Disciplinary Segregation’s Effects on Inmate Behavior: Institutional and Community Outcomes

Abstract

Disciplinary segregation (DS) is practiced in a variety of correctional settings and a growing body of research explores its subsequent effects among offenders. The present study contributes to this literature by analyzing the impact of short-term DS on violent infractions and community recidivism among a sample of inmates in Washington State. We assessed the impact of DS on these outcomes from deterrence and stain theory perspectives while controlling for social support variables such as visitations and correctional programming. Mentally ill offenders were excluded, as their abilities to make rational choices may be inconsistent with deterrence theory. Results show DS does not significantly affect post-DS infractions. Social supports significantly reduced inmates’ odds of violent infractions while incarcerated. Community models indicate no substantive differences between the DS and non-DS groups on post-prison convictions 3 years after release. Overall, DS exhibited limited effects on offenders’ institutional or community outcomes.

Keywords

disciplinary segregation violent infractions recidivism deterrence strain social support

Corrections personnel across the United States use inmate isolation techniques to maintain or reestablish institutional security (Labrecque & Mears, 2019; Mears et al., 2019; Morgan et al., 2016; Shames, Wilcox, & Subramanian, 2015). Most prisons and jails in the United States utilize some type of solitary confinement (SC) to achieve this objective. Recent institutional “snapshot” data indicate approximately 80,000 prisoners are in SC on any given day (Shames et al., 2015; Stephan, 2008), and use of SC increased by 42% between 1995 and 2005 (Shames et al., 2015). In 2010, the VERA Institute initiated research partnerships with selected state correctional authorities to more closely study the implementation of SC policy among selected U.S. prisons (VERA Institute of Justice, n.d.). These research efforts articulated two areas of concern among prisoners’ rights advocates and scholars, namely, the overreliance of an extreme form of punishment—isolation—to maintain institutional control and the negative consequences of isolation on offenders after completing SC.

Such concerns underscore distinct areas of scholarship on the effects of inmate isolation within correctional facilities. Consistent with the abovementioned themes, one area emphasizes the negative psychological effects of extreme applications of this isolation technique (e.g., Grassian & Friedman, 1986; Haney, 2003, 2018; Haney & Lynch, 1997; Rhodes, 2004; Shalev, 2009). While the implications of such intensive isolation on inmates prompt legitimate cause for concern, drawing broad conclusions from such cases is not warranted on scientific grounds. Some of these works are qualitative studies relying on small samples of individuals isolated from others for years or even decades in a few instances.

Recent research suggests that the impact of inmate isolation on mental health is modest at best when studies evaluating this relationship possess a high degree of methodological rigor (Morgan et al., 2016). In addition, criminal justice scholars’ renewed interest in investigating the effects of inmate isolation on inmates’ subsequent behaviors within prison and in the community has generated recent work on the topic. The simplicity of such an exercise belies its complexity. Inmate isolation ranges from 24 hours or less to 23 hours of isolation per day with limited physical activity (e.g., supermaximum confinement). Recent evidence indicates that correctional authorities typically segregate inmates away from the general prison population for considerably briefer periods of time than “supermax” confinement (e.g., Barak-Glantz, 1983; Beck, 2015; Labrecque & Mears, 2019; Mears, 2013; Mears & Bales, 2010); yet academic studies assessing supermax’s effects on inmates have outpaced that of more routine applications of SC (e.g., Butler, Steiner, Makarios, & Travis, 2017; Lovell, Johnson, & Cain, 2007; Mears & Bales, 2009; Mears & Reisig, 2006; Morgan et al., 2016; Pizarro & Narag, 2008; Smith, 2006).

An emerging body of scholarship shows that limited durations of inmate isolation from the general population do not significantly affect subsequent inmate misconduct (Huebner, 2003; Labrecque, 2015, 2019; Lucas & Jones, 2017; Morris, 2016). Other evidence indicates that longer term stays in SC have limited (Lovell et al., 2007; Mears & Bales, 2009) or null effects (Clark & Duwe, 2018) on recidivism in the community after release. The present study contributes to this growing body of literature by assessing the effects of short-term inmate isolation on these two outcomes. We focus on disciplinary segregation (DS), conceptualizing this type of SC as punishment for inmate misconduct, and examine its effects from deterrence and strain theory perspectives net of social support variables such as visitations and rehabilitative programming.

Literature Review

SC: Theoretical Perspectives and Prior Research

Most states define SC as the physical isolation of an inmate through segregation from the general population for a specified period of time (Metcalf et al., 2013), with heavily restricted human contact and limited physical activity (Fathi, 2015; Haney, 2003; Pizarro & Narag, 2008; Shalev, 2015; Shames et al., 2015). SC is often used interchangeably with restrictive segregation (RS), administrative segregation (AS), DS, special management unit (SMU), protective custody (PC), and security housing unit (SHU) (Fathi, 2015; Mears et al., 2019; Shames et al., 2015). Regardless of the variation in such labels or definitions, corrections officials are ultimately concerned with using isolation to maintain institutional order and facility security (Labrecque & Mears, 2019).

Theoretical perspectives: Deterrence, general strain, and social support theories

Scholars have limited theoretical understanding of how inmate isolation affects subsequent behavior. As Mears and Bales (2009) noted, studies of prisons, especially studies on DS and/or supermaxes, have not questioned the theoretical reasons why such segregated and punitive housing might either increase or decrease violent misconduct in prisons and recidivism following release. When theory is applied, deterrence or general strain theory (GST) perspectives are often used to explain the effects of inmate isolation on institutional or community outcomes (Barak-Glantz, 1983; Butler et al., 2017; Labrecque, 2015, 2019; Lucas & Jones, 2017; Mears & Bales, 2009; Morris, 2016) as well as inmate misconduct in general (e.g., Blevins, Johnson Listwan, Cullen, & Lero Jonson, 2010; Listwan, Sullivan, Agnew, Cullen, & Colvin, 2013; Mears & Bales, 2009; Morris, 2016; Morris, Carriaga, Diamond, Piquero, & Piquero, 2012). In keeping with this literature, we apply such perspectives to the current study of DS on institutional and community outcomes. Although we do not test these theoretical frameworks per se, we summarize them to provide context for understanding DS experiences and offender behavior both in and out of the prison environment.

Deterrence theory

Hypotheses emanating from deterrence theory (Beccaria, 1764/1986) predict that offenders will not engage in crime if they are certainly, swiftly, and severely punished by the criminal justice system (specific deterrence). In turn, public knowledge about official punishments of offenders is expected to deter others from engaging in the same behavior thereby deterring crime (general deterrence). Prisoners tend to have a negative view of the SC experience (Haney, 2003; Haney & Lynch, 1997; Kurki & Morris, 2001; Mears & Reisig, 2006; Valera & Kates-Benman, 2016) possibly because DS is utilized as a type of official punishment when offenders engage in misconduct or violate rules (Shames et al., 2015; Smith, 2006). Specific deterrence theory anticipates that offenders who experience sufficient pains through exposure to DS for violating regulations would not engage in subsequent misbehavior (Mears & Reisig, 2006). Given that prisons, and DS placement in particular, are structured to be a severe punishment and that certainty of punishment and the swiftness with which it is meted out would appear to be more likely in a prison environment than in the larger community, it would not be surprising if it deterred offenders from committing further infractions while incarcerated and reduced their recidivism upon release to the community.

GST

Certainly, imprisonment itself is a painful event and offenders are often exposed to experiences and harsh environments that may lead to “strain” for them (Listwan et al., 2013; Morris et al., 2012). The isolation, restrained social contact, and exposure to additional deprivations associated with SC may not enhance prison security and safety; instead, the exposure to SC may result in several unintended adverse outcomes, such as mental health problems, violence and misconduct, and victimization. In this respect, GST could be employed to explain offenders’ post-SC criminogenic engagement.

