Search results

The overall goal of this research project was to examine the impact that risk assessment has on the sentencing of racial minority youth and youth from disadvantaged neighborhoods. One of the main goals of correctional risk assessment is to reduce disparate outcomes for certain groups of youth in the juvenile justice system (e.g. Black youth). In practice, risk assessment is used with this goal in mind. However, there is very little research which shows whether or not risk assessment actually has its intended effects on sentencing. Therefore, this study set out to examine whether or not risk assessment reduces the sentencing gap seen in most research for minority youth and youth from disadvantaged neighborhoods. In addition, several other important research topics were explored to understand the role that race and socioeconomic disadvantage play in the juvenile justice system. These research topics included: (1) variation in the predictive validity of risk assessment across race, (2) variation in the predictive validity of risk assessment across neighborhood disadvantage, and (3) the moderating effects of race/gender and court dispositions on the predictive validity of risk assessment.

To achieve the research goals in this study, data was collected from a large juvenile court in a Midwestern County. Information from 4,383 youth that came into contact with the court between January 2010 and December 2016 were included in the study. Data was collected that related to youth demographics, neighborhood characteristics in which youth lived, risk assessments data measured by the Ohio Youth Assessment System (OYAS), treatment programming received, court dispositions/sentencing, and recidivism.

Teen dating violence (TDV) is an increasing criminal justice and public health issue. Due to these startling trends, federal agencies have called for increased prevention programming for this pattern of violence. In response, the proposed study tests whether a trauma-informed risk algorithm, based on modifiable risk and protective factors, can properly forecast TDV-perpetration risk in diverse adolescent samples.

High school students in Texas (N=507) and adolescents participating in a juvenile diversion program in Illinois (N=77) completed surveys at baseline, 6-months, and 12-months. Surveys measured TDV-perpetration and other violent behavior, childhood adversities, individual risk factors for TDV-perpetration, and interpersonal and culturally-relevant protective factors. A novel analytic plan involving latent growth curve modeling and evidence-based medicine (EBM) was used to contextualize each risk factor's relation to TDV-perpetration and how these findings into an empirically-based decision-making tool.

To address the relative difficulty in predicting domestic violence incidents and effectively targeting resources, the University of Chicago Crime Lab and the New York Police Department (NYPD) collaborated to develop and test a machine learning-based statistical model to predict the risk of domestic violence victimization in New York City.

Phase 1 of the project was to develop a statistical model using machine learning techniques. NYPD administrative records dated between January 2006 and January 2017 were used as input data to build and refine the tool. Due to the lack of unique identifiers for victims in the records, the research team also used data from the Chicago Police Department to create a probabilistic record linkage toolkit (Name Match) to identify which records belonged to the same person within and across data sources.

In Phase 2, the researchers aimed to field test the tool's capability to identify individuals at risk of repeated domestic violence through a large-scale randomized control trial. Measuring the effects of regular home visits of high-priority individuals thought to be at risk of serious domestic assault, the test intended to compare the selections of individuals made by officers versus those predicted by the tool.

This collection contains only the machine learning code files (R and Python) created during secondary analysis, which have been released as a zipped package. Please refer to the Data Roadmap for instructions on how to obtain the original NYPD data. To access the Name Change algorithm and documentation, please visit the Github repository.

The Long-Term Risk-Need-Responsivity (RNR) Impact Study was a pre-post, quasi-experimental study of the impact of the implementation of risk-needs assessment (RNA) and risk-need-responsivity-related (RNR) case management in 5 juvenile probation offices in 2 states. This study used 3 time points (pre-implementation, 1st year post-implementation, and 7th-year post-implementation) to examine the 7-year sustainability of impacts on system-responses (rates of informal processing, different dispositions, and out-of-home placements), youth outcomes (school and employment), and recidivism; as well as cost-effectiveness.

This study also examined whether there was a significant difference in the impacts of implementation after 7 years between probation offices that were effective versus ineffective in their first year of implementation.

