Validation of the Orange County California Probation Department Risk Assessment Instrument: Final Report

Size: px
Start display at page:

Download "Validation of the Orange County California Probation Department Risk Assessment Instrument: Final Report"

Transcription

1 Validation of the Orange County California Probation Department Risk Assessment Instrument: Final Report October 2011 Mike Eisenberg Dr. Tony Fabelo Jessica Tyler Prepared by the Council of State Governments Justice Center, with the support of the Orange County California Probation Department Council of State Governments Justice Center 100 Wall Street, 20th Floor New York, New York Montgomery Avenue, Suite 650 Bethesda, MD W. 12 th Street Austin, Texas i

2 This project was supported by the Orange County California Probation Department. Points of view or opinions in this document are those of the author and do not represent the official position of the sponsors or the Council of State Governments members. The Council of State Governments Justice Center is a national nonprofit organization that serves policymakers at the local, state, and federal levels from all branches of government. The Justice Center provides practical, nonpartisan advice and consensus-driven strategies, informed by available evidence, to increase public safety and strengthen communities by The Council of State Governments Justice Center. ii

3 Validation of the Orange County California Probation Department Risk Assessment Instrument: Final Report October 2011 Mike Eisenberg Dr. Tony Fabelo Jessica Tyler Prepared by the Council of State Governments Justice Center, with the support of the Orange County Probation Department Council of State Governments Justice Center 100 Wall Street, 20th Floor New York, New York Montgomery Avenue, Suite 650 Bethesda, MD W. 12 th Street Austin, Texas iii

4 Acknowledgements The authors of this report would like to thank Chief Deputy Probation Officer Bryan Prieto and Administrative Manager of the Research Division Marya Forster for their support and guidance throughout this project. They provided sound advice and direction, as well as sharing their experience and expertise, during the entirety of this research. Mr. Prieto s support of this project is truly appreciated. Ms. Forster was especially instrumental in the development of the database necessary for this research and her advice on research issues associated with this project. The authors would also like to thank Contracts and Purchasing Manager Lala Ragen for her guidance through the intricacies of the contracting process. Finally, the authors would like to thank the managers and administrative staff of the Orange County Probation Department for their assistance with this project. iv

5 Table of Contents Executive Summary... ix I. Introduction... 1 II. Review of Relevant Risk Assessment Issues... 2 III. Scope of Work... 4 IV. Methodology... 7 A. Sample... 7 B. Outcome Measures... 7 I. Subsequent Law Violations... 7 II. Termination Type... 7 V. Results...12 A. Overview of Validation Sample...12 B. Distribution of Populations by Risk Factors...12 C. Percent Subsequent Law Violations by Risk Factors...14 D. Distribution of Population and Subsequent New Offense Violations by Initial Risk Group Classification and Final Initial Classification...19 E. Severity of Subsequent Law Violation...23 F. Correlation of Risk Factors to Subsequent New Law Violation...24 VI. Special Analyses...26 A. Overview...26 B. Distribution of Risk Groups by Gender...26 C. Subsequent Law Violation by Risk Group by Gender...27 D. Distribution of Risk Groups by Race/Ethnicity...28 E. Subsequent Law Violation by Risk Group by Race/Ethnicity...29 F. Subsequent New Violent Law Offenses and Risk Score...30 G. Revocations and Risk Score...33 H. Analysis of Overrides...34 VII. Developing a Brief Risk Screening Instrument...36 VIII. Recommended Tasks to Modify Risk Assessment...40 A. Revise Weights for Selected Factors and Define New Risk Factors...40 B. Revise Cut-off Points...42 v

6 C. Distribution of Risk Groups and Subsequent Law Violations of Risk Groups: Current and Proposed Risk Groups...42 D. Comparison of Predictive Ability of Current and Proposed Risk Score...44 IX. Summary of Findings and Recommendations...47 A. Summary of Findings:...47 B. Summary of Recommendations...48 Bibliography...49 vi

7 Table of Figures Figure 1: Orange County Probation Department Risk Assessment Instrument... 6 Figure 2: Study Sample and Tracking Methodology... 8 Figure 3: Data Factors Examined for Validation Study... 9 Figure 4: Risk Groups by Score...10 Figure 5: Overview of Research Methodology...11 Figure 6: Distribution of Population by Percent Adjudicated for Specific Offenses...14 Figure 7: Percent Subsequent Law Violations by Number of Address Changes...18 Figure 8: Factors Meriting Re-examination...19 Figure 9: Distribution of Risk Groups by Initial Risk Score and Final Risk Classification...20 Figure 10: Percent Subsequent Law Violation by Risk Group...22 Figure 11: Severity of Subsequent Law Violation...23 Figure 12: Distribution of Risk Groups by Gender...26 Figure 13: Percent Subsequent Law Violation by Risk Group by Gender...27 Figure 14: Distribution of Risk Groups by Race/Ethnicity...28 Figure 15: Percent New Offense by Risk Groups by Race/Ethnicity...29 Figure 16: Percent Subsequent Violent Law Violation by Initial Score-based Risk Group and Final Initial Risk Group...31 Figure 17: Victim Probability Risk Factors...32 Figure 18: Victim Probability Risk Group and Subsequent Law Violation with Victim...32 Figure 19: Percent Revoked by Initial Risk and Final Initial Risk Classification...34 Figure 20: Percent Subsequent Law Violation by Overrides...35 Figure 21 : Brief Risk Screen Groups and Subsequent Law Violations...39 Figure 22: Distribution of Risk Groups: Current Risk Distribution Compared to Proposed...43 Figure 23: Percent Subsequent Law Violation by Risk Groups: Current Risk Groups Compared to Proposed...44 Figure 24: ROC Curve for Current and Proposed Revision to Orange County Risk Score...45 Figure 25: Summary of Recommendations...48 vii

8 Table of Tables Table 1: Sample Characteristics...12 Table 2: Distribution of Risk Factors...13 Table 3: Percent New Offense within 3 Years of Placement on Community Supervision by Risk Factors...15 Table 4: Subsequent Violations and Distribution of Population by Initial and Final Initial Risk Group Classification...21 Table 5: Correlation of Risk Factors with Subsequent New Law Violation...24 Table 6: Offense Category and Age at Assessment Risk Factors...36 Table 7: Risk Factors with Highest Correlation with Subsequent Law Violations...37 Table 8: Brief Risk Screen Factors and Weights...38 Table 9: Revised Risk Factors and Weights...41 Table 10: Companions Needs Item and Subsequent Law Violations...42 Table 11: Current and Proposed Risk Assessment Instrument Cut-Off Points...42 Table 12: Comparison of Current and Proposed Risk Score Predictive Ability Based on Regression Analysis R...44 Table 13: ROC Area under Curve...46 viii

9 Executive Summary The Orange County California Probation Department (OCPD) contracted with the Council of State Governments Justice Center (Justice Center) to validate the predictive accuracy of the OCPD risk assessment instrument. The risk assessment instrument is designed to differentiate the probability of probationers committing subsequent law violations after placement on probation. Based on risk of reoffending, supervision resources are allocated to provide the most intense supervision to High risk probationers. In addition to the validation study, the Department requested the development of a brief risk screening tool to quickly identify Low risk offenders not requiring a full risk assessment procedure. Projected increases in workload associated with releases from the California Department of Corrections and Rehabilitation and efficient triage of cases for supervision resources necessitated the change. The OCPD Administrative Research Manager and IT staff developed a file for the Justice Center to conduct the research. The file consisted of risk assessments completed on probationers placed on supervision in 2007 and identified subsequent law violations for this sample that occurred in the three years after placement on probation. The sample included 4,875 offenders placed on probation in The figure below details the methodology used for this research. Overview of Research Methodology OCPD Risk Assessment Instrument Validation Methodology OCPD Risk Assessment Factors and Risk Groups Relation to Subsequent Law Violation and Negative Supervision Termination Correlation of OCPD Risk Factors and Risk Groups with Subsequent Law Violations Weights of Risk Factors and Risk Groups Evaluated Validity of OCPD Risk Instrument Recommendations to Improve Predictive Ability of OCPD Risk Instrument ix

10 The summary of the recommendations are listed in the table below and followed with an overview of the results of the analysis supporting these recommendations. Summary of Recommendations 1. Implement revisions to OCPD risk instrument. a. Use new weights for selected factors. b. Add Age at Assessment as a new factor and revised Offense Group risk factor. c. Implement new cut-off scores to reflect new weighting and new factors. d. Revise data entry and database protocols to reflect revised risk instrument. 2. Review and revise override criteria and policies to reflect actual risk of reoffending. Committee to review and recommend policies for overrides incorporating risk of recidivism and risk of harm in determining use of override. a. The research literature suggests that overrides should not exceed approximately 15% of new cases or the risk instrument will be compromised. 3. Conduct pilot test of new instrument. a. Create a committee to oversee pilot test implementation. b. Establish new training protocols for revised risk instrument recommended here. c. Conduct inter-rater and intra-rater reliability testing to assure accuracy of scoring. d. Conduct an independent evaluation of pilot test to assess implementation issues before adoption. e. Reprogram data entry system. 4. Plan implementation of Brief Risk Screen. a. Develop protocols and policies for use of Brief Risk Screen to include data entry, overrides, supervision procedures and reassessments. b. Monitor implementation and evaluate impact of Brief Risk screen from workload and effectiveness perspective. 5. Evaluate change in workload requirements associated with revised risk assessment score and classification. Discussion of Results Analyses examined the distribution of the population by risk level using the current risk assessment instrument. The initial classification of probationers by the OCPD risk assessment instrument resulted in 54.8 percent of the population classified as High risk. The initial classification of a majority of the population classified as High risk is somewhat questionable as it does not appear to significantly differentiate the population when a majority of offenders are in one risk group. Subsequently, overrides result in the vast majority of offenders being classified as High risk (81.1%) at the final initial risk classification. Examination of outcomes by risk based classification and by overrides, as reported later in this study, will assist in determining the utility of the current risk classification and overrides in assigning supervision by risk of re-offending. x

11 Distribution of Risk Groups by Initial Risk Score and Final Risk Classification The figure below indicates the percent of the population committing new subsequent law violations within three years of placement on probation by initial risk score-based classification and by final initial risk classification after overrides. Approximately 42.1 percent of the sample had a subsequent law violation. While both classification methodologies result in differentiation of the population by risk, using only the risk score for classification results in clearer distinction of reoffending than the final classification used for supervision. As the figure below illustrates, outcomes based on risk score-only classification ranges from 18.8 percent of Low risk offenders committing a new law violation in the three year followup to 52.2 percent for offenders classified as High risk. This is a difference of 33.4 percent. The range of outcomes for the final initial classification groupings is only 15.3 percent (28.5% for Low risk to 43.8% for High risk). This indicates that overrides negatively impact the differentiation of population by risk, because they introduce factors unrelated to differential recidivism rates. xi

12 Percent Subsequent Law Violation by Risk Group Analyses were conducted to examine the relationship between risk factors and their ability to differentiate the population by risk of subsequent law violations. This analysis indicated a number of factors that should be re-examined due to their predictive weakness. The table below divides risk factors with predictive weaknesses into those with categories that demonstrate little difference in the percent law violation and those factors that actually do not discriminate by risk of recidivism. These factors do not contribute to the predictive accuracy of the risk instrument and may actually reduce predictive accuracy. xii

13 Risk Factors Meriting Re-examination Factor Percent Law Violations Little Difference in Outcomes by Category Number of Address Changes No Address Change = 39.7% One Address Change = 40.5% Attitude Lacking Motivation = 43.7% Not Motivated = 47.6% Time Employed Alcohol Usage Convictions for Burglary, Auto Theft or Worthless Checks 5-7 Months = 41.6% Less than 5 Months = 45.9% Factor Does Not Discriminate Accurately Occasional Abuse = 44.9% Frequent Abuse = 42.2% Burglary = 50.3% Worthless Check = 42.5% The initial risk assessment and classification procedures for determining supervision level and resources to allocate to new probationers can be labor intensive and costly. Community supervision agencies are utilizing new procedures to triage their supervision population by quickly identifying Low risk offenders, reducing supervision resources to those offenders, and re-allocating those resources to higher risk offenders. Instead of conducting a full risk assessment and classification procedure for all offenders, a brief screening tool can provide a risk assessment measure to identify those low-risk probationers who can be transferred immediately to administrative type caseloads, resulting in fewer cases requiring a full assessment. OCPD seeks to develop a brief screening tool to supplement a validated risk assessment instrument. This section describes the process and results of efforts to develop a brief risk screening instrument. The table below details the risk factors identified in a regression analysis that were the strongest predictors of subsequent law violation. These were the factors used in developing the Brief Risk Screen (BRS). Weights assigned for each factor are provided. xiii