As opposed to deterrence theory’s assumption that operating SC as a means of punishment in correctional settings would be beneficial, strain perspectives entertain the possibility that such a punishment could backfire, as such exposure might lead to the breakdown of the relationship with fellow offenders, the staff, family, and friends (Kurki & Morris, 2001). Moreover, being isolated in SC housing might exacerbate offenders’ physical and psychological harm including mental health issues, depression, self-injuries, and suicides (Dye, 2010; Listwan, Colvin, Hanley, & Flannery, 2010; Lovell, 2008; Shalev, 2015; Shames et al., 2015; Smith, 2006; Toch, 2003).

Social support theory

In addition to the potentially strain-inducing effects of SC placement, the offender also loses out on positively valued stimuli such as participation in programs and the opportunity to interact with friends and family from outside of the institution (e.g., cognitive behavioral, counseling, educational and vocational programs, religious activities, calls, letters, and visitation). Cullen and colleagues have argued that social support serves as a protective factor for reducing not only crime and recidivism but also fear of crime and victimization risk (Cullen, 1994; Cullen, Wright, & Chamlin, 1999; Hochstetler, DeLisi, & Pratt, 2010). The concept of social support contains two primary dimensions: instrumental and expressive (Cullen, 1994). Instrumental social support involves establishing relationships that enhance an individual’s capacity to achieve a desired goal (e.g., getting a job, obtaining financial assistance, receiving training). Expressive social support refers to the intrinsic quality of relationships, such as meeting one another’s needs for love, affection, and emotional validation. Both types of social support have formal and informal dimensions.

Treatment and intervention programs provided by correctional institutions are viewed by offenders as offering both formal and informal instrumental support, as well as formal and informal expressive social support. Many well-designed and well-implemented rehabilitation programs characterized as formal sources of institutional social support have demonstrated efficacy at reducing inmate misconduct and recidivism (see Andrews & Bonta, 2010; Cullen, 2013; MacKenzie, 2000; Petersilia, 2003; Smith, Gendreau, & Swartz, 2009).

Informal expressive social support such as inmates receiving calls, visits, and letters from significant others while incarcerated or having family members place money in the inmates’ commissary accounts also reduces these behaviors (Cochran & Mears, 2013; Duwe & Clark, 2013, 2014; Mears, Cochran, Siennick, & Bales, 2012; Tewksbury & DeMichele, 2005)—in some instances even more than formal institutional supports (Woo et al., 2016). Inmates experiencing short-term DS also have access to informal and formal sources of social support in prisons that are expected to decrease subsequent negative behaviors.

Research on Post-SC Behavioral Outcomes: Misconduct and Recidivism

An emerging body of scholarly research on the impact of both long-term and short-term inmate isolation on correctional and community outcomes suggests it has little to no effect on subsequent misconducts and community recidivism. Lovell et al. (2007) found that the timing of supermax inmates release to the community affected recidivism to a greater degree than supermax inmate status per se. Using a retrospective matched control design of Washington state inmates, results showed supermax inmates directly released to the community from prison had significantly higher odds of felony recidivism in the community, as well as shorter time to first reoffense in the community, relative to their non-supermax counterparts. Supermax inmates who returned to the general prison population for 90 days or more before release to the community (later-release) exhibited recidivism rates comparable with that of general population inmates.

Mears and Bales (2009) examined the effect of supermax incarceration on community-based recidivism (operationalized as a new conviction for a felony crime) to determine whether supermax incarceration affects criminal behavior after release. In their sample of Florida inmates, they employed propensity score modeling to find appropriate comparison subjects and found that supermax incarceration increased rates of violent recidivism but had no effect on other types of recidivism (e.g., property, drug). Another study employing propensity score matching (PSM) between supermax and general population inmates in Ohio also found a null effect of this type of confinement on rearrests or returns to prison up to 7 years after release (Butler et al., 2017).

Clark and Duwe (2018) evaluated the impact of inmate isolation while incarcerated on community recidivism (returns to prison due to revocation, new arrest, and new felony-level conviction) up to 3 years after release. Relative to inmates who spent no time in isolation, higher proportions of a sentence served in isolation significantly increased the incidence of revocation but did not affect new arrests or new felony convictions. In contrast to Lovell et al. (2007) but consistent with Mears and Bales (2009), direct release from the prison isolation unit to the community did not affect revocations or new crimes.

Systematic empirical study of shorter term stays in SC on subsequent inmate misconduct is also a relatively recent development in the corrections literature (e.g., Huebner, 2003; Labrecque, 2015, 2019; Lucas & Jones, 2017; Morris, 2016). Generally speaking, studies assessing short-term isolation on subsequent inmate misconduct show a null effect. For example, Huebner (2003) explored the effect of rewards and punishments through remunerative and coercive controls on subsequent offender violence based on an administrative control approach. Using a sample of 4,168 male offenders in 185 state prisons, findings indicated that DS as a coercive control strategy did not significantly predict future assaultive behavior, either on other offenders or correctional staff. Similarly, Labrecque’s (2015) analysis of DS on 14,311 offenders in the Ohio Department of Rehabilitation and Correction (ODRC) prison system indicated that DS did not influence subsequent misconduct in prison. However, offenders’ mental health statuses interacted with DS in some cases to influence certain types of subsequent misconduct in prison—but to a lesser degree than expected. The length of time an inmate spent in DS was not significantly related to subsequent misconduct in prison. Drawing on 9,016 inmates from a correctional authority in a Midwestern state, Labrecque (2019) also observed no significant effect of the number of days in isolation on either subsequent violent or non-violent inmate misconduct, but did find that inmates who served time in isolation had significantly reduced odds of a subsequent placement in SC during the study period.

Morris (2016) examined the effect of short-term DS on subsequent violent infractions among offenders incarcerated in a large southern state. Using propensity score matching to match SC offenders with non-SC offenders, findings indicated that short-term SC as a punishment for violent misconduct did not significantly affect the odds of future violent misconduct while incarcerated. Similarly, Lucas and Jones (2017) examined the effects of DS on institutional rule violations among a sample of 228 male offenders incarcerated by the Oregon Department of Corrections. They also found that DS did not significantly predict subsequent institutional misconduct.

We contribute to this emerging literature by assessing the effect of a certain type of SC—DS—on inmate misconduct for subsequent violent infractions as well as community-based recidivism after release. Our short-term time frame—anywhere from 30 to 60 days SC as a disciplinary response to an inmate’s infraction—provides an opportunity to measure the limited dosage effects of DS on offenders’ later institutional breaches and longer term law-violating behavior once they are residing in the community after release.

Generally speaking, correctional authorities determine DS placement through formal disciplinary procedures resulting from behavioral misconduct. Such disciplinary committees have discretion to determine guilt and levy sanctions to maintain prison security and safety; thus, disciplinary isolation may, as opposed to the primary purpose of the use of DS, constitute a threat to prison order and security if it intensifies prisoner maladjustment (Butler & Steiner, 2017; Howard, Winfree, Mays, Stohr, & Clason, 1994; Irwin, 1980; Sykes, 1958). On the contrary and consistent with deterrence theory, DS may decrease subsequent violent misconduct among inmates due to one’s desire to avoid such strain and stress in the future.

Both theories may be relevant. Morris (2016) noted,

. . . results suggest that exposure to solitary confinement for acts of violence tends to have no substantive effect on continuity in violent behavior among offenders engaging in an initial act of violence, but for approximately 2 percent of exposed offenders we would expect that exposure to solitary may have increased their tendency for continued violence and for an equivalent proportion, the opposite might be expected, for a total of 4 percent being affected by the solitary experience, on average. (p. 15)

Based on these mixed findings from previous studies, we hypothesize that exposure to short-term DS as a punishment will have little effect on offenders’ odds of committing a violent infraction while incarcerated and the odds of committing a new crime in the community.

