexoffenders in theWesternAustralianprison population.British Journalof Criminology, 32,54-80.

Bumby,K. M.(1996),Assessing the cognitive distortions of child mo-lestersand rapists: Development and validation of theMOLESTandRAPE scales.Sexual Abuse:A Journal ofResearchandTreatment,8,37-54.Champion, D. J. (1994).Measuring offender risk:A criminaljusticesourcebook.Westport,CT: Greenwood Press.

*Christiansen, K., Elers-Nielsen, M., Le Maire, L., & Stiirup, G. K.(1965).Recidivismamong sexual offenders.Scandinavian StudiesinCriminology, 1,55-85.

Cooper, H. M.(1989).Integrating research: A guide for literature re-views(2nded.)- NewburyPark, CA: Sage.

Cortoni, R, Heil,P., & Marshall, W. L. (1996, November).Sex as acoping mechanism and its relationship to loneliness and intimacydeficits insexual offending. Paper presented at the 15thAnnual Con-ference of the Association for the Treatment of Sexual Abusers,Chicago.

Cortoni, R, & Marshall, W. L. (1995).Coping with Sex Inventory.Un-published manuscript, Queens University, Kingston, Ontario, Canada.

*Davis, G. L., Hoffman, R.G., & Slacken, N. (1991, August). Two-4-year post-incarceration follow-up oftreatedsex offenders. Paperpresented at the 99thAnnualConvention of the American Psychologi-cal Association, San Francisco, CA.

*Dix,G. E.(1976). Differential processing of abnormal sex offenders.Journalof Criminal Law,Criminology,& Police Science, 67, 233-243.

*Doshey,L. J.(1943).The boy sex offender and hislater career.Mont-clair,NJ: Patterson Smith.

*Epperson, D. L.,Kaul,J. D., &Huot,S. J.(1995,October).Predictingriskfor recidivismfor incarceratedsex offenders: Updated develop-ment on the Sex Offender Screening Tool (SOST). Poster session

presented at the annual conference of the Association for the Treatmentof Sexual Abusers, New Orleans, LA.

Farrington, D. P., & Loeber, R. (1989). Relative improvement overchance (R1OC) and phi as measures of predictive efficiency andstrengthofassociation in 2 X 2tables.JournalofQuantitativeCrimi-nology, 5,201-213.

*Federoff, J. P., Wisner-Carlson, R., Dean, S., &Berlin,F. S.(1992).Medroxy-progesterone acetate in the treatment of paraphilic sexualdisorders. Journalof Offender Rehabilitation, 18,109-123.

*Fitch, J. H.(1962).Men convicted of sexual offences against children:A descriptive follow-upstudy.British JournalofCriminology,J, 18-37.

*Florida Department of Health and Rehabilitative Services.(1984).Sta-tusofthe sex offender treatment programs,fiscalyear 198384,andaddendum (annual report to the FloridaLegislature, Alcohol, DrugAbuse, and Mental Health Program Office). Tampa,FL:Author.

Freund, K., & Watson, R. J. (1991).Assessment of the sensitivity andspecificityof a phallometric test: An update of phallometric diagnosisof pedophilia.Psychological Assessment, 3,147-155.

*Frisbie,L. V.(1969).Another lookat sex offenders inCalifornia (Cali-forniaMentalHealth Research Monograph No. 12). Sacramento, CA:State of California Department of Mental Hygiene.

*Frisbie,L.V.,& Dondis, E. H.(1965).Recidivism among treatedsexoffenders (California Mental Health Research Monograph No. 5).Sacramento, CA:Stateof California Department of Mental Hygiene.

Furby,L., Weinrott,M.R.,&Blackshaw,L.(1989).Sexoffender recidi-vism:Areview.Psychological Bulletin, 105, 3-30.

Gendreau,P., Little, T, & Goggin, C.(1996).Ameta-analysisof thepredictors of adult offender recidivism: What works Criminology,34,575-607.

*Gibbens, T. C. N., Soothill, K.L.,& Way, C. K. (1978). Sibling andparent-child incest offences.British JournalofCriminology, 18, 40-52.