14 Brief Risk Screen Factors and Weights Weight Frequency Percent Prior Probation Violations 0 None 2, One or more 2, Age at First Conviction 0 24 or older 1, , or younger 1, Number of Prior Probation Periods 0 None 2, One or more 2, Age at Assessment 0 40 and older 1, to , Less than 25 1, Offense Type 0 Not property or drug offense 1, Property or drug offense 3, The figure below indicates the relation between BRS risk groups and subsequent law violations. The lowest risk group (0-2) had an 18.1 percent subsequent law violation rate after three years compared to 56.8 percent for the High risk group and the overall subsequent law violation rate of 42.1 percent. The table includes the percent of subsequent felony level law violations. Approximately 5.3 percent of probationers scoring 0-2 on the BRS committed a new felony offense. Approximately 12 percent of this sample were classified as Low risk (0-2 score) and 32 percent of the sample were classified as low-medium risk (4-8 score). The R score for the regression analysis was.272, which is only slightly lower than the.274 R for the current risk assessment instrument using all factors. This was achieved by using only factors that had the highest correlation with subsequent law violations and by adding two new factors significantly correlated with subsequent law violations. xiv

15 Brief Risk Screen Groups and Subsequent Law Violations Based on analysis of each current risk factor and development of new factors, revisions to the current risk instrument are proposed. The table below indicates the current and proposed distribution of the recommended supervision population as a result of the new cut-off scores and re-weighting and changing of risk factors. All risk groups are based on the initial risk score classification and do not reflect the impact of overrides. As the table shows, the percent classified as High risk (overall) declined from 54.8 percent to 34.5 percent under the proposed weights and cut-off scores, while the percent classified as Low risk went from 9.8 percent to 16.3 percent. Distribution of Risk Groups: Current Risk Distribution Compared to Proposed xv

16 The figure below indicates that even with the significant redistribution of risk groups, reoffense rates for the proposed revised score were very similar to the re-offense rates under the current distribution. As discussed above, the percent classified as Low risk went from 9.8 percent to 16.3 percent but the re-offense rate for the Low risk group went from 18.8 percent under the risk instrument to an 19.2 percent re-offense rate for the Low risk group under the proposed risk score, virtually unchanged. More Low risk offenders were identified with closely the same re-offense rate. Fewer High risk offenders were identified but with a higher re-offense rate. Under the current risk instrument, approximately 55 percent of the population was identified as High risk while the proposed risk instrument identifies 34.5 percent of the population as High risk. The re-offense rate for High risk offenders was 52.2 percent under the current instrument and under the revised risk score the re-offense rate for the High risk group is 56.6 percent, indicating increased accuracy in classifying offenders as High risk. The recidivism rate of the Medium risk group increase from 33.0 percent to 39.5 percent, which approaches the average recidivism rate of the sample (42.1%). This would appear to be more reflective of a medium recidivism rate than a rate almost 10 percent below the average. Percent Subsequent Law Violation by Risk Groups: Current Risk Groups Compared to Proposed The table below shows the comparison of current and proposed risk score predictive ability based on regression analysis. A linear regression analysis was conducted using the proposed risk factors and new weights. The previous R score for the regression analysis was.274 using the current factors compared to an R score of.288 using the proposed factors and weights, indicating an improvement in the predictive ability of the risk instrument. xvi

17 Comparison of Current and Proposed Risk Score Predictive Ability Based on Regression Analysis R Model Model Summary R R Square Current Risk Factors Proposed Risk Factors A second measure of improvement in risk classification associated with the proposed revision in the risk score is the Receiver Operating Characteristic Curve or the ROC Curve. The ROC Curve is a measure that evaluates the performance of a classification scheme in which there is one variable (Risk Score or the Revised Risk Score) with two categories (New Offense within three years or No New Offense within three years) by which subjects are classified. The Area under the Curve (AUC) represents the probability that the result of the classification for a randomly chosen positive case (prediction of re-offense that is true) will exceed the result of a randomly chosen negative case. The curve is a graphical representation of the trade-off between false negative and false positive rates. The ROC curve for both the current risk score and the proposed revised risk score are shown below. The ROC curve for the proposed revised risk score exceeds the ROC curve for the current risk score; this indicates greater accuracy of the proposed score in classifying offenders by risk. The area under the curve for the proposed revised risk score (.659) exceeds the current risk score (.642) and the lower and upper bounds of the proposed revised risk score (.644 and.674 respectively) exceed the current risk score (.626 and.657). xvii

18 ROC Curve for Current and Proposed Revision to Orange County Risk Score xviii

19 I. Introduction A validated risk assessment instrument is the foundation to implementing evidencebased practices known to reduce recidivism. A validated risk assessment instrument differentiates an offender population into groups with different probabilities of reoffending, so resources can be effectively and efficiently allocated according to risk. Research has demonstrated that allocating resources to those most likely to reoffend, and reducing resources to those offenders unlikely to reoffend, increases the probability of reducing recidivism for offenders placed on supervision (Hubbard, Travis, & Latessa, 2001). Andrews and Bonta, commenting on the importance of risk assessment, stated: The prediction of criminal behavior is perhaps one of the most central issues in the criminal justice system. From it stems community safety, prevention, treatment, ethics, and justice. Predicting who will re-offend guides police officers, judges, prison officials, and parole boards in their decision making (Andrews & Bonta, 1994). The Orange County Probation Department (OCPD) has utilized a risk/needs assessment instrument since the mid-1980s. Use of a validated actuarial assessment instrument like OCPD's is a key principle underlying the community corrections evidence-based practices model advocated by the National Institute of Corrections (Baird, Heinz and Bemus, 2009). This assessment guides supervision decisions based on the offender's level of risk and helps determine the case plan objectives according to the offender's prioritized needs. OCPD research staff initiated a preliminary assessment of the current risk instrument and determined that the risk/needs assessment instrument appeared to meet minimum levels of validity regarding classification of probationers by risk. However, several issues meriting further exploration were identified. One issue common to agencies using risk assessment instruments is the negative impact of policy overrides on risk classification, which adversely affects the predictive accuracy of the instrument. OCPD research staff noted that the override rate for the OCPD initial risk classification is approximately 30 percent, which OCPD staff note is twice that of the 15 percent and below considered an acceptable override rate by nationally recognized experts. Moreover, the OCPD preliminary assessment results showed only modest correlations with outcomes and some of the risk factors were not correlated very strongly with the recidivism outcome. The preliminary analysis also showed that the proportion of offenders classified as "High risk" has increased significantly, suggesting the need to re-examine outcomes by risk scores and to determine whether adjustments to the classification scoring ranges, particularly High risk, is warranted. 1 Finally, implementing a brief screening tool for new probationers is a practice gaining research momentum in other community corrections agencies. The purpose of this tool is to provide a quick but valid measure to identify those low-risk probationers who can be transferred immediately to administrative type caseloads, resulting in fewer cases requiring a full assessment. The OCPD seeks to develop a brief screening tool to supplement a validated risk assessment instrument. 2 1 Excerpted from Scope of Work, Consulting Services Agreement between the Council of State Governments and the County of Orange, pg Ibid. pg

20 Issues detailed above resulted in the OCPD seeking technical assistance services to validate their current risk assessment instrument and assist in developing a brief screening tool. The OCPD assessment instrument is based on a tool originally developed in Wisconsin in the 1970s. With technical assistance support from National Institute of Corrections, Chris Baird, who had pioneered the development and validation of the Wisconsin tool, consulted with OCPD in the mid-1980s to implement and validate this instrument for Orange County. Although OCPD research staff initiated and completed some preliminary revalidation work of the risk and needs assessment instrument in the last year, an independent, comprehensive revalidation study was needed to measure the instrument's effectiveness, more than 20 years after its introduction. This report reviews general issues associated with the use of risk assessment instruments in classifying offenders, presents an analysis of the OCPD risk assessment instrument and recommends changes in the instrument to improve predictive accuracy. II. Review of Relevant Risk Assessment Issues The Wisconsin Case Classification/Staff Deployment Project (Baird, Heinz, & Bemus, 1979), developed in the late 1970 s, served as the foundation of the National Institute of Corrections Model Probation and Parole Management Program. This was the early basis for probation and parole agencies to implement case classification systems across the country in the 1980 s. An actuarial risk assessment instrument, developed in Wisconsin and commonly referred to as the Wisconsin Risk Assessment, was a core component of case classification systems. In 1985, Clear and Gallagher raised questions of how well the probation risk assessment instrument, developed using Wisconsin s probation population, generalized to other populations. They recommended, classification practices need to be placed regularly under review, tested against alternative approaches, and revised where appropriate. These recommendations apply to any jurisdiction adopting risk assessment instruments, whether in Wisconsin or in other localities across the country. Validity of a risk assessment instrument is the most important supportive principle behind the proper utilization of these instruments. Namely, the instruments predictions must be supported by research showing it can identify different groups of offenders with different probabilities of reoffending. Baird (2009), one of the principal developers of the Wisconsin risk assessment instrument almost thirty years ago, recently commented that: The intent of actuarial risk assessment is to identify subgroups within an offender population who have significantly different rates of recidivism It is obviously important to identify offenders at High risk of recidivism and to devote more resources to these cases. Baird recommended that the justice field should step back and carefully review both the logic and level of evidence supporting current assessment practices. He noted that the standard for measuring the efficacy of a risk assessment model should be the level of discrimination attained by risk levels. Johnson and Hardyman (2004) establish four criteria to examine in determining validity of a risk instrument: A valid instrument identifies discrete groups of offenders who pose different levels of risk to public safety as measured by recidivism. 2

21 The risk instrument must be reliable as measured by tests of inter-rater and intra-rater reliability. Inter-rater reliability means that two different staff members would score the same offender the same way on the risk instrument and intra-rater reliability means the same staff person would score the same offender the same way repeatedly with no change in circumstances. The risk instrument should be practical, efficient, and provide utility to staff. The risk instrument is demonstrated to be fair to all offender populations such as by gender or race/ethnicity. 3

22 III. Scope of Work The contract requirements between the Justice Center and the OCPD for conducting the validation study of the OCPD Risk Assessment instrument include: Validate the ability of the OCPD Risk Assessment instrument to differentiate supervision populations into different risk levels of recidivating. Offenders placed on probation in 2007 and assessed on the OPCD risk assessment will be tracked for three years to measure the predictive ability of the risk instrument. Depending on the type of analysis that is appropriate, outcome measures will include: Indication of a new law violation after placement on probation in a threeyear follow-up period for the 2007 cohort. Indication of a new violent felony offense as defined by the California Department of Justice after placement on probation in a three-year followup period for the 2007 cohort. Termination of probation due to revocation to prison for the 2007 cohort in a three-year follow-up. 3 Determine if eliminating, adding, or reweighting factors included in the instrument can enhance the predictive ability of OCPD Risk Assessment instrument. This may also involve examining if revising cut-off scores for classifying offenders into risk level could increase the predictive accuracy of the instrument. Validate the ability of the OCPD Risk Assessment instrument to differentiate probationers into different risk levels of recidivating by gender and race/ethnicity. Develop a brief screening risk assessment with adequate predictive ability to enable identification and placement of Low risk offenders on administrative caseloads. The screening tool should identify offenders with the lowest rate of reoffending in the evaluation cohort. Conduct statistical tests to determine if differences in outcomes are statistically significant. Tests of inter-and-intra rater reliability (as described in Section II above) were not included in the scope of work for this validation study, nor can they be conducted until a validated instrument is adopted for implementation. 3 While 1.46% of all revocations were to jail, for purposes of this report revocations will be generally referenced as revocation to prison. 4

23 Figure 1 shows the Orange County Probation Department Risk Assessment Instrument form. The left side of Figure 1 details the items composing the Orange County Probation initial assessment of adult risk. 5

24 Figure 1: Orange County Probation Department Risk Assessment Instrument 6

25 IV. Methodology A. Sample Data for the OCPD risk validation study were extracted from the OCPD Integrated Case Management System (ICMS). Data provided included all initial risk assessments on felony offenders placed on probation supervision in Selecting the 2007 cohort allowed for a three-year follow-up period, considered a standard follow-up period for recidivism studies. The file contained the following data elements for each offender in the sample: Demographic variables Offense Score on each risk factor Overall initial risk score Risk override reasons Final level of supervision B. Outcome Measures Two general outcome measures were developed for this analysis. Definitions and methodologies for deriving these measures are detailed below: I. Subsequent Law Violations The OCPD Research Administrative Manager and the OCPD IT department constructed a file of subsequent law violations. These data were compiled from the ICMS and a special data request from the Orange County Superior Court to insure a complete compilation of subsequent law violations of the 2007 cohort. A final check of subsequent law violations was conducted by a manual review of the online court system. All cases without subsequent law violations indicated in ICMS were reviewed in the online court records to determine if court records indicated subsequent law violations not captured in ICMS. If subsequent law violations were found, this information was added to the database for the appropriate person. Subsequent new law violations were classified as an arrest for a new law violation and/or an arrest for a new violent law violation. The first new law violation after placement on probation was considered the recidivism event. Subsequent new law violation within three years of placement on probation (referred to in this report as recidivism unless otherwise noted) Subsequent new violent law violation within three years of placement on probation II. Termination Type 7

26 Data extracted from the OCPD ICMS included termination type used to classify an offender s termination from probation as successful or unsuccessful. Termination categories include the following: Successful termination Early termination Normal expiration Conditional termination Relief Of Supervision Ct. Granted Unsuccessful termination Revocation Prison Figure 2 depicts the sample selected and the tracking methodology. The primary outcome measure used in this analysis was an arrest for a subsequent new law violation. This is the most common outcome measure used in validation studies. Two other indicators examined were new violent law violations and terminations for revocations to prison, but both present methodological weaknesses. Subsequent new violent law violations usually have very low base rates and make violence difficult to predict (meaning their overall occurrence is relatively low compared to other type of violations). Additionally, this risk instrument was not developed to predict violence and thus the expectations for predictive accuracy should be low. Analyzing termination data also has its shortcoming as those in the termination pool include probationers with different length-of-stay under supervision. Probationers with supervision lengths of less than three years have a shorter time at risk of potentially having a negative termination than probationers with longer supervision requirements. Therefore, statistically controlling for time at risk is utilized to control for this impact on outcomes. Figure 2: Study Sample and Tracking Methodology 8

27 Figure 3 depicts the data provided by the OCPD with the basic information for conducting the validation study. These data include basic demographic descriptors, offense, probation placement data, risk assessment information, and subsequent law violation information. Figure 3: Data Factors Examined for Validation Study Main Data Elements v Demographic data of offender placed community supervision q Gender q Race/Ethnicity q Date of birth v Offense Information q Offense description q Offense type v Community Supervision Information q Initial probation officer number q Date placed on community supervision v OCPD Information q OCPD risk assessment start and completion dates q OCPD risk factors and risk scores v New Offense Information q Indicator of whether offender committed subsequent new law violation within 3 years following OCPD Risk Assessment completion data. q Indicator of whether offender committed a subsequent new violent law offense within 3 years following OCPD Risk Assessment completion date. q Indicator of whether offenders supervision revoked to prison or jail. As stated above the primary analyses conducted for this study examined the percent of offenders placed on probation who committed a subsequent law violation within three years of the completion of the OCPD risk assessment instrument. Secondary analyses examined new 9

28 violent law violations and the type of termination from supervision by OCPD risk factors (successful/revoked to prison). Analyses in this study examined outcomes by: Each risk factor comprising the OCPD risk assessment instrument. Risk groups by cut-off scores utilized to classify offenders into High, Medium, or Low risk levels. Figure 4 indicates the cut-off scores currently used by the OCPD risk instrument to classify offenders into risk levels for supervision. Cut-off scores were examined to determine if they maximize the predictive ability of the risk instrument. Adjustments are recommended for improving the predictive accuracy of the instrument as necessary. Figure 4: Risk Groups by Score A typical risk assessment instrument is composed of factors that have been determined to be associated with recidivism. The components of each risk factor are weighted to reflect its relationship with recidivism. In the OCPD risk instrument, components of each risk factor indicating low recidivism risk are given fewer points (or zero points) than components indicating high recidivism risk. For example, probationers whose age at first conviction was 24 or older receive zero points compared to four points for probationers whose age was 19 or younger at first conviction. The distribution stems from research demonstrating the younger an offender was at age of first conviction, the higher subsequent recidivism rates will be compared to offenders aged 24 or older at first conviction. By identifying risk factors known to be associated with recidivism, weighting each factor appropriate to the strength of that relationship to recidivism and adding all the scores together, a risk of recidivism can be associated with groups of offenders with those characteristics. Figure 5 depicts the research methodology used in this analysis. 10

29 Figure 5: Overview of Research Methodology OCPD Risk Assessment Instrument Validation Methodology OCPD Risk Assessment Factors and Risk Groups Relation to Subsequent Law Violation and Negative Supervision Termination Correlation of OCPD Risk Factors and Risk Groups with Tested Weights of Risk Factors and Risk Groups Evaluated Validity of OCPD Risk Instrument Recommendations to Improve Predictive Ability of OCPD Risk Instrument 11

30 V. Results A. Overview of Validation Sample Table 1 details the characteristics of the sample of OCPD probation placements tracked for three years. The sample includes 4,875 offenders placed on probation in The most significant characteristics of the population include: Approximately 79 percent of probationers placed on supervision in 2007 were male. Almost half (47.9%) of the population was white Over 50 percent of the population was 30 years of age or younger. Drug sale and drug use offenses represented 53.5 percent of the population. Table 1: Sample Characteristics Frequency Percent Gender Female 1, % Male 3, % Age Group % % % , % % % Race/Ethnicity Black % Hispanic 1, % Other % Vietnamese % White 2, % Offense Type Person % Property % Drug sales or use 2, % Other % B. Distribution of Populations by Risk Factors Table 2 details the distribution of the probation population for each risk factor that is a component of the OCPD risk assessment instrument. The table lists the risk factor, the number classified in each category and the percent of the population that is classified in that category. 12

31 Table 2: Distribution of Risk Factors Frequency Percent Number of address changes last 12 months None 1, % One 1, % Two or more 1, % Total time employed in the last 12 months 7 months or more or n/a 1, % 5 to 7 months % Less than 5 months 3, % Alcohol usage problems-past 12 months No interference with functioning 1, % Occasional abuse (some disruption) 1, % Frequent abuse (serious disruption) 1, % Drug usage problems-past 12 months No interference with functioning 1, % Occasional abuse (some disruption) % Frequent abuse (serious disruption) 3, % Number of prior felony convictions None 2, % One % Two or more 1, % Age at first conviction 24 years old or older 1, % Between 20 and 23 years old 1, % 19 years old or younger 1, % Convictions or adjudications None 4, % Burglary, theft, auto theft, or robbery % Worthless checks or forgery % Prior probation violations (Adult or Juvenile) None 2, % One or more 2, % Attitude Motivated to change 1, % Lacking in motivation, reluctant to accept responsibilities 2, % Negative, not motivated to change % Number of prior probation supervisions None 2, % 1+ prior 2, % 13

32 Analyses examined if distributions of the population on each risk factor presented any issues for the validation analysis. For example, Figure 6 illustrates the distribution of the probation population on the risk factor Convicted or Adjudicated for Burglary. The figure indicates that at placement, 89.2 percent of the probation population was convicted for offenses other than Burglary, Theft, Worthless Checks, and other offenses specified. These offenses are considered high recidivism offenses. Specifically, if the probationer had a conviction for Burglary, Theft, Auto Theft, or Robbery, the probationer receives two points on the risk score. If convicted of Worthless Checks or Forgery, the probationer receives three points on the risk score, indicating that this item is associated with a higher recidivism rate than the Burglary category. The probationer receives zero points if he has never been convicted of any of these offenses. Almost 90 percent of the offenders in the group analyzed were not convicted for any of these offenses, raising the question as to whether this item is effective in differentiating the population. Figure 6: Distribution of Population by Percent Adjudicated for Specific Offenses 100.0% 90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% 89.2% 8.2% 2.6% None Burglary, theft, auto theft, or robbery Worthless checks or forgery C. Percent Subsequent Law Violations by Risk Factors Correlations between risk factors and outcome measures are needed to determine the strength of each factor as a recidivism discriminator and its value to the risk assessment instrument in differentiating the population according to risk of re-offense (as measured by subsequent law violations). This is examined below. Table 3 examines each risk factor by the percent of probationers committing a subsequent new law violation within three years. Overall 42.1 percent (2,050 out of 4,875 offenders) of the sample committed a subsequent new law violation. 14

33 Table 3: Percent Subsequent Law Violation within 3 Years of Placement on Community Supervision by Risk Factors None One Two or more Total 7 months or more or n/a 5 to 7 months Less than 5 months No Violation Subsequent Violation in 36 Months Number of address changes in last 12 months No interference with functioning 60.3% 39.7% % 40.5% 1, % 45.0% 2,825 2, % 42.1% Total time employed in the last 12 months % 32.1% % 41.6% 1,693 1, % 45.9% Alcohol usage problems-past 12 months Occasional abuse (some disruption) Frequent abuse (serious disruption) No interference with functioning Occasional abuse (some disruption) Frequent abuse (serious disruption) % 37.8% 1, % 44.9% % 42.2% Drug usage problems-past 12 months % 25.0% % 42.5% 1,576 1, % 48.0% 15

34 None One Two or more 24 years old or older No Violation Number of prior felony convictions Subsequent Violation in 36 Months 1,762 1, % 36.3% % 45.9% % 52.6% Age at first conviction or juvenile adjudication Between 20 and 23 years old 19 years old and younger None Burglary, theft, auto theft, or robbery Worthless checks or forgery None One or more Motivated to change Convictions or adjudications 1, % 32.9% % 40.9% % 52.6% 2,554 1, % 41.3% % 50.3% % 42.5% Prior probation violations (Adult or Juvenile) Lacking in motivation, reluctant to accept responsibilities Negative, not motivated to change None 1+ prior Attitude 1, % 32.1% 1,249 1, % 51.1% 1, % 38.0% 1, % 43.7% % 47.6% Number of prior periods of probation supervision 1, % 33.5% 1,277 1, % 49.9% 16

35 As Table 3 above indicates, most risk factors identify groups representing different rates of reoffending for subsequent law violations (recidivism as defined here) consistent with valid risk prediction factors, i.e. Low risk items have lower re-offense rates than High risk items. For instance, 32.9 percent of offenders whose age at first adjudication was 24 or older had a new law violation within three years compared to 52.6 percent for offenders 19 or younger. In terms of recidivism, the greater the distinctiveness of a risk factor, the better the predictive ability of the risk item. For example, there is a 19.7 percent difference in the percent of probationers with subsequent law violation for the age at first adjudication factor (52.6% to 32.9%) compared to only a 4.4 percent difference in the percent of probationers with subsequent law violations for the alcohol usage factor (42.2% for frequent abuse and 37.8% for no abuse). The magnitude of differences in recidivism rates for each factor is one basis for weighting risk factors. The greater the difference in recidivism rates the more significant the discriminator is. A second factor that is important in this consideration is the proportion of the population classified by each factor. A risk factor could have significant differences in recidivism rates by each classification category, but if the factor only classifies a small proportion of the population on that factor, it will not be as useful a discriminator as a factor with significant differences that classifies a higher proportion of the population. Subsequent reports will examine these differences as one component of determining if re-weighting of risk factors is merited. Analyses above showed the ability of each risk factor used in the risk assessment to differentiate the population by risk of subsequent new law violations. The analysis below details the review of one risk factor as an example of subsequent analyses presented for each risk factor. Figure 7 illustrates the percent of offenders committing subsequent new law violations within three years of placement on probation by the number of address changes in the last 12 months before placement on probation. The chart indicates that 39.7 percent of offenders with no address changes in the last 12 months had a subsequent law violation in the three-year follow-up period. Approximately 45 percent of probationers who had two or more address changes had a subsequent law violation. 17

36 Figure 7: Percent Subsequent Law Violations by Number of Address Changes Outcomes for this factor are in the anticipated direction supported by the literature; the greater the number of address changes, the higher percent with new law violations. Yet, the difference in outcomes between offenders with no address changes (39.7% new violation) and offenders with one address change (40.5% new violation) is not significant (significance tests are reported later in the report). This suggests address change is not a significant discriminator in the risk score and probably merits re-weighting, rescaling, or elimination as a risk factor. As an example of how a revision might occur, this factor could be dichotomized to 0-1 Address Change and 2 or more Address Changes. This would be more consistent with outcomes reported ( No address changes has about the same recidivism rate as One address change ). Suggested changes for risk factors are recommended in the last chapter of this report. Re-weighting of this factor may also be merited. Key Finding: Most of the risk factors distinguish groups into differential rates of committing a subsequent law violation, although some risk factors are less predictive of re-offending than others 18

37 Figure 8 summarizes factors that should be re-examined due to their predictive weakness. The table divides risk factors with predictive weaknesses into those with categories that demonstrate little difference in the percent law violation and those factors that actually do not discriminate by risk of recidivism. These factors do not contribute to the predictive accuracy of the risk instrument and may actually reduce predictive accuracy. Figure 8: Factors Meriting Re-examination Factor Percent Law Violations Little Difference in Outcomes by Category No Address Change = 39.7% Number of Address Changes One Address Change = 40.5% Attitude Time Employed Alcohol Usage Convictions for Burglary, Auto Theft or Worthless Checks Lacking Motivation = 43.7% Not Motivated = 47.6% 5-7 Months = 41.6% Less than 5 Months = 45.9% Factor Does Not Discriminate Accurately Occasional Abuse = 44.9% Frequent Abuse = 42.2% Burglary = 50.3% Worthless Check = 42.5% D. Distribution of Population and Subsequent New Offense Violations by Initial Risk Group Classification and Final Initial Classification Figure 9 depicts a comparison of the distribution of offenders by risk group at initial risk classification based on risk score and final initial risk classification after overrides were applied. Risk classification based on risk score can be overridden for various policy reasons, most commonly for sex offenders or assaultive cases. When an offender has committed a serious person crime and scores as Low risk, decision makers may want to have that offender supervised at a higher level than the risk score indicates. These types of offenders may have a Low risk of reoffending, but the risk of committing another high harm offense may merit extra caution in supervision. 19

38 Figure 9: Distribution of Risk Groups by Initial Risk Score and Final Risk Classification The initial classification of probationers by the OCPD risk assessment instrument resulted in 54.8 percent of the population classified as High risk. This initial classification of a majority of the population classified as High risk is somewhat questionable as it does not appear to significantly differentiate the population when a majority of offenders are in one risk group. Subsequently, overrides result in the vast majority of offenders being classified as High risk (81.1%) at final initial risk classification. Examination of outcomes by risk based classification and by overrides, as presented later in this report, will assist in determining the utility of the current risk classification and overrides in assigning supervision by risk of re-offending. Table 4 details subsequent law violations after a three-year follow-up by initial scorebased risk groups and by the final initial risk classification for the 2007 cohort after overrides and other classification policies were applied. The table indicates that, in general, the risk assessment instrument differentiates the population by risk of subsequent law violations at the initial classification and at the final initial classification. Initial classification refers to the supervision level associated with the initial risk assessment score. Final initial classification refers to the supervision level the probationer will be initially supervised at after any overrides to the initial risk score classification are applied. The main findings indicated by Table 4 shows that 54.8 percent of the probation population is initially classified as High risk. After overrides and other policies affecting risk classification are applied, 81.1 percent of the probation population is classified as High risk. While the risk instrument does differentiate the population by risk, the differentiation and distribution are problematic as discussed in Section VII below. The differentiation of the population by risk diminishes significantly from the initial score-based risk grouping to the final initial classification 20

39 Key Finding: The high percent of offenders classified as High risk by the OCPD risk score does not support the key goal of risk assessment which is to differentiate the supervision population by risk in order to focus supervision resources by risk. The OCPD risk classification process is not providing this differentiation as well as it could. As stated above, the accuracy of classification is impacted by some risk factors that are improperly weighted or not predictive. Moreover, the use of overrides decreases the predictive accuracy of the risk score. Table 4: Subsequent Violations and Distribution of Population by Initial and Final Initial Risk Group Classification No Violation Subsequent Violation within 36 Months Total Low Medium High Total Low Medium High Total Initial Risk Classification % 18.8% 9.8% of Sample 1, , % 33.0% 35.4% of Sample 1,274 1,391 2, % 52.2% 54.8% of Sample 2,818 2,049 4, % 42.1% 100.0% Final Initial Classification % 28.5% 3% of Sample % 35.9% 16% of Sample 2,223 1,729 3, % 43.8% 81.1% of Sample 2,825 2,050 4, % 42.1% 100.0% 21

40 Figure 10 indicates the percent of the population committing new subsequent law violations within three years of placement on probation by initial risk score-based classification and by final initial risk classification after overrides. While both classification methodologies result in differentiation of the population by risk, using only the risk score for classification results in a clearer distinction than the final classification used for supervision. As the chart illustrates, the range of outcomes for risk score-only classification extends from 18.8 percent of Low risk offenders committing a new law violation in the three-year followup to 52.2 percent for offenders classified as High risk. This is a difference of 33.4 percent. The gamut of outcomes for the final initial classification groupings is only 15.3 percent (28.5% for Low risk to 43.8% for High risk). This indicates the change in risk grouping is associated with overrides and these overrides negatively impact the differentiation of population by risk, because they introduce factors unrelated to differential recidivism rates. Figure 10: Percent Subsequent Law Violation by Risk Group Key Finding: The OCPD risk assessment instrument differentiates the probation population by risk of a subsequent new law violation into groups with different rates of re-offending. The accuracy of the risk-based classification is negatively impacted by overrides. This indicates a sub-population of offenders classified as High risk and supervised at a High risk level are not actually at High risk of reoffending compared to the remaining supervision population. 22

41 E. Severity of Subsequent Law Violation Figure 11 indicates the offense severity of the subsequent law violations. Approximately 13 percent of probationers in the sample were arrested for a felony law violation in the threeyear follow-up, 29 percent were arrested for a misdemeanor, and 58 percent had no subsequent law violation. It should be recognized that only the initial new subsequent law violation was considered as the recidivism event and this was the offense captured for this study. In addition to differentiating the probability of being arrested for a subsequent law violation by risk score, the risk score is also predictive of offense severity. Only 4 percent of Low risk offenders committed a felony offense in the three year follow-up compared to 17 percent of High risk offenders committing a felony offense. Approximately 14 percent of Low risk offenders committed a misdemeanor offense compared to 35 percent of High risk offenders committing a misdemeanor offense. Figure 11: Severity of Subsequent Law Violation Key Finding: While 42% of probationers were arrested for a subsequent new law violation, only 13% of subsequent law violations were for felony offenses. The majority of subsequent law violations were for misdemeanor offenses. 23

42 F. Correlation of Risk Factors to Subsequent New Law Violation Table 5 indicates the correlation of the risk score to each of the factors composing the risk score. A Pearson correlation test is used to establish the strength of the relationship between the dependent variable (commitment of a subsequent new law violation within three years) and the independent variables (the risk factors). The higher the Pearson correlation the greater the correlation is with the dependent variable. For example, Prior probation violations is the factor with the highest correlation with committing a new offense (.193) while Alcohol usage problems has the lowest correlation with committing a new subsequent law violation (.031). Table 5: Correlation of Risk Factors with Subsequent New Law Violation Risk Factor Pearson Correlation Number of address changes in last 12 months.043** Total time employed in the last 12 months.115** Alcohol usage problems-past 12 months.031* Drug usage problems-past 12 months.182** Number of prior felony convictions.141** Age at first conviction or juvenile adjudication.176** Convictions or adjudications, past or present.039** Prior probation violations-adult or juvenile.193** Attitude.071** Number of prior periods of probation supervision.166** **. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed). The correlations of risk factors with subsequent law violations provide empirical support for analyses presented earlier indicating variation in the predictive strength of risk factors. Recommendations for re-weighting and/or replacing certain factors are presented later in this report. Another measure of the association of the risk score with subsequent new law violations utilizes a linear regression analysis that statistically examines the relationship between independent variables (risk factors) and a dependent variable (new offenses). The correlation coefficient R measures the degree to which the independent variables are related to the dependent variable. The value of R can range between 0 and 1, with 0 indicating the independent variables have little association with the dependent variable to 1 indicating a positive relationship between the independent variables and dependent variable. A linear regression analysis was conducted using the risk factors composing the OCPD risk assessment instrument and the dependent variable of subsequent new law violation within 3 years. The R score for the regression analysis was.274 which is on the lower end of R scores computed in other studies of the Wisconsin-like risk assessment instrument. As reported 24

43 in a meta-analysis of 14 studies of the Wisconsin risk instrument by Gendreau (1996), an R score for the Wisconsin risk score used in other states averaged.31. While other states adopted the Wisconsin as it was originally developed, many states have subsequently adopted weights and factors modifying the Wisconsin risk instrument based on research similar to this. Key Finding: The risk factors with the highest correlation with subsequent law violations are factors related to criminal history (prior probations, prior probation violations, prior felony convictions, drug usage problems and age at first conviction). 25

44 VI. Special Analyses A. Overview At the request of the OCPD, analyses of the OCPD risk assessment instrument by gender and race/ethnicity were conducted. Populations and sample sizes examined included: Gender o Male sample size = 3,829 o Female sample size = 1,038 Race / Ethnicity o White = 2,334 o Black = 198 o Hispanic = 1,938 o Asia/Pacific Islander/Other = 401 B. Distribution of Risk Groups by Gender Figure 12 indicates the distribution of risk groups, using the initial risk-based score, by gender. There are small differences in the distribution of the population by risk by gender. For example, 11.1 percent of females are classified as Low risk compared to the 9.5 percent of males. Approximately 50 percent of females are classified as High risk compared to 56 of males. Figure 12: Distribution of Risk Groups by Gender 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 50.30% 56.00% 38.60% 34.50% 11.10% 9.50% Female Male Low Medium High 26

45 C. Subsequent Law Violation by Risk Group by Gender Figure 13 indicates the percent of subsequent law violation during the follow-up period by risk groups and gender. While females have lower re-offense rates than males in every risk group category (a common finding in recidivism studies), the percent with subsequent law violations is consistent with their risk score classification. The OCPD risk assessment instrument effectively classifies offenders into groups having differential risk of re-offending, regardless of gender. Of probationers classified as Low risk, approximately 20 percent of males had a subsequent law violation compared to 14.8 percent of females. High risk males had a 53.1 percent violation rate compared to 48.7 percent for High risk females.. Figure 13: Percent Subsequent Law Violation by Risk Group by Gender When examining these relationships by final initial classification, the distributions are skewed toward classifying most offenders as High risk (Female = 76% High risk; Male = 82% High risk) and the differentiation of the population by risk is much weaker. For example, the difference between Low and High risk for reoffending based on risk score is exceeds 30 percent for both males (20.1% Low and 53.1% High) and females (14.8% Low and 53.1% High). The difference in reoffending using final initial classification is less than 16 percent for both males (30.8% Low and 45.0% High) and females (22.5 Low and 38.6% High) Key Finding: The OCPD risk assessment differentiates the male and female populations equitably by risk of committing a new offense into groups with different rates of committing a new offense. The differentiation is diminished at the final initial classification similarly by gender. 27

46 D. Distribution of Risk Groups by Race/Ethnicity Figure 14 details the distribution of risk classifications on the OCPD risk score by race/ethnicity. There is little variation in the distribution of the population by initial risk score by race/ethnicity. This is also true at final initial risk classification as over 80 percent of each race/ethnicity group is classified as High risk. Figure 14: Distribution of Risk Groups by Race/Ethnicity 28

47 E. Subsequent Law Violation by Risk Group by Race/Ethnicity Figure 15 indicates that, regardless of race/ethnicity, offenders are classified accurately into risk groups commensurate to their risk of re-offending. While there is variation in the percent with subsequent law violations by race/ethnicity by risk score, there is consistency in the percent of subsequent law violations by risk score. Figure 15: Percent New Offense by Risk Groups by Race/Ethnicity 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 60.3% 52.7% 51.6% 34.9% 35.0% 32.1% 14.3% 20.6% 20.1% Black Hispanic White Low Medium High Key Finding: The OCPD Risk Assessment instrument differentiates racial and ethnic populations equitably by risk of committing a new offense into groups with different rates of committing a new offense but the risk instrument classifies a majority of offenders in the High risk category for all race groups. 29

48 F. Subsequent New Violent Law Offenses and Risk Score Approximately two percent (99 out of 4,867) of probationers placed on probation in 2007 subsequently committed a new violent law violation (based on first subsequent law violation captured during the three year follow-up). Violent offenses are defined statutorily by California law. The majority of violent offenses committed by this probation cohort were felony assaults. Analyses were conducted to determine the risk assessment instrument s ability to classify probationers by risk of committing a subsequent new violent offense. It should be recognized that the OCPD risk assessment instrument was not designed as a predictor of committing a new violent offense. Most risk instruments designed and purporting to predict violent offending have identified different predictive variables than identified for risk instruments predicting new offending behavior, although there is overlap in some variables for both types of instruments. In general, predicting violence is difficult due to the relatively low base rate of violent reoffending. It is easier to predict an incident that happens 50 percent of the time than an incident that happens 20 percent of the time. In this instance, it is easier to predict the possibility of any re-offense than the possibility of re-offense that is violent. This is particularly the case when the base rate of OCPD new violent offenses is two percent of the cases committing a new law violation. Figure 16 indicates the percent of probationers with a new violent offense in the followup period based on the initial risk-based score groupings and then by the final initial risk grouping after overrides. The initial risk based groupings show no ability of the risk assessment instrument to group offenders by risk of committing a subsequent violent law violation; however, the final initial groups differentiate the population by risk of violent re-offending. The sample size is too small for these differences to be statistically significant. It should also be recognized that, even in the High risk group which has the highest percent of new violent offenses, 97.7 percent of the High risk group did not commit a new violent offense. Most violence risk predictors suffer from over-predicting violence and having too many false positives. However, further examination is merited. 30

49 Figure 16: Percent Subsequent Violent Law Violation by Initial Score-based Risk Group and Final Initial Risk Group After this initial analyses, OCPD management staff reviewed subsequent law violations of this sample to assess the violations more generally for harm or potential harm to victims as a violence measure. Each subsequent law violation was reviewed to determine if the subsequent law violation could be categorized as involving: a. Victim or potential victim b. Violent offense c. Drug sales d. Weapon involved in offense This identification of subsequent law violations provided a broad measure of subsequent law violations with victimization or victimization potential. Approximately 28 percent of the sample involved subsequent law violations with victim or victim potential. Using these classifications the research team developed a risk instrument which could be characterized as Victim Probability Predictor (VPP). Risk factors were identified by examining the correlation of the factor with the subsequent law violation involving a victim. The instrument classifies probationers into risk groups of committing a subsequent law violation involving a victim or potential victim. Figure 17 below details the risk factors, weights, and percent committing a subsequent law violation involving a victim by risk factors. 31

50 Figure 17: Victim Probability Risk Factors Weight No Victim Victim Age at First Conviction 0 24 or older 79.0% 21.0% % 28.2% 4 19 or younger 64.6% 35.4% Number of Prior Probation Periods 0 None 74.3% 25.7% 4 One or more 69.8% 30.2% Attitude 0 Motivated to change 74.6% 25.4% 1 Lacking in Motivation 70.7% 29.3% 2 Negative 69.0% 31.0% Companions 0 Good support and influence 78.8% 21.2% 2 Association with occasional negative 72.4% 27.6% 4 Association almost completely negative 68.8% 31.2% Offense Type 0 Not property 73.8% 26.2% 4 Property 62.7% 37.3% Total 72.0% 28.0% Figure 18 below indicates the relationship between the Victim Probability Risk (VPR) group and the percent committing a subsequent law violation that involved a victim or had the potential to involve a victim. Probationers classified as High risk on the VPR had twice the rate of victims or potential victims (36.9%) as the group classified as Low risk (16.0%). While the classification of victim utilizes a very general definition, it may serve as a useful identifier of a High risk group that should be subject to closer supervision than the general risk grouping of probationers who are High risk of committing any subsequent law violation. Figure 18: Victim Probability Risk Group and Subsequent Law Violation with Victim No Victim Victim Low (0-2) % % Medium (3-10) 2, % % High (11-18) % % Total 3, % 1, % 32

51 G. Revocations and Risk Score Figure 19 presents the relationship between the initial risk score classification, final initial classification, and percent revoked out of cases terminated. As discussed earlier, revocation rates as recidivism measure presents a methodological problem risk of revocation is subject to the period of supervision ordered. In other words, the revocation group does not have a uniform risk time, with some being under supervision longer than others and thus at risk of revocation for varying time periods. A group of offenders whose term of supervision is two years will have less time at-risk for revocation than a group of offenders under supervision for three years. It is likely the group under supervision for two years will have lower revocation rates than the three year group, because they cannot be revoked after that two year period, not because of any other factor such as programming, better supervision, etc. The offenders receiving a two year supervision period probably have a less extensive criminal history than offenders receiving three year supervision sentence, which is not only associated with a lower likelihood for recidivating, but also compounds the time at risk issue. Figure 19 depicts risk classifications are effective in differentiating the population by risk of revocation. As in the previous discussion, the initial risk score classification differentiates the population by risk more effectively than the final initial risk classification. The lower recidivism rate of High risk cases after the final initial classification is the result of classifying Low risk offenders as High risk. The Low risk offenders classified as High risk at Final Initial classification artificially suppresses the recidivism rate associated with the High risk population at Final Initial classification. 33

52 Figure 19: Percent Revoked by Initial Risk and Final Initial Risk Classification H. Analysis of Overrides When judicial and/or agency policies mandate an increase in supervision level from the risk instrument s recommendation, supervision agencies must override risk assessment supervision levels. High harm offenses, such as homicide or sexual assault, are common triggers of overrides. When a high harm offender scores as Low risk on the risk assessment, but offense and criminal history reflect a need for a higher level of supervision, the department will override the offender s score. Criminal justice research suggests that overriding more than 15 percent of cases can undermine the effectiveness and utilization of a risk instrument. Parole populations have a greater proportion of high harm offenders than probation populations; therefore, the use of overrides is more likely to occur with the parole population. The OCPD risk sample had 479 offenders with initial Low risk scores. In 315 cases (66% of all Low risk), the offender s supervision level was overridden to High risk. The largest offense categories of Low risk overrides to High risk included Person (n=129) and Misdemeanor Other (n=97) offenders. The OCPD risk sample had 1,723 offenders with initial Medium risk scores. In 1,005 cases (58% of all Medium risk), overrides moved the offender s supervision level to High risk. The largest offense categories of Medium risk overrides to High risk were Drug (n=304) and Person (n=291) offenders. Overall, 27 percent (1,320 out of 4,875) of probationers in the sample had their initial risk classification overridden from Low or Medium risk to High risk. Figure 20 indicates probationers with overrides had outcomes similar to their initial risk classification. In other words, a probationer whose risk score indicated he was Low risk but was subsequently classified as High risk had subsequent law violations at a rate similar to a Low risk 34

53 offender. In fact, the population with overrides performed better than the Low risk offenders without. Probationers initially classified as Low risk and supervised as Low risk had a 28.5 percent subsequent law violation rate compared to 14.5 percent for probationers initially classified as Low risk and overridden to a High risk supervision level. This same pattern occurred with Medium risk offenders. An important question is whether the overrides of Low and Medium risk offenders to High risk increases or reduces recidivism (subsequent law violations). Research suggests supervising a Low risk as a High risk can increase recidivism, but this was not the case in this study. The literature also indicates the offenders most likely to receive an override due to the high harm nature of their offense, usually have low re-offense rates. The reported results indicate this may be the case with Orange County s offenders. The population receiving overrides into the High risk category may have received more effective supervision than at their original classification. This is an issue in need of further investigation. Figure 20: Percent Subsequent Law Violation by Overrides 35

54 VII. Developing a Brief Risk Screening Instrument The initial risk assessment and classification procedures for determining supervision level and resources to allocate to new probationers can be labor intensive and costly. Community supervision agencies are utilizing new procedures to triage their supervision population by quickly identifying Low risk offenders, reducing supervision resources to those offenders and re-allocating those resources to higher risk offenders. Instead of conducting a full risk assessment and classification procedure for all offenders, a brief screening tool can provide a risk assessment measure to identify those Low risk probationers who can be transferred immediately to administrative type caseloads, resulting in fewer cases requiring a full assessment. OCPD seeks to develop a brief screening tool to supplement a validated risk assessment instrument. This section describes the process and results of efforts to develop a brief risk screening instrument. To develop the Brief Risk Screen (BRS), risk factors with the highest correlation with subsequent law violations were examined. In addition to existing risk factors with the highest correlation to subsequent law violations, two other risk factors were developed for consideration in the development of the BRS: Age at Initial Risk Assessment and Offense Category. The relationship between these two new factors and subsequent law violations is detailed in Table 6 below. Table 6: Offense Category and Age at Assessment Risk Factors Offense Type No Violation Subsequent Violation in 36 Months Total Not Property or Drug 1, , % 31.80% % Property or Drug 1,789 1,566 3,355 Age at Assessment 53.30% 46.70% % No Violation Subsequent Violation in 36 Months Total 25 or less , % 49.00% % Over 25 to 40 1, , % 41.00% % Over , % 33.40% % 36

55 Table 7 indicates the risk factors with the highest correlation with subsequent law violation considered in the development of the BRS. Utilizing regression analysis, these factors were considered in the development of the Orange County Probation Department Brief Risk Screen instrument. Table 7: Risk Factors with Highest Correlation with Subsequent Law Violations Risk Factor: Correlation with Subsequent law Violations Pearson Correlation Drug usage problems-past 12 months.182** Number of prior felony convictions.141** Age at first conviction or juvenile adjudication.176** Prior probation violations-adult or juvenile.193** Number of prior periods of probation supervision.166** Age at initial assessment.121** Drug or property offense.139** As described in the section below, the BRS was designed to utilize available data that would minimize the use of staff time devoted to screening probationers. Risk factors with the highest correlation to subsequent new law violations were available through existing computerized records with the exception of drug usage problems. The assessment of drug usage problems requires an offender interview and drug problem assessment. It was determined that the predictive ability of the screen would not be significantly impacted by not using the drug usage factor and would facilitate achieving the goal of conducting a screen with minimal staff resources. The BRS also does not utilize Prior Felony Convictions, which is highly correlated with risk of subsequent law violations. The correlation of prior felony convictions and prior periods of probation supervision is high. The omission of prior felony convictions in the BRS did not significantly impact the predictive ability of the screening instrument and aided in minimizing staff resources devoted to the BRS. 37

56 Table 8 details the risk factors identified by the regression analysis in developing the BRS. Weights assigned for each factor are provided. Table 8: Brief Risk Screen Factors and Weights Weight Frequency Percent Age at First Conviction 0 24 or older 1, % , % 4 19 or younger 1, % Number of Prior Probation Periods 0 None 2, % 4 One or more 2, % Prior Probation Violations 0 None 2, % 4 One or more 2, % Age at Assessment 0 40 or older 1, % , % 4 Under 25 1, % Offense Type 0 Not property/drug offense 1, % 2 Property/drug offense 3, % Figure 21 indicates the relation between BRS risk groups and subsequent law violations. The lowest risk group (0-2) had an 18.1 percent subsequent law violation rate after 3 years compared to the overall rate of 42.1 percent. The table also indicates the percent of subsequent law violations that were felony level offenses. Approximately 5.3 percent of probationers scoring 0-2 on the BRS committed a new felony offense. Approximately 12 percent of this sample were classified as Low risk (0-2 score) and 32 percent of the sample were classified as Low-Medium risk (4-8 score). The R score for the regression analysis was.272, which is only slightly lower than the.274 R for the current risk assessment instrument using all factors. This was achieved by using only factors that had the highest correlation with subsequent law violations and by adding two new factors with significant correlations with subsequent law violations. 38

57 Figure 21 : Brief Risk Screen Groups and Subsequent Law Violations 39

58 VIII. Recommended Tasks to Modify Risk Assessment A. Revise Weights for Selected Factors and Define New Risk Factors Table 9 illustrates the recommended weights and factors for the OCPD revised risk assessment instrument. Weight for Address Changes factor reduced due to weakness of predictive ability and categories combined. Alcohol Usage factor eliminated due to predictive inaccuracy. Weight for Attitude factor reduced due to weakness of predictive ability. Conviction or adjudication for specific offense changed to general offense categories of property and drug offenses (see specific offenses in Appendix). Weights reduced to reflect predictive accuracy. Age at Assessment factor added to Risk score and weighted equivalent to Age at Conviction factor due to strength of predictive ability. 40

59 Address Changes Table 9: Revised Risk Factors and Weights Score Subsequent Violation in 36 Months None or One % Two or more % Time Employed 7 months or more % 5 to 7 months % Less than 5 months % Drug usage No interference % Occasional abuse % Frequent abuse % Prior Felony Convictions None % One % Two or more % Age at first conviction 24 or older % 20 to % 19 or younger % Prior probation violations None % One or more % Attitude Motivated to change % Lacking in motivation % Negative % Prior periods of supervision None % One or more % Offense type Not Property or Drug % Property or Drug % Age at Assessment Over % 25.1 to % 25 or less % OCPD management staff suggested examining the needs items scored at assessment to determine if additional dynamic factors could increase the predictive ability of the risk 41

60 instrument. The Companions item was identified by OCPD management staff as highly likely to be a useful risk predictor. As Table 10 indicates, the Companions item is an excellent predictor of subsequent law violations. The correlation with subsequent law violations is.169, one of the highest correlations of all the risk factors in the risk score; however, Companions is also highly correlated with other risk factors in the risk score. The high correlation with other risk factors results in the Companion factor only adding.001 to the R. Additionally, when a regression analysis is conducted, Companions does not enter the equation as a significant factor, again due to its high correlation with other risk factors in the equation. Based on this analysis, it was determined that the Companions item would not be added to the risk score revision. Table 10: Companions Needs Item and Subsequent Law Violations Subsequent No Violation Violation within 36 Months Good support and influence % % Association with occasional negative results 1, % % Association almost completely negative % % B. Revise Cut-off Points Table 11 outlines the current and proposed risk levels by scores. The relationship between risk score and re-offense rate determine the cut-off scores for each level. Table 11: Current and Proposed Risk Assessment Instrument Cut-Off Points Risk Level Current Score Proposed Score Low Medium High 19 or greater 21 or greater C. Distribution of Risk Groups and Subsequent Law Violations of Risk Groups: Current and Proposed Risk Groups Figure 22 illustrates a comparison of current and proposed supervision population by risk level. The proposed supervision population reflects the new cut off scores and the risk factor reweights and changes. All risk groups are based on the initial risk score classification and do not reflect the impact of overrides. As Figure 23 shows, the percent classified as High risk (overall) declined from 54.8 percent to 34.5 percent under the proposed weights and cut-off scores, while the percent classified as Low risk increased from 9.8 percent to 16.3 percent. 42

