Recidivism Risk Assessment Research Articles

Predictive Validity of Tools for Assessing Recidivism Risk in Men Convicted of Sex Offending: Static-99R, Static-2002R and BARR-2002R.

Recidivism risk assessment is crucial for effective case management of men convicted of sex offending. The use of empirical actuarial risk tools has become routine in the field. However, the development of actuarial risk scales for assessing general, violent and/or sexual recidivism in these men is ongoing: The Static-99 led to the Static-99R and the Static-2002R, and the BARR-2002R emerged to assess violent recidivism risk. A study was undertaken to evaluate and compare the inter-rater and predictive validity of the Static-99R, the Static-2002R, and the BARR-2002R in a sample of 328 men convicted of sex offending released from prison in French Belgium. When the instruments were considered integrally, the two versions of the Static-the Static-99R and the Static-2002R-proved better at predicting sexual recidivism and the BARR-2002R was better at predicting violent recidivism. And, the predictive and incremental predictive validity of the factor structure identified by Brouillette-Alarie et al. (2016) was examined. Results proved consistent in that the Youthful stranger aggression and General criminality factors were better at predicting general recidivism and violent non-sexual recidivism while the Persistence/paraphilia factor was better at predicting sexual recidivism.

Vulnerability to incest – Findings from a comparative single-case study of the onset of intrafamilial child sexual abuse

Previous research about intrafamilial child sexual abuse was not able to identify specific risk factors that distinguish this unique subgroup from other sexual offending subgroups. In comparison to other groups of sexual offenders, men convicted of intrafamilial sexual child abuse (ICSA) are found to exhibit more similarities to non-offending fathers than extrafamilial sexual offenders. Consequently, the risk assessment of sexual recidivism among “incest offenders” lacks evidence-based evaluation criteria. Given the suggestion that family system factors should be included in research on the onset of ICSA, we employed the Vulnerability to Incest Model proposed by Trepper and Barrett (1989, https://doi.org/10.4324/9780203776605) in a qualitative single-case analysis. A comparative analysis of ten court evaluations of ICSA offenders revealed that all families in which ICSA has occurred demonstrated at least two vulnerability factors. The analyzed offenders exhibited comparable patterns of masculine sexual entitlement. The utility of this recently developed construct for sexual violence research is discussed and implications for further research proposed.

Intellectual Ability and Sexual Recidivism Risk Assessment: Comparing Predictive Accuracy in an Incarcerated U.S. Sample.

While studies indicate that individuals with intellectual disabilities (ID) and borderline intellectual functioning (BIF) are overrepresented among those incarcerated for sexual offenses, there is a paucity of research with respect to risk assessment for this population, particularly among incarcerated U.S. samples. This gap is concerning given that individuals perceived as high risk may be denied bail, given harsher sentences, or subject to civil confinement. The present study examined the predictive validity of the Static-99 and Static-99R, which rely on historical predictors, and the MnSOST-R, which includes a dynamic subscale. Archival records were used to compare individuals with (n = 459) and without (n = 459) ID/BIF released from New Jersey state prisons between 1996 and 2007, with an average follow-up period of 5.7 years. All three instruments demonstrated good predictive validity for those without ID/BIF, while only the MnSOST-R produced a significant AUC value for those with ID/BIF. Reincarceration rates for those with ID/BIF within high-risk groups ranged from 0% to 8%, similar to the overall recidivism base rate of 4%. Results highlight the need for additional research to determine which risk factors should be considered for those with ID/BIF.

Risk to reoffend changes over time: Improving correctional programming through progress monitoring.