*Gibbens,T C.N.,Soothill,K.L.,&Way,C. K.(1980).Child molesta-tion.InD. J.West(Ed.),Sex offenders in thecriminaljusticesystem:Paperspresentedto the12thCropwoodRound-TableConference (pp.89-99).Cambridge, England: Institute of Criminology, University ofCambridge.

*Gibbens, T. C. N., Way, C. K., & Soothill, K. L.(1977).Behaviouraltypes of rape.British JournalofPsychiatry, 130,32-42.

*Gordon,A., &Bergin, H.(1990, June).Description and evaluationof thesexual offender program at theRegional Psychiatric Centre.Paper presented at theannual meeting of theCanadian PsychologicalAssociation, Ottawa, Ontario, Canada.

*Gordon,A.,&Porporino,F. J.(1990,May).Managing the treatmentofsex offenders: ACanadian perspective(Research Brief No.B-05).Ottawa, Ontario, Canada: Research and Statistics Branch, Correc-tional Service of Canada.

*Gore,D. K.(1988).Measuring thecognitive distortionsofchildmo-lesters:Psychometric propertiesof thecognition scale.Unpublisheddoctoral dissertation, GeorgiaStateUniversity.

Gottfredson,D. M., &Tonry,M.(Eds.).(1987).Prediction and classi-fication: Criminaljustice decision making.Chicago: University ofChicago Press.

Gottfredson, M. R., & Hirschi, T.(1990).Ageneral theoryofcrime.Stanford, CA: Stanford University Press.

*Graham, K. R. (1991). The sexual addiction model in treatment ofincarcerated offenders.Sexual AddictionandCompulsivity, 1(3),1 -6.

*Gretton,H.,McBride, M., & Hare, R, D.(1995,October). Psychopa-thyinadolescentsex offenders: Afollow-up study. Paper presented attheannual conferenceof theAssociationfor theTreatmentof SexualAbusers, New Orleans, LA.

*Griinfeld,B.,&Noreik,K.(1986).Recidivism among sex offenders:


13/15

360 HANSON AND BUSSIEREA follow-up study of 541 Norwegian sex offenders. InternationalJournalof Law andPsychiatry,9,95-102.

*Hackett,T. P.(1971). Thepsychotherapy of exhibitionists in acourtclinic setting. Seminars inPsychiatry,3,297-306.

*Hall, G. C. N.(1988). Criminal behaviour as a function of clinicalandactuarialvariablesin a sexoffender population.Journalof Con-sultingand Clinical Psychology,55,773-775.

Hall. G. C. N.(1990). Prediction of sexual aggression. Clinical Psy-chology Review,10,229-245.*Hall,G. C. N.(1995a).Thepreliminarydevelopment of a theory-based

communitytreatment for sexualoffenders.Professional Psychology:ResearchandPractice,26,478-483.

Hall, G. C. N. (1995b). Sexual offender recidivism revisited: Ameta-analysisofrecent treatment studies.Journal of Consulting and Clini-calPsychology,63,802-809.

*Hall,G. C.N.,& Proctor, W. C.(1987). Criminologicalpredictors ofrecidivism in asexual offender population.JournalofConsultingandClinical Psychology,55,111-112.

Hanson, R.K.(1990).The psychological impact ofsexual victimizationon women and children.Annals ofSexResearch, 3,187-232.

Hanson, R. K. (1997).How to know what works with sex offenders.SexualAbuse:AJournalofResearchand Treatment,9,129-145.

Hanson,R. K., &Bussiere,M. T.(1996).Predictorsofsexual offenderrecidivism:Ameta-analysis (User Report96-04).Ottawa, Ontario:Department of the Solicitor General of Canada.

Hanson,R.K.,&Scott,H.(1995).Assessing perspective-takingamongsexual offenders, nonsexual criminals, and nonoffenders. SexualAbuse: A Journal of Research and Treatment, 7,259-277.

Hanson,R.K.,& Scott, H.(1996).Social networks of sexual offenders.Psychology, Crime,&Lax, 2,249-258.

Hanson, R. K., Scott,H.,&Steffy,R. A.(1995).A comparison of childmolesters and non-sexual criminals: Risk predictors and long-termrecidivism.JournalofResearch in Crime and Delinquency, 32, 325-337.

Hanson, R.K.,Steffy, R.A.,& Gauthier, R.(1992).Long-term follow-upof child molesters: Risk prediction and treatment outcome (UserReportNo.1992-02).Ottawa, Ontario: Corrections Branch, Depart-mentof the Solicitor General of Canada.

Hanson, R.K.,Steffy, R.A.,& Gauthier, R.(1993a).Long-term recidi-vism ofchildmoleslers.Journal of Consulting and Clinical Psychol-ogy, 61,646-652.

*Hanson, R.K., Steffy, R.A., & Gauthier, R. (1993b). [Long-termrecidivism ofchildmolesters].Unpublished rawdata.

Hare, R. D.,Harpur,T J.,Hakstian,A.R.,Forth, A. E., Hart, S. D., &Newman, J. P. (1990). The Revised PsychopathyChecklist:Reliabil-ityand factor structure.PsychologicalAssessment, 2,338-341.

Hathaway,S. R., & McKinley, J. C.(1983).TheMinnesota MultiphasicPersonality Inventory manual.NewYbrk:Psychological Corporation.

Hedges,L.V.,& Olkin, I.(1985).Statisticalmethodsfor meta-analysis.NewYork:AcademicPress.

Johnson, H., & Sacco, V. F. (1995). Researching violence againstwomen: Statistics Canada's national survey. Canadian Journal ofCriminology,37,281-304.

*Kahn, T.J.,& Chambers,H. J. (1991). Assessing reoffense risk withjuvenilesexoffenders.Child Welfare, 70,333-345.

Kennedy, H.G., & Grubin, D. H. (1992). Patterns of denial in sexoffenders. Psychological Medicine,22,191-196.

*Khanna, A., Brown, P., Malcolm, P. B., & Williams, S. M. (1989,March). Outcome data on sex offenses assessed and treated at Re-gional Treatment Centre (Ontario). Paper presented at the annualmeetingof the SpecialNeedsOffenders Conference, Ottawa, Ontario,Canada.

Koss,M. P.(1993a).Detecting thescopeofrape:A review of prevalenceresearch methods. Journal of Interpersonal Violence, 8,198-222.

Koss, M. P.(1993b). Rape: Scope, impact, interventions, and publicpolicyresponses. American Psychologist, 48,1062-1069.

*Lab, S. P.,Shields, G., & Schondel, C. (1993). Research note: Anevaluation ofjuvenile sexual offender treatment. Crime and Delin-quency,39,543-553.

Langevin,R.(1988). Defensivenessin sexoffenders.In R.Rogers (Ed.),Clinical assessment of malingering and deception (pp.269-290).New York:GuilfordPress.Launay, G.(1994).Thephallometricassessment of sex offenders: Someprofessional and research issues. Criminal Behaviour and MentalHealth,4,48-70.

Laws, D. R.(Eds.).(1989).Relapse prevention with sexual offenders.NewYork: Guilford Press.

Laws, D. R.(1995).Central elements inrelapseprevention procedureswith sex offenders. Psychology, Crime&Law,2,41-53.

Ley, P. (1972).Quantitative aspects of psychologicalassessment: Anintroduction.London: Duckworth.

Lisak,D.,&Roth,S.(1988).Motivational factors in nonincarceratedsexuallyaggressive men.Journal of Personality and Social Psychol-ogy, 55,795-802.

Loeber, R., & Dishion, T. J. (1983). Early predictors of male delin-quency.Psychological Bulletin,94,68-99.

Loeber,R.,& Stouthamer-Loeber, M.(1987).Prediction.In H. C.Quay(Ed.),Handbook ofjuveniledelinquency (pp.325-382).NewYork:Wiley.

