Prediction Of Sexual Recidivism Research Articles

Machine Learning and Risk Assessment: Random Forest Does Not Outperform Logistic Regression in the Prediction of Sexual Recidivism.

Although many studies supported the use of actuarial risk assessment instruments (ARAIs) because they outperformed unstructured judgments, it remains an ongoing challenge to seek potentials for improvement of their predictive performance. Machine learning (ML) algorithms, like random forests, are able to detect patterns in data useful for prediction purposes without explicitly programming them (e.g., by considering nonlinear effects between risk factors and the criterion). Therefore, the current study aims to compare conventional logistic regression analyses with the random forest algorithm on a sample of N = 511 adult male individuals convicted of sexual offenses. Data were collected at the Federal Evaluation Center for Violent and Sexual Offenders in Austria within a prospective-longitudinal research design and participants were followed-up for an average of M = 8.2 years. The Static-99, containing static risk factors, and the Stable-2007, containing stable dynamic risk factors, were included as predictors. The results demonstrated no superior predictive performance of the random forest compared with logistic regression; furthermore, methods of interpretable ML did not point to any robust nonlinear effects. Altogether, results supported the statistical use of logistic regression for the development and clinical application of ARAIs.

The Predictive Validity of the Revised Screening Scale for Pedophilic Interests (SSPI-2).

The Revised Screening Scale for Pedophilic Interests (SSPI-2) is a five-item measure that assesses for pedohebephilia (sexual attraction to prepubescent and pubescent children) based on child victim characteristics. We aimed to replicate findings by Seto, Sandler et al. (2017) by examining the predictive validity of the SSPI-2 in an independent sample of 626 men referred for a sexological assessment because of sexual offending against children. SSPI-2 scores were associated with an increased likelihood of sexual recidivism but were not significantly associated with non-sexually violent or non-violent recidivism. When they were entered together, the SSPI-2 did not contribute additional variance to the Static-99R in the prediction of sexual recidivism. Results are consistent with the findings of Seto, Sandler et al. (2017) and suggest that higher scores on the SSPI-2 may be indicative of an increased risk for sexual recidivism in individuals who have sexually offended against children.

A field study of acute dynamic risk assessment in individuals convicted of sexual offenses.

Acute dynamic risk factors can change over short periods of time, even within days or hours. When making risk assessments of individuals convicted of sexual offenses, it is necessary to consider this rapidity of change, in order to monitor appropriately the recidivism risk and to prevent future crimes. Until now, there is only one instrument-the Acute-2007-which assesses acute risk factors in individuals convicted of sexual offenses. However, the current state of research about its psychometric properties is still limited. The aim of the present study was to examine the German version of the Acute-2007 in a final sample of N = 534 adult male individuals convicted of a sexual offense in Germany. Within a prospective-longitudinal field study, appropriately trained parole officers used the Acute-2007 directly after a case was allocated to them. By using an average follow-up period of M = 3.83 years, recidivism data were drawn from the Federal Central Criminal Register of the Federal Office of Justice. Results revealed that the Acute-2007 predicted moderately general (AUC = .64), sexual (AUC = .64), and violent (AUC = .64) recidivism. By using Cox regression analysis, the Acute-2007 significantly predicted the time periods until a sexual reoffense. Furthermore, the Acute-2007 added incremental predictive validity beyond the Static-99 in the prediction of sexual recidivism. Taken together, the results of the present study support the relevance of acute dynamic risk factors in risk assessment and management of individuals convicted of sexual offenses. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Psychopathie et risque de récidive sexuelle : analyse critique de la littérature scientifique

ObjectivesFor the past twenty years, sexual violence is a major subject in the legal and criminological fields. Moreover, this phenomenon is becoming a major public health issue and becomes an important topic in termes of prevention. The main objective of this article is to clarify the current state of the litterature about the relationship between the psychopathy and the risk of sexual recidivism. Indeed, the relationship between both of them are debated among the studies right now. So, it's in this context that we wanted to expose the international studies on this topic and to discuss it. MethodsA review of literature was conducted from 1992 to 2017 from three databases: ScienceDirect, PsycInfo and PubMed. The selection process of the articles was carried out from the methodological recommendations of the « Preferred Reporting Items of Systematic reviews and Meta-Analyses » grid (PRISMA). Thus, 24 articles identifying the relationship between psychopathy and risk of sexual recidivism, or psychopathy as a predictor of sexual recidivism were selected. These 24 articles include fields studies and meta-analyzes. The diversity of the methodologies used didn’t allow the realization of meta-analysis of the results of the studies. ResultsThe results show that psychopathy appears as a risk factor for sexual recidivism or in relationship with risk of sexual recidivism or sexual recidivism only for some of the studies. The other part of them nuance these results or show an absence of relationship between psychopathy and sexual recidivism. This can be explained by the presence of other risk factors, but especially by the diversity of the methodologies used by the studies. Effectively, we can observe that researches identified in this article include different populations, that the environment of people recruted is not the same, or that different population comparisons were realised between studies. We also observe that the nature of the studies isn’t the same: some of them are longitudinal studies, others retrospective studies. This doesn’t allow us to make the same conclusions. ConclusionsTo conclude, psychopathic personality seems important to take into account in the prevention of the risk of sexual recidivism, but doesn’t appear sufficient. Indeed, it's important to consider other life factors may represent risk factors such individual characteristics, environment of the people or the nature of offenses, for example. In this perspective, the review of the literature discussed in this article aims to provide clinicians with elements of thought in the prevention of sexual recidivism. Moreover, this review aims to provide additional knowledge in psychocriminology for clinical research.

Multiple indicators of sexual interest in prepubescent or pubescent children as predictors of sexual recidivism.

Sexual interest in children is a well-known risk factor for sexual reoffending (recidivism). The present study examined the relationship between sexual interest in prepubescent children (pedophilia) or pubescent children (hebephilia) and sexual recidivism. It extended previous work by using multiple indicators of sexual interest, using a more sensitive phallometric procedure that measures change in penile blood volume, and examining both hebephilia alongside pedophilia, as well as noncontact sexual recidivism. Six hundred fifty-six men who had committed a sexual offense underwent a comprehensive sexological assessment that included multiple sexual interest measures: volumetric phallometry, self-report, and sexual behavior. Criminal record data were acquired to examine recidivism. Offenders were followed for an average of 10 years of opportunity. Indicators of hebephilia and pedophilia were significantly associated with noncontact sexual recidivism, but inconsistently associated with contact recidivism. The relationship with phallometrically assessed pedophilia and contact sexual recidivism was significant after controlling for offender age statistically. Phallometrically assessed sexual arousal to children did not add to risk appraised by an actuarial measure. Both hebephilia and pedophilia were predictive of noncontact sexual recidivism. The results for contact sexual recidivism are discussed, including a potential confound in previous work on the assessment of sexual interest in children. (PsycINFO Database Record

Acquaintance Rape: Applying Crime Scene Analysis to the Prediction of Sexual Recidivism.

The aim of the current study was to enhance the assessment and predictive accuracy of risk assessments for sexual offenders by utilizing detailed crime scene analysis (CSA). CSA was conducted on a sample of 247 male acquaintance rapists from Berlin (Germany) using a nonmetric, multidimensional scaling (MDS) Behavioral Thematic Analysis (BTA) approach. The age of the offenders at the time of the index offense ranged from 14 to 64 years (M = 32.3; SD = 11.4). The BTA procedure revealed three behavioral themes of hostility, criminality, and pseudo-intimacy, consistent with previous CSA research on stranger rape. The construct validity of the three themes was demonstrated through correlational analyses with known sexual offending measures and criminal histories. The themes of hostility and pseudo-intimacy were significant predictors of sexual recidivism. In addition, the pseudo-intimacy theme led to a significant increase in the incremental validity of the Static-99 actuarial risk assessment instrument for the prediction of sexual recidivism. The results indicate the potential utility and validity of crime scene behaviors in the applied risk assessment of sexual offenders.

A cognitive-based indicator of deviant sexual interest: Concurrent validation of the Stroop task

ABSTRACTOne of the strongest predictors of sexual recidivism is sexual deviance [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. doi:10.1037/0022-006X.66.2.348]. Phallometric testing, the most commonly used method of assessing sexual deviance, has elicited methodological and ethical criticisms, while self-report is vulnerable to social desirability and poor insight. To overcome these limitations, researchers have utilised cognitive measures, including a modified Stroop task, to measure deviant sexual interests among sex offenders using victim selection as a comparison measure. However, the results have been inconclusive. The current study explored the validity of the modified Stroop task as a measure of deviant sexual interest among a non-offending sample of 570 females and 223 male participants, using self-report as the comparison measure. The results indicated a significant gender difference in concurrent validity; there was a significant relationship between self-reported deviance and deviant word Stroop performance for male but not for female respondents, suggesting the Stroop is a viable option for assessing sexual deviance among males. Implications of these findings are discussed relative to sexual recidivism risk assessment.