Data and Method

Research Setting

A single state correctional system (Washington State Department of Corrections [WADOC]) is the focus of our analyses. WADOC (2015) considered SC “restrictive housing” or “a form of housing for individuals whose continued presence in the general population would pose a serious threat to life, property, self, staff or other inmates, or to the security or orderly operation of a correctional facility” (p. 1). WADOC employs four types of restrictive housing: AS, pre-hearing confinement, DS, and intensive management status/intensive treatment status. DS is a sanction that results from an institutional guilty finding for a serious misconduct and the only form of restrictive housing that is issued as a punishment, which by Washington State law cannot exceed 30 consecutive days (WADOC, 2013). Offenders who were placed in all types of segregation were identified through the WADOC data. While many in our sample were subject to various forms of segregation throughout their incarceration, pinpointing the exact dosage of non-DS placements in segregation proved difficult. Nonetheless, our preliminary models (not presented) included a measure of non-DS segregation, and removing the measure improved model fit considerably; thus, we do not include it in the final model (Table 3).

Data

We collected retrospective data on inmates who served DS sanctions between August 2008 and March 2016. Data elements included bed placements, misconducts, sanctions, visitation records, admission and release records, programming, demographics, and raw risk/needs assessment scores. Other data were obtained from the Washington State Administrative Office of the Courts (AOC) to measure new crimes after release.

Sample 1

We test the institutional and community effects of DS with separate samples of individuals. Sample 1 includes all individuals who were incarcerated at any point between August 1, 2008, and March 31, 2016 (N = 82,201). This timeframe was chosen for its increased accuracy in agency records due to a new records system implemented in 2008 and for the availability of the needs assessment, also implemented in 2008. Individuals convicted of capital crimes; sentenced to death, life, or life without parole; and individuals who were deported or deceased during our study period were excluded as they had no opportunity for release to the community. Individuals identified as having a mental health disorder were excluded, as their decreased abilities to make rational choices conflict considerably with core assumptions of deterrence theory. Keeping them in our sample could potentially add unnecessary measurement error to our study. After these inmates were eliminated, the resulting sample contained 43,202 individuals.

Sample 1 was then limited to offenders who recorded a violent infraction in their lifetime (n = 3,144), creating a subsample to test the relationship between the exposure to DS and subsequent violent infractions. Each offender was assigned five 6-month intervals beginning on March 31, 2013, and ending on September 30, 2015. DS days were counted when maximum custody bed start dates coincided with sanction start dates for segregation. While this does not include the entire population of those sanctioned to DS due to the pragmatics of offender management, we have no reason to believe offenders sent to DS before the sanction finding was finalized or those sent to DS even a few days after the sanction finding was finalized were any different than those sent to DS on the same date as their sanction to segregation was finalized. There were 303 offenders who experienced DS, representing 9.6% of the 3,144 in the sample. Of these 303, there were 360 six-month periods that contained at least one DS placement, and the average number of days per 6-month interval was 18.3 days.

This sample shares demographic similarities with those of other studies. Prior research has used male offenders exclusively (Huebner, 2003; Lucas & Jones, 2017), and the average age in this sample (32.8 years old) is also similar to that of other studies in this area (e.g., Butler & Steiner, 2017; Butler et al., 2017; Cochran, Toman, Mears, & Bales, 2018; Walters, 2015). The percentage of Hispanics (19%) in the sample is comparable with those in other studies (Butler & Steiner, 2017; Walters, 2015). There are differences between our sample and those in the literature, however. While the majority of DS offenders in prior research samples are Black (Butler & Steiner, 2017; Butler et al., 2017; Cochran et al., 2018; Walters, 2015), the majority of DS offenders in the current study are White.

Sample 2

In the second sample, which we use to test the effects of DS on community-based recidivism, 834 offenders who experienced DS within our full study period were identified by matching maximum custody bed start dates with start dates for applied sanctions of segregation. Next, we matched them to the remaining offenders in our sample using the three-to-one nearest neighbor method. Nearest-neighbor matching involves selecting treatment and control cases within a specified caliper, leading to similar propensities between the two groups. Each offender’s final exposure to DS during the 6-month period was selected to reduce the confounding effects of multiple in-and-out periods of DS. After matching, non-DS offenders were assigned the start and end dates of DS of their closest match. Dependent variables include convictions for new felonies and misdemeanors of five offense types (i.e., violent, property, drug, sex, and other crimes). Counts began after the end date for DS and were converted into rates accounting for time eligible to recidivate.

Measures

Consistent with previous research and WADOC policy, we operationalize DS as “confinement to an isolated cell for the overwhelming portion of each day, often 23 hours a day, with limited human interaction and minimal, if any, constructive activity” (Shames et al., 2015, p. 2). Duration of DS was calculated as the number of days spent in a maximum custody bed¹ (e.g., for cases where the maximum custody bed start dates coincided with the disciplinary sanction start date, we took the difference between that date and the bed end date). To test the institutional effects of DS, all measurements were taken on the final day of the pre-defined 6-month interval. Over 50% of the intervals included days incarcerated for offenders. Days in DS were cumulated for each 6-month period to represent our primary test variable.

The dependent variable was whether or not a violent infraction was committed in the following 6-month period after DS. The independent variables include age, African American, Hispanic, gang affiliation, familial relationship influence, drug problems (within the last 6 months), risk class based on the institution’s classification, length of time served (days), visit by blood family members, visit by non-blood family members, visit by spouse, vocational programming, rehabilitation programming, family meeting programming, volunteer work programming, and job programming. These control variables are consistent with prior research on disciplinary infractions (Huebner, 2003; Labrecque, 2015; Morris, 2016). These data are set forth in Table 1.

Table 1.

Sample 1 Descriptive Statistics: 15,720 Pooled Observations, 3,144 Offenders.

Variables	M/%	SE
Outcome measure
Violent infractions	0.05	0.00
Static predictors
Age (year)	32.87	0.08
Black	0.23	0.00
Hispanic	0.19	0.00
Gang affiliation	0.10	0.00
Risk class
Low	5.1
Moderate	22.1
High nonviolent	30.3
High violent	42.6
Dynamic predictors
Days disciplinary segregation	0.42	0.03
Familial relationship influence	0.29	0.00
Drug problems	0.37	0.00
Length of time served (days)	67.39	0.65
Visits by blood family members	0.13	0.00
Visits by non-blood family members	0.05	0.00
Visits by spouse	0.02	0.00
Rehabilitation program (hours)	10.83	0.53
Vocation program (hours)	14.13	0.50
Job program (hours)	111.82	2.17
Family meeting program (hours)	0.60	0.08
Volunteer work (hours)	0.29	0.03

Method

This study employs two separate analyses to understand how DS affects offenders’ conduct. For the first analysis, we tested institutionalized effects of SC using a generalized estimating equations (GEE) technique to examine the effects of DS on subsequent violent infractions.² The second analysis uses Sample 2 to examine new crimes committed in the community after release from prison. The effect of exposure to DS on recidivism outcomes in the analyses was assessed using propensity score modeling.³ The procedure for the PSM analysis included three steps. First, we selected 20 covariates for the matching procedure that were theoretically relevant and significantly related to DS at a bivariate level (Haney, 2003; Morris, 2016). A binary logistic regression model was then used to estimate propensity scores. A set of demographic measures were included as covariates: age (in years), race (dummy variables for each African American and Hispanic = 1; other = 0), marital status (married: yes = 1; no = 0), gang affiliation (gang: yes = 1; no = 0), visitation (ever had visits during the study period: yes = 1; no = 0), risk class based on the institution’s classification (ranged from 0 to 3 where 0 refers to low risk class and 3 refers to high violent risk class), length of time incarcerated (ranged from 1 to 8 where 1 refers to less than 1 month and 8 refers to more than 8 years time served), prior incarceration history (ranged from 0 to 4 where 0 refers to no history and 4 refers to more than 7 times), lifetime types of convictions (dummy variables for each of violent, property, drug, and sex convictions = 1; no = 0), violent infractions history (violent infractions: yes = 1; no = 0), drug problems (yes = 1; no = 0), and institutional programs (vocational, rehabilitation, job, family meeting, life skills: hours of participation in these programs). Before the match, area under the curve (AUC) estimate was .81, which showed the model was strongly predicting group assignment.