Progress monitoring is integral to evidence-based practice. Correctional settings, especially the supervision of individuals who commit sexual offenses, elicit public concern; negative outcomes can be catastrophic. Using a prospective longitudinal study of 2,939 men with a history of sexual offenses undergoing community supervision, we examined different models of progress monitoring and how they should inform the assessment of risk for sexual recidivism. We found that the most recent assessment scores of the ACUTE-2007 and STABLE-2007 sexual recidivism risk tools provided the best information about reoffending risk compared to using (a) the worst period of adjustments (i.e., highest risk score), (b) the best period of adjustments (i.e., lowest risk score), or (c) a rolling average of scores. We also found that the latest STABLE-2007 scores incrementally predicted sexual recidivism beyond baseline risk as assessed by demographic and criminal history variables (Static-99R). We conclude that the risk for sexual recidivism changes over time and that community corrections is advanced by repeated assessment of dynamic (changeable) risk factors. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

A Study on the Criminal Application of Near-term Artificial Intelligence - Taking recidivism risk assessment of domestic violence as an example -

A Study on the Criminal Application of Near-term Artificial Intelligence - Taking recidivism risk assessment of domestic violence as an example -

Pelaksanaan Asesmen Risiko dan Asesmen Kebutuhan Terhadap Narapidana Narkotika Di Lapas Narkotika Kelas III Sawahlunto

Carrying out risk and needs assessments as regulated in Article 36 Paragraph 4 of Law Number 22 of 2022 concerning Corrections is related to the need for guidance, the risk of running away, the risk of danger to other people and the mental physical and psychological health of correctional inmates. For inmates at Lapas Narkotika Kelas III Sawahlunto, this assessment is carried out by the Bapas Klas I Padang. The analysis and discussion of this research uses an empirical juridical approach using field research and interviews with informants. The research results showed that risk and needs assessments were carried out using several instruments such as the Prisoner Placement Screening Instrument (ISPN), the Indonesian Recidivism Risk Assessment Instrument (RRI) and the Criminogenic Needs Assessment. The results of the assessment are recommendations in nature but should be implemented by the prison in order to determine the prisoner development program. The obstacles faced in carrying out this risk and needs assessment are related to the distance between the Bapas Klas I Padang and the Lapas Narkotika Kelas III Sawahlunto, the lack of budget and lack of coordination, as well as the inactivity of the Bapas Post which causes the wide working area of ​​the Bapas Klas I Padang

Próba implementacji kurateli klinicznej i menedżeryzmu w dozorach probacyjnych

Introduction: At present in many countries the model of the work of probation officers has been subordinated to the case management method. In the Ministry of Justice started work on combining the resocialization treatment by probation officers with the assessment of recidivism risk in convicts classified into three groups of recidivism. In this manner, the probation methodology can shift from educational purposes to prevention and control. The principal goal of the proposed solutions was to turn probation officers into a professional service responsible for the protection of society and probationers themselves against next criminal acts. Research Aim: This is a review article. Its purpose to present the course of the reform of guardianship for adults after the political transformation. Evidence-based Facts: While preparing new solutions did not pay sufficient attention to the moment of implementation of the proposed recidivism risk management model into the practice of conducting probation supervisions, grading of control and adjustment of resocialization programs to different risk groups. The unprofessional implementation of the recidivism risk management model may result in the mass exclusion of the convicted persons from resocialization in an open environment. Summary: The probation officers were not prepared to recognize the criminogenic needs of the probationers. When formulating forecast, they use a very limited list of predictors of the future criminal behaviour of the probationers. This means that the probation officers must be equipped with appropriate, reliable, proven and structured recidivism risk estimation tools.

Оценка риска совершения преступлений лицами, освобожденными из мест лишения свободы

Introduction. While studying recidivism prediction, criminal risk assessment is considered in many countries a mandatory procedure. The paper presents the most widely known and frequently applied risk assessment tools. Special attention is paid to predicted risk assessment tools for individuals released in the Russian Federation and the Republic of Kazakhstan. Taking into account the emerging tradition of using these risk assessment tools for practical purposes and the opportunities for conducting relevant research, much attention is paid to reviewing the tools used abroad. The purpose of the study is to provide the theoretical grounds for recidivism risk assessment tools, as well as the analysis of foreign experience in application and verification of this tool. Research methods. The research uses general scientific methods: (analysis, synthesis, systematization, generalization, analogy), special methods: comparative (when studying tools for assessing the risk of recidivism), formal legal (for the study of normative legal acts). Results. The research demonstrates a wide range of approaches and patterns in the area of criminal behavior risk assessment. Criminal behavior risk assessment tools were classified, and they can be summarised in a historical perspective into four generations according to chronology. The first and least reliable approach is to assess the risk of recidivism based on the clinical opinion of professionals. At this stage the measurement was characterised by its subjectivity. The second generation was based on actuarial valuation using reliable statistical predictors and significance levels for recidivism. The third generation tended to combine risk factors based on the theory of static risk assessment. The fourth generation of tools is based on the understanding that risk should be assessed as a continuous and dynamic process related to both the risk itself and the needs and resources of the individual. This approach considers that offenders’ supervision in post-penitentiary probation is an effective means of successful risk assessment based on their individual psychological characteristics and resources.

