Program Profile: Use of Restrictive Housing in Prison to Reduce Recidivism (New Jersey)

This intervention involves the removal of incarcerated individuals from the general prison population. The program is rated No Effects. Individuals who were placed in restrictive housing in New Jersey prisons had statistically significantly more rearrests, reconvictions, and reincarcerations due to technical violations and recidivated faster, compared with individuals who did not experience restrictive housing.

A No Effects rating implies that implementing the program is unlikely to result in the intended outcome(s) and may result in a negative outcome(s).

Program Goals

Restrictive housing involves the removal of incarcerated individuals from the general prison population into restrictive housing. There are various goals of restrictive housing, including protection of certain incarcerated individuals and punishment of incarcerated individuals for their misconduct and misbehavior (Mears, 2016). A longer-term goal of restrictive housing is the reduction in recidivism once individuals leave prison.

Program Components/Target Population

There are three types of restrictive housing: 1) protective custody, 2) disciplinary segregation, and 3) administrative segregation. Protective custody targets individuals who are at a higher risk of victimization, because of their identity or the crime they committed, and separates them from the general population to provide them with safety. Disciplinary segregation, which is used as a form of punishment for prison misconduct, temporarily separates individuals from the general population. Finally, administrative segregation, which involves long-term separation (e.g., for 30 days or more) targets individuals who are considered a threat to the safety of other incarcerated individuals or prison staff.

The state of New Jersey uses restrictive housing as a tool for managing disruptive and violent incarcerated individuals. In the New Jersey Department of Corrections, restrictive housing includes both disciplinary segregation and administrative segregation (protective custody is not considered restrictive housing in New Jersey). Restrictive housing occurs in four separate units, which are located at four separate prisons. The capacity of each unit is approximately 350 individuals (a total of 1,400 individuals could be housed in restrictive housing at one time). Individuals placed in restrictive housing are generally in their cell for approximately 22 hours per day. During the 2-hour period outside of their cell, individuals are allowed to shower and have recreation time.

Receiving a restrictive housing sanction in New Jersey is a multistep process. The sanction is first imposed by the Inmate Disciplinary Officer and reviewed by the Institutional Classification Committee. An incarcerated individual can appeal the initial decision from the hearing officer during the first 48 hours. In addition to the review process, New Jersey also operates a three-tiered system within restrictive housing, ranging from level 1 to level 3. Level 1 is the most restrictive; privileges, possessions, activities, and amenities are highly restricted. By contrast, level 3 is the least restrictive but still more restrictive than the general population; level 3 also operates as a step-down unit before an individual returns to the general population. Individuals enter restrictive housing at level 1. Based on their behavior they can be promoted to level 2, and finally to level 3 before they return to the general population. Alternatively, based on behavior, individuals may also remain at their current level or be demoted to a previous level.

Program Theory

Restrictive housing uses deterrence theory as its foundation. Deterrence theory holds that, if punishment for crime or misconduct is swift, certain, and severe, then people are more likely to be dissuaded from committing additional antisocial behavior. Restrictive housing can operate as both a specific and general deterrent. As a specific deterrent, it is hypothesized that individuals who are placed in restrictive housing are discouraged from future misconduct while incarcerated to avoid facing punishment of placement in restrictive housing again. It is also believed that restrictive housing operates as a specific deterrent because it can provide individuals with ample time to reflect on their past indiscretions (Zgoba, Pizarro, and Salerno, 2020). It is further posited that restrictive housing can reduce the likelihood of recidivism once individuals are released from incarceration, as they may want to avoid reincarceration and the chance of experiencing restrictive housing again (Mears and Bales, 2009). Finally, restrictive housing can act as a general deterrent, deterring other incarcerated individuals from engaging in misconduct that may result in their placement in restrictive housing (Mears, 2016).