As shown in Table 2, our PSM diagnostics indicate a substantial reduction in selection bias, and a robust balance between the two groups was achieved on nearly all covariates in terms of bivariate difference tests, standardized differences (STD), and AUC (see Austin, 2008; Rice & Harris, 2005; Rosenbaum & Rubin, 1985).

Table 2.

Sample 2 Balancing Statistics (N = 43,202).

Variable	Before PSM				After PSM (3 to 1)
Variable	Non-DS mean	DS mean	t for difference in means	STD	Non-DS mean	DS mean	t for difference in means	STD
Age	30.41	27.26	8.80**	2.7	27.54	27.26	0.67	0.3
Black	0.16	0.25	−5.84**	54.5	0.26	0.25	0.63	5.2
Hispanic	0.12	0.19	−5.10**	52.2	0.18	0.19	−0.98	6.6
Married	0.65	0.55	5.62**	42.2	0.56	0.55	0.49	4.6
Gang affiliation	0.17	0.47	−17.61**	148.3	0.45	0.47	−0.89	8.3
Visits	0.45	0.55	−5.94**	40.4	0.58	0.55	1.55	12.2
Risk class
Low	0.10	0.05	7.78**	78.7	0.05	0.05	0.52	9.6
Moderate	0.26	0.20	4.24**	33.7	0.20	0.20	0.35	3.6
High nonviolent	0.33	0.28	3.58**	26.5	0.28	0.28	0.32	2.8
High violent	0.30	0.48	−9.94**	75.0	0.46	0.48	−0.79	6.4
Prior incarceration	0.73	0.95	−4.80**	14.3	1.00	0.95	0.99	3.0
Violent conviction	0.41	0.58	−9.95**	70.0	0.56	0.58	−1.01	8.1
Property conviction	0.30	0.35	−2.81**	17.8	0.35	0.35	−0.01	0.0
Drug conviction	0.51	0.44	3.84**	28.2	0.47	0.44	1.12	12.1
Sex conviction	0.11	0.13	−1.68	18.6	0.15	0.13	1.08	16.7
Length of time served	4.19	5.58	−22.45**	35.1	5.67	5.58	1.21	2.7
Drug problems	0.31	0.32	−0.64	4.7	0.32	0.32	0.15	0.0
Violent infraction records	0.15	0.59	−25.66**	226.9	0.56	0.59	−1.43	12.3
Vocational hours	0.82	1.28	−9.25**	19.1	1.29	1.28	0.13	0.5
Rehabilitation hours	1.97	2.30	−5.11**	10.1	2.41	2.30	1.56	3.4
Job hours	2.03	2.48	−7.08**	12.9	2.64	2.48	2.19*	4.6
Family meeting hours	0.28	0.36	−3.68**	19.1	0.36	0.36	−0.25	0.0
Life skills hours	1.20	1.97	−14.26**	34.6	1.96	1.97	−0.15	0.4
AUC		0.81				0.52
n	42,368	834			2,240	834

Note. PSM = propensity score matching; DS = disciplinary segregation; STD = standardized differences; AUC = area under the curve.

p < .05. **p < .01.

Results

Misconduct Outcomes

Findings for the binary multinomial GEE model using the first sample are presented in Table 3. The time variable was ordinal (using the first observation as the reference category) and results show no significant change in the odds of committing a violent infraction across the four intervals. In addition, days in DS did not significantly change violent infractions over time (the initial log odds = −.01, p = .152). That is, being exposed to DS did not significantly affect violent infractions in the subsequent 6-month period, a finding consistent with prior research (Huebner, 2003; Labrecque, 2015, 2019; Morris, 2016). Such a finding is inconsistent with both deterrence and strain theory perspectives of DS on subsequent violent infractions while incarcerated.

Table 3.

Binary Generalized Estimating Equations Predicting Violent Infractions (n₁ = 3,144).

Variables	b	SE	Exp(B)
Intercept	−2.15***	.22	0.12
Time points
Observation 5	−0.06	.13	0.94
Observation 4	0.09	.13	1.09
Observation 3	0.04	.13	1.04
Observation 2	−0.19	.14	0.83
Observation 1 (reference)	—	—	—
Static predictors
Age	−0.05***	.01	0.95
Black	−0.24**	.09	0.79
Hispanic	0.05	.09	1.05
Gang affiliation	−0.02	.10	0.98
Risk class	0.02	.04	1.02
Dynamic predictors
Days disciplinary segregation	−0.01	.01	0.99
Familial relationship influence	−0.11	.08	0.89
Drug problems	0.18**	.07	1.20
Length of time served (days)	0.01***	.00	1.01
Visits by blood family members	−0.21*	.10	0.81
Visits by non-blood family members	0.12	.14	1.13
Visits by spouse	−0.28	.29	0.76
Rehabilitation program	−0.16***	.04	0.86
Vocation program	−0.01	.06	0.99
Job program	−0.06**	.03	0.94
Family meeting program	−0.39	.24	0.69
Volunteer work	−0.33	.27	0.72

Note. Institutional programs were originally measured in hours, but they were converted into ordinal or dichotomous level due to high skewness of the distribution.

p < .05. **p < .01. ***p < .001.

Furthermore, age, African American race, presence of drug problems, length of time served, visits by blood family members, rehabilitation, and job program involvement significantly affected the odds of committing a subsequent violent infraction. As in other studies, we found an “aging-out effect,” in that for every year older, there was a 5% decrease in the odds of committing a violent infraction (odds ratio [OR] = .95, p < .001) (Bench & Allen, 2003; Gendreau, Goggin, & Law, 1997; Griffin & Hepburn, 2006; Howard et al., 1994; Huebner, 2003; Lahm, 2008; Steiner & Wooldredge, 2008, 2009, 2014; Walters, 2015; Wooldredge, Griffin, & Pratt, 2001).

Our measure of race indicated that African American offenders had a 21% decrease in the odds of committing a violent infraction after exposure to DS, compared with other racial or ethnic groups over time (OR = .79, p < .009). This finding differs from previous studies showing an enhanced effect of DS on violent misconduct among minority race offenders (e.g., Griffin & Hepburn, 2013; Huebner, 2003; Steiner & Wooldredge, 2008, 2009). To further explore why African American offenders had lower odds of violent infractions, we conducted difference tests on measures of risk and protective factors for inmate infractions (data not shown). Protective factors measured social support (rehabilitation program, vocation program, job program, family meeting program, volunteer work, and visitations), and risk factors measured gang affiliation, drug problems, and prior incarcerations. Results from the difference tests show that African American offenders have significantly fewer drug problems, fewer gang affiliations, and more prior incarceration experiences than other race/ethnicity groups, but did not differ from other offenders on social support measures. This intriguing finding suggests further exploration is warranted.

Offenders with drug problems had a 20% increase in the odds of committing a violent infraction in the subsequent months or years while incarcerated (OR = 1.20, p < .01). We note that this measure of drug use history from the offender needs assessment is inexact and does not mean that offenders were currently involved in drugs while incarcerated or that they had an addiction to drugs prior to incarceration, rather it was just a question about use of an array of drugs within 6 months of incarceration. We speculate that those who were using drugs prior to incarceration may be more deviant and more likely to engage in activity or behavior that would land them in DS.