The Center for Court Innovation, in partnership with independent consultants from the public defender's office of certain tribes, conducted a comprehensive survey of tribal justice system stakeholders, focused on the existing use of risk-needs assessments and similar tools, and existing data collection/technology used by tribal jurisdictions around the country. The survey results create a comprehensive portrait of tribal court system risk and need assessment, data collection, management, and challenges reported by those directly involved in managing and working with people in the system.

This study examines two counties in Oregon (Multnomah and Yamhill) that have utilized a pre-adjudication risk assessment (PAA) to inform criminal case negotiation since 2014. This study seeks to answer these core questions:

• has the introduction of a PAA into the court decision-making process impacted racial and ethnic disparities in sentencing outcomes over time in each county;
• does one PAA process appear to produce more promising results in impacting racial/ethnic disparities;
• using interviews and visual observations, how does the PAA influence case discussion and negotiation, decision-making, and workgroup norms and culture;
• does validation of the PAA tool yield significant mean score differences across racial groups and/or predictive biases?

A number of jurisdictions have turned to pre-adjudication risk assessments (PAA) as a tool to potentially lower or stabilize incarceration rates by identifying the best suitable cases for community-based supervision. Questions have been raised about using risk assessment tools to help with the negotiation of sentencing outcomes in the pre-adjudication stages of criminal cases, particularly the potential for exacerbating disparate racial/ethnic sentencing outcomes.

Because there are few existing tools for assessing the risk of recidivism for youth with a history of sexual offending that are empirically valid and reliable, knowledge and practice in this area has historically been limited. This project examined current practice and policy in the assessment, treatment, and management of juvenile sex offenders across multiple jurisdictions (Florida, New York, Oregon, Pennsylvania, and Virginia). The researchers developed a prototype assessment tool, state-specific risk assessment models, and practical guidance for building a risk assessment for sexual recidivism in juvenile justice settings.

The data file contains individual records for the full sample (n=8,035), including their risk predictors, recidivism measures, and resulting outputs (i.e., predicted probabilities of sexual recidivism) from the risk models.

This study seeks to examine important components of our nation's sex offender tracking and monitoring systems, with a focus on risk assessment and sexual recidivism (measured by re-arrest). Data were collected from 1,789 adult sex offenders in the following states.

• Florida: 500 cases
• Minnesota: 500 cases
• New Jersey: 291 cases
• South Carolina: 498 cases

The data file contains another 551 cases from the state of Massachusetts. However, due to how and when those cases were identified they were not included in the Principal Investigator's focus and analysis. There are also another 151 cases where a study participant's state is missing. Total there are 2,491 cases and 1,947 variables.

The Interpersonal Violence and Institutional Misconduct in Jails Study is a longitudinal evaluation of administrative data collected from the Los Angeles County Jail System. This study includes aggregate monthly information on the number and rate of incidents of interpersonal violence and serious institutional misconduct in the Los Angeles County Jail System over an eight-year time period (January 2010 to December 2017). This investigation also includes information on the development and validation of two separate risk assessment tools--the Inmate Risk Assessment for Perpetration (IRAP) and the Inmate Risk Assessment for Victimization (IRAV)--that were designed to help authorities proactively identify the perpetrators and victims of interpersonal violence in jail, respectively. The subjects used to construct and test these instruments were an admission cohort of all adjudicated inmates entering the Los Angeles County jail system in 2016 (N = 104,919). This population of inmates was randomly assigned into one of four groups. The first was the construction sample (n = 26,404), which was used to create the two risk assessment scales, and the other three served as cross-validation samples, which each served to evaluate the predictive accuracy and reliability of these instruments. These data include individual-level information on inmate demographics, criminal history, and other measures of institutional behavior.

This study examines the radicalization of Western women to extremist violence, both through the creation of a moral-situational-action risk model and the examination of their responses to various types of online propaganda. The Moral-Situational-Action Risk Model for Extremist Violence (MSA-RMEV) was developed using situational action theory from criminology and violence risk practice literature. The MSA-RMEV revolves around three domains reflective of propensity, mobilization, and capacity building, geared towards providing a violence risk assessment that can assist the intelligence community in preventing future acts of violence.

A sample of women who self-identified as conservative, liberal, and Muslim were exposed to jihadist, alt-right, and alt-left online propaganda. Physiological responses and self-report assessments were recorded. Eye-gaze, pupil dilation, galvanic skin response, heart rate, and facial emotions were documented, along with women's judgment of their emotional, cognitive, and arousal states, while viewing propaganda.

Based on their results, women were categorized as high-risk, medium-risk, or low-risk for violence. Additionally, numerous variables were created to identify participant's beliefs and behavior related to radicalization. Beliefs included religiosity, political affiliation, the presence of moral emotions, sacred values, developmental maturity, and militant thinking. Behaviors included group affiliations, extent of involvement in extremist activities, and presence on social media platforms such as Facebook, Reddit, and Twitter. Demographic variables such as age, marital status, number of children, race, ethnicity, country of origin, and educational status were included.

This study was a multi-site, quasi-experimental investigation of risk assessment-informed pretrial release decisions relative to decisions made without risk assessments. The study objectives were to investigate whether use of pretrial risk assessments: 1) improved risk management decisions relative to practice as usual; 2) improved the fairness of pretrial release decisions for racial minorities relative to practice as usual; and 3) revealed judge-level variation in pretrial release decisions. Secondary, administrative data were gathered from four counties in Indiana. Apart from risk assessment information, all data represent public records: court, jail and prison. Variables gathered from the data sources fall into three main categories: defendant variables, pretrial release decisions, and pretrial misconduct outcomes. Examples of specific variables include: booking and release dates, bond amounts, number and type of charges, and demographic information (age, sex and race).

To better understand the factors that may help or hinder effective implementation of risk and needs assessments, the Office of Juvenile Justice and Delinquency Prevention (OJJDP) in 2015 launched a study led by the University of Cincinnati to examine the Ohio Youth Assessment System (OYAS) in three states-Arizona, Indiana, and Ohio. The OYAS tools measure risk and needs to inform court decisions as well as the delivery of services and treatment. Researchers collected data through a variety of methods, including in-person interviews with personnel at juvenile justice agencies; web-based surveys; comprehensive case record data for youths assessed by the OYAS; and phone interviews of youth following their case disposition and/or facility release.

One of the study's key objectives was to understand how staff perceived the implementation of the OYAS. To assess the impact of the OYAS on youth outcomes--including reduced recidivism and other prosocial behavior and attitudes--the research team analyzed juvenile justice case records and conducted followup interviews with youth. Demographic variables included in this study include the race, gender, age, ethnicity, and level of education of participants.

This research uses in-depth cases studies in five counties, combining interviews and observations, interviews with state reform leaders, and a statewide survey, to examine juvenile probation officers' use of the Youth Level of Service/Case Management Inventory (YLS/CMI) in local practice. In particular, the study assessed the form YLS/CMI policies take in local Pennsylvania counties and their consistency with the Risk-Need-Responsivity (RNR) model, the extent to which officers' routine practices align with policies and RNR principles, and how state and county factors have shaped local policies and practices. In doing so, it sought to highlight promising strategies for effective RNA implementation. Study findings describe an example of an apparently well-planned statewide effort to implement and support the YLS/CMI in local counties, involving a sustained attention to attention to training, quality assurance, and problem-solving. Local county policies, while showing some variation, focused on strategies for conducting the YLS/CMI assessment and applying its results to a variety of decisions and activities--supporting, in particular, risk and need principles. Study results further suggest a substantial statewide level of practitioner adherence to a general policy model, though this pattern was stronger for some activities than others, and varied substantially by local county. Variation in implementation across counties seemed to relate, in particular, to the extent of local quality assurance processes, leadership enthusiasm for the YLS/CMI, staff enthusiasm for evidence-based practices (with implications for training and recruitment), and organizational climate.

The purpose of the Probation/Parole Officer Interactions with Women Offenders project was to better understand the nature of probation/parole agent interactions with substance-involved women offenders and the degree to which these interactions predict women's recidivism, rules violation, and changes in their crime-related needs. Other purposes were to examine whether (1) women's feelings of self-efficacy, anxiety, and reactance that may be elicited by the agent-offender interactions explained (i.e., mediated) the link between the nature of these interactions and women's outcomes; (2) the amount of communication about needs predicted outcomes; (3) initial levels of needs and quality of women's social networks moderated the relationship between agent-offender interactions and outcomes; (4) agents' and women's experience of their relationship were similar, and whose perspective (agent or offender) best predicted offender outcomes; and (5) particular agent attributes (e.g., years of experience, self-reported typical style of interacting with clients) predicted agents' average effectiveness across clients.

