Incarceration and Desistance: Evidence from a Natural Policy Experiment, North Carolina and South Dakota, 1995-2019 (ICPSR 38817)

Version Date: Apr 10, 2025

Cite this study |

Principal Investigator(s):
Shawn Bushway, RAND Corporation

https://doi.org/10.3886/ICPSR38817.v1

Version V1

Slide tabs to view more

Summary

In order to investigate whether reductions to sentence length reduce recidivism and its related costs, the research team conducted two studies under the same NIJ grant.

In Study 1, the research team used a quasi-experimental approach to examine how sentence length influenced time to rearrest. Specifically, the team estimated an instrumental variable hazard model using variation stemming from the implementation of South Dakota's Public Safety Improvement Act of 2013 (SB70). SB70 was a major sentencing reform that reduced the likelihood of incarceration and the length of sentence for certain non-violent offenses. The legislation was a response to high rates of incarceration and rising criminal justice costs with a particular focus on addressing individuals convicted of non-violent offenses. The goal was to improve public safety while reducing corrections spending. For grand theft below $5,000 and third-degree burglary, the reform reduced the severity of the felony class and the maximum allowed sentence length by 50% to 80%. Only certain offenses were subject to sentencing reductions such that unaffected offense types were included in the analysis as a comparison group.

In Study 2, the research team applied the logic of recurring event hazard models that predicted the probability of events over time with two distinct objectives: to identify desisters and to estimate the impact of life events on the longitudinal pattern of offending. Administrative data from the North Carolina Department of Public Safety (NCDPS) Division of Adult Correction were obtained to answer this research question. The team connected commonly used trajectory models to recurrent event survival models.

This collection contains only syntax files and documentation that were created for Study 1 (South Dakota). The P.I. codebook reflects the dataset that was created for analysis, though the data are not available. Please refer to the User Guide for more information. While Study 2's (North Carolina) methodology is described, please note that, excluding the NIJ Final Report, no Study 2 materials are available through ICPSR.

Citation

Bushway, Shawn. Incarceration and Desistance: Evidence from a Natural Policy Experiment, North Carolina and South Dakota, 1995-2019. Inter-university Consortium for Political and Social Research [distributor], 2025-04-10. https://doi.org/10.3886/ICPSR38817.v1

Export Citation:

  • RIS (generic format for RefWorks, EndNote, etc.)
  • EndNote

Funding

United States Department of Justice. Office of Justice Programs. National Institute of Justice (2019-R2-CX-0031)

Subject Terms

Geographic Coverage

Smallest Geographic Unit

County

Restrictions

Access to these data is restricted. Users interested in obtaining these data must complete a Restricted Data Use Agreement, specify the reason for the request, and obtain IRB approval or notice of exemption for their research.

Distributor(s)

Inter-university Consortium for Political and Social Research
Hide

Time Period(s)

2009-01-01 -- 2015-12-30 (Study 1 analysis), 1995-01 -- 2019-11 (Study 2 analysis)

Data Collection Notes

  1. Due to data sharing agreements with the South Dakota Office of the Attorney General, the South Dakota criminal history data (Study 1) are unable to be shared. The North Carolina data (Study 2) were obtained from the North Carolina Department of Public Safety (NCDPS) Offender Public Information Search.
Hide

Study Purpose

The purpose of the project was to investigate whether reduced sentence length impacts recidivism in two distinct ways: using a natural experiment approach with legislation reducing sentence time for certain crimes (South Dakota's SB70), and predicting the probability of desisting from crime over the life course.

Study Design

Study 1. Using a quasi-experimental design, the research team used an instrumental variable hazard model to determine whether time to next arrest changed when individuals received shorter incarceration sentences. The implementation of South Dakota's Public Safety Improvement Act of 2013 (SB70) was used as the instrumental variable to measure if there was variation in sentence length following sentencing reform. The sample included arrests with sentences that were impacted by SB70 and a comparison group of arrests resulting in conviction for similar offenses with sentences not impacted by SB70.

Study 2. To estimate the impact of life events on likelihood of recidivism, the research team used a recurring event hazards model that predicted the probability of recidivism/desistance among a 1979 birth cohort with criminal histories in North Carolina. The team used the publicly available Offender Profile, Sentence Component, and Sentence Computation data files from the North Carolina Department of Public Safety (NCDPS) Offender Public Information Search/Inmate Locator website. The data downloaded for analysis contained records associated with convictions from 1972 (earliest available) through April 8, 2021 (download date).

The initial research design relied on using data from South Dakota to conduct Study 2. However, after discovering the data were missing information on incarceration spells, the research team pivoted to using data from North Carolina, as they had access to it from other projects.

Sample

Study 1. The main analysis sample consisted of 2,616 arrests in South Dakota occurring between January 1, 2009 and December 31, 2015 that resulted in a conviction for at least one SB70-eligible grand theft or third-degree burglary offense without homicide (experimental/treated group), or at least one other burglary offense not impacted by SB70 (comparison group). 1,727 were charged with a treated offense, while 889 were charged with other burglary.

Study 2. The analytic sample was limited to individuals born in 1979 and risk periods following convictions from January 1995 to November 2019. Age of eligibility for conviction in the adult criminal justice system changed after November 2019, and later data were also impacted by the COVID-19 pandemic. Convictions occurring during incarceration were excluded.

Time Method

Longitudinal: Cohort / Event-based

Universe

  • Arrests occurring in South Dakota between 2009 and 2015.
  • Convictions, sentences, and incarceration periods in North Carolina between 1995 and 2019.

Unit(s) of Observation

Event/Process (Arrest), Individual, Time Unit (Risk Period)

Data Source

South Dakota Computerized Criminal History System

North Carolina Department of Public Safety (NCDPS) Offender Public Information Search/Inmate Locator

Data Type(s)

Description of Variables

Study 1. The analytic dataset contained an "index" arrest for each individual (arrests occurring between 2009-2015). For each index arrest, characteristics such as the number and type of offenses on the arrest, felony status and felony conviction, and time sentenced were present. Other variables included number of prior arrests, prior convictions, prior prison time served, arrest date, county of arrest, race/ethnicity, sex, and birthdate. Time to next arrest in days was the outcome measure. Exposure to SB70 was defined as an arrest that occurred after July 1, 2013 and that resulted in conviction for at least one property offense impacted by SB70.

Study 2. Independent variables included demographic information, frequency and variety of prior convictions, and measures of prior incarceration periods. The outcome variable and unit of analysis was length of risk period, i.e., the number of years between prison release and the next conviction.

Hide

Original Release Date

2025-04-10

Hide

Notes

  • 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.

  • The public-use data files in this collection are available for access by the general public. Access does not require affiliation with an ICPSR member institution.

  • ICPSR usually offers files in multiple formats for researchers to be able to access data and documentation in formats that work well within their needs. If you have questions about the accessibility of materials distributed by ICPSR or require further assistance, please visit ICPSR’s Accessibility Center.

  • One or more files in this data collection have special restrictions. Restricted data files are not available for direct download from the website; click on the Restricted Data button to learn more.