A comparative user study of human predictions in algorithm-supported recidivism risk assessment

AbstractIn this paper, we study the effects of using an algorithm-based risk assessment instrument (RAI) to support the prediction of risk of violent recidivism upon release. The instrument we used is a machine learning version of RiskCanvi used by the Justice Department of Catalonia, Spain. It was hypothesized that people can improve their performance on defining the risk of recidivism when assisted with a RAI. Also, that professionals can perform better than non-experts on the domain. Participants had to predict whether a person who has been released from prison will commit a new crime leading to re-incarceration, within the next two years. This user study is done with (1) general participants from diverse backgrounds recruited through a crowdsourcing platform, (2) targeted participants who are students and practitioners of data science, criminology, or social work and professionals who work with RisCanvi. We also run focus groups with participants of the targeted study, including people who use RisCanvi in a professional capacity, to interpret the quantitative results. Among other findings, we observe that algorithmic support systematically leads to more accurate predictions from all participants, but that statistically significant gains are only seen in the performance of targeted participants with respect to that of crowdsourced participants. Among other comments, professional participants indicate that they would not foresee using a fully-automated system in criminal risk assessment, but do consider it valuable for training, standardization, and to fine-tune or double-check their predictions on particularly difficult cases. We found that the revised prediction by using a RAI increases the performance of all groups, while professionals show a better performance in general. And, a RAI can be considered for extending professional capacities and skills along their careers.

Examining Risk Factor and Recidivism Rate Differences Between Youth Adjudicated for Sex and Non-Sex Offenses: A Propensity Score Matching Approach

ABSTRACT Youth adjudicated for sex offenses are frequently overridden to high-risk on general recidivism risk assessments. Little research supports this use of overrides. This study seeks to examine the extent to which those adjudicated for sex and non-sex offenses differ on recidivism rates in a sample of 1,380 sex and non-sex offending youth from a Midwestern state juvenile court. Propensity score matching was used to examine general recidivism rate differences between those adjudicated for sex and non-sex offenses. After matching youth on race, age, and eight criminogenic risk factors, youth adjudicated for sexual offenses recidivated at a significantly lower rate.

Liens entre les compétences émotionnelles, les difficultés de régulation émotionnelle et la qualité de vie professionnelle des psychologues médico-légaux francophones