Length of time served was found to be a significant predictor of committing a subsequent violent infraction. That is, for every day incarcerated, offenders had a 1% increase in the odds of engaging in a violent infraction (OR = 1.01, p < .001). This finding is consistent with many prior studies (Cunningham, Sorensen, & Reidy, 2005; Gover, Pérez, & Jennings, 2008; Sorensen & Wrinkle, 1996; Steiner, Butler, & Ellison, 2014; Zamble, 1992).

Surprisingly, offenders who were identified as affiliated with a gang were not significantly more likely to commit a violent infraction in the subsequent months or years while incarcerated (log odds = −.02, p = .832), a finding inconsistent with prior studies (see Griffin & Hepburn, 2006; Huebner, 2003; Worrall & Morris, 2012). For example, Griffin and Hepburn (2006) and Worrall and Morris (2012) found that gang affiliation is associated with violence among offenders. It is possible that this finding might be explained by an overuse of the appellation “gang member” or by the fact that if it did apply at one time, it no longer does. On the contrary, we used “gang affiliates” and not solely gang members, which may have different meanings and/or be measured differently in other jurisdictions.

Findings also indicated that social support—measured as visits by blood-related family members, participating in rehabilitation programs, and participating in job programs—significantly reduced the odds of committing a violent infraction. Offenders who had at least one visit by blood-related family members had a 19% decrease in the odds of engaging in a violent infraction (OR = 0.81, p = .041), offenders who participated in rehabilitation had 14% fewer odds of engaging in a violent infraction (OR = 0.86, p < .001), and offenders who attended a job program had a 6% reduction in the odds of engaging in a violent infraction (OR = 0.94, p < .05).

Scholars have suggested the use of remunerative controls (e.g., work assignments, prison programming) rather than the use of coercive controls (e.g., segregation) to enhance inmates’ adherence to prison rules and garner their support of the prison system (Butler & Steiner, 2017; Colvin, 1992; Huebner, 2003). Provisions of instrumental as well as expressive social supports such as clinical interventions and rehabilitative programs may also enhance inmates’ willingness to recognize the legitimacy of authority figures inside prisons. These findings are consistent with prior studies of visitation, rehabilitation, and job programs on future offending (Cullen & Jonson, 2012; Duwe & Clark, 2013; Mears et al., 2012; Mears, Wang, & Bales, 2014; Wilson, Bouffard, & MacKenzie, 2005).

Recidivism Outcomes

We used our second sample (described above) to assess the effects of DS on community-based recidivism. We compared mean differences in conviction rate outcomes between the two groups following the PSM match using Student’s t tests. A 3-year follow-up period was established based on prior literature that has shown that recidivism occurs primarily within the first 3 years following release (Griffin & Hepburn, 2006; Petersilia, 2003).⁴ As shown in Table 4 and consistent with findings from other research (see Hamilton & Campbell, 2013; Petersilia, 2003), the highest risk period for conviction rates is within the initial year following release and the trend gradually decreased over time.

Table 4.

Cumulative Yearly Conviction Rate Outcome Comparisons by Study Group (n₂ = 3,074).

Outcome	Year 1%			Year 2%			Year 3%
Outcome	Non-DS	DS	p value	Non-DS	DS	p value	Non-DS	DS	p value
Felony	1.05	1.10	.853	0.82	0.65	.451	0.46	0.49	.800
Violent Felony	0.25	0.39	.146	0.24	0.20	.660	0.14	0.16	.753
Property Felony	0.38	0.40	.906	0.22	0.18	.504	0.14	0.20	.558
Drug Felony	0.25	0.16	.051	0.15	0.15	.930	0.10	0.07	.321
Sex Felony	0.01	0.00	.496	0.01	0.00	.046*	0.00	0.00	.549
Violent Msdr	0.14	0.12	.568	0.09	0.09	.861	0.07	0.09	.419
Property Msdr	0.19	0.24	.284	0.12	0.15	.256	0.08	0.14	.106
Drug Msdr	0.19	0.021	.605	0.12	0.18	.051	0.09	0.15	.064
Sex Msdr	0.00	0.00	.529	0.00	0.00	.085	0.00	0.00	.426
Other Msdr	0.12	0.14	.474	0.07	0.13	.011*	0.05	0.09	.149

Note. “Felony” represents the sum of violent, property, drug, and sex felony convictions. DS = disciplinary segregation; Msdr = misdemeanor.

p < .05.

Results show that exposure to DS did not affect offender conviction rates within the 3-year follow-up period when compared with non-SC subjects. Aside from felony sex crimes and “other” misdemeanors measured at Year 2, DS and non-DS offenders did not significantly differ for all remaining convictions. Generally speaking, exposure to DS while incarcerated did not affect subsequent offending among former inmates once they re-enter their communities. These results are consistent with other studies showing institutional segregation of inmates has little effect on recidivism after release (e.g., Clark & Duwe, 2018), but contradict others showing higher rates of some types of felony recidivism among offenders who experienced SC while incarcerated (Lovell et al., 2007; Mears & Bales, 2009). Our findings are not consistent with either deterrence or strain theory perspectives.

Discussion and Conclusion

Scholarly debate on the impact of inmate segregation has a lengthy history in corrections research (King, 1999; Listwan, Jonson, Cullen, & Latessa, 2008; Morgan et al., 2016; Smith, 2006). Drawing on studies suggesting a deleterious effect on offenders’ physical and mental health (Arrigo & Bullock, 2008; Haney, 2003; Haney, 2018; Haney & Lynch, 1997), scholars have initiated new studies to more thoroughly examine the effects of isolation on offenders in correctional institutions (Browne, Cambier, & Agha, 2011). We assessed the short-term and long-term effects of DS on institutional and community outcomes among a cohort of offenders incarcerated by the WADOC. Using two samples of offenders, we employed a multi-level regression technique (GEE) and a propensity score model to examine whether the exposure to DS affects institutional outcomes such as violent infractions and community outcomes such as felony and misdemeanor conviction rates. Results indicate that exposure to DS does not significantly affect involvement in violent infractions, a finding consistent with similar investigations of isolation on subsequent inmate misconduct (Huebner, 2003; Labrecque, 2015, 2019; Lucas & Jones, 2017; Morris, 2016). In terms of institutional behavior, being older, African American, participating in visits with family, and participating in rehabilitation and jobs programming reduced the odds of committing a violent infraction over time; possession of a drug history increased those odds. For community outcomes, results showed no difference on subsequent felony conviction rates in the community between the DS group and non-DS groups, with the exception of those convicted of sex offenses and misdemeanors (and only in the second year of release to the community). This finding is consistent with other research showing a limited effect of inmate isolation on recidivism (Clark & Duwe, 2018; Lovell et al., 2007; Mears & Bales, 2009).

In addition to these findings, our results are consistent with research showing that visitation reduces offenders’ infractions while incarcerated and improves their outcomes when re-entering the community (Cochran & Mears, 2013; Duwe & Clark, 2013, 2014; Mears et al., 2012; Tewksbury & DeMichele, 2005). A recent meta-analysis examining the predictors of restrictive housing shows that offenders receiving this type of segregation are likely to be younger and have more extensive criminal records, poorer institutional behavior, and more criminogenic needs than the general prison population (Labrecque, 2018a). In light of our findings, offenders in DS potentially need more rehabilitation and social support than general population offenders to enhance prosocial adaptations to prison and increase facility safety (Labrecque & Smith, 2019). In sum, DS exhibited mostly null effects on both institutional and most community outcomes, while social supports—both instrumental and expressive—significantly reduced the odds of subsequent infractions for violence while incarcerated as well as recidivism in the community.

