criminal legal system, including arrest, revocation, conviction, and incarceration. These measures therefore reflect the interaction between individuals and the criminal legal system and not necessarily engagement in criminal behavior that may go undetected by criminal legal system actors. Nor do such measures account for the greater likelihood that some individuals may face arrest due to their identity or location and others may be arrested and convicted despite being innocent. In addition, as discussed in Chapter 2, recidivism rates differ depending on whether they are based on the records of those entering or leaving prison. Most persons who have been to prison do not return to prison, while roughly half of those released from prison at a given time, many of whom have been in prison multiple times, are likely to return. Finally, recidivism rates typically include technical violations of the conditions of community supervision that may not constitute criminal behavior for the general population (e.g., missing a parole meeting or failing a drug test). In these ways, current recidivism measures risk being both under- and over-inclusive.

Referring generally to a “recidivism rate” based solely on administrative data sources invites misinterpretation and policy responses that are not appropriately tailored to the actual circumstances of reoffending or to the specific purposes of research and interventions. One common misinterpretation has been mentioned: assuming that persons who have been in prison will probably return. Policy responses may be misinformed by common recidivism measures, which do not record changes in the frequency or seriousness of criminal activity by persons released from prison—the focus of our first recommendation.

A robust body of literature on desistance has demonstrated that the cessation of criminal activity, like other behavioral changes, is incremental and may involve setbacks. Despite this, recidivism is typically measured in a binary manner that distinguishes between people who are and those who are not rearrested, reconvicted, or reincarcerated after release from prison without adequate description of consequential changes in post-release criminal involvement. Measures of desistance offer a more nuanced, complete, and realistic view of the cessation of criminal behavior.

Viewing any return to crime as failure obscures how cessation of criminal behavior actually occurs and fails to recognize positive progress, such as longer time elapsing between crimes or fewer serious crimes committed. For example, using existing measures of recidivism, someone previously convicted of armed robbery who is arrested or convicted on a misdemeanor charge for shoplifting would be labeled a recidivist. An alternative measure from a desistance perspective would acknowledge setbacks in the process of desistance and count less serious criminal activity as a potential indicator of progress. Moreover, focusing only on criminal activity neglects other signifiers of progress, including change in life circumstances, self-view, and

feelings of hope that can result in reduced involvement in crime over time, eventually leading to the cessation of criminal behavior.

Conclusion 1: Recidivism rates based on administrative records reflect the interaction between individuals and the criminal legal system. These measures reflect decisions by legal authorities and not necessarily an individual’s return to criminal activity.

Conclusion 2: Because cessation or reduction in criminal behavior often occurs as part of a gradual process that may involve setbacks, measures of desistance from crime offer a more realistic account of an individual’s reduction in criminal activity.

Recommendation 1: To ensure more precise and accurate use of the construct of recidivism, researchers, policy makers, and practitioners should (a) specify the exact actions taken by legal authorities (arrest, revocation, conviction, incarceration) included in their measures, (b) clarify the limitations of the data used to measure these actions, and (c) supplement binary recidivism measures with measures of desistance from crime, such as the frequency and seriousness of offenses.

MEASURING SUCCESS

As emphasized in our conversations with formerly incarcerated individuals, correctional officials, and service providers for crime victims and survivors, recidivism does not capture important, positive post-release outcomes that facilitate social integration and individual well-being. A more meaningful conception of success views post-release outcomes through the lens of healthy adult development across multiple life domains in addition to criminal involvement: education, employment, housing, family and social support, and mental and physical health. Such a broad conception of success involves a heightened sense of personal well-being, which is not generally captured in administrative records. As such, the measurement of post-release success should include reliance on self-report data and standardized psychological instruments that provide indicators of post-release success not contained in official records. Examples of existing instruments discussed in Chapter 4 include the Cantril Self-Anchoring Striving Scale, 100 Million Healthier Lives, and Well-being in the Nation instruments. Other measures include the National Institutes of Health’s PhenX Toolkit.

Conclusion 3: Post-release success involves multiple life domains (e.g., health, employment, housing, civic engagement) and not simply involvement in the criminal legal system. Success entails a heightened

sense of personal well-being, which is best measured on the basis of self-report surveys and validated assessment instruments.

Recommendation 2: Researchers should review existing measures and, as needed, develop and validate new measures to evaluate post-release success in multiple domains, including personal well-being, education, employment, housing, family and social supports, health, civic and community engagement, and legal involvement.

Barriers to Success

Success following imprisonment cannot be understood without attention to the social context into which people return. An individual’s environment after release can support or undermine their ability to successfully return to society. What resources and supports, if any, are available in an individual’s neighborhood to facilitate success? In what ways are the organizations that provide these resources and supports attentive to or dismissive of the needs of returning individuals? How do local or state policies and practices impede progress and eventual desistance from criminal behavior?

Given persistent racial and ethnic inequalities in health, housing, education, and employment—and the disproportionate incarceration of Black, Brown, and Indigenous populations—how is post-release success for historically marginalized groups shaped by structural inequalities within each of these systems and the community at large? Without understanding how community contextual factors and existing policies support or hinder an individual’s return from prison, opportunities to identify and potentiate success after incarceration are missed.

Measuring structural barriers and systemic inequalities that impede success can include recording the residential address (or other geographic identifiers) in intervention studies and program evaluations of individuals returning from prison. Linking the geographic data to existing small-area measures, such as the Area Deprivation Index, would enhance understanding of how community socioeconomic conditions affect an individual’s potential for success following release from prison. Researchers can also track whether individuals are subject to particular state or local regulations that restrict employment, housing, or public assistance for those with criminal records, local or state variations in what constitutes a technical parole violation, and the level of police surveillance in their community.