ObjectivesDespite being less recognized as a specific subdomain of Psychology in French-speaking countries, forensic psychologists do face the same challenges as their Anglo-Saxon counterparts. Among these, the literature identifies the quality of life in the workplace, which can be divided into Compassion Satisfaction, and Compassion Fatigue, the latter being composed of Burn-out and Secondary Traumatic Stress. Previous research has acknowledged the role of emotional competence, especially the management of emotions, as a central factor of healthcare professionals’ relationship with their work assignments. However, Pirelli and collaborators recently highlighted the lack of empirical research among forensic psychologists. As an answer to their call, this study investigates the association and predictive factors of emotional competence, difficulties in emotional management, and the quality of professional life. Materials and methodsAn online survey (Lime Survey) was used to collect data regarding socio-demographic variables such as age, gender, years of experience, and training. The following self-questionnaires were then given to the subjects: Profile of Emotional Competence, Difficulties in Emotion Regulation Scale, and Professional Quality of Life. Descriptive, correlational, and regressive analyses were undertaken to, respectively: (1) describe our sample and compare it to available norms (Student's t), (2) associate (Bravais-Pearson r) the mean scores of the three scales, and (3) evaluate the prediction (simple and multiple linear regression) of emotional competence and emotional dysfunction on the quality of professional life. These predictions were then reversed. Both regression analyses used the backward method. ResultsThe sample was composed of 44 French-speaking forensic psychologists (79.50% female), median age 34.20 (SD=7.96). They were mainly trained in clinical psychology (84.10%), followed by forensic psychology (18.20%). The majority of the sample worked as carceral psychologists (59.10%), and their main missions were support and guidance (78.70%), followed by recidivism risk assessment (52.30%) and crisis management (50.00 %). Descriptive results. The participants exhibited significantly higher emotional dysfunction scores (total score, p ≤ .001; Lack of Emotional Clarity, p=.10; Impulse Control Difficulties,p ≤ .001; Difficulties Engaging in Goal-Directed Behavior,p ≤ .001) when compared to available norms. Regarding their perceived quality of life, all scores were either significantly lower or higher based on the norms used. A similar analysis was not possible for emotional competence as available norms were gender-specific, and the sample gender distribution was clearly unbalanced. Associations and regression relations were confirmed between emotional competence, emotional dysfunction, and the quality of professional life, especially between Compassion Satisfaction and emotional competence (r=.33–.56) on the one hand and between Burn-out and emotional dysfunction on the other hand (r=.41–.65). Unexpectedly, Secondary Traumatic Stress was only predicted by emotional dysfunction (Impulse Control Difficulties, β=.39, t=2.71, R2=.13, p=.010). ConclusionsTaken as a whole, the results indicate that (a) Compassion Satisfaction and Burn-out are key elements to assess in the workplace, confirming previous studies among other health professionals, and (b) although correlated to specific emotional dysfunction such as Difficulties Engaging in Goal-Directed Behaviors and Impulse Control Difficulties, none of these factors predict nor is predicted by Secondary Traumatic Stress.

Exploratory rasch analysis of a static-99R clinical cohort assessment.

Modern test theory supplements the more prevalent classic methods for assessing test properties. However, such an assessment of the commonly used sexual recidivism risk assessment instrument, Static-99R, has yet to be attempted. This study evaluated the psychometric properties of said instrument using Rasch analysis. The clinical cohort assessed consisted of individuals with mental disorders convicted of a sexual offense (N = 146). Results showed that the original ten-item Static‑99R did not meet the Rasch model requirements, and revisions of the instrument with seven and nine items each only marginally improved performance. More reliable results could likely have been obtained with a larger, non-clinical sample and a more randomized distribution of missing data. Despite the consistently poor performance of item 3 ("Index non-sexual violence") in all three analyses, reliability was slightly improved by dichotomizing the only two polytomous items in the Static-99R; items 1 ("Age at release from index offense") and 5 ("Prior sexual offenses"). These results may be of interest considering the significant change of splitting the formerly dichotomous item 1 into four different response categories in the revision of Static-99 to Static-99R.

Interpretable algorithmic forensics

One of the most troubling trends in criminal investigations is the growing use of “black box” technology, in which law enforcement rely on artificial intelligence (AI) models or algorithms that are either too complex for people to understand or they simply conceal how it functions. In criminal cases, black box systems have proliferated in forensic areas such as DNA mixture interpretation, facial recognition, and recidivism risk assessments. The champions and critics of AI argue, mistakenly, that we face a catch 22: While black box AI is not understandable by people, they assume that it produces more accurate forensic evidence. In this Article, we question this assertion, which has so powerfully affected judges, policymakers, and academics. We describe a mature body of computer science research showing how “glass box” AI—designed to be interpretable—can be more accurate than black box alternatives. Indeed, black box AI performs predictably worse in settings like the criminal system. Debunking the black box performance myth has implications for forensic evidence, constitutional criminal procedure rights, and legislative policy. Absent some compelling—or even credible—government interest in keeping AI as a black box, and given the constitutional rights and public safety interests at stake, we argue that a substantial burden rests on the government to justify black box AI in criminal cases. We conclude by calling for judicial rulings and legislation to safeguard a right to interpretable forensic AI.