Study Limitations

As with all studies, ours possesses some weaknesses that must be acknowledged. The use of the GEE technique strengthened the rigor of the study by allowing assessment of repeated measures over a period of time—providing the opportunity to review effects multiple times. However, GEE only provides an average of effects over the number of intervals and that might mask findings associated with individual intervals. Moreover, the selection of 6-month intervals for this study was based on the need to ensure that a risk assessment had been completed for each interval, but it might be that shorter or longer periods between intervals would have resulted in different findings.

Furthermore, this study uses a quasi-experimental design rather than a true experimental design; thus, there may be variables that should have been included in our covariate matrix. While the propensity score modeling technique mitigates the vulnerabilities of a quasi-experimental design, the technique has some challenges. Foremost among them is unavoidable selection bias due to lack of appropriate measures for all offenders, excluding some inmates from participating in our study. Nonetheless, the large sample size, use of PSM to provide valid matched cases to our treatment subjects, and 7-year study period underscore a robust set of findings that will enrich the debate on the use of DS.

Finally, we excluded inmates with a designated mental health condition as their cognitive abilities were potentially mismatched with the rationality assumptions of deterrence theory. While this may have introduced some bias into our findings (recent research shows that inmates with psychological problems comprise up to 30% of inmates in some restrictive housing units—see Beck, 2015), the results of our study are consistent with those of similar research designs including mentally ill offenders (e.g., Clark & Duwe, 2018; Labrecque, 2015, 2019) as well as those that did not include these offenders (e.g., Butler et al., 2017; Lucas & Jones, 2017; Mears & Bales, 2009; Morris, 2016). Prior research indicates that the greater the methodological rigor of study designs assessing SC and subsequent inmate behavior, the weaker its effects (Morgan et al., 2016) and our results are consistent with this finding. Yet, if serious mental illness (SMI) functions as a risk factor for entering SC (Labrecque & Smith, 2019), counterfactual research designs excluding inmates with these conditions may present an incomplete assessment of the relationship between SC and subsequent inmate misconduct. Future research should explore the causal linkages between mental illness, correctional confinement, and prison adjustment to advise correctional authorities as to best practices with this vulnerable population (Morgan et al., 2016) and researchers should adjust study designs accordingly.

Organizational Practice and Policies

Our findings replicate other studies showing inmate segregation does not substantively affect inmates’ subsequent infractions or reconvictions once in the community (e.g., Clark & Duwe, 2018; Labrecque, 2015, 2019; Lucas & Jones, 2017; Morris, 2016). Taken together, this emerging literature demonstrates that limited applications of inmate isolation from the general population of prisoners—30 days at a time, for example—satisfy short-term institutional goals such as facility and staff safety without posing substantive long-term psychological damage to offenders (Morgan et al., 2016). Moreover, our research validates prior studies showing offenders’ decreased risks of infractions and community recidivism with participation in social support activities (Duwe & Clark, 2013, 2014; Mears et al., 2012; Woo et al., 2016).

Corrections policies sanctioning offenders by curtailing their abilities to receive visits from family or participate in employment or other rehabilitative skill building programs inadvertently weaken a safe environment for staff and inmates. Instead, correctional authorities applying the principles of effective treatment among those currently within SC (Butler, Solomon, & Spohn, 2018; Labrecque, 2018b) or those at highest risk for this sanction (Labrecque & Smith, 2019) could enhance facility safety and inmates’ adjustment to prison in the following ways: (a) adjusting treatment dosage for these high risk offenders, (b) engaging criminogenic risk through rehabilitative programming, and (c) delivering services in a way congruent with neuroatypical cognitive functioning. Our study suggests that holding inmates accountable in prison while maintaining institutional control promoting facility safety is the product of both proximate responses to inmate misconduct (like short-term DS) and distal mechanisms such as sustained contact with social support activities enhancing their prosocial behaviors.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Laurie Drapela

Notes

Author Biographies

Youngki Woo is an assistant professor in the Department of Criminal Justice at The University of Texas Rio Grande Valley. His research interests include theory testing, correctional institutions, procedural justice and legitimacy of legal authorities, and quantitative methods. His work appears in journal such as Crime & Delinquency, Journal of Criminal Justice, and Journal of Offender Rehabilitation.

Laurie Drapela is an associate professor of Criminal Justice at Washington State University Vancouver. Her research areas include juvenile justice, therapeutic specialty court programs, and intersections of behavioral health, mental health, and criminal justice. Her work appears in Justice Quarterly, Crime & Delinquency, Youth & Society, Journal of Youth & Adolescence, and Deviant Behavior.

Michael Campagna, PhD, is a research associate for the Nebraska Center for Justice Research, and a faculty member at the University of Nebraska Omaha. He received his Ph.D. from WSU in 2017 and his MA from Indiana University of Pennsylvania in 2009. He has experience as a juvenile social worker and his research interests include correctional policy, risk and needs assessment, mental health, prevention, and reentry. His work appears in Criminal Justice and Behavior, The Prison Journal, and Journal of Law and Criminal Justice.

Mary K. Stohr, PhD, is a professor in the Department of Criminal Justice and Criminology at Washington State University. She has published over 90 academic works, including seven books and over 50 journal articles, primarily in the area of corrections. Her current research concerns the effects of solitary confinement and of marijuana law implementation. She served as Executive Director of the Academy of Criminal Justice Sciences for five years, and received the Founder’s Award from ACJS in 2009. Many moons ago she worked as a correctional officer and then a counselor in an adult male prison.

Zachary K. Hamilton, PhD, is an associate professor of Criminal Justice and Criminology at Washington State University and director of the Washington State Institute for Criminal Justice (WSICJ). His research interests include corrections, offender reentry, assessment, substance abuse treatment, and quantitative methods. His work appears in Justice Quarterly, Experimental Criminology, Criminal Justice and Behavior, and Criminology and Public Policy.

Xiaohan Mei is a PhD candidate in Criminal Justice and Criminology at Washington State University and senior research associate of the Washington State Institute for Criminal Justice (WSICJ). His research interests include the study of deviance and criminology, risk-needs assessment instrument validation, core correctional practices and program fidelity, management in corrections and organizational studies, quantitative methodology and psychometric analytic approaches. His work appears in The Prison Journal, Journal of Offender Rehabilitation and International Journal of Offender Therapy and Comparative Criminology.

Elizabeth Thompson Tollefsbol received her PhD from the Department of Criminal Justice and Criminology at Washington State University in December 2018. Her areas of interest include incarcerated veterans, gender responsivity and pathways research, offender assessment, and policy effectiveness within the justice system. She is also a Senior Research Associate with the Washington State Institute for Criminal Justice where, through funding by the U.S. Department of Justice, she has examined the implementation of Smart Justice programming in Washington State. Her work appears in Justice Quarterly and The Prison Journal.

References

Andrews

D. A.

Bonta

(2010). The psychology of criminal conduct (5th ed.). Los Angeles, CA: Anderson/LexisNexis.

Arrigo

B. A.

Bullock

J. L.

(2008). The psychological effects of solitary confinement on prisoners in supermax units: Reviewing what we know and recommending what should change. International Journal of Offender Therapy and Comparative Criminology, 52, 622-640.

Austin

P. C.

(2008). A critical appraisal of propensity-score matching in the medical literature between 1996 and 2003. Statistics in Medicine, 27, 2037-2049.

Barak-Glantz

I. L.

(1983). Who’s in the “hole?” Criminal Justice Review, 8, 29-37.

Beccaria

(1986). On crimes and punishments. Indianapolis, IN: Hackett. (Original work published 1764)

Beck

A. J.

(2015). Use of restrictive housing in US prisons and jails, 2011–12. Washington, DC: U.S. Department of Justice.

Bench

L. L.

Allen

T. D.

(2003). Investigating the stigma of prison classification: An experimental design. The Prison Journal, 83, 367-382.

Blevins

K. R.

Johnson Listwan

Cullen

F. T.

Lero Jonson

(2010). A general strain theory of prison violence and misconduct: An integrated model of inmate behavior. Journal of Contemporary Criminal Justice, 26, 148-166.