*Mair,K.J.,&Stevens,R. H.(1994).Offending historiesandoffendingbehaviour. A ten year follow-up of sex offenders tried by sheriff anddistrict courts inGrampian, Scotland.Psychology, Crime and Law,1,83-92.

*Mair, K.J., & Wilson, D.(1995).Sexual reconviction over timeinScottish sex offenders identified through court records.Unpublishedmanuscript.

Malamuth,N. M., & Brown, L. M.(1994). Sexuallyaggressivemen'sperceptionsof women's communications: Testing three explanations.Journal of Personality and Social Psychology, 67,699-712.

*Malcolm,P. B., Andrews, D. A., & Quinsey, V. L. (1993).Discrimi-nantand predictive validityof phallometric measured sexual age andgender preference. Journal ofInterpersonal Violence,S,486-501.

*Maletzky, B. M. (1980).Assisted covert sensitization.In D. J. Cox &R. J. Daitzman (Eds.), Exhibitionism: Description, assessment andtreatment (pp. 187-251).NewYork:Garland.

*Maletzky,B. M. (1991).Treating thesexual offender. Newbury Park,CA: Sage.

*Maletzky, B. M. (1993). Factors associated with success and failurein thebehavioral andcognitive treatment ofsexualoffenders. AnnalsofSex Research, 6,241-258.

"Marques,J.K.,Day,D. M.,Nelson,C., &West,M. A. (1993). Effectsof cognitive-behavioral treatment on sexoffenders' recidivism: Pre-liminary results of a longitudinal study.Criminal Justice and Behavior,21,28-54.

*Marques, J.K., Nelson, C.,West,M.A., &Day,D. M.(1994).Therelationship between treatment goalsandrecidivism amongchildmo-lesters.BehaviourResearch and Therapy, 32,577-588.

*Marshall, P. (1994). Reconviction of imprisoned sexual offenders.Research Bulletin,36,23-29.

Marshall, W. L. (1996).Assessment, treatment, and theorizing aboutsexoffenders: Developments during thepast twenty yearsand futuredirections. Criminal JusticeandBehavior,23,162-199.

"Marshall,W. L., &Barbaree,H. E.(1988).Thelong-term evaluationof a behavioral treatment program. Behavior Researchand Therapy,26,499-511.

Marshall, W.L., Jones, R., Ward, T., Johnston, P., & Barbaree, H. E.(1991).Treatment outcome withsexoffenders. Clinical PsychologyReview,11,465-485.


14/15


Marshall,W.L.,&Pithers,W.D.(1994).Areconsideration oftreatmentoutcome with sex offenders.Criminal Justice and Behavior, 21, 10-27.

*McBride,M., Gretton,H., & Hare, R. D. (1995,October).Familialriskfactors, psychopathy andrecidivisminadolescent sexual offend-ers.Paper presented at the annual conference of the Association forthe Treatment of Sexual Abusers, New Orleans, LA.

*McConaghy, N., Blaszczynski, A., Armstrong, M. S., & Kidson,W.(1989).Resistance to treatment ofadolescentsex offenders.ArchivesofSexual Behavior, 18,97-107.

McKibben,A.,Proulx, J., &Lusignan,R.(1994).Relationships betweenconflict,affect anddeviantsexual behaviors in rapists and child mo-lesters.Behaviour ResearchandTherapy, 32,571-575.

*Meyer,W. J.,Cole,C., &Emory,E.(1992).Depo Provera treatmentfor sexoffending behavior:Anevaluationofoutcome.BulletinoftheAmerican Academy ofPsychiatryandLaw,20,249-259,

*Meyers, L.C.,& Romero, J.(1980).A ten-year follow-up ofsexof-fender recidivism.Philadelphia,PA: J. J.Peters Institute.

*Mohr,J.,Turner, R.. & Jerry, M.(1964).Pedophilia and exhibitionism.Toronto Ontario,Canada:Universityof Toronto Press.

*Money, I, & Bennett,R. G. (1981). Post-adolescent paraphilicsexoffenders:Antiandrogenicand counselling therapyfollow-up.Interna-tional JournalofMental Health, 10,122-133.

