Level Of Service Inventory-Revised Research Articles

Examining the Utility of the LSI-R in a Sample of Women who have Sexually Offended

The ability to identify factors predictive of recidivism among individuals who have offended is integral to properly targeting supervision and treatment focus. Instruments such as the Level of Service Inventory–Revised (LSI-R) show promise with general justice-involved females, but no study to our knowledge has assessed it with women who have sexually offended. The purpose of this study is to examine the utility of the LSI-R in a sample of 243 women from the State of Texas who have sexually offended. Results indicate that the total score was significantly related to criminal recidivism (Harrell’s C = .67) and any recidivism (Harrell’s C = .65). Analyses of the subscales indicated that criminal history and alcohol/drug subscales were significantly related to all forms of recidivism, while the education/employment and financial scales were significantly related to criminal recidivism and any recidivism. These results provide insight into risk assessment and treatment for women who have sexually offended.

Difference in Response to Feedback and Gender in Three Therapeutic Community Units.

Therapeutic communities (TCs) for substance abuse incorporate a system of peer feedback through written affirmations and corrections. Previous research has found that TC residents show a response to affirmations that is detectable for roughly 8 weeks, with response to corrections being of shorter duration and weaker overall. It is not clear whether and to what extent response to feedback in TCs varies between men and women. Previous research in other settings suggests that women should be more responsive to feedback than men. In order to test this hypothesis we draw on a large dataset of affirmations and corrections sent and received in three 80 bed TC units, two of which house men and one of which houses women. The analysis uses a multilevel negative binomial model, treating affirmations and corrections that TC residents receive as predictors of affirmations that they send over a 9 week period (week 0, the week during which affirmations and corrections are actually sent, and eight subsequent weeks). The model controls for gender, age, race, unit and scores on the Level of Service Inventory-Revised (LSI-R). The relationship between affirmations received and those sent is stronger for women during the initial week and on lags 1-2 and 5-8. The relationship between corrections received and affirmations sent is stronger for women on lags 2 and 8. Graphs suggest that response to affirmations falls off in an exponential curve, while that to corrections appears to include a periodic element. These results indicate that both men and women respond to feedback, but that the strength of the women's response is somewhat greater. These results suggest that any difference in suitability by gender to the feedback approach that characterizes TCs may favor women.

Moderating effects of age, employment status and level of education on the predictive utility of the LSI-R for sexual offenders

Using a Midwestern sample of sex offenders, the current study reports findings on the utility of the Level of Service Inventory-Revised (LSI-R) in predicting recidivism for offenders with a history of sexual crimes. The current study includes data from a sample of 250 sexual offenders over a 36-month period. This study aims to investigate how three variables – level of education, age and employment status – correlate to recidivism, and if they moderate the effectiveness of the tool in predicting future offending. This study found a statistically significant relationship between total LSI-R score and general recidivism, supporting the predictive utility of the LSI-R for sex offenders. These findings imply the need for intervention programmes focused on young sex offenders to reduce their potential for reoffending. The present study contributes to the available research by increasing the knowledge regarding sexual offender recidivism rates. Keywords: Sex offenders, recidivism, age, level of education, employment

Diagnostic Moderators of the Risk-Recidivism Relationship for Offenders with Mental Illness.

Recent research suggests that early onset of criminality is a potential marker for a subgroup of offenders with mental illness (OMI) for whom general risk/need factors are more relevant. The present study examines whether the presence of certain psychiatric diagnostic categories (i.e., Psychotic Disorders, Mood Disorders, Personality Disorders, Substance use Disorders) moderates the prediction of arrest outcomes using Level of Service Inventory-Revised (LSI-R) total scores in an archival sample of N = 179 OMI probationers. Moderator analyses demonstrated that LSI-R total scores predicted arrest outcomes significantly better for those with a diagnosis of Personality Disorder using path and survival analytic methods. Hierarchical regressions using LSI-R subscales as predictors revealed lower pro-criminal Attitudes for Mood Disordered participants, while more severe Criminal History and Family/Marital problems characterized those diagnosed with Personality Disorders. This study adds to the developing literature on subgroups of OMI for whom general risk/need factors are particularly important and highlights personality disturbance as another potential characterizing feature.

Weights matter: Improving the predictive validity of risk assessments for criminal offenders

This article explored whether the use of optimal weights can improve the predictive validity of risk assessments for criminal offenders, an issue the literature has long been divided on. By analyzing the unweighted instrument Level of Service Inventory – Revised (LSI-R), the study provided evidence in favor of optimal weights, showing that a weighted instrument clearly outperforms an unweighted one in predicting recidivism. The magnitude of the difference in performance would not be expected by some of the literature. The study also showed that objective and easily measurable items of such instruments were associated with larger weights than subjective items.

Racial Bias and LSI-R Assessments in Probation Sentencing and Outcomes

Risk assessments are now implemented in correctional settings across the United States as an evidence-based strategy to inform sentencing and supervision decisions. Despite growing research examining racial bias in the predictive validity of risk assessments, few studies have investigated racial bias in the context of judicial decision-making. We investigated the interactive contributions of race and Level of Service Inventory–Revised (LSI-R) risk assessments in predicting sentence length and probation outcomes in 11,792 Black and White probationers. Results showed White probationers at low-risk levels received longer sentences relative to Black probationers classified at the same risk levels. However, there were few differences at higher risk levels and no evidence of racial bias in the predictive accuracy of LSI-R assessments for other probation outcomes. Findings highlight the need for prospective and carefully controlled investigations into whether risk assessments improve the fairness and accuracy of sentencing and other risk management decisions.

Prison Climate and Its Role in Reducing Dynamic Risk Factors During Offender Treatment.

Although several offender treatment experts have suggested that therapeutic relationships play an important role in offender treatment, empirical finding supporting those arguments are scarce. The present study has therefore examined the relationship between prison climate, treatment motivation, and their influence on changes in risk factors in N = 215 inmates and detainees in four correctional facilities in Berlin, Germany. The inmates' perception of prison climate significantly correlated with their attitudes towards treatment. More positive climate in terms of therapeutic hold and more positive treatment attitude in terms of trust in therapy were also the best predictors of stronger decreases in dynamic risk factors measured by the Level of Service Inventory-Revised (LSI-R). Our results support the importance of treatment relationship factors within the course of offender rehabilitation.

Ecological Context, Criminal Propensity, and Recidivism

This study examines whether the relationship between individual-level risk and recidivism varies according to ecological context, measured at the census tract level. It is hypothesized that high-risk offenders—as measured by Minnesota Screening Tool Assessing Recidivism Risk (MnSTARR) 2.0 and Level of Service Inventory–Revised (LSI-R)—will have elevated risk of recidivism when living in disadvantaged neighborhoods and lower risk of recidivism when living in affluent neighborhoods. These hypotheses are tested with hierarchical logistic models predicting rearrest and revocation for a technical violation among a sample of approximately 3,000 offenders released from Minnesota state prisons in 2009. Rearrest was positively related to neighborhood disadvantage and negatively related to neighborhood affluence, while revocation was positively related to neighborhood urbanism. Further, neighborhood disadvantage moderated the association between LSI-R and rearrest; however, this interaction was not in the hypothesized direction. The results contradict prior literature examining similar relationships at the county level.

Risk Factors for Absconding Among Adult Parolees in Colorado.

Studies that have examined absconding have pointed to the relevance of factors related to demographics, situational, and social characteristics, as well as criminal history variables as predictive of offenders' decision to abscond. This study contributes to this literature by extending factors that have been explored for criminal justice outcomes such as recidivism to predict the likelihood of absconding. In particular, this study incorporates the Level of Service Inventory-Revised (LSI-R) risk assessment instrument and types of violations to examine their impact on the likelihood of absconding. Using a large sample of parolees from Colorado ( n = 30,181), the findings in this study indicate that the largest predictors of absconding were substance abuse problems and education. In addition, variables measuring the components of the LSI-R and specific nonabsconding parole violations moderately predict absconding. The implications of the findings and recommendations for future research are discussed.

Evaluating concurrent validity of criminal justice and clinical assessments among women on probation

BackgroundWomen in the criminal justice (CJ) system experience complex and comorbid medical, psychiatric, and substance use disorders, which often contribute to CJ involvement. To identify intersections between CJ and health needs, we calculated Spearman r correlations between concurrent CJ and clinical assessments from women on probation in Connecticut who were enrolled in a clinical trial. We examined longitudinal trends in CJ risk scores over 9 years of observation (2005–2014), modeling time to probation recidivism with shared gamma frailty models and comparing contiguous time points by Wilcoxon matched-pairs signed rank tests.ResultsWomen (N = 31) were predominantly white (67.7%) with at least some high school education (58.1%) and mostly unemployed (77.4%) and unstably housed (83.9%). Most met clinical criteria for severe substance use and/or psychiatric disorders. Concurrent measures of substance use, mental health, social support, partnerships, and risk by the Level of Service Inventory-Revised (LSI-R) and clinical assessments were not significantly correlated. The LSI-R personal/emotional sub-score, however, positively correlated with the Addiction Severity Index psychiatric composite score (r = 0.40, 95% CI 0.03–0.68, p = 0.03). After adjusting for age, race and number of previous events, having some high school education versus none marginally decreased the hazard for probation recidivism and having > 5 inpatient psychiatric admissions versus none increased the hazard of probation recidivism 7-fold (HR 7.49, 95% CI 1.33–42.12, p = 0.022). Women with 0–1 recurrent probation terms (n = 16) had a significantly lower mean LSI-R score than those with 2–4 recurrent probation terms (35.9 [SD 6.4] versus 39.2 [SD 3.0], p = 0.019), but repeated LSI-R scores did not change over time, nor vary significantly beyond the group mean.ConclusionsIn this small, quantitative study of women on probation, widely used CJ assessment tools poorly reflected women’s comorbid medical, psychiatric, and substance use needs and varied minimally over time. Findings illustrate the limitations of contemporary CJ assessment tools for women with complex needs. The field requires more comprehensive assessments of women’s social and health needs to develop individualized targeted case plans that simultaneously improve health and CJ outcomes.

The belief in redeemability – version 2 (BiR-2) scale and its relation to desistance

Purpose The purpose of this paper is to test the relationship between the belief in redeemability – Version 2 (BIR-2) Scale and desistance from crime. It also seeks to explore how patterns of responding on the BIR-2 with offenders compare to previous patterns of responding in the general public. Design/methodology/approach The authors report the results of a study of offenders using the belief in redeemability – Version 2 (BiR-2) scale. In total, 180 offenders under the supervision of the Community Corrections Service (formerly the Probation and Parole Service) of New South Wales completed the ten-item questionnaire and when these data were combined with demographic and reoffending data collected by Corrective Services New South Wales, 168 sets of useable data were collected. Scores on the BIR-2 scale were compared to Level of Service Inventory – Revised (LSI-R) score, Most Serious Offence category, Aboriginal and Torres Strait Islander status, number of custodial sentences in previous five years, age, gender and reoffending. Findings Results showed that the sample overall was closely representative of the caseload from which the study sample was drawn (a metropolitan community corrections office) and that BIR-2 scores showed a small, significant, negative correlation with LSI-R scores. Analysis of re-offending data indicated a small, positive, but non-significant correlation with BIR-2. Implications of this are discussed and future research outlined. Practical implications The paper suggests that it is worth attempting to measure belief in redeemability in the broader context of a narrative approach to desistance. Originality/value This is the first time that a scale has been used to test the importance of a belief in redeemability quantitatively and to permit the use of multivariate analysis.

A Person-centered Approach to Risk and Need Classification in Drug Court

A sample of 265 New York City drug court participants completed the Level of Service Inventory—Revised (LSI-R) and Texas Christian University Drug Screen II (TCUDS). Three participant clusters were identified through a person-centered analysis of their LSI-R and TCUDS responses: low risk (LR), criminogenic risk (CR), and complex behavioral health needs (CBHN). Although CBHN scored higher than CR and LR on the LSI-R and TCUDS, they were no more likely to be re-arrested at 24 months and no higher in their rate of positive drug tests. The CR cluster predicted re-arrest beyond the LSI-R and rate of positive drug tests beyond the LSI-R and TCUDS. CBHN participants placed in a residential (vs. non-residential) setting were disproportionately likely to be re-arrested. Results point to a sub-population of drug court participants not captured in variable-centered summary risk scores, who might require intensive case management or referral to suitable treatment.

Reliability and Validity of START and LSI-R Assessments in Mental Health Jail Diversion Clients.

Risk assessment instruments are increasingly used in mental health jail diversion programs. This study examined the reliability and validity of Short-Term Assessment of Risk and Treatability (START) and Level of Service Inventory-Revised (LSI-R) assessments overall and by client race. Research assistants completed START and LSI-R assessments for 95 diversion clients. Arrests and jail days were collected via official records and self-report 3, 6, 9, 12, and 18 months after baseline. Assessments demonstrated good interrater reliability and convergent validity. START strength total scores and LSI-R risk estimates were the strongest predictors of recidivism. Total scores and risk estimates did not differ as a function of client race, but there were some differences in accuracy of START vulnerability and LSI-R total scores and risk estimates in predicting jail days (but not arrests), over shorter follow-ups. No such differences were found for START strength total scores across any follow-up period or recidivism measure.

Age Group Differences in the Predictive Validity of the Level of Service Inventory–Revised

The Level of Service Inventory–Revised (LSI-R) is a widely used risk assessment tool with demonstrated predictive validity among incarcerated offenders. However, researchers have yet to examine the predictive validity of the LSI-R with adult inmates of different ages. We examined the predictive validity of the LSI-R and its 10 subcomponents across developmentally defined age groups of men confined in prisons in a Midwestern state. We found that the LSI-R predicted the prevalence of misconduct similarly across the groups, but there were a number of differences in the magnitude of the effects when we examined the incidence of prison misconduct.

Socioeconomic marginality in sentencing: The built-in bias in risk assessment tools and the reproduction of social inequality

This article develops a sociological analysis and critique of including socioeconomic factors such as education, employment, income and housing in risk assessment tools that inform sentencing decisions. In widely used risk assessment tools such as the Level of Service Inventory-Revised (LSI-R) (Canada, US), the Correctional Offender Management Profiling for Alternative Sanctions (COMPAS) (US), the Offender Assessment System (OASys) (UK) and the Recidive InschattingsSchalen (RISc) (the Netherlands), socioeconomic marginality contributes to a higher risk score, which increases the likelihood of a (longer) custodial sentence for underprivileged offenders compared to their more privileged counterparts. While this has been problematized in relation to gender and racial/ethnic bias, the problem of socioeconomic bias in itself has received little attention. Given the already marginalized position of many justice involved individuals and longstanding concerns about such disparities, and the adverse effects of imprisonment on socioeconomic opportunities, it is essential to evaluate the unintended social consequences of assessing socioeconomic marginality as ‘risk factor’. Elaborating on earlier critiques, I conceptualize risk-based sentencing as a meaning-making process through which (access to) resources and recognition are distributed among offender populations. Through tracing in detail two cultural processes – stigmatization and rationalization – I analyse how risk assessment is likely to produce sentencing disparities as well as to reproduce, and possibly exacerbate, social inequalities more generally.

Assessment of Sex Offenders With the Minnesota Multiphasic Personality /Inventory–2–Restructured Form

This study examined the association between scores on the Minnesota Multiphasic Personality Inventory-2-Restructured Form (MMPI-2-RF) with static and dynamic risk assessment instruments, including the STATIC-99 and Level of Service Inventory-Revised (LSI-R). The sample included 304 male adults who were convicted of sexual offenses against children and were referred to a sex offender treatment program. On average, the sample had a Low-Moderate risk of re-offending according to the STATIC-99 and LSI-R. The results indicated that MMPI-2-RF scale scores in this setting are characterized by relatively high levels of under-reporting and externalizing psychopathology compared with the normative sample. We also found that scale scores in this sample produced reliability estimates that were similar to the normative sample. Finally, external correlations between the MMPI-2-RF scales and the risk assessment instruments indicated that the test was associated in expected ways with constructs measured by these instruments. Correlations were most robust among scales in the externalizing/behavioral dysfunction domain of the MMPI-2-RF. Overall, the results of the study support and guide use of the test in this population.

Identifying the Characteristics of Chronic Drunk Drivers Using a Criminological Lens

The supervision of driving while intoxicated/impaired (DWI) offenders composes a large share of probation officer’s caseloads. Crime and justice researchers have made tremendous advances in the ability to classify offenders according to general recidivism, with the Level of Service Inventory–Revised (LSI-R) being one of the most tested instruments used by community corrections officers. There has yet to emerge a risk assessment tool designed to classify DWI offenders according to their likelihood to be arrested for multiple DWIs. This article estimates a series of multinomial logistic regression models to identify differences between one-time DWI offenders and chronic DWI offenders. The central independent variables are the items on the LSI-R and a popular alcohol and substance abuse screener (Adult Substance Use Survey [ASUS]), while controlling for age, race, gender, and marital status. These instruments are important tools to classify DWI offenders, and the results suggest that the community corrections field (including the offenders supervised) would benefit from a risk assessment tool to classify DWI offenders.

The Validity of the Level of Service Inventory–Revised at the Intersection of Race and Gender

Although the Level of Service Inventory–Revised (LSI-R) has been validated across genders and races, few studies have investigated the validity of the instrument for different combinations of races and genders. Using data from a sample of offenders in New Jersey ( N = 9,454), we measure the LSI-R’s validity in predicting recidivism within 1 year of prison release. Our findings indicate that the LSI-R is a valid predictor of recidivism across different race and gender combinations; however, the instrument performs poorly at distinguishing recidivists from non-recidivists and exhibits the lowest discriminant capacity for Black males. Policy implications are discussed.

A Jack of All Trades But a Master of None? Evaluating the Performance of the Level of Service Inventory–Revised (LSI-R) in the Assessment of Risk and Need

ABSTRACTWe evaluate the performance of the LSI-R in assessing risk and need among 26,001 offenders released from Minnesota prisons between 2003 and 2011. Using multiple performance metrics, the results reveal the LSI-R incorrectly classified 41 percent of the offenders, and its area under the curve (AUC) was 0.628. Drawing on the extensive program evaluation research conducted on Minnesota prisoners over the last decade, we evaluated the LSI-R’s assessment of need by examining the relationship between LSI-R domain scores, participation in effective interventions, and recidivism outcomes. The findings provide, at best, moderate support for the LSI-R’s performance in assessing need. Interventions targeting the education/employment, alcohol/drug, and emotional/personal domains generally yielded better recidivism outcomes among offenders with higher scores in these domains.

The Effect of Cognitive-Behavioral Therapy on Driving While Intoxicated Recidivism

This longitudinal study examined differences in driving while intoxicated (DWI) recidivism among defendants with multiple DWI offenses after receiving a 16-week cognitive-behavioral therapy (CBT) program compared with those who received standard services. Defendants on probation supervision with histories of repeated DWI offenses ( N = 286; male = 240, female = 46) were referred to CBT. There was a significant difference in DWI recidivism 3 years after CBT among participants (11%) and the New York State (25%) and national (30%) rates; the comparison group reoffended at the rate of 25%. Level of Service Inventory–Revised (LSI-R) scores of the CBT group also decreased significantly compared with the comparison group, and CBT participants showed significant improvement in overall functioning measured by Global Assessment of Functioning (GAF). CBT offered in this innovative manner was more effective with recalcitrant, hard to treat, DWI offenders. It may also benefit precontemplative offenders by helping them recognize that their drinking is negatively impacting their lives and may warrant more traditional alcohol treatment.