Expeditious Disclosure of Sexual History via Polygraph Testing: Treatment Outcome and Sex Offense Recidivism

Self-disclosure of problematic behavioral history and the timeliness of full disclosure were explored as indicators of treatment responsivity and possible predictors of sexual recidivism. Specifically, this study evaluated the relationships among the variables of expeditious full disclosure of sexual history (measured via polygraph testing), treatment completion status, and sexual recidivism within a sample of 192 convicted sex offenders on probation/parole. The study compared a group of non-expeditious disclosers with expeditious disclosers across several variables (i.e., risk score/level, age of full disclosure, presence of sexual deviance, psychopathy, and denial). The independent variable of expeditious disclosure was defined as having disclosed sexually deviant behavioral history sufficient to produce a “no significant reaction” result on a sexual history polygraph examination (within 12 months of treatment program intake). A 5-year sexual recidivism follow-up period was utilized. Univariate analyses revealed that though expeditious disclosure was positively correlated with treatment completion, it was not a statistically significant factor in predicting sexual recidivism. The study provides emerging evidence that polygraph-facilitated full disclosure of sexual history is related treatment outcome and that it is worthy of additional study with larger samples in regard to any relationship with sexual recidivism.

What does it mean when age is related to recidivism among sex offenders?

Age is a robust predictor of recidivism and an item on actuarial tools commonly used to predict sexual violent recidivism among sex offenders. However, little is known about whether or how much offenders' risk diminishes as a result of aging. In the first of two studies, we examined the sexual and violent recidivism of 533 sex offenders who were over age 50 on release. Age at index offense was at least as good at predicting both outcomes as was age at release, and age at index offense provided at least as much incremental validity in the prediction of violent recidivism to scores on a brief static actuarial tool. Neither age added incrementally to static score in the prediction of sexual recidivism. The second study examined how well age at first offense, age at index offense, and age at release predicted violent recidivism among 527 sex offenders aged 13 to 79 at release. Age at first offense predicted best. When age was removed from score on the Sex Offender Risk Appraisal Guide, all ages added incrementally but age at release least to SORAG score. When participants were divided into quartiles based on age at index offense, there was no evidence from any quartile that age at release predicted violent recidivism better than age at first offense. The authors concluded that age at release is a poor index of within-subject changes in risk of sexual or violent recidivism. No adjustment to a sex offender's score on a comprehensive actuarial tool that includes age at first or index offense should be made simply because the offender is older.

Psychopathy and the combination of psychopathy and sexual deviance as predictors of sexual recidivism: Meta-analytic findings using the Psychopathy Checklist—Revised.

Clinicians routinely administer Hare's (2003) Psychopathy Checklist-Revised (PCL-R) to sex offenders and report PCL-R scores as meaningful predictors of recidivism risk. Although a 2005 meta-analysis reported a small (d=0.29) association between PCL-R scores and sexual recidivism (Hanson & Morton-Bourgon, 2005), no meta-analysis has examined effects for PCL-R factors and facets, the widely cited combination of high PCL-R and high sexual deviance scores, or potential moderators of the PCL-R/recidivism relation. We conducted a meta-analysis of effects from all available studies examining the relation between PCL-R scores and sexual recidivism (k=20, N=5,239). The effect for PCL-R Total scores predicting sexual recidivism was d=0.40, which falls beyond the upper end of the 2005 confidence interval [.20, .38]. Effects were stronger for Factor 2 (d=0.44) and Facet 4 (d=0.40) scores than other factor or facet scores (ds=0.01 to 0.17). Effects tended to be stronger for scores calculated for research (d=0.44) compared to those calculated for clinical use (d=0.28). Offenders who scored high on both the PCL-R and a measure of sexual deviance were more likely to reoffend sexually than other offenders (odds ratio=2.80 to 3.21, k=6). Results indicate that PCL-R scores, particularly combined with a measure of sexual deviance, are potentially relevant to sex offender risk. But results also underscore several practical challenges to implementing these findings in routine clinical practice.