Study 1

Zgoba, Pizarro, and Salerno (2020) investigated the use of restrictive housing in New Jersey, focusing on whether the recidivism rates of those placed in restrictive housing differed from those who did not serve time in restrictive housing while in prison. The sample included individuals released from the New Jersey Department of Corrections prisons during 2012 and 2013, which included 18,917 individuals. Of these individuals, 1,278 were women, 62 of whom had served time in restrictive housing. Because of the small sample size, these women were subsequently dropped from the sample, as were 192 individuals who had missing data. This left a total eligible sample of 17,447 individuals released during 2012 and 2013, of whom 4,022 served time in restrictive housing and 13,425 did not serve time in restrictive housing.

Those placed in restrictive housing served an average of 211.43 days within the sanction (the median was 112 days). Time in restrictive housing ranged from fewer than 30 days (23.2 percent) to 2,995 days. Over half the sample (55.5. percent) served more than 90 days in restrictive housing, while 15 percent served from 61 to 90 days, and 6 percent served from 31 to 60 days.

Demographic and institutional data were obtained from the incarcerated individuals’ prison files. Criminal history and recidivism data, including all criminal justice–related events, were provided by the New Jersey State Police. Any criminal justice–related event that occurred outside the state of New Jersey was obtained by reviewing the Interstate Identification Index reports.

Propensity score matching was used to compare individuals who were placed in restrictive housing at least once during their prison sentence with those who did not serve any time in restrictive housing. Groups were matched using covariates that are known to predict both placement in restrictive housing and recidivism, such as age at release, race, prior criminal history, and offense of commitment. Criminal history measured whether the individual had any prior criminal justice events that occurred before the incarceration for which they were released in 2012 or 2013. Offense of commitment was coded as five separate dichotomous variables for each offense category: violent offenses (0,1); weapons offenses (0,1); property offenses (0,1); drug offenses (0,1); and other offenses (0,1). Using these covariates, propensity scores were calculated for each individual (n = 17,447). Of those individuals, 4,022 served time in restrictive housing and 13,425 did not. This resulted in propensity scores ranging from 0.04 to 0.99, with balance not achieved. To improve balance, the size of the treatment group (those who experienced restrictive housing) and the size of the untreated comparison group (those who did not experience restrictive housing) were each reduced by 50 percent. This resulted in a treatment group of 2,011 and an untreated comparison group of 6,713. Additionally, the model was specified to include only age, race, prior criminal history, and violent offense of commitment. Propensity scores were recalculated and ranged from 0.05 to 0.48, achieving balance. The treatment and untreated comparison groups were matched using one-to-one nearest-neighbor matching without replacement (this strategy ensures that only one untreated case is matched to a treated case). This resulted in 2,011 matched cases, for a total sample of 4,022. Before matching, there were statistically significant differences between mean propensity scores for race, age at release, and violent commitment offense. Following matching, these differences were no longer statistically significant. The two groups were approximately 17 percent white, with a mean age of 34 at release, and nearly half were incarcerated for a violent offense (48.4 percent for the treatment group, and 46.8 percent for the untreated comparison group).

To examine the impact of restrictive housing on postrelease offending behaviors, the study authors measured recidivism within 3 years of release. Recidivism was defined as rearrest, reconviction, reincarceration, reincarceration for a technical parole violation, and reincarceration for a new commitment. Both counts of the criminal justice–related event and time to event were measured. Reincarcerations were also categorized based on the offense type (e.g., violent, weapons, property, drug, other, and community supervision violation). Individuals who were reincarcerated for more than one offense were categorized into the highest offenses category (for example, if an individual was reincarcerated for a property and violent offense, they were categorized as a violent offense). The CrimeSolutions review of this study focused on the effects of the restrictive housing on number of rearrests, number of reconvictions, number of reincarcerations due to technical violations, number of reincarcerations due to a new offense, and time to rearrest. No subgroup analysis was conducted.

These sources were used in the development of the program profile:

These sources were used in the development of the program profile: