A Deceptively Simple Measure in a Complex and Contentious System
by Sumita (Mira) Das and Jeffrey A. Butts
John Jay College Research and Evaluation Center
June 2026
Introduction
Recidivism is the default measure of effectiveness for interventions operated or sponsored by criminal justice agencies. It’s a deceptively straightforward premise: individuals who relapse more often and more quickly into old behavior patterns are likely to keep repeating the behavior and cause more harm. Of course, simple ideas often get messy in practice. Researchers and practitioners should address several questions before relying on recidivism as a measure for assessing the effectiveness of justice interventions. Can the outcomes of the criminal justice system be measured accurately using only data curated by the system itself? Are the criminal justice actions that typically serve as recidivism indicators (i.e., rearrest, reincarceration) closely aligned with every intervention program’s theory of change? Is it sufficient to measure outcomes of failure, or should programs track success? Are all sources of failure and success clearly identified and appropriately located?
Recidivism is a partial and quite imperfect outcome measure for guiding the development of public safety policies and practices. The research challenge for any social intervention is to rely on outcomes that meet George Doran’s (1981) SMART criteria — i.e., specific, measurable, achievable, relevant, and time-bound. Researchers in the justice sector could add two more letters to Doran’s acronym, making it SMART(er). Outcome measures in criminal justice must also be equitable and responsive to the cultural, socio-economic, and demographic contexts that affect operations and decision-making in the American justice system. Recidivism is SMART(er) as an outcome measure when those who rely on it consider the longstanding limitations of criminal justice systems and how those limitations shape basic performance metrics.
The Concept of Recidivism
The Merriam-Webster dictionary defines recidivism as the “tendency to relapse into a previous condition or mode of behavior, especially: relapse into criminal behavior.” Does the word “tendency” imply that public safety policies should account for the mere possibility of criminal behavior, rather than only actual behavior? How long should relapses be monitored before it is appropriate to conclude that a “tendency” has abated? Specifying relapses into a “previous condition or mode of behavior” suggests that recidivism occurs not only when individuals behave as they did before, but also when some other aspect of their lives resembles a previous time. If a behavior reappears but causes less harm, does that still qualify as recidivism? If a person once convicted of a violent felony is arrested for a property misdemeanor, is that their same “mode” of behavior? Is it a return to the previous condition? Recidivism is a social construct with multiple sources and causes, not a simple event that can be counted. Recidivism is shaped by many factors and rarely captured as a single event. There is no universal standard.
What are the most trusted and reliable recidivism indicators? Different agencies and stakeholders rely on different measures. Depending on the organization responsible and the stage of criminal justice processing involved, recidivism metrics reflect different behaviors and domains. Law enforcement agencies track rearrests. Probation offices follow revocations. Prison administrators track readmissions. Agencies monitor whatever data they can readily access. The various definitions and methods may be confusing for lawmakers and the public. Knowing how one agency measures recidivism can start a conversation, but not end one (Lai et al. 2022; Maltz 1984; Zara and Farrington 2016).
Measuring Recidivism
Recidivism will likely remain a key feature of policy and practice in the public safety sector. It provides a simple, familiar metric for assessing the effectiveness of crime prevention efforts and criminal justice interventions for individuals who were previously arrested and/or convicted. Pointing out the logical flaws in recidivism will not diminish its salience to audiences long accustomed to seeing recidivism as the key indicator of criminal justice efficacy, but maintaining awareness of those flaws is essential for justice officials, researchers, and lawmakers (Butts and Schiraldi 2018).
To address potential biases in recidivism measures, researchers may ask justice-involved individuals to report their own behavior, either confidentially or anonymously. Self-reported measures ask respondents to acknowledge their criminal offenses and harmful acts, whether or not they are known to justice agencies and government officials. The advantage of such measures is that they capture a wider range of data than the criminal justice system will ever be able to ascertain. The obvious weakness, however, is accuracy. Depending on the stage of their involvement in the justice system and the settings in which such questions are asked, people may not be able to answer accurately or honestly (Li 2026). Researchers have traditionally noted the potential for error in all measures of recidivism, whether self-reported or official, and they advise caution when interpreting recidivism regardless of the source (Griswold 1978; Maltz 1984). While self-report data will always play an important role in research and evaluation, recidivism measures for policy and practice largely depend on official records maintained by the organizations comprising the criminal and youth justice systems.