Browne

Cambier

Agha

(2011). Prisons within prisons: The use of segregation in the United States. Federal Sentencing Reporter, 24, 46-49.

10.

Butler

H. D.

Solomon

Spohn

(2018). Programming in restrictive housing: Considerations for improving outcome evaluations. Criminal Justice and Behavior, 45, 1174-1191.

11.

Butler

H. D.

Steiner

(2017). Examining the use of disciplinary segregation within and across prisons. Justice Quarterly, 34, 248-271.

12.

Butler

H. D.

Steiner

Makarios

M. D.

Travis

L. F.

III . (2017). Assessing the effects of exposure to supermax confinement on offender postrelease behaviors. The Prison Journal, 97, 275-295.

13.

Clark

V. A.

Duwe

(2018). From solitary to the streets: The effect of restrictive housing on recidivism. Corrections. Advance online publication. doi:10.1080/23774657.2017.1416318

14.

Cochran

J. C.

Mears

D. P.

(2013). Social isolation and inmate behavior: A conceptual framework for theorizing prison visitation and guiding and assessing research. Journal of Criminal Justice, 41, 252-261.

15.

Cochran

J. C.

Toman

E. L.

Mears

D. P.

Bales

W. D.

(2018). Solitary confinement as punishment: Examining in-prison sanctioning disparities. Justice Quarterly, 35, 381-411.

16.

Colvin

(1992). The penitentiary in crisis: From accommodation to riot in New Mexico. Albany: State University of New York Press.

17.

Cullen

F. T.

(1994). Social support as an organizing concept for criminology: Presidential address to the academy of criminal justice sciences. Justice Quarterly, 11, 527-559.

18.

Cullen

F. T.

(2013). Rehabilitation: Beyond nothing works. Crime and Justice, 42, 299-376.

19.

Cullen

F. T.

Jonson

C. L.

(2012). Correctional theory: Context and consequences. Thousand Oaks, CA: Sage.

20.

Cullen

F. T.

Wright

J. P.

Chamlin

M. B.

(1999). Social support and social reform: A progressive crime control agenda. Crime & Delinquency, 45, 188-207.

21.

Cunningham

M. D.

Sorensen

J. R.

Reidy

T. J.

(2005). An actuarial model for assessment of prison violence risk among maximum security inmates. Assessment, 12, 40-49.

22.

Duwe

Clark

(2013). Blessed be the social tie that binds: The effects of prison visitation on offender recidivism. Criminal Justice Policy Review, 24, 271-296.

23.

Duwe

Clark

(2014). The effects of prison-based educational programming on recidivism and employment. The Prison Journal, 94, 454-478.

24.

Dye

M. H.

(2010). Deprivation, importation, and prison suicide: Combined effects of institutional conditions and inmate composition. Journal of Criminal Justice, 38, 796-806.

25.

Fathi

D. C.

(2015). United States: Turning the corner on solitary confinement. Canadian Journal of Human Rights, 4, 167-177.

26.

Gendreau

Goggin

C. E.

Law

M. A.

(1997). Predicting prison misconducts. Criminal Justice and Behavior, 24, 414-431.

27.

Gover

A. R.

Pérez

D. M.

Jennings

W. G.

(2008). Gender differences in factors contributing to institutional misconduct. The Prison Journal, 88, 378-403.

28.

Grassian

Friedman

(1986). Effects of sensory deprivation in psychiatric seclusion and solitary confinement. International Journal of Law and Psychiatry, 8, 49-65.

29.

Griffin

M. L.

Hepburn

J. R.

(2006). The effect of gang affiliation on violent misconduct among inmates during the early years of confinement. Criminal Justice and Behavior, 33, 419-466.

30.

Griffin

M. L.

Hepburn

J. R.

(2013). Inmate misconduct and the institutional capacity for control. Criminal Justice and Behavior, 40, 270-288.

31.

Guo

Fraser

M. W.

(2015). Propensity score analysis: Statistical methods and applications (2nd ed.). Thousand Oaks, CA: Sage.

32.

Hamilton

Z. K.

Campbell

C. M.

(2013). A dark figure of corrections: Failure by way of participation. Criminal Justice and Behavior, 40, 180-202.

33.

Haney

(2003). Mental health issues in long-term solitary and “supermax” confinement. Crime & Delinquency, 49, 124-156.

34.

Haney

(2018). Restricting the use of solitary confinement. Annual Review of Criminology, 1, 285-310.

35.

Haney

Lynch

(1997). Regulating prisons of the future: A psychological analysis of supermax and solitary confinement. New York University Review of Law & Social Change, 23, 477-570.

36.

Hochstetler

DeLisi

Pratt

T. C.

(2010). Social support and feelings of hostility among released inmates. Crime & Delinquency, 56, 588-607.

37.

Howard

Winfree

L. T.

Jr. Mays

G. L.

Stohr

M. K.

Clason

D. L.

(1994). Processing inmate disciplinary infractions in a federal correctional institution: Legal and extralegal correlates of prison-based legal decisions. The Prison Journal, 74, 5-31.

38.

Huebner

B. M.

(2003). Administrative determinants of inmate violence: A multilevel analysis. Journal of Criminal Justice, 31, 107-117.

39.

Irwin

(1980). Prisons in turmoil. Boston, MA: Little Brown.

40.

King

R. D.

(1999). The rise and rise of supermax: An American solution in search of a problem? Punishment & Society, 1, 163-186.

41.

Kurki

Morris

(2001). The purposes, practices, and problems of supermax prisons. Crime and Justice, 28, 385-424.

42.

Labrecque

R. M.

(2015). The effect of solitary confinement on institutional misconduct: A longitudinal evaluation (Doctoral Thesis). University of Cincinnati, USA. NCJ # 249013. Retrieved from https://www.ncjrs.gov/pdffiles1/nij/grants/249013.pdf

43.

Labrecque

R. M.

(2018a). Taking stock: A meta-analysis of the predictors of restrictive housing. Victims & Offenders, 13, 675-692.

44.

Labrecque

R. M.

(2018b). Specialized or segregated housing units: Implementing the principles of risk, needs, and responsivity. In Dobson

(Ed.), Routledge handbook on offenders with special needs (pp. 69-83). New York, NY: Routledge.

45.

Labrecque

R. M.

(2019). Assessing the impact of time spent in restrictive housing confinement on subsequent measures of institutional adjustment among men in prison. Criminal Justice and Behavior. Advance online publication. doi:10.1177/0093854818824371

46.

Labrecque

R. M.

Mears

D. P.

(2019). Prison system versus critics’ views on the use of restrictive housing: Objective risk classification or ascriptive assignment? The Prison Journal, 99, 194-218.

47.

Labrecque

R. M.

Smith

(2019). Reducing institutional disorder: Using the inmate risk assessment for segregation placement to triage treatment services at the front end of prison sentences. Crime & Delinquency, 65, 3-25.

48.

Lahm

K. F.

(2008). Inmate-on-inmate assault: A multilevel examination of prison violence. Criminal Justice and Behavior, 35, 120-137.

49.

Liang

K. Y.

Zeger

S. L.

(1986). Longitudinal data analysis using generalized linear models. Biometrika, 73, 13-22.

50.

Listwan

S. J.

Colvin

Hanley

Flannery

(2010). Victimization, social support and psychological well-being: A study of recently released prisoners. Criminal Justice and Behavior, 37, 1140-1159.

51.

Listwan

S. J.

Jonson

C. L.

Cullen

F. T.

Latessa

E. J.

(2008). Cracks in the penal harm movement: Evidence from the field. Criminology & Public Policy, 7, 423-465.

52.