Seventy-three probation and parole agents with women-specific caseloads were sampled and recruited from 16 Michigan counties; and then 402 offenders meeting participant criteria were recruited from the caseloads. At recruitment of agents, a survey collected information about their characteristics. Two or three months later, client needs and social networks were measured in an interview with the offenders. At month 5, agent and woman reports of relationship and communication were measured, and woman reports of reactions to these were measured. A survey was used to collect agent data, and an interview was used to collect offender data. At month 8, psychosocial outcomes, treatment engagement, crime-causing needs, and official outcomes (e.g., arrest) were measured. For the 24 months from the start of supervision, official data were collected on arrests and convictions.

These data are part of NACJD's Fast Track Release and are distributed as they were received from the data depositor. The files have been zipped by NACJD for release, but not checked or processed except for the removal of direct identifiers. Users should refer to the accompanying readme file for a brief description of the files available with this collection and consult the investigator(s) if further information is needed.

This study sought to: (1) document the spectrum of injuries and injury characteristics observed among physically-abused older adults reported to Adult Protective Services (APS) and compare those findings to injuries found among non-abused older adults, (2) identify observable injury characteristics and abuse circumstances that healthcare providers, law enforcement and prosecutors consider to be key forensic markers of physical abuse, (3) document information and evidence integral for achieving successful criminal prosecution, and (4) describe approaches that community-based frontline workers can employ to better document evidence of physical abuse.

The data included in this collection were collected under a National Institute of Justice-funded project that sought to document the spectrum and characteristics of injuries observed among physically abused, community-dwelling APS clients.

The collection includes two SAS datasets: injury.sas7bdat (with 47 variables and 403 cases) and subjectleveldata.sas7bdat (with 122 variables and 165 cases); and three SAS System Program Files: analysis-code-v1.sas, cleaned-injury-datasets-v1.sas, and formats.sas. Demographic variables in the collection are in the subjectleveldata.sas7bdat dataset, and include age, year of birth, gender, race, language, and level of education.

The Sri Lankan Environmental and Agricultural Decision-making Survey (SEADS) collected quantitative data from paddy farming households in 24 pre-selected villages in the dry zone of Sri Lanka. These data include information about livelihoods, economic activity, household characteristics, cultivation, and experiences with water scarcity and environmental stress. The objective of SEADS was to collect high quality data that could be used to:

• Understand the impacts of water scarcity on farming households throughout the dry zone, and to assess vulnerability to climate change impacts.
• Document methods of coping with water scarcity utilized by farmers and communities throughout the dry zone.
• Estimate the rate of adoption of agricultural adaptations promoted by governmental and non-governmental organizations and community leaders (e.g., parachute method, low flood depth irrigation, short duration seed varieties.)
• Understand what programs or policies may help support farmers to cope with water scarcity in the future.

Demographic variables collected include age, gender, religion, ethnicity, district of origin, education level, and occupation.

These data are part of NACJD's Fast Track Release and are distributed as they were received from the data depositor. The files have been zipped by NACJD for release, but not checked or processed except for the removal of direct identifiers. Users should refer to the accompanying readme file for a brief description of the files available with this collection and consult the investigator(s) if further information is needed.

This study examined the validity, reliability, equity and cost of nine juvenile justice risk assessment instruments. It was designed to provide a comprehensive examination of how several risk assessments perform in practice. This study posed the following questions:

1. Is each risk assessment instrument sufficiently reliable (i.e., inter-rater reliability) to ensure that decisions regarding level of risk and identified service needs are consistent across the organization?
2. What specific risk assessment items are associated with less reliability? What items are rated reliably by staff?
3. Is each risk assessment instrument valid? Specifically, what degree of discrimination is attained between assigned risk levels? Could the instrument be improved by adding or deleting specific factors and/or altering cut-off scores?
4. Is each risk assessment instrument valid for population subgroups: White/Caucasian, Black/African American, Hispanic/Latino, females, probationers, and youth in aftercare status? Could equity be improved by adding or deleting specific factors or altering cut-off scores?
5. What costs are associated with each risk assessment instrument?