61 Figure 22: Distribution of Risk Groups: Current Risk Distribution Compared to Proposed Figure 23 below indicates re-offense rates for the proposed revised score and the current distribution were very similar, even with the significant redistribution of risk groups. As discussed above, the percent classified as Low risk increased from 9.8 percent to 16.3 percent, while the re-offense rate for the Low risk group had a smaller increase, moving from 18.8 percent 19.2 percent. The re-offense rate for the Low risk group under the proposed risk score is virtually unchanged. More Low risk offenders were identified with closely the same re-offense rate. The new risk classification identified fewer High risk offenders; however, the offenders had higher re-offense rates. Under the current risk instrument, approximately 55 percent of the population was identified as High risk, but proposed risk instrument identifies 34.5 percent of the population as High risk. The current instrument has a High risk re-offense rate of 52.2 percent; the revised risk score generates a group with a 56.6 percent re-offense rate. The rate gain indicates increased accuracy in classifying offenders as High risk. The recidivism rate of the medium risk group increase from 33.0 percent to 39.5 percent, which approaches the average recidivism rate of the sample (42.1%). This is more reflective of a medium recidivism rate than the current rate, which is 10 percent below the average. 43

62 Figure 23: Percent Subsequent Law Violation by Risk Groups: Current Risk Groups Compared to Proposed D. Comparison of Predictive Ability of Current and Proposed Risk Score Table 12 shows the comparison of current and proposed risk score predictive ability based on regression analysis. A linear regression analysis was conducted using the proposed risk factors and new weights. The previous R score for the regression analysis was.274 using the current factors compared to an R score of.288 using the proposed factors and weights, indicating an improvement in the predictive ability of the risk instrument. Table 12: Comparison of Current and Proposed Risk Score Predictive Ability Based on Regression Analysis R Model Model Summary R R Square Current Risk Factors Proposed Risk Factors A second measure of improvement in risk classification associated with the proposed revision in the risk score is the Receiver Operating Characteristic Curve or the ROC Curve. The ROC Curve is a measure that evaluates the performance of a classification scheme in which there is one variable (Risk Score or the Revised Risk Score) with two categories (New Offense 44

63 within three years or No New Offense within three years) by which subjects are classified. The Area under the Curve (AUC) represents the probability that the result of the classification for a randomly chosen positive case (prediction of re-offense that is true) will exceed the result of a randomly chosen negative case. The curve is a graphical representation of the trade-off between false negative and false positive rates. The ROC curve for both the current risk score and the proposed revised risk score are shown below in Figure 24. The ROC curve for the proposed revised risk score exceeds the ROC curve for the current risk score; this indicates greater accuracy of the proposed score in classifying offenders by risk. The area under the curve for the proposed revised risk score (.659) exceeds the current risk score (.642) and the lower and upper bounds of the proposed revised risk score (.644 and.674 respectively) exceed the current risk score (.626 and.657). Figure 24: ROC Curve for Current and Proposed Revision to Orange County Risk Score 45

The Substance Abuse Felony Punishment Program: Evaluation and Recommendations

The Substance Abuse Felony Punishment Program: Evaluation and Recommendations The Substance Abuse Felony Punishment Program: Evaluation and Recommendations Criminal Justice Policy Council Prepared for the 77 th Texas Legislature, 2001 Tony Fabelo, Ph.D. Executive Director The Substance

More information

CJI/CPOC OCTOBER WORK SESSION Organizational Development and Change Management

CJI/CPOC OCTOBER WORK SESSION Organizational Development and Change Management October 12, 2011 CJI/CPOC OCTOBER WORK SESSION Organizational Development and Change Management Wendy S. Still David Koch Chief Adult Probation Officer Chief Deputy Probation Officer San Francisco Adult

More information

2009 Florida Prison Recidivism Study Releases From 2001 to 2008

2009 Florida Prison Recidivism Study Releases From 2001 to 2008 2009 Florida Prison Recidivism Study Releases From 2001 to 2008 May 2010 Florida Department of Corrections Walter A. McNeil, Secretary Bureau of Research and Data Analysis dcresearch@mail.dc.state.fl.us

More information

The Impact of Arizona s Probation Reforms in 2010

The Impact of Arizona s Probation Reforms in 2010 Issue Brief Project Public Safety NamePerformance Project The Impact of Arizona s Probation Reforms An analysis of trends in Arizona s prison system in 2008 estimated that the inmate population would increase

More information

Overall, 67.8% of the 404,638 state

Overall, 67.8% of the 404,638 state U.S. Department of Justice Office of Justice Programs Bureau of Justice Statistics Special Report April 2014 ncj 244205 Recidivism of Prisoners Released in 30 States in 2005: Patterns from 2005 to 2010

More information

Statistics on Women in the Justice System. January, 2014

Statistics on Women in the Justice System. January, 2014 Statistics on Women in the Justice System January, 2014 All material is available though the web site of the Bureau of Justice Statistics (BJS): http://www.bjs.gov/ unless otherwise cited. Note that correctional

More information

Re-validation of the Nonviolent Offender Risk Assessment Instrument: Study Update

Re-validation of the Nonviolent Offender Risk Assessment Instrument: Study Update Re-validation of the Nonviolent Offender Risk Assessment Instrument: Study Update Current Instrument 3 Refined Risk Assessment Instrument: Significant Factors in Assessing Risk Relative Degree of Importance

More information

Working Paper # 2009-09 March 4, 2009. Prisoner Reentry and Rochester s Neighborhoods John Klofas

Working Paper # 2009-09 March 4, 2009. Prisoner Reentry and Rochester s Neighborhoods John Klofas Center for Public Safety Initiatives Rochester Institute of Technology John M. Klofas, Ph.D, Director Chris Delaney, Deputy Director John.klofas@rit.edu (585) 475-2423 Working Paper # 2009-09 March 4,

More information

PRETRIAL CRIMINAL JUSTICE RESEARCH

PRETRIAL CRIMINAL JUSTICE RESEARCH NOVEMBER 2013 RESEARCH SUMMARY Pretrial criminal justice research commissioned by the Laura and John Arnold Foundation (LJAF) has thrown new light on how critical the earliest decisions made in the criminal

More information

LEGISLATIVE BUDGET BOARD. Statewide Criminal Justice Recidivism and Revocation Rates

LEGISLATIVE BUDGET BOARD. Statewide Criminal Justice Recidivism and Revocation Rates LEGISLATIVE BUDGET BOARD Statewide Criminal Justice Recidivism and Revocation Rates LEGISLATIVE BUDGET BOARD STAFF SUBMITTED TO THE 83RD TEXAS LEGISLATURE JANUARY 2013 STATEWIDE CRIMINAL JUSTICE RECIDIVISM

More information

RUTLAND COUNTY TREATMENT COURT

RUTLAND COUNTY TREATMENT COURT Data Driven Decisions RUTLAND COUNTY TREATMENT COURT CONTROL GROUP EVALUATION FINAL REPORT Submitted to: Karen Gennette State Treatment Court Coordinator Vermont Court Administrator s Office Submitted

More information

Contra Costa County: A Model for Managing Local Corrections

Contra Costa County: A Model for Managing Local Corrections JFA Institute Conducting Justice and Corrections Research for Effective Policy Making Contra Costa County: A Model for Managing Local Corrections Prepared by James Austin, Ph.D. Robin Allen Roger Ocker

More information

Published annually by the California Department of Justice California Justice Information Services Division Bureau of Criminal Information and

Published annually by the California Department of Justice California Justice Information Services Division Bureau of Criminal Information and Published annually by the California Department of Justice California Justice Information Services Division Bureau of Criminal Information and Analysis Criminal Justice Statistics Center 2011 Juvenile

More information

In many jurisdictions, state and local government

In many jurisdictions, state and local government September 2012 States Report Reductions in Recidivism In many jurisdictions, state and local government officials have intensified their efforts to reduce recidivism. As policymakers are under tremendous

More information

6/16/2010. Participant. Public Defender Assistant District Attorney. Assessor Counselor

6/16/2010. Participant. Public Defender Assistant District Attorney. Assessor Counselor Evidence-Based Practices Courts Hon. John C. Creuzot Dallas, Texas Teresa May-Williams, Ph.D. Assistant Director Dallas County CSCD Texas Definition of Courts Texas law allows probation even with prior

More information

Fresno County Public Safety Realignment: One Year of Data. Interim Report

Fresno County Public Safety Realignment: One Year of Data. Interim Report Fresno County Public Safety Realignment: One Year of Data Interim Report Prepared by Owen Research & Evaluation September 23, 203 ORE Owen Research & Evaluation Introduction and Overview This report presents

More information

This chapter shall be known and may be cited as the Alyce Griffin Clarke Drug Court Act.

This chapter shall be known and may be cited as the Alyce Griffin Clarke Drug Court Act. 9-23-1. Short title This chapter shall be known and may be cited as the Alyce Griffin Clarke Drug Court Act. HISTORY: SOURCES: Laws, 2003, ch. 515, 1, eff from and after July 1, 2003. 9-23-3. Legislative

More information

Probation is a penalty ordered by the court that permits the offender to

Probation is a penalty ordered by the court that permits the offender to Probation and Parole: A Primer for Law Enforcement Officers Bureau of Justice Assistance U.S. Department of Justice At the end of 2008, there were 4.3 million adults on probation supervision and over 800,000

More information

Delinquent Youth Committed to the Department of Youth Rehabilitation Services 2004-2011

Delinquent Youth Committed to the Department of Youth Rehabilitation Services 2004-2011 Delinquent Youth Committed to the Department of Youth Rehabilitation Services 2004-2011 Akiva M. Liberman, Ph.D. Jennifer Yahner, M.A. John K. Roman, Ph.D. August 2012 2012. The Urban Institute. All rights

More information

The Juvenile Justice Commission s Stabilization and Reintegration Program: An Updated Recidivism Analysis

The Juvenile Justice Commission s Stabilization and Reintegration Program: An Updated Recidivism Analysis The Juvenile Justice Commission s Stabilization and Reintegration Program: An Updated Recidivism Analysis Submitted by: New Jersey Department of Law & Public Safety Juvenile Justice Commission Office of

More information

CORRELATES AND COSTS

CORRELATES AND COSTS ANOTHER LOOK AT MENTAL ILLNESS AND CRIMINAL JUSTICE INVOLVEMENT IN TEXAS: CORRELATES AND COSTS Decision Support Unit Mental Health and Substance Abuse Services Another Look at Mental Illness and Criminal

More information

Pierce County. Drug Court. Established September 2004

Pierce County. Drug Court. Established September 2004 Pierce County Drug Court Established September 2004 Policies and Procedures Updated September 2013 TABLE OF CONTENTS I. Drug Court Team II. Mission Statement III. The Drug Court Model IV. Target Population

More information

Risk/Needs Assessment 101: Science Reveals New Tools to Manage Offenders

Risk/Needs Assessment 101: Science Reveals New Tools to Manage Offenders Issue Brief Project PUBLIC SAFETY NamePERFORMANCE PROJECT Risk/Needs Assessment 101: Science Reveals New Tools to Manage Offenders Every day, criminal justice officials make decisions that have enormous

More information

Criminal Justice Research Report

Criminal Justice Research Report Division of Criminal Justice Services Office of Justice Research and Performance Criminal Justice Research Report Andrew M. Cuomo Governor Michael C. Green September 2012 Executive Deputy Commissioner

More information

North Carolina Criminal Justice Performance Measures

North Carolina Criminal Justice Performance Measures North Carolina Criminal Justice Performance Measures November 17, 214 Prison Population North Carolina Prison Population: 1997 214 45, 4, 35, 31,581 4,824 41,3 37,665 3, 25, 2, 15, 1, 5, 1997-2 Growth:

More information

Con-Quest Residential Substance Abuse Treatment Program Outcome Evaluation. February 2004

Con-Quest Residential Substance Abuse Treatment Program Outcome Evaluation. February 2004 Residential Substance Abuse Treatment Program Outcome Evaluation February 2004 Introduction The link between drugs and crime in the United States is widely accepted. Drug users frequently commit crime

More information

Domestic Violence Offenders in Missouri

Domestic Violence Offenders in Missouri MISSOURI STATE HIGHWAY PATROL Domestic Violence Offenders in Missouri A Study on Recidivism Eric Chambers & Mark Krispin March 2011 Contents I. BACKGROUND... 5 II. METHODOLOGY... 5 III. ANALYSIS OF ARRESTS

More information

Title 34-A: CORRECTIONS

Title 34-A: CORRECTIONS Title 34-A: CORRECTIONS Chapter 5: PROBATION AND PAROLE Table of Contents Subchapter 1. GENERAL PROVISIONS... 3 Section 5001. DEFINITIONS... 3 Section 5002. PARDONS BY THE GOVERNOR... 4 Section 5003. PROHIBITED

More information

Lifetime Likelihood of Going to State or Federal Prison

Lifetime Likelihood of Going to State or Federal Prison U.S. Department of Justice Office of Justice Programs Bureau of Justice Statistics Special Report March 1997, NCJ-160092 Lifetime Likelihood of Going to State or Federal Prison By Thomas P. Bonczar and