*Motiuk,L.L.,& Brown, S. L. (1993).Survival time until suspensionfor sex offenders on conditional release(Research ReportNo. R-31).Ottawa, Ontario: Correctional Service ofCanada.

Motiuk, L.L., &Porporino,F. J. (1989).Field test of the communityrisk/needs management scale:Astudy of offenders oncaseload (Re-search Report No. R-06). Ottawa, Ontario: Correctional Service ofCanada.

*Nutbrown,V.,&Stasiak,E.(1987).Aretrospective analysisofO.C.I.cost effectiveness 1977-81(Ontario Correctional Institute ResearchMonograph No.2).Brampton, Ontario, Canada: Ontario Ministry ofCorrectionalServices.

*Pacht,A.R.,& Roberts, L. M. (1968). Factors related to parole experi-encesand thedeviatedsexoffender.Correctional Psychologist,3, 8-9,34-35.

Pedhazur,E. J.(1982).Multiple regression in behavioral research(2nded.).NewYork:Holt,Rinehart &Winston.

*Perkins, D. (1987). A psychological treatment programme for sexoffenders. In B. J. McGurk, D. M. Thornton, & M. Williams (Eds.),Applyingpsychology to imprisonment:Theoryand practice (pp.191-217). London: Her Majesty's Stationary Office.

Peters, S., Wyatt, G., & Finkelhor, D.(1986). Prevalence. In D. Fin-kelhor &Associates (Eds.),Sourcebookonchild sexual abuse (pp.15-59).Newbury Park,CA:Sage.

*Pierson,T. A.(1989).1989updatefigures for MissouriSex OffenderStudy. Jefferson City: Missouri Department of Corrections.

Pithers, W.D.,Beal, L.S.,Armstrong,J.,& Petty, J.(1989).Identifica-tionofrisk factors throughclinical interviewsandanalysisofrecords.InD. R. Laws(Ed.),Relapse preventionwithsex offenders (pp. 77-87). NewYork:Guilford Press.Pithers,W. D., Kashima, K.,dimming,G. F, Beal,L.S.,&Buell,M.(1988). Relapse prevention of sexual aggression. In R. Prentky & V.Quinsey(Eds.),Human sexual aggression:Currentperspectives (pp.244-260).NewYork:NewYorkAcademyofSciences.

*Prentky,R.A.,Knight, R.A.,& Lee, A. F.(1994,November).A 24-year follow-up of (variably) treatedsex offenders: Methodologicalissues in determining recidivism rates.Paper presentation at the annualconference of the Association for the Treatment of Sexual Abusers,SanFrancisco, CA.

*Prentky, R. A., Knight, R.A.,& Lee, A. F.(1997).Risk factorsassoci-atedwith recidivism among extrafamilial childmolesters.JournalofConsultingand Clinical Psychology, 65,141-149.

Prentky,R.A.,Knight,R.A.,Lee,A.F.,&Cerce,D. D.(1995). Pre-dictive validity of lifestyle impulsivity for rapists. Criminal JusticeandBehavior,22(2),106-128.

Proulx,J.,McKibben, A.,&Lusignan,R.(1996).Relationships betweenaffectivecomponents and sexual behaviors in sexualaggressors.Sex-ualAbuse:AJournalofResearchandTreatment,8,279-289.

*Proulx,J.,Pellerin,B.,McKibben,A.,Aubut,J.,&Ouimet,M.(1995).[Static anddynamic predictors of recidivism in sexual aggressors].Unpublished raw data.

Proulx, J.,Pellerin,B.,Paradis,Y., McKibben,A.,Aubut,J., & Ouimet,M. (1997). Static and dynamic predictors of recidivism in sexualaggressors.Sexual Abuse:AJournal ofResearchandTreatment,9,7-27.

Quinsey,V. L.(1984).Sexualaggression:Studies of offenders againstwomen. In D. N. Weisstub (Ed.),Law and mental health: Interna-tional perspectives (Vol. 1, pp.84-121).NewYork:Pergamon Press.