The assessment of dynamic risk and recidivism in a sample of special needs sexual offenders

The predictive validity of four risk assessment instruments: the RRASOR, SVR-20, RM2000-V and the ARMIDILO-Stable and -Acute dynamic client subscales were assessed on a sample of 88 offenders: 44 mainstream and 44 sexual offenders with special needs, who had been matched on risk items within the RRASOR tool. Instruments were coded retrospectively from file information. Sexual reconviction data was used, in conjunction with sexual recidivism data based on unofficial data sources, over a mean follow-up period of 8.8 years. The results of this study found that the ARMIDILO instrument was the best predictor for sexual reconviction among offenders with special needs (ARMIDILO-Stable, AUC=0.60; ARMIDILO-Acute, AUC=0.73), while the predictive validities of the RRASOR (AUC=0.53) and the RM2000-V (AUC=0.50) were little better than chance. In contrast, the SVR-20 yielded a higher score (AUC=0.73) for the non-ID sample, than for the intellectually disabled sample (AUC=0.45). Within the special needs group, the ARMIDILO-Acute, SVR-20 Psychosocial Affect, and Overall scales were better predictors of sexual recidivism for the intellectually disabled subgroup (AUCs ranging from 0.75 to 0.88). These results are discussed in the context of current practice.

Are Actuarial Risk Data Used to Make Determinations of Sex Offender Risk Classification?

This study examined whether evaluators use actuarial risk scores and risk information to make determinations about sex offender risk status for the purpose of enhanced registration and notification. Although it was expected that sexual offenders selected for enhanced registration and notification would have higher scores on actuarial risk assessment tools than those who were not selected, few differences were found between groups with regard to risk factors associated with sexual offense recidivism. Given that actuarial tools enhance the prediction of sexual recidivism, this study may shed light on problems in the implementation of sex offender policy measures. Results are discussed as they pertain to the assessment and application of registration and community notification statutes for sexual offenders.

Sex Offenses and Offenders Reconsidered

Analyses of national and state-based data suggest that the rate of reported sexual offenses has declined in the past 25 years. This decline began prior to the implementation of registration and notification laws in the mid-1990s and continued to decline after the laws’ passage. It remains unclear, however, whether registration and notification laws have produced a qualitative change in the characteristics of the offender population and/or the predictors of sexual recidivism. This study investigates key differences among offenders pre- and postimplementation of community notification laws across an 11-year period (1990-2000) using a random sample of 550 offenders in New Jersey. All offenders were followed for 6.5 years, and reoffenses were analyzed across a range of demographic, clinical, and criminal offense predictors.

Predicting sex offender recidivism. I. Correcting for item overselection and accuracy overestimation in scale development. II. Sampling error-induced attenuation of predictive validity over base rate information.

The authors demonstrate a statistical bootstrapping method for obtaining unbiased item selection and predictive validity estimates from a scale development sample, using data (N = 256) of Epperson et al. [2003 Minnesota Sex Offender Screening Tool-Revised (MnSOST-R) technical paper: Development, validation, and recommended risk level cut scores. Retrieved November 18, 2006 from Iowa State University Department of Psychology web site: http://www.psychology.iastate.edu/ approximately dle/mnsost_download.htm] from which the Minnesota Sex Offender Screening Tool-Revised (MnSOST-R) was developed. Validity (area under receiver operating characteristic curve) reported by Epperson et al. was .77 with 16 items selected. The present analysis yielded an asymptotically unbiased estimator AUC = .58. The present article also focused on the degree to which sampling error renders estimated cutting scores (appropriate to local [varying] recidivism base rates) nonoptimal, so that the long-run performance (measured by correct fraction, the total proportion of correct classifications) of these estimated cutting scores is poor, when they are applied to their parent populations (having assumed values for AUC and recidivism rate). This was investigated by Monte Carlo simulation over a range of AUC and recidivism rate values. Results indicate that, except for the AUC values higher than have ever been cross-validated, in combination with recidivism base rates severalfold higher than the literature average [Hanson and Morton-Bourgon, 2004, Predictors of sexual recidivism: An updated meta-analysis. (User report 2004-02.). Ottawa: Public Safety and Emergency Preparedness Canada], the user of an instrument similar in performance to the MnSOST-R cannot expect to achieve correct fraction performance notably in excess of what is achievable from knowing the population recidivism rate alone. The authors discuss the legal implications of their findings for procedural and substantive due process in relation to state sexually violent person commitment statutes and the Supreme Court's Kansas v. Hendricks decision regarding the constitutionality of such statutes.

Predictive validity of the SVR-20 and Static-99 in a Dutch sample of treated sex offenders.