Many criminal justice organizations create official recidivism measures, including local jails, probation and parole offices, prosecutors and court systems, state-level corrections departments, allied service providers, and various oversight bodies. Common metrics are rearrest, reconviction, and reincarceration. Less common measures include revocations of community supervision (parole or probation). All recidivism measures from official sources have limits. They depend on administrative offices to maintain accurate records linked over time to create event sequences, making them vulnerable to undercounting, underreporting, and various inconsistencies across jurisdictions (Blumstein and Larson 1971; Maltz 1984).
Official metrics vary widely across jurisdictions and agency types. Agencies are especially inconsistent in reporting arrests below some minimum level of crime severity, and states vary in their policies for diverting rather than charging individuals subject to rearrest, including young people and those with mental illnesses (Villeneuve and Quinsey 1995). Corrections and parole agencies typically report annual arrests, convictions, and incarcerations, and lawmakers review these figures as they make funding and policy decisions.
Most Frequently Used Official Measures of Recidivism
Researchers and policymakers must be cautious when responding to recidivism reports, especially when they are used to compare the effectiveness of programs and agencies. Notable differences in law, policy, and procedure, as well as place, time, and population, will affect recidivism comparisons (Maltz 1984). Recidivism rates differ even between well-established federal datasets maintained by the Administrative Office of U.S. Courts, the United States Sentencing Commission, and the Bureau of Justice Statistics in the U.S. Department of Justice (Johnson 2017). Federal agencies monitor recidivism among prison release cohorts at the national level using various measures, including in-state and out-of-state rearrests, reconvictions, and returns to prison for new crimes or technical violations (Durose and Antenangeli 2021; Rosenfeld and Grigg 2022).
State agencies use different practices for reporting probation/parole revocations for technical violations versus new criminal offenses. Recidivism that results from absconding during community supervision can be defined in different ways. Some jurisdictions traditionally counted any supervisee who failed to report to a supervising agency for six months or more, while others counted only absences resulting in a return to prison (Maltz 1984). The Council of State Governments (2022) found 32 of 50 states that reported actual returns to prison as a key recidivism metric, while 29 included probation and parole revocations that could potentially lead to prison.
The timeframe used to generate recidivism metrics is another major source of variation in recidivism rates. An agency reporting rearrests for 24 months following an intervention will almost always show a higher recidivism rate than one reporting new arrests in just the first six months. Researchers at the U.S. Sentencing Commission tracked rearrests among offenders placed on federal probation in 2005, and found that just over a third of the probationers were rearrested within eight years (Hunt and Dumville 2016). The median time to rearrest was 21 months. When half of all eventually rearrested probationers avoid arrest for nearly two years, is it appropriate to use recidivism to infer something about the effectiveness of probation? What if the definition of recidivism were reconviction rather than arrest? Court processing can be time-consuming. Delayed reporting means delayed measurement of outcomes, further complicating comparisons of recidivism over time and between jurisdictions (Ringland 2013; Wormith and Goldstone 1984).
Comparing recidivism across jurisdictions and interventions serving different populations is always complicated by the need for common denominators of time and place (Maltz 1984). Is recidivism the same for a person with two new arrests at 12 months versus someone with a single arrest at six months? When a person is arrested one year after being released from incarceration, are they a greater risk to public safety than someone arrested after 18 months? A judge may decide that three arrests in one year would indicate a high-risk defendant. Should the assessment be different if those three arrests are spread evenly across a full year rather than occurring within the same month?
Population-level risks are another major source of variation in official recidivism reporting. Parolees (i.e., post-incarceration supervision) tend to be higher risks than probationers. Thus, parole agency reports will likely show a higher incidence of recidivism. Should this be used to infer something about the effectiveness of parole vis-à-vis probation? Young adults show higher recidivism rates and more substance use disorders, with higher rates among young people with prior histories of violent victimization and trauma (Perker et al. 2019). Is it ethical to view interventions as less effective or to impose more severe sanctions on people due to age?
Offense categories matter as well (Durose and Antenangeli 2021). People convicted of property crimes recidivate at higher rates than those convicted of violent crimes. Thus, an agency serving property offenders will naturally report higher recidivism rates. A jurisdiction measuring recidivism with felony reconvictions will report lower rates than one measuring recidivism as any reconviction. Blumstein and Larson (1971) recommended that recidivism reports should always include sample characteristics (e.g., prior arrests, age, prosocial orientation, interest in rehabilitation) and then control for such differences. Maltz (1984) recommended a deeper examination of all correlates of recidivism (e.g., age-adjusted criminal career paths).
Predicting Recidivism
Risk assessment instruments are statistical decision-making tools that justice agencies and policymakers use to predict the likelihood that an individual will reoffend. Many risk assessment tools are available from commercial vendors, and their presence in the criminal justice sector continues to grow. Risk-based predictive instruments can draw on a combination of official records and clinically observed data. They can be administered by trained case managers, clinicians, or supervision officers. Whether drawing on official data, self-report measures, or the judgment of justice professionals, public agencies across the United States have increasingly turned to algorithm-based risk assessment tools to predict recidivism (Fazel et al. 2022; Kroner and Loza 2001; Loza and Loza-Fanous 2001; Viljoen et al. 2025). Justice agencies use statistical predictions to determine the restrictiveness or severity of criminal sentences, the focus of treatment efforts, and the terms of community supervision.
Researchers advise that risk instruments work best when they involve multiple measures, are locally validated, and regularly updated. Justice intervention programs use risk assessments in two circumstances: (1) to assess participants at intake to determine their service needs and set appropriate intervention levels, and (2) to measure change over time as an indicator of program effectiveness and case outcomes. Harris et al. (1993) emphasized the importance of clinical judgment alongside formal instruments — particularly attention to dynamic risk and protective factors such as marital status, substance abuse, psychopathy, and offending history. Mills and Kroner (2006) similarly recommended combining static and dynamic instruments, with clinicians playing a key role in accurately rating dynamic risks.
Decades of research evidence point to the importance of including dynamic criminogenic needs in risk-based measures. In a meta-analysis of 131 studies, Gendreau et al. (1996) found that risk assessment instruments can be useful for predicting recidivism outcomes with dynamic factors (e.g., criminogenic needs, companions, antisocial personality, social achievement, and substance abuse) as well as static factors (e.g., criminal history, pre-adult antisocial behavior history, age, race, and gender). The analysis suggested that preventing recidivism requires more than treatment to reduce personal disorders and emotional distress.
Gendreau et al. (1996) also recommend using risk instruments designed for specific offender populations and at designated points in justice systems. Their caution was echoed broadly in the literature. Researchers advise that predictive instruments for assessing recidivism risk be applied only with populations and time periods for which they have been validated (Rice and Harris 1995; Villeneuve and Quinsey 1995; Yesberg et al. 2015). Rosenfeld and Grigg (2022) reminded researchers that the value and utility of recidivism measures depend on the population samples from which data are drawn, the specific follow-up or “time to failure” periods involved, and the populations to which recidivism analyses are expected to generalize.
Researchers report that the accuracy of risk assessment algorithms may be overstated. In one study of a widely used risk assessment tool (COMPAS), researchers compared its accuracy to predictions made by a group of people with little to no criminal justice expertise (Dressel and Farid 2018). Participants were asked to review anonymous descriptions of defendants that notably did not include race or ethnicity, and to predict each subject’s likely recidivism using far less detail for each case than available through the assessment tool — i.e., 7 characteristics versus 137. The recidivism predictions of the study participants were roughly as accurate as those derived statistically from the assessment tool.
Other studies examining the predictive performance of risk assessment tools find their accuracy to be mixed, ranging from poor to moderate. Some of the most positive findings came from studies that used very small samples or included co-authors who were involved in developing the risk assessment tools being tested (Fazel et al. 2022). The most thorough review of tools for predicting recidivism cautioned policymakers and practitioners to avoid an uninformed fascination with predictive algorithms. Based on a meta-analysis of 31 previous studies, researchers suggested that risk assessment tools may complement, but never replace, the informed clinical judgment of staff members and leaders of justice intervention programs (Viljoen et al. 2025).
Alternatives to Recidivism
Recidivism is a measure of failure. It assesses the rate at which people previously charged with crimes go on to commit additional crimes and become re-involved in the justice system. A different way to judge the effectiveness of justice interventions would be to ask how often people succeed after participating in an intervention. Rather than assessing only the risk of returning to crime, researchers could also assess the probability of avoiding or “desisting” from crime. Modern desistance approaches utilize multi-measure longitudinal frameworks that include both risk and protective factors, rather than analyzing risks alone.
Kazemian (2015) argued that justice programs should support desistance, treating crime reduction as a complex process of change in which individuals learn to be law-abiding over time. Where most traditional recidivism measures produce a binary outcome — individuals either recidivate or do not — desistance offers a more nuanced alternative. It monitors behavior over time and notes various indicators of a person’s success while allowing for occasional setbacks. Rosenfeld and Grigg (2022) described desistance as trackable through patterns or trajectories of events, particularly the declining frequency or severity of law violations. Under a desistance framework, a former armed robber who commits a misdemeanor can represent a setback while maintaining the possibility of general improvement.
Risk Factors, Protective Factors, and Desistance Metrics
Desistance-based programs support prosocial pathways to reducing recidivism. Rather than tracking justice events alone, agencies could compile data about various protective factors that benefit justice-involved individuals, such as employment, housing, and education. Coupland and Olver (2020) found that protective factors were significantly associated with reductions in violent recidivism and community-level safety. Butts and Schiraldi (2018) recommended shifting justice effectiveness measures toward desistance, social integration, and community well-being, as traditional metrics like arrest and incarceration may reinforce the racial and class biases already embedded in justice system operations.
The key challenge in desistance-based effectiveness monitoring is the availability of data. Rather than relying solely on metrics created by criminal justice agencies, measuring desistance requires routine access to individual-level, identifiable data from a range of organizations across health, education, housing, employment, and many other sectors. Agencies in these sectors are not known for readily and competently sharing detailed information about individuals across state and local jurisdictions. Furthermore, when organizations requesting such information are part of the criminal justice system, sharing detailed data could pose serious legal risks for the agencies that provide it.
Recidivism in Policy and Practice
Decades of research on criminal justice point to key principles regarding the role of recidivism in criminal justice policy and practice.
(1) Official recidivism both absorbs and reflects the racial and class disparities discernible at multiple stages of the criminal justice process.
(2) Using recidivism as the principal measure of effectiveness for justice interventions results in a systematic devaluation of programs that serve high-risk individuals and high-surveillance communities.
(3) Algorithmic tools built with recidivism data perpetuate racially disparate impacts even if they are technically accurate in establishing statistical relationships between past and future events generated by the justice system.
(4) The underlying problem of bias is inherent to any decision-making structure that relies on recidivism as the key effectiveness outcome for criminal justice interventions.
Recidivism is a complex social phenomenon shaped by many factors, including individual, socio-economic, and environmental factors. Younger people are more likely than older people to violate the law and be arrested. Males are arrested and imprisoned more often than females. Individuals lacking employment, regular income, educational credentials, and adequate housing are more likely to come into contact with law enforcement and become involved in the justice system. When people are affected by mental and behavioral disorders, previous trauma, and family instability, the risks of justice involvement escalate dramatically. Beyond individual factors, anyone living in a disadvantaged neighborhood is exposed to more prevalent firearm possession, illegal drug markets, and group violence, all of which may increase their chances of contact with law enforcement.
Basing the sequence of criminal justice responses on system-generated recidivism measures ignores the broad web of individual and social factors that combine to produce the cumulative inequality that sustains racial disparities throughout criminal justice processing (Kurlychek and Johnson 2019). The challenge facing policymakers, practitioners, and researchers is how to acknowledge that recidivism obscures structural problems in the justice system while using recidivism measures effectively and equitably for both public safety and social justice (Freeman et al. 2021).
The justice system traditionally relies on four recidivism indicators: rearrest, reconviction, reincarceration, and the revocation of community supervision. As indicators of risk to public safety, each has its place, and each can be misused. They track justice system metrics but not individual change or progress. They are vulnerable to underreporting, inconsistencies in policy and practice, variations in clinical assessment, and potentially harmful differences in community understanding and fairness.
Recidivism can be one item in a diverse menu of performance metrics. It is not a wholly sufficient indicator of the well-being and behavior of individuals previously involved in the criminal justice system. High rates of recidivism also illuminate the failure of social structures and community resources. The consequences of a toxic social environment are readily observed among individuals exposed to it. When public safety problems derive from a combination of social and individual pathologies, any effort to prevent them must involve both social and individual interventions. When policy focuses on recidivism with no attention to its social correlates, the goal of community safety is ill-served.
A more effective approach would be to monitor both desistance and recidivism. Positive indicators would include the absence of new arrests among known individuals, as well as community rates of housing stability, employment, social service utilization, and community cohesion. Intervention programs could start by understanding the specific risks and needs of the communities where their clients tend to reside, and then design a menu of strategies that address the full array of causal factors.
Evaluation researchers could assess an intervention’s efficacy by tracking changes at both the neighborhood and individual levels. Data would have to be collected directly from neighborhood residents rather than relying exclusively on justice system metrics. Measures could rely on self-reports from individuals and their acquaintances, reports from neighborhood residents, and interviews with the leaders of local organizations. By including a broader range of metrics, policymakers and the general public would begin to see community safety risks as a complex social phenomenon rather than the simple aggregation of bad behaviors attributed to bad people.
Conclusion
To make the application of recidivism SMART(er), policymakers and practitioners must acknowledge the social, political, and economic environments in which recidivism is measured and deployed. Establishing the effectiveness of justice interventions is not a simple task. It requires a suite of indicators that measure the impact of any policies or practices designed to support public safety and help residents avoid behaviors that harm themselves, their families, and their communities. A diverse menu of outcomes should focus on more than negative indicators. Systems tracking SMART(er) outcomes would attend to equity and respond to the broader cultural and social contexts of criminal justice policy. The task is not simply to track changes among people already involved in justice systems and those disproportionately from disadvantaged communities. It is also to track changes in communities in ways that matter to their residents.
The foundations of recidivism as an outcome measure presume that interventions by the criminal justice system reduce future criminality. As interventions become more severe, reaching a pinnacle with incarceration, the effects on criminality should increase. Yet, criminological research casts doubt on this basic premise. The research literature is replete with negative findings about the deterrent effects of incarceration, including many studies that even find criminogenic effects (Loeffler and Nagin 2022). The pattern of evidence would seem to undermine the essential policy logic underlying recidivism as an outcome measure. If the strongest interventions deployed by criminal justice systems do not reliably reduce recidivism, does it make sense to use recidivism as the principal benchmark of effectiveness? Recidivism outcomes are known to be unreliable, and they can impose social costs that fall disproportionately on already-disadvantaged populations. Measuring the outcomes of criminal justice interventions should include recidivism, but exclusive reliance on recidivism is ill-advised and potentially reckless.
Acknowledgments
The authors appreciate the contributions of other researchers who reviewed drafts of this report. In particular, we appreciate the assistance of Kathleen Tomberg and Rebecca Balletto, our colleagues at the Research and Evaluation Center. The authors are also indebted to the many policymakers, agency staff, and community members who have shared their knowledge and experience over the years.
All photographs were licensed from Dreamstime.com
Funding
Developed with funding provided by the Research Foundation of the City University of New York through indirect cost recoveries generated by previous projects at the Research and Evaluation Center. All conclusions are those of the authors. Funders and partners of the Research and Evaluation Center are not responsible for any findings or opinions presented in Center publications.