The study collection includes 31 SPSS data files

1. all_jais_combined.sav (n=1,141; 6 variables)
2. ar_fire_final_file_ojjdp-ICPSR.sav (n=119; 205 variables)
3. AR_yls_irr_FINAL-ICPSR.sav (n=155; 136 variables)
4. azaoc_FINALFILE-ICPSR.sav (n=7,589; 438 variables)
5. AZAOC_irr_FINAL-ICPSR.sav (n=458; 101 variables)
6. AZDJC_FINAL_FILE-ICPSR.sav (n=1,265; 1,290 variables)
7. AZDJC_irr_FINAL-ICPSR.sav (n=55; 120 variables)
8. COMMITMENT_FINAL_SAMPLE2-ICPSR.sav (n=11,154; 719 variables)
9. FinalDJJReleasesWithRecid_BothYears2-ICPSR.sav (n=90,818; 31 variables)
10. FIRE_NE_COMM_FINAL_FILE_OJJDP-ICPSR.sav (n=597; 174 variables)
11. fire_ne_probation_final-ICPSR.sav (n=1,077; 237 variables)
12. FL_irr_FINAL-ICPSR.sav (n=519; 140 variables)
13. GA_irr_FINAL-ICPSR.sav (n=509; 263 variables)
14. gafire_boyscommunityALL_FINAL-ICPSR.sav (n=5,009; 781 variables)
15. gafire_communityALLforretrofit2-ICPSR.sav (n=6,943; 666 variables)
16. gafire_finalsampforanalysis_all-ICPSR.sav (n=7,412; 642 variables)
17. gafire_finalsampforanalysis_girls-ICPSR.sav (n=2,005; 768 variables)
18. jais_boys_wk_1-ICPSR.sav (n=1,989; 484 variables)
19. jais_girls_wk_1-ICPSR.sav (n=745; 484 variables)
20. NE_irr_FINAL-ICPSR.sav (n=727; 160 variables)
21. OR_irr_FINAL-ICPSR.sav (n=477; 112 variables)
22. ORE_FIRE_final-ICPSR.sav (n=12,370; 340 variables)
23. PROBATION_FINAL_BOYS_ALL-ICPSR.sav (n=20,621; 837 variables)
24. PROBATION_FINAL_GIRLS_ALL-ICPSR.sav (n=6,748; 849 variables)
25. va_boyssample-ICPSR.sav (n=1,106; 1,273 variables)
26. va_final_sample_fullscreen-ICPSR.sav (n=1,439; 1,237 variables)
27. va_girlssample-ICPSR.sav (n=333; 1,256 variables)
28. VA_irr_expert_FINAL-ICPSR.sav (n=10; 308 variables)
29. VA_irr_worker_FINAL-ICPSR.sav (n=685; 308 variables)
30. vafinalsample-ICPSR.sav (n=1,919; 1,200 variables)
31. workersurveyfinal-ICPSR.sav (n=400; 69 variables)

These data are part of NACJD's Fast Track Release and are distributed as they were received from the data depositor. The files have been zipped by NACJD for release, but not checked or processed except for the removal of direct identifiers. Users should refer to the accompanying readme file for a brief description of the files available with this collection and consult the investigator(s) if further information is needed.

This study examines criminal background checks that were carried out by the New York State Department of Health (DOH) for individuals with criminal history records who have been provisionally hired to work as unlicensed direct care employees in the long term care industry. These individuals applied for jobs in this industry for the first time in 2008 and 2009.

This information was then augmented with criminal history information provided by the NY Division of Criminal Justice Services (DCJS).

The collection contains 1 Stata data file (Criminal_background_check.dta (n=7209; 40 variables)).

These data are part of NACJD's Fast Track Release and are distributed as they were received from the data depositor. The files have been zipped by NACJD for release, but not checked or processed except for the removal of direct identifiers. Users should refer to the accompanying readme file for a brief description of the files available with this collection and consult the investigator(s) if further information is needed.