인공지능(AI)을 활용한 보호관찰대상자 분류 연구

TAs a research method, a literature study was conducted focusing on the 'Recidivism Risk Assessment Tool for Adults Subject to Probation' (KPRAI-R), which is used by the Crime Prevention Policy Bureau of the Ministry of Justice. In particular, the classification of probation subjects was divided into 'initial classification' and 'reclassification', and the limitations of each were derived, and the necessity of using artificial intelligence was divided into initial classification and reclassification and considered. Through this, it was proposed to improve the initial classification and reclassification through the establishment of an 'artificial intelligence (AI)-based automatic classification system'.
 As for the specific improvement measures, first, in relation to the initial classification, ① a more sophisticated AI-based classification system was derived by combining the information analyzed during the initial classification with the evaluation method of the recidivism risk assessment tool (KPRAI-R), ② the career and A plan to use artificial intelligence to minimize the deviation according to inclination and expertise was presented. Second, in relation to reclassification, ① prepare an artificial intelligence-based automatic reclassification system based on post-mortem information analysis after the start of probation, ② prepare an alarm system to automatically recognize the risk of recidivism and prevent recidivism by analyzing additional information collected during probation presented.

Protective factors in the prediction of criminal outcomes for youth with sexual offenses using tools developed for adults and adolescents: Tests of direct effects and moderation of risk.

Interest in protective factors in risk assessment work with adjudicated populations is increasing and evidence suggests that protective factors in structured professional judgment (SPJ) tools predict the absence of one or more types of recidivism with some evidence also of incremental validity in recidivism-desistance prediction models with risk scales. But, there is little evidence of interactions, demonstrated using formal tests of moderation, between scores on risk- and protective factor-focused applied assessment tools, despite the documentation of interactive protective effects with nonadjudicated populations. In this study, with 273 justice-involved male youth and a fixed 3-year follow-up, direct effects of medium size were found for sexual recidivism, violent (including sexual) recidivism, and any new offense with totals for tools developed for adult offending populations (modified versions of the actuarial risk-focused Static-99 and the SPJ protective factor-focused Structured Assessment of PROtective Factor [SAPROF]) and tools developed for adolescent offending populations (the actuarial risk-focused Juvenile Sexual Offense Recidivism Risk Assessment Tool-II [JSORRAT-II] and the SPJ protective factor-focused DASH-13). As well, incremental validity and interactive protective effects, in the small-to-medium size range, were found for the prediction of violent (including sexual) recidivism using various combinations of these tools. The value-added information provided by strengths-focused tools indicated by these findings suggest their inclusion in comprehensive risk assessments in applied practice has promise for improving prediction and also intervention and management planning with justice-involved youth. The findings also highlight the need for further research on developmental considerations and practical questions about how to integrate strengths with risks to inform such work empirically. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

A comparison of the predictive accuracy of structured and unstructured risk assessment methods for the prediction of recidivism in individuals convicted of sexual and violent offense.

One of the most commonly replicated results in the research area of recidivism risk assessment is the superiority of structured and standardized prediction methods in comparison to unstructured, subjective, intuitive, or impressionistic clinical judgments. However, the quality of evidence supporting this conclusion is partly still controversially discussed because studies including direct comparisons of the predictive accuracy of unstructured and structured risk assessment methods have been relatively rarely conducted. Therefore, we examined in the present study retrospectively N = 416 expert witness reports written about individuals convicted of violent and/or sexual offenses in Germany between 1999 and 2015. The predictive accuracy of different methodological approaches of risk assessment (subjective clinical [i.e., unstructured clinical judgment; UCJ], structured professional judgment [SPJ], actuarial risk assessment instruments [ARAIs], and combinations of ARAIs-/SPJ-based risk assessments) was compared by analyzing the actual reoffenses according to the Federal Central Register (average follow-up period M = 7.08 years). In accordance with previously published results, the results indicated a higher predictive accuracy for structured compared to unstructured risk assessment approaches for the prediction of general, violent, and sexual recidivism. Taken together, the findings underline the limited accuracy of UCJs and provided further support for the use of structured and standardized risk assessment procedures in the area of crime and delinquency. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Međuposmatračka pouzdanost instrumenta za procenu rizika od recidivizma kod visokorizičnih osuđenih