Quinsey,V. L.(1986).Men who have sexwithchildren. In D. N. Weiss-tub (Ed.),Law and mental health: International perspectives (Vol.2,pp.140-172).NewYork:Pergamon Press.

Quinsey, V.L., Harris, G. X, Rice, M.E., &Lalumiere, M.(1993).Assessing treatment efficacy in outcome studies of sex offenders.Journal ofInterpersonal Violence, 8,512-523.

*Quinsey,V.L.,Lalumiere,M.L.,Rice,M.E.,&Harris,G. T.(1995).Predicting sexual offenses.In J. C.Campbell (Ed.),Assessing danger-ousness: Violenceby sexual offenders, batterers,andchild abusers(pp.114-137).Thousand Oaks,CA:Sage.

*Quinsey, V.L.,Rice, M. E., &Harris,G. T.(1990). Psychopathology,sexialdeviance, andrecidivism among sex offenders released froma maximum security psychiatric institution (Res. Rep. No. VD.-I).Penetanguishene, Ontario, Canada: Mental Health Centre.

*Quinsey, V.L.,Rice, M.E.,&Harris,G. T.(1995).Actuarialpredic-tionof sexual recidivism.JournalofInterpersonal Violence,W,85105.

*Radzinowicz, L. (1957).Sexual offenses. London: MacMillan."Reddon,J. R.(1995).[Recidivism datafromthe Pheonix Program for

sex offendertreatment].Unpublished raw data.*Rice, M.E.,& Harris, G. T.(1997).Cross-validation and extensionof the Violence Risk Appraisal Guide for child molesters and rapists.

Law andHuman Behavior, 21, 2 3 1 2 4 1 .*Rice, M.E., Harris, G.T., & Quinsey, V. L. (1990).A follow-up of

rapists assessed in a maximum-security psychiatric facility. Journalof Interpersonal Violence,5, 435448.

*Rice,M.E.,Quinsey, V.L.,& Harris, G. T.(1989).Predicting sexualrecidivismamong treatedanduntreatedextrafamilial child molestersreleasedfrom amaximum security psychiatric institution (Res. Rep.No.VI-JJ1).Penetanguishene, Ontario, Canada:MentalHealth Centre.

*Rice, M.E.,Quinsey, V. L., &Harris,G. T.(1991).Sexual recidivismamong child molesters releasedfromamaximum security institution.Journal of Consulting andClinical Psychology,59,381-386.

*Romero, J. J., & Williams, L. M. (1983). Group psychotherapy andintensive probation supervision with sex offenders: A comparativestudy.Federal Probation,47, 3642.

*Rooth,R.G.,& Marks, I. M.(1974).Persistent exhibitionism:Short-term responseto aversion, self-regulation, and relaxation treatments.Archives ofSexual Behavior,3,227-247.

Rosenthal, R. (1991).Meta-analytic procedures for social research.Newbury Park, CA: Sage.

Rosenthal, R.(1995).Writingmeta-analytic reviews.Psychological Bul-letin,US, 183-192.

*Ryan,G.,&Miyoshi,T.(1990).Summaryof apilot follow-up studyof adolescent sexualperpetratorsafter treatment.Interchange, 1, 6-8.

Schmidt, F. L. (1996). Statistical significance testing and cumulative


15/15

362 HANSON AND BUSSIEREknowledge in psychology: Implications for training of researchers.Psychological Methods, I,115-129.

*Schram, D. D., Milloy. C. D., &Rowe, W. E. (1991). Juvenile sexoffenders: Afollow-up studyofreoffense behavior.Olympia: Washing-tonState Institute for Public Policy.

Seidman, B. T., Marshall, W. L., Hudson, S, M., & Robertson, P. J.(1994).An examination of intimacy andlonelinessin sex offenders.Journal ofInterpersonal Violence, 9,518-534.Shrout, P. E., & Fleiss,J. L. (1979). Intraclass correlations: Uses inassessing rater reliability.Psychological Bulletin, 86,420-428.