In this retrospective study, the interrater reliability and predictive validity of 2 risk assessment instruments for sexual violence are presented. The SVR-20, an instrument for structured professional judgment, and the Static-99, an actuarial risk assessment instrument, were coded from file information of 122 sex offenders who were admitted to a Dutch forensic psychiatric hospital between 1974 and 1996 (average follow-up period 140 months). Recidivism data (reconvictions) from the Ministry of Justice were related to the risk assessments. The base rate for sexual recidivism was 39%, for nonsexual violent offenses 46%, and for general offenses 74%. Predictive validity of the SVR-20 was good (total score: r = .50, AUC = .80; final risk judgment: r = .60, AUC = .83), of the Static-99 moderate (total score: r = .38, AUC =.71; risk category: r = .30, AUC = .66). The SVR-20 final risk judgment was a significantly better predictor of sexual recidivism than the Static-99 risk category.

Additional Flaws in the Minnesota Sex Offender Screening Tool-Revised

Abstract Wollert's (2002) paper on actuarial tests that are used for the prediction of sexual recidivism (ATSRs) showed that the recidivism rates for the score groups in the MnSOST-R developmental sample were inflated, and that they “shrank” by as much as 44 percentage points on cross-validation. Doren and Dow (2002) present an analysis of Wollert's (2002) research and an alternative study that they claim demonstrated a lack of shrinkage. This commentary presents a critique of Doren and Dow's response. Various considerations indicate that the original cross-validation was appropriate. Doren and Dow's alternative was also found to be unacceptably flawed. Additional analyses cast doubt on the MnSOST-R's validity coefficients and its relevance for identifying likely recidivists, and a consensus was apparent that the MnSOST-R experience table for the developmental sample should not be used in civil commitment hearings. Finally, a number of guidelines for the development of ATSRs were reviewed. In particular, ...

Risk assessment of sexual offenders

With the frequency of acts of sexual aggression in our community increasing, clinicians are now under greater pressure to make predictions concerning the propensity of sexual offenders to re‐offend. Thus, the development of accurate and valid tools for assessing sexual recidivism, along with the identification of predictive variables on which such tools are based, has become a focus of research. This article reviews the literature identifying predictors of sexual recidivism and examines the various approaches that may be employed by mental health professionals in the assessment of risk of sexual offenders.

Predictors of sexual recidivism: did meta-analysis clarify the role and relevance of denial?

Seven studies selected for meta-analysis of the denial variable in a large-scale meta-analysis of predictors of sexual recidivism are briefly reviewed and a number of methodologic issues are identified that relate to interpretation of the failure of meta-analysis to find an effect for the denial variable. These include variability in definition of the denial variable, variation in treatment access and exclusion for deniers, low base rates of recidivism, small sample sizes, low power, and high probability of type II error across individual studies. Despite the apparent objectivity and power of this methodology, a careful examination of the individual studies suggests that meta-analysis did not clarify the role and relevance of denial as a predictor of sexual recidivism. Further work is necessary to clarify the role and relevance of denial in treatment success, risk assessment, and risk prediction.

Cross-validation and extension of the Violence Risk Appraisal Guide for child molesters and rapists.

A cross-validation of the Violence Risk Appraisal Guide was performed on a sample of 159 child molesters and rapists followed for an average of 10 years at risk. The performance of the instrument was also examined on a 10-yr followup of 288 sex offenders that included both those in the original construction sample for the VRAG and the validation sample. The instrument performed as well as it had in construction for predicting violent recidivism in both the cross-validation and extended followup samples, and moderately well in the prediction of sexual recidivism. Survival analyses showed that child molesters exhibited higher risk of sexual recidivism than rapists or offenders against both children and adults, whereas the opposite was true for violent recidivism. As predicted, psychopathy and phallometrically determined sexual deviance exhibited a multiplicative interaction effect on sexual recidivism. Proportional hazards event history analyses supported the use of the VRAG for the prediction of violent recidivism among sex offenders.

Actuarial Prediction of Sexual Recidivism

An actuarial approach to predicting sexual recidivism among known sex offenders is illustrated with a reanalysis of follow-up data of rapists and child molesters. One hundred seventy-eight sex offenders who had been assessed at a maximum security psychiatric facility were followed for an average of 59 months of opportunity to reoffend. Twenty-eight percent were convicted of a new sex offense and 40% were arrested, convicted, or returned to the psychiatric facility for a violent (including sex) offense. Rapists were more likely to recidivate than child molesters. Psychopathy, measures of previous criminal history, and phallometric indexes of deviant sexual interests were found to be useful predictors of sexual recidivism. A predictor scale was formed by weighting each predictor that was significant in the regression analyses by a number reflecting its univariate correlation with the criterion. A linear relationship (r = .45) was found between scores on the predictor scale and reconviction for a sexual offense.