Recommended Citation
Das, Sumita (Mira) and Jeffrey A. Butts (2026). Recidivism: Use with Caution. New York, NY: Research and Evaluation Center, John Jay College of Criminal Justice, City University of New York.
References
Berg, Mark T. and Beth M. Huebner (2011). Reentry and the ties that bind: An examination of social ties, employment, and recidivism. Justice Quarterly 28(2): 382-410.
Bhuller, Manudeep, Gordon B. Dahl, Katrine V. Løken and Magne Mogstad (2020). Incarceration, recidivism, and employment. Journal of Political Economy 128(4): 1269-1324.
Blumstein, Alfred and Richard C. Larson (1971). Problems in modeling and measuring recidivism. Journal of Research in Crime and Delinquency 8(2): 124-132.
Bronson, Jennifer, Jessica Stroop, Stephanie Zimmer and Marcus Berzofsky (2020). Drug use, dependence, and abuse among state prisoners and jail inmates, 2007-2009. Special Report. Washington, DC: Bureau of Justice Statistics, U.S. Department of Justice.
Butts, Jeffrey A. and Vincent N. Schiraldi (2018). Recidivism Reconsidered: Preserving the Community Justice Mission of Community Corrections. Cambridge, MA: Program in Criminal Justice Policy and Management, Kennedy School, Harvard University.
Chen, Elsa Y. and Sophie E. Meyer (2020). Beyond recidivism: Toward accurate, meaningful, and comprehensive data collection on the progress of individuals reentering society. In Andrea Leverentz, Elsa Y. Chen and Johnna Christian (Editors), Beyond Recidivism: New Approaches to Research on Prisoner Reentry and Reintegration. New York, USA: New York University Press, pp. 13-38.
Cloyes, Kristin G., Bob Wong, Seth Latimer and Jose Abarca (2010). Time to prison return for offenders with serious mental illness released from prison: A survival analysis. Criminal Justice and Behavior, 37(2): 175-187.
Council of State Governments Justice Center (2022). 50-state report on public safety.
Coupland, Richard B.A. and Mark E. Olver (2020). Assessing protective factors in treated violent offenders: Associations with recidivism reduction and positive community outcomes. Psychological Assessment, 32(5): 493-508.
Doran, George T. (1981). There’s a S.M.A.R.T. way to write management’s goals and objectives. Management Review, 70(11): 35-36.
Drawve, Grant and Susan McNeeley (2021). Recidivism and community context: Integrating the environmental backcloth. Journal of Criminal Justice, 73: 101786.
Dressel, Julia and Hany Farid (2018). The accuracy, fairness, and limits of predicting recidivism. Science Advances, 4(1).
Durose, Matthew R. and Leonardo Antenangeli (2021). Recidivism of Prisoners Released in 34 States in 2012: A 5-year Follow-up Period (2012–2017). Washington, DC: Bureau of Justice Statistics, U.S. Department of Justice.
Fazel, Seena, Matthias Burghart, Thomas Fanshawe, Sharon Danielle Gil, John Monahan and Rongqin Yu (2022). The predictive performance of criminal risk assessment tools used at sentencing: Systematic review of validation studies. Journal of Criminal Justice, 81: 101902.
Fontaine, Jocelyn and Jennifer Biess (2012). Housing as a Platform for Formerly Incarcerated Persons. Washington, DC: Urban Institute.
Freeman, Kelly Roberts, Cathy Hu and Jesse Jannetta (2021). Racial Equity and Criminal Justice Risk Assessment. Washington, DC: Urban Institute.
Gendreau, Paul, Tracy Little and Claire Goggin (1996). A meta‐analysis of the predictors of adult offender recidivism: What works! Criminology, 34(4): 575-608.
Griswold, David B. (1978). A comparison of recidivism measures. Journal of Criminal Justice, 6(3): 247-252.
Han, Woojae and Allison D. Relich (2016). The impact of community treatment on recidivism among mental health court participants. Psychiatric Services, 67(4): 384-390.
Harding, David J., Jessica J.B. Wyse, Cheyney Dobson and Jeffrey D. Morenoff (2014). Making ends meet after prison. Journal of Policy Analysis and Management, 33(2): 440–470.
