Conditions of Contact: Reexamining the Relationship between Prison Visitation and Recidivism

Abstract

While in prison, incarcerated persons can be subjected to qualitatively different conditions of confinement ranging from minimum to maximum security settings. In this study, we use data on 17,542 incarcerated men to assess whether the relationship between visitation and recidivism varies across the types of settings (i.e., minimum, medium, close, and maximum custody units) in which individuals are housed and receive their visits. We find that the relationship between prison visitation and rearrest varies across conditions of confinement, and that this association is highly attenuated for individuals confined in close and maximum security units. We also find that these patterns are robust to alternate various measures of recidivism (reconviction and return to prison) and visitation (frequency and timing). These results raise questions about the nature and effectiveness of visitation in highly restrictive settings, and suggest that an ecological framework should be applied to future studies of prison visitation and recidivism.

Full text

Running Head: Conditions of Contact
Conditions of Contact:
Reexamining the Relationship between Prison Visitation and Recidivism
Jillian J. Turanovic
Melinda Tasca
Article accepted at Justice Quarterly
June 2021
Funding:
This material is based upon work supported by the National Science Foundation under Grant No.
153531. Any opinions, findings, and conclusions or recommendations expressed in this article
are those of the authors and do not necessarily reflect the views of the National Science
Foundation.
Data availability statement:
The data that support the findings of this study are available on request from the corresponding
author. The data are not publicly available due to privacy restrictions.
Author bios:
Jillian J. Turanovic is an associate professor in the College of Criminology and Criminal Justice
at Florida State University. Her research is aimed at examining various issues in criminological
theory and correctional policy, with a special focus on victimization, violence, and health.
Melinda Tasca is an associate professor in the Department of Criminal Justice at the University
of Texas at El Paso. Her research focuses on corrections, the collateral consequences of
incarceration, and disparities in the criminal justice system.
Corresponding author:
Please address correspondence to Jillian Turanovic, College of Criminology and Criminal
Justice, Florida State University, Eppes Hall, 112 S. Copeland Street, Tallahassee, FL 32306-
1273. Phone: (850) 645-0375. Email: jturanovic@fsu.edu.

1
Conditions of Contact:
Reexamining the Relationship between Prison Visitation and Recidivism
Abstract
While in prison, incarcerated persons can be subjected to qualitatively different conditions of
confinement ranging from minimum to maximum security settings. In this study, we use data on
17,542 incarcerated men to assess whether the relationship between visitation and recidivism
varies across the types of settings (i.e., minimum, medium, close, and maximum custody units) in
which individuals are housed and receive their visits. We find that the relationship between prison
visitation and rearrest varies across conditions of confinement, and that this association is highly
attenuated for individuals confined in close and maximum security units. We also find that these
patterns are robust to alternate various measures of recidivism (reconviction and return to prison)
and visitation (frequency and timing). These results raise questions about the nature and
effectiveness of visitation in highly restrictive settings, and suggest that an ecological framework
should be applied to future studies of prison visitation and recidivism.
Key words: Prison visitation, recidivism, conditions of confinement

2
The tough-on-crime movement of the 1980s and 1990s gave rise to a host of punitive
policies that transformed the landscape of American corrections. Widespread increases in
mandatory minimum sentencing laws, harsher punishments, and “supermax” prison facilities
resulted in more people serving time behind bars, for longer periods of time, and in more severe
settings. Correctional rehabilitation fell out of favor for “coddling” offenders (Cullen, 2005; Pratt,
2019), and the rhetoric of “let ‘em rot” became popular among the American public (DiIulio,
1994). However, the policies aimed at “getting tough” on prisoners ultimately did little to enhance
public safety or reduce recidivism—to be sure, 79 percent of released prisoners are estimated to
be rearrested within six years (Alper et al., 2018). There is now a growing consensus that we may
need to ease (rather than enhance) the pains of imprisonment to ensure that those in prison are not
returning to their communities worse off than they were before.
One way to do so is through prison visitation. Visits represent “a sort of breach in the wall”
between prison and the outside world (Conover, 2001, p. 151), and they provide a valuable
opportunity for individuals to connect face-to-face with their loved ones during confinement.
Visits are beneficial as they allow incarcerated persons to build, repair, and maintain their social
ties to friends and family—ties that can help them find housing, secure employment, and access
treatment services upon release (Berg & Huebner, 2011; Cochran, 2014; Mears & Cochran, 2014;
Visher & Travis, 2003). During visits, loved ones can offer support via expressions of love,
encouragement, and active listening, and provide a much-needed reprieve from the loneliness and
isolation of prison life (Brunton-Smith & McCarthy, 2017; Hickert et al., 2019). Existing research
supports that incarcerated persons who receive visits have smoother transitions back to society,
and that visitation is often associated with a lower risk of recidivism (Bales & Mears, 2008;
Cochran, 2014; Mitchell et al., 2016).
And yet, as some have noted, there is variability in how visitation is experienced (Maruna
& Toch, 2005; Turanovic & Tasca, 2019), and associations with recidivism are not the same for

3
everyone (Cochran & Mears, 2013; McNeeley & Duwe, 2020). But for the most part, we have
struggled to explain why this variation in the relationship between visitation and recidivism exists.
Some explanations have focused on factors such as age, sex, and race and ethnicity, given their
prominence to criminological and recidivism analyses more generally (Bales & Mears, 2008;
Cochran et al., 2020). Others have focused on the patterning of visits, recognizing that individuals
who receive more visits, more consistently, might have more social support upon reentry to the
community (Cochran, 2014). And still, others have emphasized that who visits matters, where
visits from family (or specific family members) are thought to be more beneficial (Duwe & Clark,
2013; McNeeley & Duwe, 2020; Mears et al., 2012). Although this work has been both innovative
and informative, it has admittedly placed a heavy focus on the individual characteristics of
incarcerated persons and their visitors in explaining variation in the relationship between visitation
and recidivism. It might therefore be useful to take a step back and consider the broader contexts
under which the visits themselves take place.
During their time behind bars, individuals can experience qualitatively different conditions
of confinement—ranging from minimum to maximum security settings—that shape in
fundamental ways the nature of prison visitation. For instance, in lower security settings, visits are
typically held in cafeteria-style rooms in which incarcerated persons and their visitors can walk
around or sit next to each other, share snacks, play games, and embrace (even if briefly) (Comfort,
2002, 2003). Alternatively, in many maximum security environments around the country, visits
more often occur behind glass and under the watchful eye of armed guards, where confined persons
are shackled, conversations are closely monitored, and there are no opportunities for physical
contact (Comfort, 2008; Cramer et al., 2017; Metcalf et al., 2013). Visitors to maximum security
facilities are also subject to more punitive and distressing security protocols, and visits are
restricted in duration and frequency (Boudin et al., 2013; Wall, 2013). The broader stresses of
confinement in these settings for incarcerated persons and visitors may limit the extent to which

4
visits are supportive and allow for the strengthening of social ties—issues that can undermine the
protective benefits of visitation for recidivism.
Even though conditions of confinement are of central importance to many areas of
correctional research (Haney, 2018; Mears & Bales, 2009; Wildeman et al., 2018), the physical
settings that incarcerated persons and their visitors occupy have been largely overlooked when
studying the link between visitation and recidivism (Moran, 2013). Accordingly, the purpose of
this study is to determine whether the relationship between prison visitation and recidivism varies
across conditions of confinement—that is, across the minimum, medium, close, and maximum
custody settings in which individuals are confined and receive their visits. We analyze
administrative data from a large and diverse population of incarcerated men from a single state
prison system and examine recidivism 24-months post-release. In carrying out this research, our
broader purpose is to gain a deeper understanding of how conditions of confinement shape the
relationship between in-prison experiences and post-release behavior.
Prison Visitation and Recidivism
Isolation from family and friends is one of the most painful features of life behind bars
(Adams, 1992; Sykes, 1958). And though it is part and parcel of the prison experience, separation
from loved ones is a major source of maladjustment (Jiang & Winfree, 2006; La Vigne et al., 2005;
Maruna, 2001). Research has routinely shown that detachment from social networks and the
“removal of sources of support” is distressing and can lead to numerous challenges while
incarcerated (Liebling, 1999, p. 326). Prison visitation, however, offers a temporary relief from
this separation, and affords incarcerated persons the chance to connect directly with family and
friends in the prison setting. While other forms of contact, such as letters and phone calls, are
certainly meaningful, the ability to visit, face-to-face, over the course of several hours is considered
a vital bonding opportunity for confined persons and their loved ones.

5
In addition to easing the pains of imprisonment, visits can also help incarcerated persons
establish, strengthen, and maintain their supportive ties to others (Barrick et al., 2014; Mowen &
Visher, 2016; Visher & Travis, 2003). Upon release, these social ties can help individuals find
housing, access treatment services, overcome material hardship, and secure employment (Berg &
Huebner, 2011; Harding et al., 2014). Without strong or supportive social ties, however,
individuals may have difficulties reintegrating back into the community and “negotiating the
common tasks of conventional living” (Jacoby & Kozie-Peak, 1997, p. 483; Western et al., 2015).
As such, several studies have linked prison visitation to reductions in recidivism (Bales &
Mears, 2008; Duwe & Clark, 2013; Mears et al., 2012). And though not all studies have reached
similar conclusions—with some more recent work showing modest or nonexistent effects (Atkin-
Plunk & Armstrong, 2018; Cochran et al., 2020)—research has linked visitation to decreases in
the odds of recidivism that hover around 26 percent (Mitchell et al., 2016). But regardless of the
overall magnitude of the visitation-recidivism relationship, one of the most striking and consistent
findings within this literature is that the relationship between visitation and reoffending is
heterogeneous. How this heterogeneity is explained, however, varies from study to study,
depending on whether the focus is on the characteristics of incarcerated persons, the types of
visitors, or the timing and patterns of visitation over time (see Cochran & Mears, 2013). For
instance, Bales and Mears (2008) found that visitation was more likely to reduce recidivism among
men, non-White persons, and those who had been incarcerated more than once; and, that visits
closer to release, and visits from spouses, were more likely to be associated with reductions in
recidivism. Alternatively, Duwe and Clark (2013) found that visits from spouses—particularly ex-
spouses—were not helpful for reducing recidivism; and Cochran (2014) showed that visits that
occurred closer to admission to prison were more protective against recidivism than visits closer
to release.