Recidivism risk assessment in the penal system has significantly advanced, and today assessment instruments are in use that should contribute to more effective treatment of convicts. However, despite the efforts invested in the development and enhancement of the instruments, there are still certain problems in their application, which researchers and practitioners strive to eliminate or minimize. In accordance with the aforementioned, the aim of this paper is to examine the interrater reliability of the Questionnaire for Assessing Risk, Capacity and Needs of convicts in the prison system of the Republic of Serbia. The study was conducted in three penitentiaries (Požarevac, Sremska Mitrovica and Niš) on a sample of 54 convicted persons who were assessed as having a high or extremely high recidivism risk in the reception department and 35 employed professional staff members. The obtained data indicate a moderate agreement (0.637) measured by Cohen's Kappa coefficient, and an absolute agreement among of the evaluators in 88.9% of cases. On the other hand, absolute agreement of the evaluators, when looking at the recidivsm risk score, exists in only 7 cases. It was found that the largest number of items have a moderate agreement, as well as that the static recidivism risk factors were most consistently evaluated. The findings of our research show that the instrument has moderate rater agreement, on a sample of a high recidivism risk offenders, but also that there are certain problems that affect interobserver reliability.

Understanding risk with FOTRES?

The present paper examines the recidivism risk assessment instrument FOTRES, addressing the questions whether FOTRES provides us with an adequate understanding of risk, whether we actually understand FOTRES itself, and whether FOTRES is fair. The evaluation of FOTRES uses the criteria of empirical accuracy, representational accuracy, domain of validity, intelligibility, and fairness. This evaluation is compared to that of COMPAS, a different, much-discussed risk assessment instrument. The paper argues that FOTRES performs poorly in comparison to COMPAS with respect to some of the criteria, and that both FOTRES and COMPAS do not show a satisfactory performance with respect to other criteria.

Predicting Recidivism Risk Meets AI Act

Quantitative recidivism risk assessment can be used at the pretrial detention, trial, sentencing, and / or parole stage in the justice system. It has been criticized for what is measured, whether the predictions are more accurate than those made by humans, whether it creates or increases inequality and discrimination, and whether it compromises or violates other aspects of fairness. This criticism becomes even more topical with the arrival of the Artificial Intelligence (AI) Act. This article identifies and applies the relevant rules of the proposed AI Act in relation to quantitative recidivism risk assessment. It does so by focusing on the proposed rules for the quality of the data and the models used, on biases, and on the human oversight. It is concluded that legislation may consider requiring providers of high-risk AI systems to demonstrate that their solution performs significantly better than risk assessments based on simple models, and better than human assessment. Furthermore, there is no single answer to evaluate the performance of quantitative recidivism risk assessment tools that are or may be deployed in practice. Finally, three approaches of human oversight are discussed to correct for the negative effects of quantitative risk assessment: the optional, benchmark, and feedback approach.

Attending to the Positive: A Retrospective Validation of the Structured Assessment of Protective Factors-Sexual Offence Version.

Sexual recidivism risk assessment tools focus almost exclusively on risk factors associated with increased rates of recidivism and do not attend to protective factors that might mitigate reoffense risk. The present study investigated the predictive validity of the Structured Assessment of Protective Factors - Sexual Offence version (SAPROF-SO), developed to assess hypothesised protective factors against sexual recidivism in adult males. The SAPROF-SO pilot version contains 24 items across two domains: Personal and Professionally Provided Support. SAPROF-SO scores were rated retrospectively from a review of archived case files of 210 men with convictions for child sexual offenses, using the SAPROF-SO pilot manual and a supplementary retrospective scoring guide developed for the current study. SAPROF-SO Total and Personal domain scores were significantly predictive of sexual recidivism after an average follow-up period of 12.24years (AUC = .81), and to a lesser extent, violent and general recidivism. SAPROF-SO Total and Personal scores additionally provided significant incremental validity over Static-99R scores in the prediction of sexual recidivism. Results support the predictive validity of protective factors for reduced sexual recidivism and invite future research examining how to integrate the SAPROF-SO alongside contemporary sexual recidivism risk assessment tools.