Harris, Grant T., Marnie E. Rice and Vernon L. Quinsey (1993). Violent recidivism of mentally disordered offenders: The development of a statistical prediction instrument. Criminal Justice and Behavior, 20(4): 315-335.
Huebner, Beth M. and Jennifer Cobbina (2007). The effect of drug use, drug treatment participation, and treatment completion on probationer recidivism. Journal of Drug Issues, 37(3): 619-641.
Hunt, Kim Steven and Robert Dumville (2016). Recidivism Among Federal Offenders: A Comprehensive Overview. Washington, DC: United States Sentencing Commission.
Jacobs, Leah A. and Aaron Gottlieb (2020). The effect of housing circumstances on recidivism: Evidence from a sample of people on probation in San Francisco. Criminal Justice and Behavior, 47(9): 1097–1115.
James, Doris J. and Lauren E. Glaze (2006). Mental health problems of prison and jail inmates. Bureau of Justice Statistics Special Report (NCJ 213600). Washington, DC: Bureau of Justice Statistics, U.S. Department of Justice.
Johnson, James L. (2017). Comparison of recidivism studies: AOUSC, USSC, and BJS. Federal Probation, 81(1): 52-54.
Kazemian, Lila (2015). Straight Lives: The Balance between Human Dignity, Public Safety, and Desistance from Crime. New York, NY: John Jay College of Criminal Justice, Research and Evaluation Center.
Kendall, Sacha, Sarah Redshaw, Stephen Ward, Sarah Wayland and Elizabeth Sullivan (2018). Systematic review of qualitative evaluations of reentry programs addressing problematic drug use and mental health disorders amongst people transitioning from prison to communities. Health and Justice, 6(4): 1-11.
King, Ryan and Brian Elderbroom (2014). Improving Recidivism as a Performance Measure. Washington, DC: Urban Institute.
Klingele, Cecelia M. (2019). Measuring change: From rates of recidivism to markers of desistance. Journal of Criminal Law and Criminology, 109(4): 769–817.
Kroner, Daryl G. and Wagdy Loza (2001). Evidence for the efficacy of self-report in predicting nonviolent and violent criminal recidivism. Journal of Interpersonal Violence, 16(2): 168-177.
Kurlychek, Megan C. and Brian D. Johnson (2019). Cumulative disadvantage in the American criminal justice system. Annual Review of Criminology, 2: 291-319.
Lai, Ijun, Jillian Stein, Christian Geckeler and Ellie Pasternack (2022). Common Indicators of Recidivism Used in Program and Policy Evaluations. Princeton, NJ: Mathematica and Social Policy Research Associates.
Li, Nan (2026). How valid are self-reports of delinquent and criminal behavior? A meta-analysis comparing self-report and official record measures. Journal of Criminal Justice, 103: 102610.
Link, Nathan Wong and Leah K. Hamilton (2017). The reciprocal lagged effects of substance use and recidivism in a prisoner reentry context. Health & Justice, 5(8): 1-14.
Loeffler, Charles E. and Daniel S. Nagin (2022). The impact of incarceration on recidivism. Annual Review of Criminology, 5: 133–152.
Loza, Wagdy and Amel Loza-Fanous (1999). Anger and prediction of violent and nonviolent offenders’ recidivism. Journal of Interpersonal Violence, 14(10): 1014-1029.
Loza, Wagdy and Amel Loza-Fanous (2001). The effectiveness of the Self-Appraisal Questionnaire in predicting offenders’ postrelease outcome: A comparison study. Criminal Justice and Behavior, 28(1): 105-121.
Maltz, Michael D. ([1984] 2001). Recidivism. Orlando, FL: Academic Press, Inc.
McGuire, James, Charlotte A.L. Bilby, Ruth M. Hatcher, Clive R. Hollin, Juliet Hounsome and Emma J. Palmer (2008). Evaluation of structured cognitive–behavioural treatment programmes in reducing criminal recidivism. Journal of Experimental Criminology, 4: 21-40.