Slavin,R. E.(1995).Bestevidencesynthesis:Anintelligent alternativetometa-analysis. Journal of Clinical Epidemiology, 48, 9-18.

*Smiley, C., & Mulloy, R.(1995). Treatment of sex offenders: Doesinterventionwork?Unpublished manuscript, Regional Treatment Cen-tre, Correctional Service of Canada, Abbotsford, British Columbia,Canada.

*Smith, W.R., & Monastersky, C.(1986). Assessing juvenile sexualoffenders' risk for reoffending. Criminal Justice and Behavior, 13,115-140.

*Song,L.,&Lieb,R.(1995).WashingtonStatesexoffenders: Overviewofrecidivism studies.Olympia: Washington State Institute for PublicPolicy.

*Soothill,K. L.,Jack,A.,&Gibbens,T.(1976).Rape:22-yearcohortstudy.Medicine, Scienceand theLaw,16,62-69.

Spitzer, W. O. (Ed.). (1995).The Potsdam International ConsultationonMeta-Analysis [Specialissue].Journal of Clinical Epidemiology,48(1}.

Studer, L. H., Reddon, J. R., Roper, V, &Estrada, L.(1996).Phoenix:Aninpatient treatment programfor sexoffenders.Journalof OffenderRehabilitation, 23,91-97.

*Sturgeon,V,H.,& Taylor, J. (1980). Report of a five-year follow-upstudyof mentally disordered sexoffenders releasedfrom AtascaderoState Hospitalin 1973. Criminal Justice Journal,4,31-63.

*Sturup,G. K.(1953).Sexual offenders and their treatment inDenmarkandother Scandinavian countries. International Review of CriminalPolicy.4, 1-19.

*Sttirup, G. K.(1960).Sexoffenses:TheScandinavian experience.Lawand Contemporary Problems,25,361-375.

*Stiirup,G. K.(1961).Correctional treatment and thecriminal sexualoffender. Canadian Journal of Corrections, 3,250-265.

Templeman, T. L., & Stinnett, R. D.(1991).Patterns of sexual arousalandhistory in a normal sample of young men.Archives ofSexualBehavior,20,137-150.

*Thornton, D. (1995). [A 16-year follow-up of 244 sexual offendersfrom HM Prison Service].Unpublished raw data.*Tracy,R, Donnelly,H.,Morgenbesser,L., & MacDonald,D. (1983).

Program evaluation: Recidivism research involving sexoffenders. InJ. G. Greer & I. R. Stuart (Eds.), The sexual aggressor: Currentperspectives on treatment (pp. 198-213).NewYork:VanNostrandReinhold,

*VermontCenter forPreventionandTreatmentofSexual Abuse,(1991).Vermont treatment program for sexual aggressors:Programevalua-tion.(Available from 18 Blair Park Road, P.O. Box 606, Villiston,VT05495-0606).

*Weaver,C.,&Fox, C.(1984).TheBerkeleySexOffender Group:Aseven year follow-up.Probation Journal,31,143-146.

Webster,C.D., Harris,G. T,Rice,M.E.,Cormier, C.,&Quinsey,V. L.(1994).The violenceprediction scheme.Toronto, Ontario, Canada:Centre of Criminology, University of Toronto.Wierzbicki,M.,&Pekarik,G.(1993).A meta-analysis of psychotherapydropout.Professional Psychology: Research andPractice,24, 190-195.

*Williams,S.M.,&Khanna,A.(1987).Evaluation ofthe sex offendertreatment program, Regional TreatmentCentreOntario.Kingston:OntarioRegional TreatmentCentre, Correctional Serviceof Canada.

*Wing,W. A. (1984).Astudy of reoffenses bytreated sexoffenders atWestern State Hospital. Unpublished manuscript.

*Wonnith, J.S.,& Ruhl, M, (1987).Preventive detention in Canada.Journal of Interpersonal \rwlence,1,399-430.

ReceivedMarch 3, 1997Revision received June 19, 1997

Accepted July7,1997

Hanson Full

Documents