Listwan

S. J.

Sullivan

C. J.

Agnew

Cullen

F. T.

Colvin

(2013). The pains of imprisonment revisited: The impact of strain on inmate recidivism. Justice Quarterly, 30, 144-168.

53.

Lovell

(2008). Patterns of disturbed behavior in a supermax population. Criminal Justice and Behavior, 35, 985-1004.

54.

Lovell

Johnson

L. C.

Cain

K. C.

(2007). Recidivism of supermax prisoners in Washington State. Crime & Delinquency, 53, 633-656.

55.

Lucas

J. W.

Jones

M. A.

(2017). An analysis of the deterrent effects of disciplinary segregation on institutional rule violation rates. Criminal Justice Policy Review, 30, 765-787.

56.

MacKenzie

D. L.

(2000). Evidence-based corrections: Identifying what works. Crime & Delinquency, 46, 457-471.

57.

Mears

D. P.

(2013). Supermax prisons. Criminology & Public Policy, 12, 681-719.

58.

Mears

D. P.

Bales

W. D.

(2009). Supermax incarceration and recidivism. Criminology, 47, 1131-1166.

59.

Mears

D. P.

Bales

W. D.

(2010). Supermax housing: Placement, duration, and time to reentry. Journal of Criminal Justice, 38, 545-554.

60.

Mears

D. P.

Cochran

J. C.

Siennick

S. E.

Bales

W. D.

(2012). Prison visitation and recidivism. Justice Quarterly, 29, 888-918.

61.

Mears

D. P.

Hughes

Pesta

G. B.

Bales

W. D.

Brown

J. M.

Cochran

J. C.

Wooldredge

(2019). The new solitary confinement? A conceptual framework for guiding and assessing research and policy on “restrictive housing.” Criminal Justice and Behavior. Advance online publication. doi:10.1177/0093854819852770

62.

Mears

D. P.

Reisig

M. D.

(2006). The theory and practice of supermax prisons. Punishment & Society, 8, 33-57.

63.

Mears

D. P.

Wang

Bales

W. D.

(2014). Does a rising tide lift all boats? Labor market changes and their effects on the recidivism of released prisoners. Justice Quarterly, 31, 822-851.

64.

Metcalf

Morgan

J. N.

Oliker-Friedland

Resnik

Spiegel

Tae

. . . Holbrook

(2013). Administrative segregation, degrees of isolation, and incarceration: A national overview of state and federal correctional policies. New Haven, CT: Liman Public Interest Program at Yale Law School.

65.

Morgan

R. D.

Gendreau

Smith

Gray

A. L.

Labrecque

R. M.

MacLean

. . . Mills

J. F.

(2016). Quantitative syntheses of the effects of administrative segregation on inmates’ well-being. Psychology, Public Policy, and Law, 22, 439-461.

66.

Morris

R. G.

(2016). Exploring the effect of exposure to short-term solitary confinement among violent prison inmates. Journal of Quantitative Criminology, 32, 1-22.

67.

Morris

R. G.

Carriaga

M. L.

Diamond

Piquero

N. L.

Piquero

A. R.

(2012). Does prison strain lead to prison misbehavior? An application of general strain theory to inmate misconduct. Journal of Criminal Justice, 40, 194-201.

68.

Petersilia

(2003). When prisoners come home: Parole and prisoner reentry. Oxford, UK: Oxford University Press.

69.

Pizarro

J. M.

Narag

R. E.

(2008). Supermax prisons: What we know, what we do not know, and where we are going. The Prison Journal, 88, 23-42.

70.

Rhodes

(2004). Total confinement: Madness and reason in the maximum security prison. Los Angeles: University of California Press.

71.

Rice

M. E.

Harris

G. T.

(2005). Comparing effect sizes in follow-up studies: ROC area, Cohen’s d, and r. Law and Human Behavior, 29, 615-620.

72.

Rosenbaum

P. R.

Rubin

D. B.

(1985). Constructing a control group using multivariate matched sampling methods that incorporate the propensity score. The American Statistician, 39, 33-38.

73.

Shalev

(2009). Supermax: Controlling risk through solitary confinement. London, England: Willan.

74.

Shalev

(2015). Solitary confinement: The view from Europe. Canadian Journal of Human Rights, 4, 143-165.

75.

Shames

Wilcox

Subramanian

(2015). Solitary confinement: Common misconceptions and emerging safe alternatives (Report). New York, NY: VERA Institute of Justice.

76.

Smith

P. S.

(2006). The effects of solitary confinement on prison inmates: A brief history and review of the literature. Crime and Justice, 34, 441-528.

77.

Smith

P. S.

Gendreau

Swartz

(2009). Validating the principles of effective intervention: A systematic review of the contributions of meta-analysis in the field of corrections. Victims & Offenders, 4, 148-169.

78.

Sorensen

Wrinkle

R. D.

(1996). No hope for parole: Disciplinary infractions among death-sentenced and life-without-parole inmates. Criminal Justice and Behavior, 23, 542-552.

79.

Steiner

Butler

H. D.

Ellison

J. M.

(2014). Causes and correlates of prison inmate misconduct: A systematic review of the evidence. Journal of Criminal Justice, 42, 462-470.

80.

Steiner

Wooldredge

(2008). Inmate versus environmental effects on prison rule violations. Criminal Justice and Behavior, 35, 438-456.

81.

Steiner

Wooldredge

(2009). The relevance of inmate race/ethnicity versus population composition for understanding prison rule violations. Punishment & Society, 11, 459-489.

82.

Steiner

Wooldredge

(2014). Sex differences in the predictors of prisoner misconduct. Criminal Justice and Behavior, 41, 433-452.

83.

Stephan

J. J.

(2008). Census of state and federal correctional facilities, 2005 (NCJ 222182). Washington, DC: U.S. Department of Justice Office of Justice Programs.

84.

Sykes

G. M.

(1958). The society of captives. Princeton, NJ: Princeton University Press.

85.

Tewksbury

DeMichele

(2005). Going to prison: A prison visitation program. The Prison Journal, 85, 292-310.

86.

Toch

(2003). The contemporary relevance of early experiments with supermax reform. The Prison Journal, 83, 221-228.

87.

Valera

Kates-Benman

C. L.

(2016). Exploring the use of special housing units by men released from New York correctional facilities: A small mixed-methods study. American Journal of Men’s Health, 10, 466-473.

88.

VERA Institute of Justice. (n.d.). Segregation reduction project. Available from www.vera.org

89.

Walters

G. D.

(2015). Criminal thinking as a predictor of prison misconduct and mediator of the static risk–infractions relationship. The Prison Journal, 95, 353-369.

90.

Washington State Department of Corrections. (2013). Restrictive housing: Terminology and practice. Tumwater, WA: Author.

91.

Washington State Department of Corrections. (2015). Restrictive housing policy (DOC 320.255). Retrieved from https://www.doc.wa.gov/information/policies/files/320255.pdf

92.

Wilson

D. B.

Bouffard

L. A.

MacKenzie

D. L.

(2005). A quantitative review of structured, group-oriented, cognitive-behavioral programs for offenders. Criminal Justice and Behavior, 32, 172-204.

93.

Woo

Stohr

M. K.

Hemmens

Lutze

Hamilton

Yoon

O. K.

(2016). An empirical test of the social support paradigm on male inmate society. International Journal of Comparative and Applied Criminal Justice, 40, 145-169.

94.

Wooldredge

Griffin

Pratt

(2001). Considering hierarchical models for research on inmate behavior: Predicting misconduct with multilevel data. Justice Quarterly, 18, 203-231.

95.

Worrall

J. L.

Morris

R. G.

(2012). Prison gang integration and inmate violence. Journal of Criminal Justice, 40, 425-432.

96.

Zamble

(1992). Behavior and adaptation in long-term prison inmates: Descriptive longitudinal results. Criminal Justice and Behavior, 19, 409-425.