Mills, Jeremy F. and Daryl G. Kroner (2006). The effect of discordance among violence and general recidivism risk estimates on predictive accuracy. Criminal Behaviour and Mental Health, 16(3): 155-166.
Olson, David E. and Arthur J. Lurigio (2000). Predicting probation outcomes: Factors associated with probation rearrest, revocations, and technical violations during supervision. Justice Research and Policy, 2(1): 73-86.
Pelletier, Emily and Douglas Evans (2019). Beyond recidivism: Positive outcomes from higher education programs in prisons. Journal of Correctional Education, 70(2): 49-68.
Perker, Selen Siringil, Lael E.H. Chester and Vincent Schiraldi (2019). Emerging adult justice in Illinois: Towards an age-appropriate approach. New York, NY: Columbia University, Justice Lab.
Rice, Marnie E. and Grant T. Harris (1995). Violent recidivism: Assessing predictive validity. Journal of Consulting and Clinical Psychology, 63(5): 737-748.
Ringland, Clare (2013). Measuring recidivism: Police versus court data. Crime and Justice Bulletin. NSW Bureau of Crime Statistics and Research. Sydney, New South Wales: Author.
Rosenfeld, Richard and Amanda Grigg (Editors). (2022). The Limits of Recidivism: Measuring Success After Prison. Washington, DC: National Academies of Sciences, Engineering, and Medicine. Committee on Law and Justice, Division of Behavioral and Social Sciences and Education.
Skeem, Jennifer L. and Christopher T. Lowenkamp (2016). Risk, race, and recidivism: Predictive bias and disparate impact. Criminology, 54(4): 680–712.
Steadman, Henry J., Fred C. Osher, Pamela Clark Robbins, Brian Case and Steven Samuels (2009). Prevalence of serious mental illness among jail inmates. Psychiatric services, 60(6): 761-765.
Tong, L.S. Joy and David P. Farrington (2006). How effective is the “Reasoning and Rehabilitation” programme in reducing reoffending? A meta-analysis of evaluations in four countries. Psychology, Crime & Law, 12(1): 3-24.
Tripodi, Stephen J., Johnny S. Kim and Kimberly Bender (2010). Is employment associated with reduced recidivism? The complex relationship between employment and crime. International Journal of Offender Therapy and Comparative Criminology, 54(5): 706–720.
Viljoen, Jodi L., Ilvy Goossens, Sanam Monjazeb, Dana M. Cochrane, Lee M. Vargen, Melissa R. Jonnson, Adam J. E. Blanchard, Shanna M.Y. Li and Jourdan R. Jackson (2025). Are risk assessment tools more accurate than unstructured judgments in predicting violent, any, and sexual offending? A meta-analysis of direct comparison studies. Behavioral Sciences & the Law, 43(1): 75-113.
Villeneuve, David B. and Vernon L. Quinsey (1995). Predictors of general and violent recidivism among mentally disordered inmates. Criminal Justice and Behavior, 22(4): 397-410.
Wilson, David B., Catherine A. Gallagher and Doris L. MacKenzie (2000). A meta-analysis of corrections-based education, vocation, and work programs for adult offenders. Journal of Research in Crime and Delinquency, 37(4): 347-368.
Wilson, David B., Leana Allen Bouffard and Doris L. MacKenzie (2005). A quantitative review of structured, group-oriented, cognitive-behavioral programs for offenders. Criminal Justice and Behavior, 32(2): 172-204.
Wormith, J. Stephen and Colin S. Goldstone (1984). The clinical and statistical prediction of recidivism. Criminal Justice and Behavior, 11(1): 3-34.
Yesberg, Julia A., Jessica M. Scanlan, Laura J. Hanby, Ralph C. Serin and Devon L.L Polaschek (2015). Predicting women’s recidivism: Validating a dynamic community-based ‘gender-neutral’ tool. Probation Journal, 62(1): 33-48.
Zara, Georgia and David P. Farrington (2016). Criminal Recidivism: Explanation, Prediction and Prevention. Oxfordshire, England: Routledge, an imprint of Taylor & Francis Group.