6
Consequently, questions remain about the conditions under which, and for whom, visits
are most beneficial. The work produced thus far is certainly a good start, but two key features merit
attention. First, with some exceptions (e.g., Arditti, 2003; Comfort, 2002, 2008; Moran, 2013),
research rarely considers the different security settings in which visits occur, and how these
settings alter the visitation experience. This likely means that the average associations between
visitation and recidivism reported in most studies are driven by visitation in low security settings
(where most individuals serve their time and receive their visits) and that the consequences of
visits in higher security environments are less known. Second, while studies have considered how
individual features, such as incarcerated persons’ age, sex, race, or frequency of visits, might
moderate the visitation-recidivism relationship, researchers have yet to consider how conditions
of confinement moderate the association as well. This is a core oversight that, if addressed, could
shed light on a potentially important conditional factor that can help explain how visits affect
behavior upon release.
Visitation and Conditions of Confinement
Prison facilities vary greatly in terms of their conditions of confinement, which typically
reflect differences in the populations of individuals that they house. For example, maximum and
close security facilities are more likely to confine individuals with extensive or violent criminal
histories, lengthy records of institutional misconduct, gang affiliations, and serious mental health
or behavioral problems (Butler et al., 2013; Mears et al., 2019; Pizarro & Stenius, 2004).
Administratively, such individuals are thought to pose more of a threat to the safety and security
of institutions, to be more challenging to manage, or to be disruptive to daily operations (Labrecque
& Mears, 2019; Mears & Castro, 2006; Rhodes, 2004). While in maximum security housing,
incarcerated persons are locked alone in their cells for most of the day with limited opportunities
for recreation, work, or socialization with others, and they are usually escorted in wrist and ankle

7
restraints when moving throughout the institution (Foster, 2016; Labrecque, 2016; Metcalf et al.,
2013).
Indeed, maximum security facilities attempt to “create and maintain total or almost total
social control” (Sykes, 1958, p. xiv, emphasis in original). Typified by concrete, steel, and little
natural light, such facilities have been described as “painful and demeaning” (Sundt et al., 2008,
p. 101). While some restrictions are eased in close security settings (e.g., allowing individuals to
leave their cells for short periods of time, housing two persons in each cell), opportunities to take
part in programming, training, or employment are limited, as movements throughout these
facilities remain constrained. This is in stark contrast to lower custody facilities where individuals
are more likely to be housed in open dormitory-style arrangements, and where they routinely leave
their housing units (unrestrained) to eat their meals, attend programming, spend time on the yard,
and work within the prison.
These differences in conditions of confinement also carry over to visitation. In lower-
custody settings (such as in minimum or medium security facilities), visitation rooms tend to be
arranged cafeteria-style, where incarcerated persons and their visitors can sit and chat together at
tables. Confined persons are permitted to greet their visitors with an embrace, to hold hands, to
hold small children on their lap, and to walk around with their visitors in designated spaces
(Comfort, 2002; Tasca, 2014). They can share snacks and food from vending machines, read
books, and play cards or board games (Boudin et al., 2013). This is not to say that visitation in
these settings is necessarily easy, carefree, or pleasant—it is, after all, still prison (Arditti, 2005;
Tasca et al., 2016)—but it is less restrictive than in maximum or close security facilities.
To be sure, in maximum custody, visitation looks different and has different rules. For one,
visits are typically non-contact, meaning that communication occurs through Plexiglas barriers,
and incarcerated persons cannot have physical contact with their visitors (Metcalf et al., 2013;
Rhodes, 2004). They must remain seated on their side of their visitation booth and, in some

8
correctional systems, remain in wrist and ankle restraints for the duration of the visit. Generally,
there are no games, no sharing of snacks, no hugs, and no handholding. These strict procedures
reflect prison administrators’ view that visitation introduces “an element of unpredictability” to
highly controlled institutions—one that can lead to “disruption in the routines” of the facility and
“security threats such as the conveyance of contraband” (Wall, 2013, p. 201). There is a looming
concern by staff that visitors will try to “smuggle in forbidden and even dangerous items,” such as
drugs and weapons, that could lead to “all manner of security breaches and a destabilization of the
inmate climate” (Wall, 2013, p. 201-202). Especially in maximum security prisons, and to a lesser
degree in close security settings, correctional officers are “expected to exercise a very high level
of scrutiny throughout the visiting process” (Wall, 2013, p. 201). Visiting hours are more limited
in both maximum and close security settings, appointments must be scheduled in advance, and
only a few individuals can receive visits at a given time (Boudin et al., 2013; Metcalf et al., 2013).
Although at any facility, regardless of security level, visitors may face long, frustrating
waits, poor treatment by staff, and crowded, noisy conditions (Christian, 2005; Sturges & Al-
Khattar, 2009), maximum security institutions can be particularly stressful and intimidating to
visit. Complete with armed guards and a secure perimeter, such facilities are built to be imposing
(Sykes, 1958; Mears, 2013; Rhodes, 2004). Their more elaborate security protocols can also make
some visitors feel like “quasi-inmates” (Arditti, 2012; Comfort, 2003; Hairston, 1988). This
process is referred to as “secondary prisonization,” whereby visitors are subjected to the same
elaborate regulations and levels of intense surveillance that incarcerated persons face (Aiello &
McCorkel, 2018; Arditti, 2012; Poehlmann et al., 2010). Generally, the prison staff charged with
maintaining order and control in maximum security facilities are concerned with institutional
safety above all else. As one administrator put it, “the idea that visiting can help lead to better
outcomes for public safety down the road can seem highly attenuated when compared to the risks
and threats it raises in the here and now” (Wall, 2013, p. 201, emphasis added). These elements

9
carry consequences for visits, where visitors (and incarcerated persons) might be distraught by the
overly restrictive conditions of confinement, and the environment might make it more challenging
to have helpful, calm, and supportive interactions (Arditti, 2005; Comfort, 2008; Tasca, 2014).
Thus, the protective link between visitation and recidivism may be weakened in such settings.
Along those lines, some research shows that visitation experiences can be negatively
impacted by overly restrictive security procedures and harsher conditions of confinement. For
example, visitors tend to get upset if they are surveilled aggressively, such as when their demeanor,
clothing, and undergarments are too closely scrutinized (Comfort, 2003). Visiting a secure facility
has also been described as “intense and terrifying,” especially for children who fear being “crushed
like bugs” by the heavy metal doors (Aiello & McCorkel, 2018, p. 366). Some recent work has
even shown that incarcerated persons tend to feel more negatively about visitation in maximum
security settings. Indeed, Turanovic and Tasca (2019) found that in maximum security units,
incarcerated persons felt stressed and guilty during visits, and that in minimum security units,
individuals were less likely to report negative interactions with visitors. Findings like these suggest
that the benefits of visitation might be negated in more restrictive conditions of confinement.
Current Focus
Prison visitation is assumed to be protective against recidivism insomuch as it can help
incarcerated persons strengthen their social ties, receive support and guidance, and have productive
discussions about reentry. In highly restrictive prison environments, however, it is unclear whether
these goals can be achieved. Even if visits provide incarcerated persons a temporary reprieve from
lengthy periods of isolation in high-custody settings, the punitive conditions under which contact
occurs and the broader stresses of confinement within these settings (for both incarcerated
individuals and visitors) may undermine the protective nature of visitation for reentry.
Accordingly, the purpose of the current study is to examine whether the relationship between
prison visitation and recidivism is moderated by conditions of confinement—that is, whether the

10
strength of the association between prison visitation and recidivism varies across the minimum,
medium, close, and maximum security settings in which individuals are housed and receive their
visits. We use data from a large and diverse population of incarcerated men from a single state
prison system, and measure recidivism within two years following release. In carrying out this
study, our broader goal is to determine whether the prison environment can affect the link between
visitation and post-release behavior.
Methods
Data
This study relies on administrative data on all incarcerated men serving sentences of 12
months or more who were released from the Department of Corrections (DOC) in a large state
between July 1st, 2010, and June 30th, 2013 (N = 17,542).
1
This state has 14 male prison facilities
containing upwards of 30,000 men, and it houses one of the most racially and ethnically diverse
prison populations in the United States (Carson, 2020). The DOC data contain details on
incarcerated persons across every facility in the state, including information on risk classification,
mental health needs, disciplinary reports, demographic characteristics (e.g., age, sex, and
race/ethnicity), and various aspects of criminal justice involvement before and after incarceration.
In addition, the data provide information on both prison visitation and housing type.
The DOC used a unit-based custody system comprised of four security levels: minimum,
medium, close, and maximum. In a unit-based custody system, individuals’ security classifications
are synonymous with the type of housing unit in which they are placed (i.e., maximum custody
persons are housed in maximum security units, close-custody persons are housed in close security
units, medium-custody persons in medium security units, and minimum-custody persons in
minimum security units). Each unit had its own recreation and visitation areas, and none were
1
Individuals serving time in prison for parole violations were not included in the sample.

11
“mixed-use”—meaning that there were no general population units where a variety of incarcerated
individuals with differing security classifications were housed together.
During the study period, across the 14 male prison facilities, there were 23 minimum
security units, 34 medium security units, 22 close security units, and 35 maximum security units.
Depending on the unit in which they were housed, individuals had different privileges and were
subject to different levels of monitoring and restrictions. For instance, while housed in maximum
security units, individuals had limited work opportunities, had to be escorted in restraints when
moved, and had limited visitation and phone privileges. The amount of time that individuals
remained in maximum security housing was based on the discretion of DOC personnel, and
reviews occurred every 180 days. DOC policies stated that housing placements were based
primarily on in-prison behaviors and risk assessment scores. Individuals could be transferred up
from lower custody levels if they engaged in misconduct (e.g., committing or leading others to
commit disruptive or violent behaviors) or when deemed a threat to the security of the institution.
Correctional officers made initial recommendations for placement that were elevated through an
administrative chain of command for approval.
2
According to DOC policies, individuals in maximum security units were allowed only one
two-hour visit per week, where they communicated with their visitors through Plexiglas barriers
while remaining in wrist and ankle restraints. All conversations were recorded and closely
monitored. This contrasts with visitation within lower custody units, where incarcerated persons
could have physical contact with visitors, share food, converse privately, and walk around together
within indoor and outdoor spaces. Lower custody units also allowed for more visitation—up to
four, four-hour blocks of contact visits per week in minimum and medium security units, and up
to three, four-hour blocks per week in close security units. On visiting days, a variety of security
protocols were in place to check visitors and their vehicles for contraband, including the use of
2
All DOC policies described are specific to the time period in which the sample was incarcerated. Some policies have
since changed.