Using multiple data time points this study investigated the prospective validity of a secondary gang prevention program called Gang Risk of Entry Factors (GREF) assessment. At Time 1 of the study interview cut-points were established for high and low risk on nine risk factors that were included on the assessment. Those who scored high risk on four or more risk factors were determined eligible for secondary prevention. At time 2 each participate was then classified into one of four levels of gang membership. The goal of this was to investigate how successful the GREF was in identifying the youth (in the absence of a program) who become associated with a street gang in the 12 to 18 months of the study time frame .

The Oklahoma School Readiness Reach-by-Risk study builds on the development of the Oklahoma School Readiness Risk Index (SRRI) in 2011 by including data on the scope of early childhood programs in each of the state's 77 counties. This research project seeks to analyze the prevalence of particular socio-demographic indicators of school readiness, as well as promote informed policy funding decisions related to early childhood education. The purpose of this study is to highlight counties where the risk for starting school unprepared to learn is high, yet opportunities for quality early childhood programs and services are low.

Within this study the researchers focused on prekindergarten students (with Hispanic background) who displayed signs of being unprepared to learn by examining two social constructs: risk and reach (identified by early childhood programs). Risk measures were assessed through a risk index comprised of 11 socioeconomic and demographic indicators found by empirical research to increase a child's risk of being unprepared for school.

The Reach measures were created by gathering data from early education programs, the state's universal prekindergarten program, early childhood home visitation programs, and child care services. Early childhood program reach was determined by analyzing the county-level service density of early childhood education and home visitation programs in addition to several aspects of child care services, such as provider quality ratings and enrollment of children with child care subsidies in quality facilities.

Police Departments' Use of Lethality Assessments: An Experimental Evaluation examined the effectiveness of the Lethality Assessment Protocol (LAP), a tool used to gauge the severity of danger to victims of intimate partner violence (IPV) and determine whether to immediately connect victims with additional resources and safety options. Specifically, the evaluation focused on the effectiveness of the LAP at decreasing the rates of repeat, lethal and near lethal violence and increasing the rates of emergency safety planning and help seeking among women who experienced IPV and called the police. Additionally, the predictive and concurrent validity of the screening portion of the LAP was evaluated, as were the implementation of the LAP by officers and IPV victims' satisfaction with the police responses they experienced.

The study consisted of two groups: (1) a comparison group, which included women who were victims of IPV and were referred to the study by a police officer; and (2) an intervention group which consisted of victims of IPV who were administered the LAP by police. Both groups were contacted for baseline and follow-up phone interview surveys that recorded the victims' self-reported demographic information (age, race, income, education marital status), information about the status of their relationships with their partners, as well as the type of abuse they had endured and how this affected their behavior.

Since 2001, Vermont Department of Corrections (DOC) sex offender treatment providers and probation and parole officers have scored every adult male sex offender under community supervision on three measures of static (unchangeable) risk at intake (i.e., Static-99, RRASOR-Rapid Risk Assessment for Sex Offence Recidivism, and VASOR-Vermont Assessment of Sex Offender Risk) and one measure of dynamic (changeable) risk (i.e, SOTNPS) at intake and then every six months thereafter. This project conducted record reviews to verify the accuracy of the risk assessment scores and examined how scores on the Sex Offender Treatment Needs and Progress Scale (SOTNPS) and one or more of these static risk instruments can be combined into an overall model of risk assessment. An empirically derived decision-making model was created to assist correctional administrators, probation and parole officers, and treatment providers in allocating and delivering supervision and treatment services based on an individual's treatment needs and risk to sexually re-offend.

Three hypotheses were tested. First, it was expected that one or more static risk measures (Static-99R, Static-2002R and VASOR) would predict sexual recidivism with moderate accuracy in the sample. Second, a dynamic risk measure, Sex Offender Treatment Needs and Progress Scale (SOTNPS), or a subset of risk factors contained in this measure would also predict sexual recidivism with moderate accuracy and be sensitive to the changes in dynamic risk over time. Third, a combined static and dynamic risk measure would predict sexual recidivism more accurate than either measure alone.