Understanding how institutions and organizations hinder success or facilitate it is equally important, especially concerning how systems that provide health care, food, transportation, education, and employment support the needs of individuals released from prison. Linking data from

correctional systems to other administrative data from state and local government could increase understanding of how different sectors support the success of individuals following release. For example, some studies have linked data from correctional systems to substance use treatment, opioid overdose, and cancer tumor databases. These studies demonstrate how linkages with health systems that compile data on the social determinants of health can better serve people who are being released from prison.

Moreover, linkages can be made between administrative data from various systems and data from well-being inventories and ongoing longitudinal surveys such as those discussed in Chapter 4. To be sure, such data linkages would need to be carefully designed and monitored, with input from individuals with a history of incarceration, both to protect privacy and to avoid coercive surveillance leading to repeat encounters with the criminal legal system.¹

Conclusion 4: The existence of community-level and policy facilitators of and barriers to success can be documented in studies that link data on post-release success and local socioeconomic conditions, policies that restrict access to employment, housing, and public benefits, and structural inequalities that disproportionately affect persons of color.

Recommendation 3: Researchers should review and, as needed, develop new measures of facilitators of and structural barriers to post-release success that link data across multiple domains, including personal well-being, education, employment, housing, family and social supports, health, civic and community engagement, and legal involvement. These measures should reflect the particular needs and experiences of historically marginalized groups.

NATIONAL STANDARDS FOR MEASURING POST-RELEASE SUCCESS

Linking data across multiple recordkeeping systems would facilitate the development of national standards to measure success among persons released from prison. By establishing a standardized, minimum set of demographic, social, economic, and legal data to be collected by local, state, and federal agencies, national standards would enhance the comparability of evaluations of post-release outcomes and the quality and utility of administrative data for monitoring success across multiple policy domains.

___________________

¹ See the two-volume National Academies (2017) report Federal Statistics, Multiple Data Sources, and Privacy Protection (https://www.nap.edu/catalog/24893/federal-statistics-multiple-datasources-and-privacy-protection-next-steps).

These standardized uniform success measures would supplement, not replace, the local measures correctional agencies use to monitor their performance. As noted in Chapter 2, the local measures can yield best practices in the measurement of success to inform the development and refinement of the national standards.

A model for this kind of standardized data repository is the Uniform Crime Reports (recently superseded by National Incident-Based Reporting System), which provides harmonized crime classification and coding procedures for local law enforcement agencies. In addition, national standards for measuring post-release success would benefit from the kind of continuous updating by the World Health Organization of its International Classification of Diseases (ICD) codes used in research, treatment, and disease prevention efforts. The ICD also harmonizes data across different systems and levels of government.

More immediately, the academic community could develop a website of core success measures, instruments, and validation studies from multiple administrative domains that is accessible to researchers, practitioners, and policymakers. The website and accompanying communication among researchers and practitioners would begin to establish norms around common measurement and provide guidance on how to address gaps, errors, and other data issues when measuring success across diverse information sources. Such a toolkit could be developed by a combination of partnerships between private foundations and government agencies.

Conclusion 5: National standards for measuring success among individuals released from prison would augment the comparability of program evaluations and the utility of administrative and other data across multiple policy domains. The development of a website containing core measures and instruments would hasten the eventual development of national measurement standards. These efforts can be supported by federal agencies and private foundations committed to improving success for persons released from prison.

Recommendation 4: The National Institute of Justice, Bureau of Justice Statistics, Bureau of Justice Assistance, National Institutes of Health, and other federal agencies and centers whose missions are central to the success of persons released from prison should (1) convene interdisciplinary research advisory panels to assess data, methods, and recommendations for measuring post-release success; (2) request grant proposals from researchers and practitioners, in collaboration with formerly incarcerated persons, to review existing measures of success and develop and validate new measures as needed; and (3) consider questions relevant to the measurement of post-release success in existing

survey protocols such as the American Community Survey and data collection efforts in other domains such as education, labor, and health. In order to carry out such activities, additional funding will be required. Private foundations committed to improving success among persons released from prison should support this evaluation independently or in partnership with federal agencies. Governmental and private support should be directed, at a minimum, to the following issues:

1. The quality of records from legal and other social institutions used to monitor post-release success;
2. The utility and feasibility of linking records across multiple administrative domains;
3. The utility and feasibility of linking existing administrative data with instruments measuring personal well-being;
4. The development of a website containing core measures of success across multiple administrative domains and the role of qualitative as well as quantitative research in the development of these measures; and
5. The eventual development of uniform national standards for measuring post-release success.

CONCLUSION

Five overarching themes emerged from the committee’s deliberations on improving the measurement of success among persons released from prison:

1. Current binary measures of recidivism do not adequately reflect the continuation or cessation of criminal behavior and should be augmented by measures of desistance from crime that, at a minimum, account for changes in the frequency and seriousness of criminal activity.
2. Post-release success is multifaceted and cannot be adequately measured by indicators of criminal involvement alone.
3. Persistent group inequalities require that measures of post-release success take into account the needs and experiences of historically marginalized populations.
4. The perspectives of persons with lived experience of incarceration and practitioners need to be part of all efforts to improve and implement new measures of post-release success.
5. Improving measures of post-release success will benefit formerly incarcerated persons, the communities to which they return, and society as a whole by supporting policies to facilitate post-release social integration, enhance personal well-being, and improve public safety.