12
search dogs, metal detectors, and pat downs. Dress codes were strictly enforced, and visitors were
only allowed to bring limited personal belongings into facilities. In all units, regardless of security
level, correctional staff would pat down incarcerated persons before every visit and strip-search
them afterwards. Even though these processes are common across correctional systems, they can
be distressing for incarcerated individuals and their visitors.
Measures
Recidivism
The dependent variable for this study is recidivism, measured as any rearrest for a new
offense committed within 24 months of release from prison (1 = rearrest, 0 = no rearrest). Arrest
information came from the state’s Department of Public Safety and was merged with DOC data.
Approximately 45.5% of the sample was rearrested for a new offense within 24 months of release.
Of those who were rearrested, 21.4% were arrested for a violent crime, 23.3% for a property crime,
18.3% for a drug crime, and 37.0% for a public order crime (e.g., loitering, disorderly conduct,
and nuisance crimes) or another miscellaneous offense (e.g., gambling, unlawful solicitation, and
other criminal acts). We recognize that rearrest is only one way to capture recidivism, and therefore
in additional analyses we measure recidivism using reconviction and return to prison. These
analyses are presented and described later. Descriptive statistics on all variables used in the study
can be found in Table 1.
--Insert Table 1 about here--
Prison visitation
The key independent variable, prison visitation, was measured as any visit within 12
months of release from prison (1 = visited, 0 = not visited). We focus specifically on visitation
during the year prior to release given that this is a time when incarcerated individuals and their
visitors are planning for their return to the community. Approximately half of the sample (49.6%)
received at least one visit in the 12 months prior to release. We began by using a binary indicator

13
of visitation given that most of the variation was at the prevalence margin, not the count margin
(see also Hickert et al., 2018). But note that we also present additional analyses using several
continuous measures of visitation (total count of visits, count of visits received in each type of
housing unit, number of months that a visit was received, and recency of visits).
3
Conditions of confinement
The moderators of interest are conditions of confinement. These conditions were measured
using a categorical variable for the type of unit where individuals were housed the longest in the
12 months prior to release from prison. These conditions reflect minimum security, medium
security, close security, and maximum security housing units. Categories were mutually exclusive,
and minimum security serves as the reference category. Most individuals were confined primarily
in minimum security (41.0%) or medium security (41.3%) housing units in the year prior to release,
and fewer served the majority of their time in close security (10.0%) or maximum security (7.7%)
housing units.
4
On average, individuals served 10 consecutive months in a housing unit at their
designated security level in the year prior to release (mean = 297 days; SD = 78.5). Less than 10%
of the sample served fewer than 6 months in a unit at their designated security level; the fewest
number of days being 100 (0.6%). Removing these individuals from the sample had no impact on
the results.
As noted, the DOC used a unit-based custody system, where housing units were
synonymous with individuals’ security classifications. As such, minimum security units confined
individuals who were at low risk to the institution and to the general public, whereas medium
security units confined those who represented a moderate risk. The conditions of confinement in
minimum and medium security units were similar, the key difference being that while in medium
custody housing, individuals had more limited movements within the institution and they could
3
Visitation did not capture meetings with attorneys or prison clergy, and the DOC did not allow conjugal visits.
4
Unlike in some state prison systems, individuals were not automatically reclassified or allowed to step down to a
lower custody unit as they neared release from prison. Indeed, 80.6% of the sample who served most of their time in
maximum security housing in the year prior to reentry were released directly from maximum custody units.

14
not work outside of prison grounds (e.g., on community work crews). Minimum and medium
security units housed larger populations and were designed campus-style, with freestanding
buildings for sleep, programming, and visitation, and larger open spaces for recreation (Johnston,
2000). Approximately 82% of the state’s total prison population was housed in minimum or
medium security units.
Close and maximum security units were designed to confine the high- and highest-risk
individuals, respectively. Close and maximum security units were similar in architecture, and they
consisted of several smaller housing units (or “pods”) that allowed for enhanced supervision. The
conditions of confinement varied somewhat between close and maximum security in that, in close
security units, individuals were allowed brief time outside of their cells each day to socialize with
other incarcerated persons (either within their pod or outside in a small recreation pen), and they
could be housed in double-occupancy cells. Also, unlike in maximum security units, persons
confined in close custody were not always required to be escorted in full restraints within the
institution, and they were eligible to receive contact visits (i.e., without Plexiglas barriers). Contact
visitation in close security units, however, was far more restricted and monitored than in lower
custody units. Rooms were smaller, fewer individuals could be visited at one time, and they were
constrained from walking around or going outside with their visitors. Even though living
conditions across housing units of the same security level could vary slightly (e.g., some were
louder, some were older, some were more populated, and some had better heating and cooling
systems), the conditions of visitation were uniform within each security level.
A notable limitation of the data is that the DOC does not record valid or consistent
information on each visitor’s relationship to the incarcerated person (e.g., mother, father, sister,
brother, spouse, etc.). Therefore, we can only reliably assess whether incarcerated persons were
visited, but not differences in visitor type or the nature of individuals’ relationships with visitors.

15
Assessing how the association between visitation and recidivism varies across prison settings
depending on who visits is an important avenue for future research.
Control variables
In this DOC, housing classifications were based on individuals’ institutional behaviors,
treatment needs, and risk scores. Thus, it was critical to adjust for these differences in the analyses.
As such, a lengthy roster of control variables was included. These covariates explain individuals’
placements into different custody levels and can help account for between-person differences in
the likelihood of receiving visits and the risk of recidivism.
First, we controlled for the number of housing unit transfers each person experienced in
the year prior to release. Each time someone was moved to a new unit for at least 24 hours it was
recorded as a transfer. On average, individuals experienced two housing unit transfers in the year
prior to release, and nearly half the sample experienced no transfers. The vast majority of transfers
(76.2%) were for routine operations, which included population adjustments, lateral transfers, and
temporary placements for scheduled treatments (e.g., dental care, medical care, mental health
programming). A small portion of the sample (1.6%) experienced 10 or more transfers in the year
prior to release. The exclusion of these high-transfer individuals had no impact on the results.
Second, we included a measure of distance from prison to home. This represents the
distance in miles from each individual’s home address to the prison complex where the most time
was served in the year prior to release. This measure ranged up to 550 miles and was logged to
reduce the undue influence of outliers on the results.
5
Distance from home has been documented
as one of the most substantial barriers to visitation in prior work (Christian, 2005; Cochran et al.,
2016; Young & Turanovic, 2020).
We also included a measure of each incarcerated person’s risk score. As a part of the intake
classification process, a structured assessment tool was used to quantify the level of risk (or
5
Less than 0.8% of individuals in the sample were incarcerated more than 550 miles from home.

16
“dangerousness”) each person posed to the general public and to the institution. Risk scores were
determined by various factors such as the severity of current and prior offenses, escape history,
completion of major programming, age at intake, and histories of institutional and community
violence. Risk scores ranged from 2 to 10, with higher scores indicating greater risk.
Individuals’ most recent program phase level prior to release (range = 1 – 3) was also
controlled for. Consistent with an earned incentive program in this DOC, individuals were eligible
to receive visits in certain frequencies and lengths according to their program phase. As people
moved up to higher phase levels, they had greater visitation privileges. For example, “phase 1”
individuals in minimum and medium security units were allowed visitors for only one, four-hour
block per week, whereas “phase 3” individuals were allowed visits for up to four, four-hour blocks
per week. Phase levels were determined by in-prison behaviors and the completion of
programming. This is an important covariate given that individuals at higher phase levels have
more opportunities for visitation. Note, however, that regardless of what phase individuals were
in, those housed in maximum security units were only allowed non-contact visits for one, two-
hour block per week.
To account for criminal history, the number of prior felony convictions was included
(capped at 15 prior felonies), as well as a categorical variable for the most serious current offense
(with categories for violent, property, drug, and other non-violent offenses, where “violent
offense” serves as the reference category). Sentence length was also accounted for by the number
of years in prison served.
Institutional misconduct was controlled for using the number of violent, property, drug,
threat/intimidation, security, defiance, and other non-violent disciplinary reports incurred while
ever incarcerated in this DOC.
6
Each measure ranged from 0 to 10. Individuals’ full misconduct
6
Disciplinary histories were categorized as follows: violent infractions (e.g., homicide, assault on an officer or on
another incarcerated person), property infractions (e.g., possession of a communication device, theft, possession of
stolen property), drug infractions (e.g., possession of drug paraphernalia, drugs, or narcotics), threat/intimidation
infractions (e.g., intimidating, harassment, stalking), security infractions (e.g., rioting, escape, promoting prison

17
histories (including those from prior incarceration terms) were captured as they affect custody
placements, risk scores, programming, and eligibility for visitation. The results were not sensitive
to whether disciplinary histories specific to the current incarceration term were used or not.
Dichotomous variables for whether an individual was classified by the DOC as being gang
affiliated or having a history of attempted suicide were also included. Gang affiliations included
both prison and street gangs, and the DOC determined whether individuals were either suspected
or validated street and/or prison gang affiliates (1 = yes, 0 = no). Suicide histories included any
suicide attempt, in or out of prison (1 = yes, 0 = no). Additionally, mental health scores obtained
from structured screenings conducted at intake were also incorporated. These screenings
determined the level of mental health treatment needs (i.e., no need, low need, moderate need, and
high need). Mental health scores range from 1 to 4, where “1” reflects no mental health treatment
needs, and “4” reflects high needs.
In addition, several demographic variables were included, such as age at release,
race/ethnicity (with categories for white, Latino, Black, Native American, and other race, where
“white” served as the reference category), and undocumented immigrant status (1 = yes, 0 = no).
We also included measures for low educational attainment (1 = no high school or graduate
equivalency degree, 0 = otherwise), marital status (1 = married, 0 = not married), and the number
of dependents (e.g., children, spouses, and other household members) recorded for each individual
(up to 15).
Lastly, a categorical variable for the prison complex where individuals served the most
time in the year prior to release was included in the analyses. This variable contains 14 different
categories for each of the male prison complexes in the state, and “complex #1”—a prison complex
that is made up largely of minimum security housing units—served as the reference category.
contraband), defiance infractions (e.g., disorderly conduct, resisting or disobeying a verbal or written order), and other
non-violent infractions (e.g., conspiracy or attempt to commit an offense).

18
Analytic Strategy
The analyses proceeded as follows. First, logistic regression models were specified to
assess the overall association between visitation and recidivism. Second, we included interaction
terms in the model to assess how the relationship between visitation and recidivism varied across
conditions of confinement. Third, a series of additional tests were conducted to determine the
robustness of the findings to different measures of recidivism (reconviction and return to prison)
and visitation (frequency and timing of visits).
7
Fourth, descriptive analyses were used to explore
how heterogeneous the association between visitation and recidivism was across housing units of
similar conditions of confinement. All analyses were conducted in Stata 16 and were specified
using robust standard errors corrected for the clustering of individuals within housing units.
8
Note
that the findings presented are correlational, and causal inference cannot be established.
Results
The results of logistic regression analyses assessing the relationship between prison
visitation and recidivism are displayed in model 1 of Table 2. On average, visitation was associated
with a reduced likelihood of recidivism. More specifically, the odds of rearrest were 24.6% lower
among individuals who were visited (OR = .754). The coefficient for visitation was substantively
large and statistically significant (b = -.283, p < .001). It can also be seen in model 1 that conditions
of confinement were related to recidivism. Specifically, compared to confinement in minimum
7
Prior to estimating multivariate models, various diagnostics were examined to rule out the presence of harmful
collinearity. The highest variance inflation factor was 2.12 (average = 1.33) and bivariate correlations between
independent variables did not exceed an absolute value of .495. The two strongest correlations were between violent
discipline reports and other discipline reports. The key findings regarding the relationship between visitation and
recidivism were not sensitive to the inclusion/exclusion of these variables.
8
As an alternative to clustered standard errors, we specified multilevel logistic regression models that nested
individuals within housing units, within prison complexes. In the unconditional model, variance components indicated
that the likelihood of rearrest varied across housing units (.223, p < .001) but not prison complexes (.022, n.s.). The
intraclass correlation coefficient showed that 6.35% of the variation in rearrest existed between housing units, and
0.66% existed between prison complexes. Once all covariates were added to the model, the variance component at the
housing unit level was reduced to non-significance (.0017, n.s.). Further, the regression coefficients in the fully
specified model were nearly identical in terms of magnitude and statistical significance as those produced from logistic
regression with clustered standard errors. Given the similarity in findings, and the non-significant variance
components in the fully specified multilevel models, we proceed with presenting logistic regression models.

19
security housing, the odds of rearrest were 22.9% higher for individuals confined in maximum
security housing.
--Insert Table 2 about here--
The next step was to identify whether the relationship between prison visitation and
recidivism varied across housing conditions. To do so, two-way interaction terms between prison
visitation and conditions of confinement (medium, close, and maximum security) were included
in model 2 of Table 2. Focusing on the interaction terms, it can be seen that the relationship
between prison visitation and recidivism varied across individuals’ conditions of confinement.
Specifically, two substantively and statistically significant interactions emerged: one between
visitation and close security (OR = 1.433, p <.001) and another between visitation and maximum
security (OR = 1.375, p < .001). These coefficients suggest that, relative to individuals incarcerated
in minimum security units, visited individuals confined in close and maximum custody housing
units were more likely to be rearrested upon release.
To better interpret these interaction terms, predicted probabilities are presented in Figure
1. Looking across conditions of confinement, Figure 1 shows that visitation had the strongest
association with rearrest for individuals housed in minimum and medium security units.
Specifically, individuals in minimum security units who were visited had a probability of rearrest
that was 7.9 percentage points lower than those were not visited (41.2 vs. 49.1); and for individuals
confined in medium security units, the probability of rearrest was 6.5 percentage points lower
among those who were visited (41.6 vs. 48.1). In contrast, individuals in close security units who
were visited had a probability of rearrest that was only 0.2 percentage points lower than for those
who were not visited (46.7 vs. 46.9); and for individuals in maximum security units, the probability
of rearrest was just 1.1 percentage point lower among visited persons (49.1 vs. 50.2). These
findings, although correlational, suggest that visitation may do less to reduce recidivism for

20
individuals confined in close and maximum security units than for individuals confined in
minimum and medium security units.
--Insert Figure 1 about here--
Alternate Measures of Recidivism and Visitation
Next, we respecified the models using alternate measures of recidivism to ensure that the
findings were not an artifact of focusing solely on rearrest (King & Elderbroom, 2014). We
included indicators of reconviction for a new offense and return to prison within 24-months of
release (see Table 3). 42.8% of the sample were reconvicted, and 37.1% returned to prison within
24 months of release (8.6% of individuals in the sample returned to prison for a technical parole
violation). Consistent with the findings presented previously, these analyses show that individuals
who were visited in the year prior to release were less likely to be reconvicted and less likely to
return to prison, and that these associations varied across conditions of confinement. Specifically,
the interaction term visitation x close security was positive and statistically significant for
reconviction and return to prison; as was the interaction term visitation x maximum security for
reconviction and return to prison. These findings are consistent with those presented in Table 2
and suggest that visitation may be less protective against recidivism for individuals confined in
close and maximum security units.
--Insert Table 3 about here--
We also respecified the models using frequency measures of visitation, and these analyses
are presented in Table 4. First, we assessed the number of visits received in the year prior to release.
This measure ranged up to 48 visits, which was the maximum number of visits allowed per year
in the highest custody units.
9
Again, as seen in models 1 and 2 of Table 4, the results were generally
consistent: visitation was associated with lower odds of rearrest, and the interaction term visitation
9
Models were also estimated with a measure that ranged up to 103 visits (the maximum number of visits received by
anybody in the sample) and the results were the same. Less than 1% of the sample received more than 48 visits in the
year prior to release.

21
x close security was positive and statistically significant (OR = 1.015, p = .006). The one difference
was that the interaction term visitation x maximum security reached only a marginal level of
statistical significance (p = .093), even though it was positive with an odds ratio of 1.013.
-- Insert Table 4 about here --
Because these findings were somewhat less consistent with those presented thus far, to
better examine how the frequency of visitation influenced recidivism across conditions of
confinement, we also included measures for the number of times each person was visited
specifically while housed in a minimum, medium, close, and maximum custody unit (range per
each 0-48). These variables were not mutually exclusive, since it was possible for individuals to
be visited in more than one type of housing unit in the year prior to release (e.g., if someone spent
time in multiple housing units of different security levels). As can be seen in model 3 of Table 4,
more frequent visitation in minimum and medium security housing units was associated with a
lower likelihood of recidivism, whereas more frequent visitation in close and maximum custody
units was not. Specifically, each additional visit in a minimum or medium security unit was
associated with a 1.7% and 1.9% average decrease in the odds of rearrest, respectively. In contrast,
the frequency of visitation in close or maximum custody units had virtually no associations with
recidivism, with slope coefficients that were effectively zero. For visits in close security housing,
the unrounded coefficient was .00035 (OR = 1.00035), and for visits in maximum custody housing,
it was -.00043 (OR = .99957). Thus, more visits in close and maximum security units were not
linked to decreases in recidivism. These findings reaffirm the results presented previously in that
the correlation between visitation and recidivism is diminished in higher custody settings.
In addition to the frequency of visitation, in the next set of models, we also assessed how
the timing of visits (consistency and recency) influenced recidivism. First, for visitation
consistency, we created an indicator for the number of months in which individuals received at

22
least one visit in the year prior to release (range 0-12). Less than 4% of the sample were visited
every month. As seen in model 1 of Table 5, more months of visits were associated with lower
odds of rearrest. And as shown in model 2, the interaction terms visitation x close security and
visitation x maximum security were positive and statistically significant—a pattern wholly
consistent with what has been presented thus far. Unlike in minimum security units, more months
of visitation for individuals in close and maximum custody units did not correlate with lower odds
of recidivism.
-- Insert Table 5 about here --
Finally, following Bales and Mears (2008), a measure of visitation recency was created
that gave more weight to visitation in the months closer to release. For the month prior to release,
full weight was given if a visit occurred that month (a score of “1” was assigned). For two months
prior to release, visitation received 11/12th weight (1 x [11/12]), and so on, where a visit 12 months
prior to release received 1/12th weight (1 x [1/12]). The 12 weighted values were then summed to
create the recency measure (range 0-6.5). Once again, the findings remained the same. Model 3 of
Table 5 shows that visits closer to release were correlated with reduced recidivism, and model 4
indicates that this relationship, too, was negated within higher custody units. The interaction terms
visitation x close security and visitation x maximum security were both positive and statistically
significant, indicating that more recently visited persons in close and maximum custody units still
had greater odds of recidivism than individuals in minimum security units. Taken together, the
results appear to be quite stable. Even when using different measures of visitation and recidivism,
the same pattern emerges—the relationship between visitation and recidivism is attenuated for
individuals housed in higher custody units.
Further Examining Heterogeneity Across Housing Units
Despite the consistency in findings, there is still additional heterogeneity to explore across
housing units of the same security level. It is possible that different housing units, despite having

23
similar structural conditions of confinement, may vary in other ways that influence the relationship
between visitation and recidivism. So, using the same model specification from Table 2, we
generated predicted probabilities of rearrest, by visitation, for each housing unit. These are
displayed in Figure 2. Units are rank ordered from left to right in each graph, beginning with the
unit where the correlation between visitation and recidivism was the strongest.
-- Insert Figure 2 about here --
Figure 2 shows that, although the predicted probability of rearrest varies across units within
conditions of confinement (ranging from around .20 to .70 in medium security units, and .40 to .80
in maximum security units, for example), the strength of the association between visitation and
rearrest—which reflects the size of the difference between visited and non-visited categories—is
relatively uniform across housing units of the same custody level. Specifically, across minimum
security units, percentage point differences in the probability of rearrest between visited and non-
visited persons ranged from 7.3 to 8.6. Across medium security units, differences ranged from 4.1
to 7.5. In contrast, in close security units, the percentage point difference in the probability of
rearrest between visited and non-visited persons ranged from just 0.1 to 1.1; and in maximum
security units, this percentage point difference ranged from 0.3 to 1.3. There was no close or
maximum custody housing unit that presented a correlation between visitation and recidivism that
overlapped with even the weakest association evident within minimum and medium security
housing units. This level of uniformity may be due to the extensive covariates used in the analysis,
as well as the rigid housing policies of this state prison system—where units were not “mixed use,”
conditions of confinement were similar within security levels, and individuals’ custody
classifications were synonymous with the types of housing in which they served their time.
Taken together, the findings revealed three important patterns: (1) prison visitation had a
consistent, negative relationship with recidivism, (2) the strength of the association between prison
visitation and recidivism varied across conditions of confinement, and (3) the association between

24
prison visitation and recidivism was strongest for individuals confined in minimum and medium
security units, and weakest for those confined in close and maximum security units. These results
and their implications are discussed in more detail below.
Discussion
Prisons are not pleasant places, and formerly incarcerated persons often return to their
communities worse off than they were before. Given this reality, visitation has become an
important tool to help with the reentry process—one that can strengthen individuals’ social ties
and allow them to receive support and guidance in preparation for release. Despite growing
research on the topic, questions remained about the extent to which variation in prison settings
might shape the visitation experience. The conditions in which individuals are confined can set the
parameters for contact with loved ones, thereby affecting the nature, quality, and frequency of
interactions. Just as “some prisons are more survivable than others” (Liebling, 2011, p. 530), some
conditions of confinement may be more conducive to beneficial visits than others. Our work here
examined how the relationship between prison visitation and recidivism varied across conditions
of confinement, as captured through contact in maximum, close, medium, and minimum security
settings. Considering the results that were presented, three broad conclusions are warranted.
First, although we found that prison visitation, on average, was correlated with lower odds
of recidivism, there was considerable variation in this association. In particular, the strength of the
relationship between visitation and recidivism varied across conditions of confinement and was
highly attenuated in more restrictive settings. We attribute these findings, in part, to the different
contexts of visitation within close and maximum security units. Because the individuals confined
within these units were deemed to pose greater threats to institutional safety and security
(Labrecque & Mears, 2019; Mears et al., 2019), there were many enhanced restrictions and
surveillance protocols that carried over to visitation. For example, in the state prison system under
study, protocols in maximum custody units dictated that individuals be escorted from their cells to

25
non-contact visitation areas in shackles, where they communicated with visitors through glass or
steel partitions, and their conversations were closely surveilled by staff. Close and maximum
custody units also had imposing physical features, including highly secured perimeters,
computerized gates, and cage-like cells for living and visiting (Rhodes, 2004)—none of which sent
particularly welcoming messages to incarcerated persons or their visitors.
In contrast, in lower security environments, policies allowed for incarcerated persons to
access the visiting area free from restraints, and visitors experienced less invasive monitoring by
staff (Codd, 2008). Further, because visits in these units allowed for individuals to sit together at
tables, their ability to listen and communicate was not muffled by physical barriers or antiquated
speakers against a backdrop of institutional noise (Toch, 2001). Without the added institutional
stressors tied to overly restrictive conditions of confinement, it may be easier to have productive
discussions related to housing, employment, parenting, and interpersonal relationships (Turanovic
& Tasca, 2019). Incarcerated persons and visitors could have more private conversations, spend
longer periods of time together, and share physical touch. Simply put, in lower security settings
there may be more opportunities to achieve the goals of visitation. In overly restrictive conditions
of confinement, the benefits of visitation for reentry may be negated.
Second, a key finding was that visitation in close security units was no more protective
against recidivism than visitation in maximum security units. As such, our results should be
considered alongside the push for reforms that seek to expand the use of close custody
confinement. In the face of ongoing litigation over inhumane and overly restrictive housing
conditions (Haney, 2018), many state prison systems have proposed reducing the reliance on
maximum security imprisonment in favor of close security housing (Bertsch et al., 2018). In the
state prison system under study, however, close and maximum security units were nearly identical
in structure. Individuals confined in close custody, despite having some additional privileges, still
required their out-of-cell movements to be sparse and controlled, they were monitored intensely