More information

Crime in Missouri 2012

Crime in Missouri 2012 Crime in Missouri MISSOURI STATE HIGHWAY PATROL RESEARCH AND DEVELOPEMENT DIVISION STATISTICAL ANALYSIS CENTER FOREWORD This publication is produced by the Missouri State Highway Patrol, Statistical Analysis

More information

Arrests in Wisconsin 2010

Arrests in Wisconsin 2010 Wisconsin Office of Justice Assistance 1 S. Pinckney Street, Suite 615 Madison, WI 53703 Scott Walker Governor Arrests in Wisconsin 2010 July 2011 Arrests in Wisconsin 2010 i 07/2011 Wisconsin Statistical

More information

SEATTLE MUNICIPAL COMMUNITY COURT

SEATTLE MUNICIPAL COMMUNITY COURT SEATTLE MUNICIPAL COMMUNITY COURT Outcome Evaluation Final Report October 2009 By: M. Elaine Nugent Borakove TABLE OF CONTENTS Introduction...1 Evaluation Design...2 Principal Findings...3 Defendant Characteristics...4

More information

The Hamilton County Drug Court: Outcome Evaluation Findings

The Hamilton County Drug Court: Outcome Evaluation Findings The Hamilton County Drug Court: Outcome Evaluation Findings Final Report Submitted by: Shelley Johnson, M.S. Project Director and Edward J. Latessa, Ph.D. Principal Investigator University of Cincinnati

More information

Mercyhurst College Civic Institute. An Overview of the Erie County Criminal Justice System

Mercyhurst College Civic Institute. An Overview of the Erie County Criminal Justice System Mercyhurst College Civic Institute An Overview of the Criminal Justice System January 2005 Erika Brown, Research Analyst Art Amann, Director Table of Contents Table of Contents...1 Introduction...2 Methodology...2

More information

Evaluation of the San Diego County Community Corrections Programs

Evaluation of the San Diego County Community Corrections Programs JFA Institute Conducting Justice and Corrections Research for Effective Policy Making Evaluation of the San Diego County Community Corrections Programs Prepared by James Austin, Ph.D. Roger Ocker Robert

More information

Using Data to Inform Evidence-Based Decision Making. January 8, 2013

Using Data to Inform Evidence-Based Decision Making. January 8, 2013 Using Data to Inform Evidence-Based Decision Making January 8, 2013 1 New York State Criminal Justice System 500 local police agencies arrest 600,000 annually. 62 locally elected county district attorneys

More information

Community-Based Program Evaluation Series: Overview of Community-Based Juvenile Probation Programs. Part 1

Community-Based Program Evaluation Series: Overview of Community-Based Juvenile Probation Programs. Part 1 Community-Based Program Evaluation Series: Overview of Community-Based Juvenile Probation Programs Part 1 June 2013 Introduction The Texas Juvenile Justice Department (TJJD) Program & Services Registry

More information

Trends Related to the Certification of Juveniles as Adults

Trends Related to the Certification of Juveniles as Adults Trends Related to the Certification of Juveniles as Adults Prepared for the Texas House of Representatives Corrections Committee 82 nd Legislative Session Interim March 6, 2012 Overview Certification is

More information

Governor Jennifer M. Granholm,

Governor Jennifer M. Granholm, Analyses of Crime, Community Corrections, and Sentencing Policies Governor Jennifer M. Granholm, Senate Majority Leader Michael D. Bishop, and Speaker of the House Andy Dillon requested intensive technical

More information

SORNA Substantial Implementation Review State of Tennessee

SORNA Substantial Implementation Review State of Tennessee September 2011 SORNA Substantial Implementation Review State of Tennessee The U.S. Department of Justice, Office of Justice Programs, Office of Sex Offender Sentencing, Monitoring, Apprehending, Registering,

More information

Evidence Based Correctional Practices

Evidence Based Correctional Practices Evidence Based Correctional Practices What works in corrections is not a program or a single intervention but rather a body of knowledge that is accessible to criminal justice professionals. 1 The National

More information

Chapter 938 of the Wisconsin statutes is entitled the Juvenile Justice Code.

Chapter 938 of the Wisconsin statutes is entitled the Juvenile Justice Code. Juvenile Justice in Wisconsin by Christina Carmichael Fiscal Analyst Wisconsin Chapter 938 of the Wisconsin statutes is entitled the Juvenile Justice Code. Statute 938.1 of the chapter states that it is

More information

2012 Iowa Board of Parole Risk Assessment Validation

2012 Iowa Board of Parole Risk Assessment Validation 2012 Iowa Board of Parole Risk Assessment Validation Iowa Department of Human Rights Division of Criminal and Juvenile Justice Planning Statistical Analysis Center March 20, 2012 Cheryl Davidson, Analyst

More information

Crime Rates and Youth Incarceration in Texas and California Compared: Public Safety or Public Waste?

Crime Rates and Youth Incarceration in Texas and California Compared: Public Safety or Public Waste? CENTER ON JUVENILE AND CRIMINAL JUSTICE JUNE 2007 www.cjcj.org Crime Rates and Youth Incarceration in Texas and California Compared: Public Safety or Public Waste? By Mike Males PhD, Christina Stahlkopf

More information

Evaluation of the Performance of the Texas Department of Criminal Justice Rehabilitation Tier Programs

Evaluation of the Performance of the Texas Department of Criminal Justice Rehabilitation Tier Programs Evaluation of the Performance of the Texas Department of Criminal Justice Rehabilitation Tier Programs Criminal Justice Policy Council Prepared for the 77 th Texas Legislature, 2001 Tony Fabelo, Ph.D.

More information

How To Change The Way A Prison System Works

How To Change The Way A Prison System Works Moving the California Department of Corrections and Rehabilitation from an Offense- Based to Risk-Based System Susan Turner University of California, Irvine September 2010 Presentation Will Cover Size

More information

Justice Reinvestment in New Hampshire

Justice Reinvestment in New Hampshire January 2010 Justice Reinvestment in New Hampshire Analyses & Policy Options to Reduce Spending on Corrections & Increase Public Safety Background IN JUNE 2009, GOVERNOR JOHN LYNCH, Supreme Court Chief

More information

How To Know More About Waldo County

How To Know More About Waldo County WALDO COUNTY Justice Data USM Muskie School of Public Service Maine Statistical Analysis Center Waldo County Facts Waldo County, incorporated in 1827, is in mid coast Maine along Penobscot Bay. The county

More information

REDUCING STATEWIDE RECIDIVISM: CHECKLIST FOR STATE REENTRY COORDINATORS

REDUCING STATEWIDE RECIDIVISM: CHECKLIST FOR STATE REENTRY COORDINATORS STATE REENTRY COORDINATORS are uniquely positioned to ensure that the policies and practices that go into a recidivism-reduction initiative are being implemented effectively, both in corrections settings

More information

FACT SHEET. Views from the National Council on Crime and Delinquency. Youth Under Age 18 in the Adult Criminal Justice System. Christopher Hartney

FACT SHEET. Views from the National Council on Crime and Delinquency. Youth Under Age 18 in the Adult Criminal Justice System. Christopher Hartney June 26 FACT SHEET Views from the National Council on Crime and Delinquency Youth Under Age 18 in the Adult Criminal Justice System Christopher Hartney Negative Impacts on Youth Processed in the Adult

More information

The Start of a Criminal Career: Does the Type of Debut Offence Predict Future Offending? Research Report 77. Natalie Owen & Christine Cooper

The Start of a Criminal Career: Does the Type of Debut Offence Predict Future Offending? Research Report 77. Natalie Owen & Christine Cooper The Start of a Criminal Career: Does the Type of Debut Offence Predict Future Offending? Research Report 77 Natalie Owen & Christine Cooper November 2013 Contents Executive Summary... 3 Introduction...

More information

LCCJAB Departmental Presentation, Detailed Review Lebanon County Adult Probation & Parole April 21 st, 2009

LCCJAB Departmental Presentation, Detailed Review Lebanon County Adult Probation & Parole April 21 st, 2009 LCCJAB Departmental Presentation, Detailed Review Lebanon County Adult Probation & Parole April 21 st, 2009 Presenter: Sally A. Barry Lebanon County Chief Adult Probation & Parole Officer General Information:

More information

Ready for Reform? Public Opinion on Criminal Justice in Massachusetts

Ready for Reform? Public Opinion on Criminal Justice in Massachusetts SURVEY BRIEF FEBRUARY 2014 Ready for Reform? Public Opinion on Criminal Justice in Massachusetts The non-partisan MassINC Polling Group recently concluded a public opinion research project on criminal

More information

Proposition 5. Nonviolent Offenders. Sentencing, Parole and Rehabilitation. Statute.

Proposition 5. Nonviolent Offenders. Sentencing, Parole and Rehabilitation. Statute. Proposition 5 Nonviolent Offenders. Sentencing, Parole and Rehabilitation. Statute. SUMMARY This measure (1) expands drug treatment diversion programs for criminal offenders, (2) modifies parole supervision

More information

Over the last several years, the importance of the risk principle has been

Over the last several years, the importance of the risk principle has been Understanding the Risk Principle: How and Why Correctional Interventions Can Harm Low-Risk Offenders Over the last several years, the importance of the risk principle has been well established in many

More information

Colorado Pretrial Assessment Tool (CPAT) FAQs

Colorado Pretrial Assessment Tool (CPAT) FAQs Colorado Pretrial Assessment Tool (CPAT) FAQs Prepared by the Pretrial Justice Institute Updated: October, 2012 What else besides the CPAT itself does a jurisdiction need to do to implement the CPAT and

More information

Security Classification Using the Custody Rating Scale

Security Classification Using the Custody Rating Scale Security Classification Using the Custody Rating Scale Brian A. Grant and Fred Luciani Research Branch Correctional Service of Canada February, 1998 This report is part of a series of 24 research/evaluation

More information

Interagency Council on Intermediate Sanctions Substance Abuse Providers Correctional Program Checklist (CPC) Statistical Information May 21, 2010

Interagency Council on Intermediate Sanctions Substance Abuse Providers Correctional Program Checklist (CPC) Statistical Information May 21, 2010 Interagency Council on Intermediate Sanctions Substance Abuse Providers Correctional Program Checklist (CPC) Statistical Information May 21, 2010 Prepared by Timothy Wong Department of the Attorney General

More information

Probation Semi-Annual Report January-June, 2015

Probation Semi-Annual Report January-June, 2015 Probation Semi-Annual Report January-June, 2015 Nebraska Supreme Court Office of Probation Administration Statement of Purpose This report was completed in compliance with Neb. Rev. Stat. 29-2252.01 (Appended

More information

Adult Plea Negotiation Guidelines

Adult Plea Negotiation Guidelines From the office of the Rice County Attorney: Adult Plea Negotiation Guidelines Revision June, 2004 1. These guidelines apply to any adult felony defendant case prosecuted by this office, which is not disposed

More information

Community Based Corrections Substance Abuse Treatment For the Higher Risk Offender

Community Based Corrections Substance Abuse Treatment For the Higher Risk Offender CHESTER J. CULVER, GOVERNOR PATTY JUDGE, LT. GOVERNOR DEPARTMENT OF CORRECTIONS JOHN R. BALDWIN, DIRECTOR Iowa Department of Corrections Performance Audit Report Community Based Corrections Substance Abuse

More information

The Creation and Validation of the Ohio Risk Assessment System (ORAS)*

The Creation and Validation of the Ohio Risk Assessment System (ORAS)* 16 FEDERAL PROBATION Volume 74 Number 1 The Creation and Validation of the Ohio Risk Assessment System (ORAS)* Edward J. imtessa, University of Cincinnati Richard Lemke, University of West Georgia Matthew

More information

Orange County, Texas Adult Criminal Justice Data Sheet

Orange County, Texas Adult Criminal Justice Data Sheet Orange County, Texas Adult Criminal Justice Data Sheet For more information, contact Dr. Ana Yáñez- Correa at acorrea@texascjc.org, or (512) 587-7010. The Texas Criminal Justice Coalition seeks the implementation

More information

The Erie County Drug Court: Outcome Evaluation Findings

The Erie County Drug Court: Outcome Evaluation Findings The Erie County Drug Court: Outcome Evaluation Findings February 2001 By: Shelley Johnson Listwan, M.S. Co-Project Director Deborah Koetzle Shaffer, M.A. Co-Project Director and Edward J. Latessa, Ph.D.