26
by staff, and they faced numerous restrictions that carried over to visitation. Thus, downgrading
from maximum to close custody may not necessarily mean that the settings in which individuals
are housed and interact with visitors become appreciably less restrictive. Subsequent research
should further examine the consequences of confinement in close security prisons.
Third, given that the results were only correlational in nature, they raise important
questions about the causal processes underlying the associations observed. A critical next step will
be to examine how visitation is subjectively experienced by incarcerated persons and their visitors
across different prison settings. With administrative data, we were unable to measure directly the
quality or nature of interactions with visitors that can explain differences in the effectiveness of
visitation across custody levels. And due to the limitations with how the visitation data were
recorded by the prison system, we were also unable to examine the influence of who visited and
why (e.g., a current or former romantic partner, parent, sibling, friend, child) across conditions of
confinement. Ideally, a moderated mediation model should be tested that incorporates
characteristics of visitors and visit dynamics to explain the associations observed.
Notably, individuals housed in close and maximum custody units were less likely to be
married or have a high school degree compared to persons confined in minimum or medium
security settings. They were also more likely to be gang affiliated, to be serving time for a violent
crime, to have a history of attempted suicide, and to have more significant mental health treatment
needs. Given these histories, such individuals may have more strained and complicated
relationships with family members, which could make their visits more volatile, upsetting, and
unproductive (Beckmeyer & Arditti, 2014; Tasca et al., 2016; Turanovic & Tasca, 2019). These
dynamics may be particularly stressful when they play out within overly restrictive conditions of
confinement. As such, we recognize that some of the moderating relationships we uncovered might
be attributed, in part, to characteristics of high-risk persons, and not solely the higher custody
settings in which they received their visits. Even though we accounted for individual differences

27
in security classifications as rigorously as we could, there is a need for original data collection
efforts that better document incarcerated persons’ lived experiences within varying conditions of
confinement. Such data would allow researchers to better disentangle individual characteristics
from the features of prison settings that shape visitation dynamics across custody levels.
Given that there is much left to explore, an ecological approach—such as the one advanced
by Arditti (2005)—should be used as a guide for future work. An ecological model explicitly
recognizes that multiple systemic contexts shape the nature and quality of visitation, placement
into higher security facilities, and reentry outcomes. Ecological theory typically focuses on four
nested systems: the microsystem, mesosystem, exosystem, and macrosystem (Bronfenbrenner,
1979). Although there is a need to further identify the micro and mesosystemic processes that
shape the impacts of visitation on recidivism across conditions of confinement—such as by
examining features of visitation rooms, interactions between visitors and incarcerated persons,
treatment by prison staff, and the culture within housing units (Auty & Liebling, 2020; Moran,
2013)—there is also a need to recognize the influence of broader systemic processes. That is, the
associations between visitation, exposure to restrictive prison conditions, and recidivism, are
embedded in a larger socio-structural context. As such, they will further be influenced by the kinds
of ecological processes (e.g., poverty, racial inequality, weakened institutions of social support,
low collective efficacy) that set the stage for both the causes and consequences of incarceration
(Clear, 2007; Sampson & Loeffler, 2010; Wakefield & Uggen, 2010). The daunting—yet
important—task for future work will be to further determine how exosystemic and macrosystemic
factors, that include correctional policies, concentrated disadvantage, and structural inequality,
shape the relationships we observed here.
Finally, there are several limitations worth noting. For one, due to data constraints, causal
inferences could not be made. With other data sources, it may be possible to use alternate methods,
such as an instrumental variable approach, to draw firmer conclusions (Cochran et al., 2020).

28
Additionally, the findings are reflective of only one state prison system and may not generalize
elsewhere. In systems that use different custody protocols, have mixed-use units, or allow for less
restrictive visits in close and maximum security settings, the association between prison visitation
and recidivism may not vary as much across conditions of confinement. Relatedly, we could not
determine whether other benefits of visitation, such as reduced stress, lower misconduct, and
improved perceptions of support, were also undermined in higher security settings; or, if the
patterns we found were limited to recidivism. If so, it may suggest that visits, though beneficial in
other ways, may not be enough to help those confined in close and maximum security settings
overcome the harms of incarceration and the many challenges of reentry (Chen & Shapiro, 2007;
Western et al., 2015). Further, this study focused only on incarcerated men, and it is unknown if
the findings would be the same among women. In the system under study, conditions of
confinement and visitation policies were similar across men’s and women’s prisons, but few
women were housed in close and maximum security units. Little is known about the experiences
of women who serve their time and receive visits in highly restrictive settings (Aranda-Hughes et
al., 2021; Butler, 2019), and future research should delve deeper into this.
In the end, our work echoes calls for researchers to reignite an empirical focus on prison
conditions and prison life (Arditti, 2005; Kreager & Kruttschnitt, 2018; Wacquant, 2002)—similar
to the way things used to be done in early correctional research (Clemmer, 1940; Irwin & Cressey,
1962; Sykes, 1958). There is much left to learn about how people’s lived experiences in prison
might ultimately affect what happens to them after they get out (Wildeman et al., 2018). To better
understand the consequences of imprisonment—and to potentially reduce associated harms—
researchers must delve more deeply into individuals’ interactions with penal institutions across the
custody continuum. We hope that assessing variability across conditions of confinement, as we
have done here, is one step in that direction.

29
References
Adams, K. (1992). Adjusting to prison life. Crime and Justice, 16, 275-359.
Alper, M., Durose, M.R., & Markman, J. (2018). 2018 update on prisoner recidivism: A 9-year
follow-up period (2005-2014). Washington, DC: U.S. Department of Justice, Bureau of
Justice Statistics.
Aiello, B.L., & McCorkel, J.A. (2018). “It will crush you like a bug”: Maternal incarceration,
secondary prisonization, and children’s visitation. Punishment and Society, 20, 351-374.
Aranda-Hughes, V., Turanovic, J.J., Mears, D.P., & Pesta, G.B. (2021). Women in solitary
confinement: Relationships, pseudofamilies, and the limits of control. Feminist
Criminology, 16, 47-72.
Arditti, J.A. (2003). Locked doors and glass walls: Family visiting at a local jail. Journal of Loss
and Trauma, 8, 115-138.
Arditti, J.A. (2005). Families and incarceration: An ecological approach. Families in Society, 86,
251-260.
Arditti, J.A. (2012). Parental incarceration and the family: Psychological and social effects of
imprisonment on children, parents, and caregivers. New York: New York University
Press.
Atkin-Plunk, C.A., & Armstrong, G.S. (2018). Disentangling the relationship between social ties,
prison visitation, and recidivism. Criminal Justice and Behavior, 45, 1507-1526.
Auty, K.M., & Liebling, A. (2020). Exploring the relationship between prison social climate and
reoffending. Justice Quarterly, 37, 358-381.
Bales, W.D., & Mears, D.P. (2008). Inmate social ties and the transition to society: Does
visitation reduce recidivism? Journal of Research in Crime and Delinquency, 45, 287-
321.
Barrick, K., Lattimore, P.K., & Visher, C.A. (2014). Reentering women: The impact of social
ties on long-term recidivism. The Prison Journal, 94, 279-304.
Beckmeyer, J.J., & Arditti, J.A. (2014). Implications of in-person visits for incarcerated parents’
family relationships and parenting experience. Journal of Offender Rehabilitation, 53,
129-151.
Berg, M.T., & Huebner, B.M. (2011). Reentry and the ties that bind: An examination of social
ties, employment, and recidivism. Justice Quarterly, 28, 382-410.
Bertsch, L., Kempf, K., Mohr, G., Raemisch, R., Choinski, W., Resnik, J.,…Yordy, K. (2018).
Working to limit restrictive housing: Efforts in four jurisdictions to make changes.
Association of State Correctional Administrators. New Haven, CT: The Liman Center for
Public Interest Law, Yale Law School.

30
Boudin, C., Stutz, T., & Littman, A. (2013). Prison visitation policies: A fifty-state survey. Yale
Law and Policy Review, 32, 149-189.
Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and
design. Cambridge, MA: Harvard University Press.
Brunton-Smith, I., & McCarthy, D.J. (2017). The effects of prisoner attachment to family on re-
entry outcomes: A longitudinal assessment. The British Journal of Criminology, 57, 463-
482.
Butler, H.D. (2019). Understanding How in-Prison Experiences Influence Female Offenders’
Maladjustment to Prison. Justice Quarterly. Advance online publication.
Butler, H.D., Griffin, H., & Johnson, W.W. (2013). What makes you the “worst of the worst?”
An examination of state policies during supermaximum confinement. Criminal Justice
Policy Review, 24, 676-694.
Carson, E.A. (2020). Prisoners in 2019. Washington, DC: U.S. Department of Justice, Bureau of
Justice Statistics.
Chen, M.K., & Shapiro, J.M. (2007). Do harsher prison conditions reduce recidivism? A
Discontinuity-based approach. American Law and Economics Review, 9, 1-29.
Christian, J. (2005). Riding the bus: Barriers to prison visitation and family management
strategies. Journal of Contemporary Criminal Justice, 21, 34-48.
Clear, T.R. (2007). Imprisoning communities: How mass incarceration makes disadvantaged
neighborhoods worse. New York: Oxford University Press.
Clemmer, D. (1940). The prison community. New Braunfels, TX, US: Christopher Publishing
House.
Cochran, J.C. (2014). Breaches in the wall: Imprisonment, social support, and recidivism.
Journal of Research in Crime and Delinquency, 51, 200-229.
Cochran, J. C., Barnes, J. C., Mears, D. P., & Bales, W. D. (2020). Revisiting the effect of
visitation on recidivism. Justice Quarterly, 37(2), 304-331.
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.
Cochran, J.C., Mears, D.P., Bales, W.D., & Stewart, E.A. (2016). Spatial distance, community
disadvantage, and racial and ethnic variation in prison inmate access to social ties.
Journal of Research in Crime and Delinquency, 53, 220-254.
Codd, H. (2008). In the shadow of prison: Families, imprisonment, and criminal justice.
Portland, OR: Willan.
Comfort, M.L. (2002). “Papa’s house”: The prison as domestic and social satellite. Ethnography
3, 497-499.

31
Comfort, M.L. (2003). In the tube at San Quentin: The “secondary prisonization” of women
visiting inmates. Journal of Contemporary Ethnography, 32, 77-107.
Comfort, M.L. (2008). Doing time together: Love and family in the shadow of prison. Chicago,
IL: University of Chicago Press.
Conover, T. (2001). Newjack: Guarding Sing Sing. New York: Random House.
Cramer, L., Goff, M., Peterson, B., & Sandstrom, H. (2017). Parent-child visiting practices in
prisons and jails: A synthesis of research and practice. Washington, DC: Urban Institute.
Cullen, F.T. (2005). The twelve people who saved rehabilitation: How the science of
criminology made a difference. Criminology, 43, 1-42.
DiIulio, J.J. (1994). Let ‘em rot. The Wall Street Journal, January 26, A-14.
Duwe, G., & Clark, V. (2013). Blessed be the social tie that binds: The effects of prison
visitation on offender recidivism. Criminal Justice Policy Review, 24, 271-296.
Foster, H. (2016). The conditions of confinement in restrictive housing. In Restrictive housing in
the U.S.: Issues, challenges, and future directions (pp. 85-116). Washington, DC: U.S.
Department of Justice.
Hairston, C.F. (1988). Family ties during imprisonment: Do they influence future criminal
activity? Federal Probation, 52, 48-52.
Haney, C. (2018). Restricting the use of solitary confinement. Annual Review of Criminology, 1,
285-310.
Harding, D.J., Wyse, J.J.B, Dobson, C., & Morenoff, J.D. (2014). Making ends meet after
prison. Journal of Policy Analysis and Management, 33, 440-470.
Hickert, A., Tahamont, S., & Bushway, S. (2018). A tale of two margins: Exploring the
probabilistic processes that generate prison visits in the first two years of incarceration.
Journal of Quantitative Criminology, 34, 691-716.
Hickert, A., Palmen, H., Dirkzwager, A., & Nieuwbeerta, P. (2019). Receiving social support
after short-term confinement: How support pre- and during-confinement contribute.
Journal of Research in Crime and Delinquency, 56, 563-604.
Irwin, J., & Cressey, D.R. (1962). Thieves, convicts and the inmate culture. Social Problems, 10,
142-155.
Jacoby, J.E., & Kozie-Peak, B. (1997). The Benefits of social support for mentally ill offenders:
Prison-to-community transitions. Behavioral Sciences and the Law, 15, 483-501.
Jiang, S., & Winfree, L.T. (2006). Social support, gender, and inmate adjustment to prison life:
Insights from a national sample. The Prison Journal, 86, 32-55.
Johnston, N. (2000). Forms of constraint: A history of prison architecture. Chicago: University
of Illinois Press.

32
King, R.S., & Elderbroom, B. (2014). Improving recidivism as a performance measure.
Washington, DC: Urban Institute.
Kreager, D.A., & Kruttschnitt, C. (2018). Inmate society in the era of mass incarceration. Annual
Review of Criminology, 1, 261-283.
Labrecque, R.M. (2016). The use of administrative segregation and its function in the
institutional setting. In Restrictive housing in the U.S.: Issues, challenges, and future
directions (pp. 49-84). Washington, DC: U.S. Department of Justice.
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.
La Vigne, N.G., Naser, R.L., Brooks, L.E., & Castro, J.L. (2005). Examining the effect of
incarceration and in-prison family contact on prisoners’ family relationships. Journal of
Contemporary Criminal Justice, 21, 314-335.
Liebling, A. (1999). Prison suicide and prisoner coping. Crime and Justice, 26, 283-359.
Liebling, A. (2011). Moral performance, inhuman and degrading treatment and prison pain.
Punishment and Society, 13, 530-550.
Maruna, S. (2001). Making good: How ex-convicts reform and rebuild their lives. Washington,
DC: American Psychological Association.
Maruna, S., & Toch, H. (2005). The impact of imprisonment on the desistance process. In J.
Travis & C. Visher (Eds.), Prisoner reentry and crime in America (pp. 139-178). New
York: Cambridge University Press.
McNeeley, S., & Duwe, G. (2020). Keep your friends close and your enemies closer: Prison
visitation, spatial distance, and concentrated disadvantage of visitor neighborhoods, and
offender recidivism. Justice Quarterly, 37, 571-589.
Mears, D.P. (2013). Supermax prisons: The policy and the evidence. Criminology and Public
Policy, 12, 681-719.
Mears, D.P., & Bales, W.D. (2009). Supermax incarceration and recidivism. Criminology, 47,
1131-1166.
Mears, D.P., & Castro, J.L. (2006). Wardens’ views on the wisdom of supermax prisons. Crime
and Delinquency, 52, 398-431.
Mears, D.P., & Cochran, J.C. (2014). Prisoner reentry in the era of mass incarceration.
Thousand Oaks, CA: Sage.
Mears, D.P., Cochran, J.C., Siennick, S.E., & Bales, W.D. (2012). Prison visitation and
recidivism. Justice Quarterly, 29, 888-918.

33
Mears, D.P., Hughes, V., Pesta, G.B., Bales, W.D., Brown, J.M., Cochran, J.C., & Wooldredge.
J. (2019). The new solitary confinement? A conceptual framework for guiding and
assessing research and policy on ‘restrictive housing. Criminal Justice and Behavior, 46,
1427-1444.
Metcalf, H., Morgan, J., Oliker-Friedland, S., Resnick, J., Spiegel, J., Tae, H., Work, A., &
Holbrook, B. (2013). Administrative segregation, degrees of isolation, and incarceration:
A national overview of state and federal correctional policies. New Haven, CT: Liman
Public Interest Program, Yale Law School.
Mitchell, M.M., Spooner, K., Jia, D., & Zhang, Y. (2016). The effect of prison visitation on
reentry success: A meta-analysis. Journal of Criminal Justice, 47, 74-83.
Moran, D. (2013). Carceral geography and the spatialities of prison visiting: Visitation,
recidivism, and hyperincarceration. Environment and Planning D: Society and Space, 31,
174-190.
Mowen, T.J., & Visher, C.A. (2016). Changing the ties that bind: How incarceration impacts
family relationships. Criminology and Public Policy, 15, 503-528.
Pizarro, J., & Stenius, V.M.K. (2004). Supermax prisons: Their rise, current practices, and effect
on inmates. The Prison Journal, 84, 248-264.
Poehlmann, J., Dallaire, D., Loper, A.B., & Shear, L.D. (2010). Children’s contact with their
incarcerated parents: Research findings and recommendations. American
Psychologist, 65, 575-598.
Pratt, T.C. (2019). Addicted to incarceration: Corrections policy and the politics of
misinformation in the United States (2nd Ed). Thousand Oaks, CA: Sage.
Rhodes, L.A. (2004). Total confinement: Madness and reason in the maximum security prison.
Los Angeles: University of California Press.
Sampson, R.J., & Loeffler, C. (2010). Punishment's place: the local concentration of mass
incarceration. Daedalus, 139, 20-31.
Sturges, J.E., & Al-Khattar, A.M. (2009). Survey of jail visitors about visitation policies. The
Prison Journal, 89, 482-496.
Sundt, J.L., Castellano, T.C., & Briggs, C.S. (2008). The sociopolitical context of prison
violence and its control: A case study of supermax and its effect in Illinois. The Prison
Journal, 88, 94-122.
Sykes, G.M. (1958). The society of captives: A study of a maximum security prison. Princeton,
NJ: Princeton University Press.
Tasca, M. (2014). “It’s not all cupcakes and lollipops”: An investigation of the predictors and
effects of prison visitation for children during maternal and paternal incarceration.
Doctoral dissertation, Arizona State University.

34
Tasca, M., Mulvey, P., & Rodriguez, N. (2016). Families coming together in prison: An
Examination of visitation encounters. Punishment and Society, 18, 459-478.
Toch, H. (2001). The future of supermax confinement. The Prison Journal, 81, 376-388.
Turanovic, J.J., & Tasca, M. (2019). Inmates’ experiences with prison visitation. Justice
Quarterly, 36, 287-322.
Visher, C.A., & Travis, J. (2003). Transitions from prison to community: Understanding
individual pathways. Annual Review of Sociology, 29, 89-113.
Wacquant, L. (2002). The curious eclipse of prison ethnography in the age of mass
incarceration. Ethnography, 3, 371-397.
Wakefield, S., & Uggen, C. (2010). Incarceration and stratification. Annual Review of
Sociology, 36, 387-406.
Wall, A.T. (2013). Why do they do it that way? A response to Prison Visitation Policies: A Fifty-
State Survey. Yale Law and Policy Review, 32, 199-203.
Western, B., Braga, A.A., Davis, J., & Sirois, C. (2015). Stress and hardship after
prison. American Journal of Sociology, 120, 1512-1547.
Wildeman, C., Fitzpatrick, M.D., & Goldman, A.W. (2018). Conditions of confinement in
American prisons and jails. Annual Review of Law and Social Science, 14, 29-47.
Young, B.C., & Turanovic, J.J. (2020). Spatial distance as a barrier to visitation for incarcerated
youth and why families overcome it. Justice Quarterly. Advance online publication.

35
Table 1. Descriptive Statistics
Mean or % (SD)
Range
Dependent Variables
Rearrest
45.5%
0-1
Reconviction
42.8%
0-1
Return to prison
37.1%
0-1
Key Independent Variables
Visitation
Any visitation
49.6%
0-1
Visitation count
4.7 (9.2)
0-48
Visits in minimum security
2.2 (6.7)
0-48
Visits in medium security
1.8 (5.6)
0-48
Visits in close security
0.5 (3.1)
0-48
Visits in maximum security
0.3 (2.2)
0-48
Visitation consistency (months)
2.2 (3.6)
0-12
Visitation recency
1.1 (1.9)
0-6.5
Conditions of confinement
Minimum security housinga
41.0%
0-1
Medium security housing
41.3%
0-1
Close security housing
10.0%
0-1
Maximum security housing
7.7%
0-1
Control Variables
Housing unit transfers
2.0 (2.7)
0-49
Distance from prison to home (logged)
4.8 (0.8)
0.6-6.3
Risk score
5.1 (1.3)
3-10
Phase level
2.5 (0.8)
1-3
Prior felony convictions
6.4 (5.0)
0-15
Current offense
Violenta
32.8%
0-1
Property
29.3%
0-1
Drug
22.5%
0-1
Other non-violent
15.4%
0-1
Years in prison
3.3 (2.6)
1.0-21.3
Violent discipline reports
0.5 (1.2)
0-10
Property discipline reports
0.2 (0.7)
0-10
Drug discipline reports
0.4 (0.8)
0-10
Threat/intimidation reports
0.4 (1.1)
0-10
Security discipline reports
0.3 (0.6)
0-10
Defiance discipline reports
0.8 (1.5)
0-10
Other discipline reports
0.5 (1.2)
0-10
Gang affiliated
28.8%
0-1
History of attempted suicide
6.4%
0-1
Mental health score
1.6 (0.8)
1-4
Age at release
33.0 (10.6)
18-87
Race/ethnicity
Whitea
42.6%
0-1
Latino
35.6%
0-1
Black
14.5%
0-1
Native American
6.4%
0-1
Other race
0.9%
0-1
Undocumented immigrant
1.0%
0-1
Low educational attainment
26.9%
0-1
Married
15.8%
0-1
Number of dependents
1.3 (1.6)
0-15
Prison complex
Prison #1a
10.2%
0-1

36
Prison #2
9.8%
0-1
Prison #3
12.5%
0-1
Prison #4
0.9%
0-1
Prison #5
12.6%
0-1
Prison #6
6.1%
0-1
Prison #7
12.9%
0-1
Prison #8
5.2%
0-1
Prison #9
10.9%
0-1
Prison #10
2.2%
0-1
Prison #11
2.2%
0-1
Prison #12
2.3%
0-1
Prison #13
10.4%
0-1
Prison #14
2.0%
0-1
Notes: N = 17,542.
a Indicates the reference category.

37
Table 2. Logistic Regression Estimates, Rearrest
Model 1
Model 2
Variables
OR
(SE)
OR
(SE)
Key Independent Variables
Visitation
.754***
(.023)
.690***
(.033)
Visitation x Medium security housing
1.066
(.065)
Visitation x Close security housing
1.433***
(.139)
Visitation x Maximum security housing
1.375***
(.116)
Medium security housing
.986
(.046)
.953
(.048)
Close security housing
1.084
(.077)
.902
(.080)
Maximum security housing
1.229**
(.097)
1.052
(.102)
Control Variables
Housing unit transfers
1.015*
(.006)
1.016*
(.006)
Distance from prison to home
.965
(.038)
.963
(.037)
Risk score
1.044*
(.020)
1.045*
(.020)
Phase level
.821***
(.020)
.822***
(.020)
Prior felony convictions
1.083***
(.004)
1.083***
(.004)
Current offense
Property offense
1.157**
(.050)
1.157**
(.050)
Drug offense
.923
(.046)
.922
(.047)
Other non-violent offense
1.034
(.073)
1.037
(.074)
Years in prison
.862***
(.007)
.861***
(.007)
Violent discipline reports
1.003
(.023)
1.005
(.023)
Property discipline reports
1.051
(.029)
1.052
(.029)
Drug discipline reports
1.027
(.027)
1.027
(.027)
Threat/intimidation reports
1.088***
(.020)
1.087***
(.020)
Security discipline reports
1.073*
(.032)
1.070*
(.032)
Defiance discipline reports
1.032*
(.016)
1.034*
(.016)
Other discipline reports
1.044*
(.019)
1.045*
(.019)
Gang affiliated
1.421***
(.065)
1.417***
(.064)
History of attempted suicide
.903
(.054)
.905
(.054)
Mental health score
1.062*
(.030)
1.063*
(.030)
Age at release
.960***
(.002)
.960***
(.002)
Race/ethnicity
Latino
1.044
(.043)
1.044
(.043)
Black
1.226**
(.083)
1.223**
(.083)
Native American
1.327**
(.120)
1.330**
(.120)
Other race
.884
(.170)
.882
(.170)
Undocumented immigrant
.709
(.141)
.707
(.140)
Low educational attainment
1.019
(.033)
1.019
(.033)
Married
.881**
(.040)
.885**
(.040)
Number of dependents
1.012
(.012)
1.012
(.012)
Intercept
3.446***
(.839)
3.648***
(.890)
Notes: N = 17,542. Entries represent odds ratios (OR) and robust standard errors (SE) adjusted for the clustering of
individuals within housing units. Models also include variables for the prison complex where individuals served their
time in the year prior to release.
*p < .05; **p < .01; ***p < .001 (two-tailed test).

38
Table 3. Logistic Regression Estimates, Reconviction and Return to Prison
Reconviction
Return to Prison
Model 1
Model 2
Model 3
Model 4
Variables
OR
(SE)
OR
(SE)
OR
(SE)
OR
(SE)
Key Independent Variables
Visitation
.753***
(.023)
.700***
(.035)
.576***
(.018)
.543***
(.026)
Visitation x Medium security
1.050
(.072)
1.008
(.060)
Visitation x Close security
1.336**
(.138)
1.391**
(.140)
Visitation x Maximum security
1.321**
(.110)
1.249**
(.103)
Medium security housing
.967
(.044)
.941
(.048)
1.065
(.046)
1.060
(.055)
Close security housing
1.039
(.070)
.896
(.075)
.952
(.064)
.898
(.072)
Maximum security housing
1.159
(.090)
1.012
(.092)
1.210
(.164)
1.087
(.166)
Control Variables
Housing unit transfers
1.017*
(.007)
1.017*
(.007)
.989
(.007)
.990
(.008)
Distance from prison to home
.966
(.033)
.965
(.033)
.966
(.035)
.966
(.035)
Risk score
1.044*
(.019)
1.045*
(.019)
1.070***
(.021)
1.071***
(.021)
Phase level
.828***
(.020)
.828***
(.020)
.801***
(.021)
.802***
(.021)
Prior felony convictions
1.083***
(.005)
1.083***
(.005)
1.045***
(.004)
1.045***
(.004)
Current offense type
Property offense
1.158***
(.048)
1.158***
(.047)
1.430***
(.056)
1.429***
(.056)
Drug offense
.940
(.049)
.939
(.049)
1.071
(.056)
1.070
(.056)
Other non-violent offense
1.060
(.070)
1.063
(.071)
.898
(.064)
.901
(.064)
Years in prison
.860***
(.008)
.859***
(.008)
.969**
(.010)
.969**
(.010)
Violent discipline reports
1.004
(.022)
1.006
(.022)
.989
(.023)
.991
(.023)
Property discipline reports
1.064*
(.028)
1.065*
(.028)
1.040
(.030)
1.040
(.030)
Drug discipline reports
1.027
(.026)
1.027
(.026)
1.145***
(.028)
1.146***
(.028)
Threat/intimidation reports
1.079***
(.021)
1.078***
(.021)
1.147***
(.023)
1.146***
(.023)
Security discipline reports
1.069*
(.028)
1.067*
(.027)
1.038
(.026)
1.035
(.026)
Defiance discipline reports
1.029
(.016)
1.030
(.016)
1.069***
(.014)
1.070***
(.014)
Other discipline reports
1.051*
(.021)
1.051*
(.021)
1.075***
(.022)
1.076***
(.022)
Gang affiliated
1.463***
(.061)
1.460***
(.061)
1.504***
(.066)
1.502***
(.065)
History of attempted suicide
.882*
(.053)
.884*
(.052)
1.235**
(.095)
1.237**
(.094)
Mental health score
1.079**
(.029)
1.080**
(.029)
.732***
(.034)
.733***
(.034)
Age at release
.960***
(.002)
.959***
(.002)
.982***
(.002)
.982***
(.002)
Race/ethnicity
Latino
1.036
(.048)
1.036
(.048)
.813***
(.033)
.813***
(.033)
Black
1.236**
(.081)
1.234**
(.081)
1.159**
(.065)
1.156**
(.064)
Native American
1.382**
(.129)
1.385***
(.129)
1.103
(.078)
1.104
(.078)
Other race
.917
(.187)
.915
(.187)
.935
(.181)
.933
(.181)
Undocumented immigrant
.661*
(.124)
.659*
(.123)
.700
(.193)
.698
(.193)
Low educational attainment
1.015
(.035)
1.015
(.035)
1.070
(.039)
1.070
(.039)
Married
.872**
(.042)
.875**
(.042)
.785***
(.043)
.788***
(.043)
Number of dependents
1.007
(.011)
1.007
(.011)
.992
(.009)
.991
(.009)
Intercept
2.957***
(.705)
3.100***
(.727)
2.378**
(.708)
2.458**
(.726)
Notes: N = 17,542. Entries represent odds ratios (OR) and robust standard errors (SE) adjusted for the clustering of
individuals within housing units. Models also include variables for the prison complex where individuals served their
time in the year prior to release.
*p < .05; **p < .01; ***p < .001 (two-tailed test).

39
Table 4. Logistic Regression Estimates, Rearrest (Visitation Frequency)
Number of Visits
Visits per Condition
Model 1
Model 2
Model 3
Variables
OR
(SE)
OR
(SE)
OR
(SE)
Key Independent Variables
Visitation count
.985***
(.003)
.982***
(.004)
Visitation x Medium security
1.004
(.006)
Visitation x Close security
1.015**
(.005)
Visitation x Maximum security
1.013
(.008)
Visits in minimum security
.983***
(.004)
Visits in medium security
.981***
(.005)
Visits in close security
1.000
(.004)
Visits in maximum security
1.000
(.008)
Medium security housing
.993
(.046)
.975
(.053)
Close security housing
1.092
(.076)
1.016
(.078)
Maximum security housing
1.238**
(.098)
1.167
(.096)
Control Variables
Housing unit transfers
1.014*
(.006)
1.015*
(.006)
1.017**
(.006)
Distance from prison to home
.934
(.036)
.931
(.036)
.931
(.035)
Risk score
1.046*
(.020)
1.046*
(.020)
1.045*
(.019)
Phase level
.820***
(.020)
.821***
(.020)
.824***
(.020)
Prior felony convictions
1.082***
(.004)
1.082***
(.004)
1.082***
(.004)
Current offense
Property offense
1.155**
(.050)
1.156**
(.051)
1.157**
(.052)
Drug offense
.925
(.046)
.926
(.046)
.927
(.047)
Other non-violent offense
1.027
(.073)
1.028
(.074)
1.030
(.074)
Years in prison
.860***
(.007)
.860***
(.007)
.860***
(.007)
Violent discipline reports
1.005
(.023)
1.005
(.023)
1.010
(.012)
Property discipline reports
1.049
(.029)
1.052
(.029)
1.055
(.029)
Drug discipline reports
1.025
(.027)
1.026
(.027)
1.027
(.027)
Threat/intimidation reports
1.089***
(.021)
1.087***
(.021)
1.087***
(.021)
Security discipline reports
1.076*
(.032)
1.075*
(.032)
1.082**
(.031)
Defiance discipline reports
1.032*
(.016)
1.033*
(.016)
1.036*
(.016)
Other discipline reports
1.046*
(.019)
1.047*
(.019)
1.045*
(.019)
Gang affiliated
1.415***
(.064)
1.415***
(.064)
1.425***
(.062)
History of attempted suicide
.903
(.053)
.904
(.053)
.898
(.052)
Mental health score
1.064*
(.030)
1.065*
(.030)
1.065*
(.030)
Age at release
.961***
(.002)
.961***
(.002)
.961***
(.002)
Race/ethnicity
Latino
1.037
(.042)
1.036
(.043)
1.036
(.043)
Black
1.241**
(.085)
1.240**
(.085)
1.234**
(.084)
Native American
1.328**
(.119)
1.328**
(.119)
1.334**
(.117)
Other race
.887
(.168)
.885
(.168)
.885
(.168)
Undocumented immigrant
.706
(.139)
.707
(.139)
.706
(.139)
Low educational attainment
1.028
(.033)
1.030
(.033)
1.034
(.033)
Married
.878**
(.040)
.880**
(.040)
.881**
(.040)
Number of dependents
1.011
(.012)
1.011
(.012)
1.011
(.012)
Intercept
3.485***
(.864)
3.601***
(.888)
3.570***
(.879)
Notes: N = 17,542. Entries represent odds ratios (OR) and robust standard errors (SE) adjusted for the clustering of
individuals within housing units. Models also include variables for the prison complex where individuals served the
majority of their time in the year prior to release. Models 1 and 2 measure visitation using an overall count of visits,
and Model 3 measures visitation using the count of visits received in each condition of confinement.
*p < .05; **p < .01; ***p < .001 (two-tailed test).

40
Table 5. Logistic Regression Estimates, Rearrest (Visitation Timing)
Visitation Consistency (Months)
Visitation Recency
Model 1
Model 2
Model 3
Model 4
Variables
OR
(SE)
OR
(SE)
OR
(SE)
OR
(SE)
Key Independent Variables
Visitation consistency
.970***
(.006)
.955***
(.007)
Visitation recency
.946***
(.012)
.920***
(.014)
Visitation x Med. security
1.021
(.014)
1.034
(.027)
Visitation x Close security
1.045***
(.012)
1.099***
(.025)
Visitation x Max. security
1.044**
(.015)
1.079**
(.029)
Medium security housing
.992
(.045)
.946
(.051)
.993
(.045)
.955
(.052)
Close security housing
1.089
(.076)
.983
(.070)
1.092
(.076)
.975
(.070)
Maximum security housing
1.243**
(.100)
1.128
(.089)
1.246**
(.101)
1.139
(.090)
Control Variables
Housing unit transfers
1.014*
(.006)
1.015*
(.006)
1.014*
(.006)
1.015*
(.006)
Distance from home
.955
(.036)
.952
(.035)
.953
(.036)
.949
(.035)
Risk score
1.046*
(.020)
1.046*
(.020)
1.046*
(.020)
1.046*
(.020)
Phase level
.819***
(.020)
.819***
(.020)
.819***
(.020)
.819***
(.020)
Prior felony convictions
1.083***
(.004)
1.083***
(.004)
1.083***
(.004)
1.083***
(.004)
Current offense
Property offense
1.161**
(.051)
1.160**
(.051)
1.162**
(.051)
1.162**
(.051)
Drug offense
.927
(.047)
.926
(.047)
.927
(.047)
.927
(.047)
Other non-violent offense
1.026
(.073)
1.025
(.073)
1.027
(.073)
1.026
(.073)
Years in prison
.861***
(.007)
.860***
(.007)
.861***
(.007)
.860***
(.007)
Violent discipline reports
1.004
(.023)
1.004
(.023)
1.004
(.023)
1.004
(.023)
Property discipline reports
1.052
(.029)
1.055
(.029)
1.052
(.029)
1.055
(.029)
Drug discipline reports
1.026
(.027)
1.027
(.027)
1.026
(.027)
1.027
(.027)
Threat/intimidation reports
1.088***
(.021)
1.087***
(.021)
1.088***
(.021)
1.086***
(.021)
Security discipline reports
1.074*
(.032)
1.074*
(.032)
1.074*
(.032)
1.073*
(.032)
Defiance discipline reports
1.033*
(.017)
1.035*
(.016)
1.034*
(.017)
1.035*
(.016)
Other discipline reports
1.045*
(.019)
1.045*
(.019)
1.045*
(.019)
1.046*
(.019)
Gang affiliated
1.416***
(.064)
1.416***
(.065)
1.416***
(.064)
1.417***
(.065)
History of attempted suicide
.909
(.054)
.910
(.054)
.910
(.054)
.910
(.054)
Mental health score
1.063*
(.030)
1.065*
(.030)
1.063*
(.030)
1.064*
(.030)
Age at release
.962***
(.002)
.961***
(.002)
.962***
(.002)
.962***
(.002)
Race/ethnicity
Latino
1.036
(.042)
1.036
(.043)
1.035
(.042)
1.035
(.042)
Black
1.238**
(.085)
1.237**
(.085)
1.238**
(.085)
1.237**
(.085)
Native American
1.336**
(.121)
1.341**
(.121)
1.337**
(.122)
1.342**
(.122)
Other race
.882
(.168)
.886
(.168)
.881
(.168)
.885
(.168)
Undocumented immigrant
.714
(.141)
.716
(.140)
.715
(.140)
.715
(.139)
Low educational attainment
1.029
(.033)
1.032
(.033)
1.031
(.033)
1.034
(.033)
Married
.873**
(.040)
.876**
(.040)
.871**
(.040)
.875**
(.040)
Number of dependents
1.011
(.012)
1.010
(.012)
1.011
(.012)
1.010
(.012)
Intercept
3.173***
(.770)
3.347***
(.801)
3.156***
(.768)
3.325***
(.797)
Notes: N = 17,542. Entries represent odds ratios (OR) and robust standard errors (SE) adjusted for the clustering of
individuals within housing units. Models also include variables for the prison complex where individuals served the
majority of their time in the year prior to release. Models 1 and 2 measure visitation using the number of months a
visit was received, and Models 3 and 4 measure visitation using a weighted recency measure.
*p < .05; **p < .01; ***p < .001 (two-tailed test).

41
Figure 1. Predicted Probabilities of Rearrest, by Conditions of Confinement
.491 .481 .469
.502
.412 .416
.467
.491
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.55
Minimum Security Medium Security Close Security Maximum Security
Not visited Visited

42
Figure 2. Predicted Probabilities of Rearrest, by Housing Unit and Conditions of Confinement
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Minimum Security
Not Visited Visited
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33
Medium Security
Not Visited Visited
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Close Security
Not Visited Visited
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35
Maximum Security
Not Visited Visited

43
Appendix A. Descriptive Statistics by Conditions of Confinement
Minimum Security
Housing
Medium Security
Housing
Close Security
Housing
Maximum Security
Housing
Mean or % (SD)
Mean or % (SD)
Mean or % (SD)
Mean or % (SD)
Range
Dependent Variables
Rearrest**
39.3%
44.9%
59.1%
63.8%
0-1
Reconviction**
36.9%
42.0%
55.7%
60.7%
0-1
Return to prison**
29.3%
38.3%
48.1%
58.1%
0-1
Key Independent Variables
Any visitation**
53.7%
46.8%
49.5%
43.5%
0-1
Visitation count**
5.3 (9.9)
4.3 (8.7)
4.8 (9.0)
3.9 (8.3)
0-48
Visits in minimum security**
5.0 (9.5)
0.3 (1.6)
0.2 (1.2)
0.2 (1.6)
0-48
Visits in medium security**
0.2 (1.3)
3.9 (8.2)
0.4 (1.7)
0.2 (1.6)
0-48
Visits in close security**
0.0 (0.7)
0.1 (1.0)
4.2 (8.4)
0.4 (2.2)
0-48
Visits in maximum security**
0.0 (0.3)
0.0 (0.6)
0.1 (0.8)
3.1 (7.1)
0-48
Visitation consistency (months)**
2.5 (3.8)
2.0 (3.4)
2.2 (3.5)
1.9 (3.3)
0-12
Visitation recency**
1.3 (2.0)
1.0 (1.8)
1.1 (1.9)
1.0 (1.8)
0-6.5
Control Variables
Housing unit transfers**
1.4 (2.0)
2.0 (2.6)
2.8 (3.0)
3.8 (4.9)
0-49
Distance from prison to home (log)**
4.9 (0.8)
4.9 (0.8)
4.7 (0.8)
4.4 (0.7)
0.6-6.3
Risk score**
4.5 (1.2)
5.3 (1.2)
5.9 (1.2)
6.3 (1.3)
3-10
Phase level**
2.6 (0.7)
2.4 (0.8)
2.2 (0.8)
2.2 (0.8)
1-3
Prior felony convictions**
6.6 (4.9)
6.2 (5.0)
6.6 (5.1)
6.6 (5.0)
0-15
Current offense**
Violenta
27.5%
47.9%
46.6%
53.9%
0-1
Property
32.3%
27.9%
35.2%
31.8%
0-1
Drug
25.2%
17.4%
12.8%
9.7%
0-1
Other non-violent
15.0%
6.8%
5.4%
4.6%
0-1
Years in prison**
2.8 (2.2)
3.4 (2.8)
3.7 (2.6)
4.4 (3.0)
1.0-21.3
Violent discipline reports**
0.2 (0.7)
0.4 (1.1)
0.8 (1.4)
1.9 (2.2)
0-10
Property discipline reports**
0.1 (0.4)
0.2 (0.6)
0.5 (0.9)
0.8 (1.5)
0-10
Drug discipline reports**
0.2 (0.6)
0.4 (0.8)
0.7 (1.2)
0.6 (1.1)
0-10
Threat/intimidation reports**
0.2 (0.8)
0.4 (1.1)
0.8 (1.5)
1.2 (1.8)
0-10
Security discipline reports**
0.1 (0.4)
0.2 (0.6)
0.6 (0.9)
0.8 (1.1)
0-10
Defiance discipline reports**
0.3 (0.9)
0.8 (1.4)
1.8 (2.1)
2.4 (2.5)
0-10
Other discipline reports**
0.3 (0.8)
0.5 (1.1)
0.8 (1.5)
1.2 (1.9)
0-10
Gang affiliated**
13.5%
30.8%
53.3%
66.9%
0-1
History of attempted suicide**
4.4%
7.4%
8.3%
8.6%
0-1

44
Mental health score**
1.5 (0.7)
1.7 (0.8)
1.8 (0.9)
1.9 (0.9)
1-4
Age at release**
35.2 (10.4)
33.2 (10.9)
28.7 (9.0)
26.8 (7.5)
18-87
Race/ethnicity**
Whitea
46.5%
43.1%
34.7%
29.8%
0-1
Latino
32.4%
34.5%
43.9%
47.0%
0-1
Black
13.4%
16.0%
15.0%
11.5%
0-1
Native American
6.5%
5.6%
5.8%
11.2%
0-1
Other race
1.2%
0.8%
0.6%
0.5%
0-1
Undocumented immigrant*
1.2%
0.9%
0.5%
0.3%
0-1
Low educational attainment**
22.5%
26.6%
33.2%
43.7%
0-1
Married**
16.9%
16.3%
12.5%
11.5%
0-1
Number of dependents*
1.4 (1.6)
1.3 (1.7)
1.3 (1.6)
1.2 (1.6)
0-15
Prison complex **
Prison #1a
19.0%
5.7%
1.4%
1.0%
0-1
Prison #2
0.0%
13.7%
0.0%
53.5%
0-1
Prison #3
13.9%
9.6%
0.0%
34.3%
0-1
Prison #4
0.0%
1.6%
0.0%
2.9%
0-1
Prison #5
4.3%
13.6%
49.5%
2.4%
0-1
Prison #6
13.1%
1.8%
0.0%
0.0%
0-1
Prison #7
12.9%
10.6%
29.8%
3.1%
0-1
Prison #8
8.5%
4.0%
0.0%
0.6%
0-1
Prison #9
12.6%
9.0%
19.3%
1.4%
0-1
Prison #10
0.0%
5.4%
0.0%
0.1%
0-1
Prison #11
5.3%
0.0%
0.0%
0.1%
0-1
Prison #12
5.5%
0.0%
0.0%
0.0%
0-1
Prison #13
0.0%
25.0%
0.0%
0.1%
0-1
Prison #14
4.9%
0.0%
0.0%
0.6%
0-1
N
7,187
7,248
1,762
1,345
Notes: All measures varied significantly across conditions of confinement. Differences were assessed using chi-square tests (for categorical variables) and
one-way ANOVA (for continuous variables).
a Indicates the reference category.
*p < .01; **p < .001 (two-tailed test).