More information

Three Year Recidivism Tracking of Offenders Participating in Substance Abuse Treatment Programs

Three Year Recidivism Tracking of Offenders Participating in Substance Abuse Treatment Programs Three Year Recidivism Tracking of Offenders Participating in Substance Abuse Treatment Programs Prepared for the 76 th Texas Legislature, 1999 Criminal Justice Policy Council Tony Fabelo, Ph.D. Executive

More information

Collection of Data on Juvenile Justice Ms. Creem and Mr. Tarr moved that the bill be amended by inserting, after section, the following new section:-

Collection of Data on Juvenile Justice Ms. Creem and Mr. Tarr moved that the bill be amended by inserting, after section, the following new section:- Budget Amendment ID: FY2014-S3-90 Collection of Data on Juvenile Justice Ms. Creem and Mr. Tarr moved that the bill be amended by inserting, after section, the following new section:- JUD 90 SECTION. (a)

More information

TEXAS DEPARTMENT OF CRIMINAL JUSTICE PAROLE DIVISION

TEXAS DEPARTMENT OF CRIMINAL JUSTICE PAROLE DIVISION TEXAS DEPARTMENT OF CRIMINAL JUSTICE PAROLE DIVISION NUMBER: PD/POP-3.6.4 DATE: 12/10/13 POLICY AND OPERATING PROCEDURE PAGE: 1 of 15 SUPERSEDES: 09/01/11 SUBJECT: SEX OFFENDER REGISTRATION PROGRAM AUTHORITY:

More information

The Second Chance Act Frequently Asked Questions

The Second Chance Act Frequently Asked Questions The Second Chance Act Frequently Asked Questions What does the Second Chance Act do? The Second Chance Act primarily authorizes federal funding for state and federal reentry programs. It also directs but

More information

January 2014 Report No. 14-02

January 2014 Report No. 14-02 January 2014 Report No. 14-02 Expansion Drug Courts Can Produce Positive Outcomes Through Prison Diversion and Reduced Recidivism at a glance Overall, the eight expansion drug courts appear to divert non-violent

More information

An Overview Of The Texas Youth Commission s Specialized Treatment Programs

An Overview Of The Texas Youth Commission s Specialized Treatment Programs An Overview Of The Texas Youth Commission s Specialized Treatment Programs Criminal Justice Policy Council Prepared for the 76 th Texas Legislature Tony Fabelo, Ph.D. Executive Director An Overview Of

More information

Revised 4/15/03 th. Highlights. 68% of State prison inmates did not receive a high school diploma. and 53% of Hispanics

Revised 4/15/03 th. Highlights. 68% of State prison inmates did not receive a high school diploma. and 53% of Hispanics U.S. Department of Justice Office of Justice Programs Revised 4/15/03 th Bureau of Justice Statistics Special Report January 2003, NCJ 195670 Education and Correctional Populations By Caroline Wolf Harlow,

More information

DEVELOPING A NATIONAL MODEL FOR PRETRIAL RISK ASSESSMENT

DEVELOPING A NATIONAL MODEL FOR PRETRIAL RISK ASSESSMENT NOVEMBER 2013 RESEARCH SUMMARY DEVELOPING A NATIONAL MODEL FOR PRETRIAL RISK ASSESSMENT Every day in America, judges have to answer a critical question again and again: What are the chances that a recently

More information

Educational Achievement of Inmates in the Windham School District

Educational Achievement of Inmates in the Windham School District Educational Achievement of Inmates in the Windham School District Criminal Justice Policy Council April 2000 Tony Fabelo, Ph.D. Executive Director Educational Achievement of Inmates in the Windham School

More information

It s all apples and oranges. January 31, 2012 Nathan Brady OLRGC

It s all apples and oranges. January 31, 2012 Nathan Brady OLRGC It s all apples and oranges. January 31, 2012 Nathan Brady OLRGC What is recidivism and what is the impact on the state? How does Utah compare nationally? What is Utah doing to address inmate recidivism

More information

Program Guidelines and Procedures for Adult Transitional Case Management

Program Guidelines and Procedures for Adult Transitional Case Management Texas Department of Criminal Justice TCOOMMI Program Guidelines and Procedures for Adult Transitional Case Management Number: PGP 01.07 September Date: 1, 2013 Page: 1 of 5 Supersedes: January 3, 2011

More information

Alameda County Probation Department A Look into Probation Monthly Report July 2013

Alameda County Probation Department A Look into Probation Monthly Report July 2013 Alameda County Probation Department A Look into Probation Monthly Report July 2013 Chief LaDonna M. Harris Chief Probation Officer 400 Broadway Oakland, California 94607 510-268-7233 1 Executive Summary

More information

Criminal Justice 101. The Criminal Justice System in Colorado and the Impact on Individuals with Mental Illness. April 2009

Criminal Justice 101. The Criminal Justice System in Colorado and the Impact on Individuals with Mental Illness. April 2009 Criminal Justice 101 The Criminal Justice System in Colorado and the Impact on Individuals with Mental Illness April 2009 Acronyms DOC = Department of Corrections DYC = Division of Youth Corrections DCJ

More information

Services to At-Risk Youth (STAR) Program Evaluation

Services to At-Risk Youth (STAR) Program Evaluation Services to At-Risk Youth (STAR) Program Evaluation Criminal Justice Policy Council March 2003 Tony Fabelo, Ph.D. Executive Director Services to At-Risk Youth (STAR) Program Evaluation To view or download

More information

DUI Arrests, BAC at the Time of Arrest and Offender Assessment Test Results for Alcohol Problems

DUI Arrests, BAC at the Time of Arrest and Offender Assessment Test Results for Alcohol Problems DUI Arrests, BAC at the Time of Arrest and Offender Assessment Test Results for Alcohol Problems 2001 Donald D. Davignon, Ph.D. Abstract Many DUI/DWI offenders have drinking problems. To further reduce

More information

Criminal Justice Professionals Attitudes Towards Offenders: Assessing the Link between Global Orientations and Specific Attributions

Criminal Justice Professionals Attitudes Towards Offenders: Assessing the Link between Global Orientations and Specific Attributions Criminal Justice Professionals Attitudes Towards s: Assessing the Link between Global Orientations and Specific Attributions Prepared by: Dale Willits, M.A. Lisa Broidy, Ph.D. Christopher Lyons, Ph.D.

More information

O H I O DRUG COURT EVALUATION

O H I O DRUG COURT EVALUATION O H I O DRUG COURT EVALUATION Bob Taft, Governor Maureen O Connor, Lt. Governor Domingo S. Herraiz, Director Ohio Office of Criminal Justice Services O H I O DRUG COURT EVALUATION Bob Taft, Governor Maureen

More information

LANCASTER COUNTY ADULT DRUG COURT

LANCASTER COUNTY ADULT DRUG COURT LANCASTER COUNTY ADULT DRUG COURT Administered by the Lancaster County Department of Community Corrections Judicial Oversight by the Lancaster County District Court www.lancaster.ne.gov keyword: drug court

More information

PROPOSAL. Expansion of Drug Treatment Diversion Programs. December 18, 2007

PROPOSAL. Expansion of Drug Treatment Diversion Programs. December 18, 2007 December 18, 2007 Hon. Edmund G. Brown Jr. Attorney General 1300 I Street, 17 th Floor Sacramento, California 95814 Attention: Ms. Krystal Paris Initiative Coordinator Dear Attorney General Brown: Pursuant

More information

Virginia State Crime Commission s Sex Offender Task Force

Virginia State Crime Commission s Sex Offender Task Force Virginia State Crime Commission s Sex Offender Task Force Formation of Virginia s Sex Offender Task Force Heinous crimes against children committed by sex offenders in other states have been well publicized

More information

Long-term Impact Evaluation of Specialized Sex Offender Probation Programs In Lake, DuPage and Winnebago Counties

Long-term Impact Evaluation of Specialized Sex Offender Probation Programs In Lake, DuPage and Winnebago Counties Long-term Impact Evaluation of Specialized Sex Offender Probation Programs In Lake, DuPage and Winnebago Counties Prepared for the The Illinois Criminal Justice Information Authority By: Loretta J. Stalans,

More information

Corrections Rehabilitative Programs Effective, But Serve Only a Portion of the Eligible Population

Corrections Rehabilitative Programs Effective, But Serve Only a Portion of the Eligible Population February 2007 Report No. 07-14 Corrections Rehabilitative Programs Effective, But Serve Only a Portion of the Eligible Population at a glance The department s rehabilitative programs serve only a small

More information

Using a Proxy Score to Pre-screen Offenders for Risk to Reoffend February 1, 2006 Brad Bogue, William Woodward, Lore Joplin

Using a Proxy Score to Pre-screen Offenders for Risk to Reoffend February 1, 2006 Brad Bogue, William Woodward, Lore Joplin Using a Proxy Score to Pre-screen Offenders for Risk to Reoffend February 1, 2006 Brad Bogue, William Woodward, Lore Joplin Background: Evidence-based practice requires that offenders who are of higher

More information

RHODE ISLAND SEX-OFFENDER REGISTRATION AND NOTIFICATION

RHODE ISLAND SEX-OFFENDER REGISTRATION AND NOTIFICATION RHODE ISLAND SEX-OFFENDER REGISTRATION AND NOTIFICATION CONTACT INFORMATION Rhode Island State Police Bureau of Criminal Investigation 150 South Main Street Providence, RI 02903-2907 Telephone: 401-421-5268

More information

New Mexico Sentencing Commission

New Mexico Sentencing Commission New Mexico Sentencing Commission Dan Cathey and Chris Miller November 2007 Summary The New Mexico Corrections Department is the only state agency in our sample handling both probation and parole services.

More information

County of San Diego SB 618 Reentry Program. May 3, 2007

County of San Diego SB 618 Reentry Program. May 3, 2007 County of San Diego SB 618 Reentry Program May 3, 2007 1 California Recidivism - The Highest Return to Prison Rate in the Nation In FY 2006-07 it is estimated that San Diego County will convict over 16,000

More information

Crime and Justice in Colorado 2006

Crime and Justice in Colorado 2006 Crime and Justice in Colorado 2006 June 2007 Prepared by: Office of Research and Statistics Kerry Lowden, Report Project Manager Kim English, Research Director Linda Harrison, Senior Statistical Analyst

More information

Options for Policymakers Considering a Justice Reinvestment Initiative in Louisiana

Options for Policymakers Considering a Justice Reinvestment Initiative in Louisiana Options for Policymakers Considering a Justice Reinvestment Initiative in Louisiana Submitted to: National Center for Urban Communities at Tulane and Xavier Universities Council of State Governments Criminal

More information

Department of Community and Human Services Mental Health, Chemical Abuse and Dependency Services Division

Department of Community and Human Services Mental Health, Chemical Abuse and Dependency Services Division Criminal Justice Initiative Community Center for Alternative Programs Intensive Outpatient Chemical Dependency Treatment Program Two Year Outcomes Subsequent to Program Changes Department of Community

More information

Nebraska s Youth Rehabilitation. and Treatment Centers. Nebraska YRTCs Issue Brief. A Publication of. www.voicesforchildren.com

Nebraska s Youth Rehabilitation. and Treatment Centers. Nebraska YRTCs Issue Brief. A Publication of. www.voicesforchildren.com Nebraska YRTCs Issue Brief Nebraska s Youth Rehabilitation and Treatment Centers Every year, Nebraska s courts send a number of serious and not-so-serious juvenile offenders to the Youth Residential Treatment

More information

Responsible for prosecuting all criminal and traffic cases within Mecklenburg County

Responsible for prosecuting all criminal and traffic cases within Mecklenburg County Memo from District Attorney Andrew Murray on Impacts Related to Reductions or Elimination of City funded services to the District Attorney s Office and Court System At the May 13, 2015 Budget Adjustments

More information

SENTENCING REFORM FOR NONVIOLENT OFFENSES: BENEFITS AND ESTIMATED SAVINGS FOR ILLINOIS

SENTENCING REFORM FOR NONVIOLENT OFFENSES: BENEFITS AND ESTIMATED SAVINGS FOR ILLINOIS SENTENCING REFORM FOR NONVIOLENT OFFENSES: BENEFITS AND ESTIMATED SAVINGS FOR ILLINOIS LISE MCKEAN, PH.D. SUSAN K. SHAPIRO CENTER FOR IMPACT RESEARCH OCTOBER 2004 ACKNOWLEDGEMENTS PROJECT FUNDER Chicago

More information

When incarceration rates increase 10%, research shows that crime rates

When incarceration rates increase 10%, research shows that crime rates Evidence-Based Public Policy Options to Reduce Criminal Justice Costs and Crime Rates by Assistant Director Washington State Institute for Public Policy When incarceration rates increase 10%, research

More information

Redirection as Effective as Residential Delinquency Programs, Achieved Substantial Cost Avoidance

Redirection as Effective as Residential Delinquency Programs, Achieved Substantial Cost Avoidance March 2006 Report No. 06-34 Redirection as Effective as Residential Delinquency Programs, Achieved Substantial Cost Avoidance at a glance Although it experienced several start-up challenges, the Redirection

More information

Fourth Judicial District of Minnesota Pretrial Evaluation: Scale Validation Study

Fourth Judicial District of Minnesota Pretrial Evaluation: Scale Validation Study Fourth Judicial District of Minnesota Pretrial Evaluation: Scale Validation Study Fourth Judicial District Research Division Marcy R. Podkopacz, Ph.D., Research Director October, 2006 www.mncourts.gov/district/4

More information