Abstract
Evidence suggests that visits act as a mechanism to restrain antisocial behavior within prison settings. However, evidence that visits may promote prosocial behavior, such as program engagement, is less prevalent. Using a random sample of incarcerated men, the current study uses multilevel logistic regression models to determine whether the presence and quality of visits during a period of program enrollment is associated with an increase in the odds of completing the enrollment period. Results suggest that experiencing a visit during enrollment, especially when visits involve multiple persons, increases the likelihood of completing the enrollment period. These results hold even when accounting for prior visitation experiences. Implications of these findings support widening the availability of visit opportunities and modalities.
Introduction
The crisis of mass incarceration has produced an abundance of negative life outcomes for incarcerated persons (IP) and the communities they are absent from, including the severance of familial and friendship ties, long-term damage to opportunity and employment prospects, and perpetuation of social and financial disadvantage (Western & Pettit, 2010). Recent decades have evidenced a shift away from the punitive, get-tough ideals which characterized carceral systems of the 1990s and early 2000s. Rather, administrators and policymakers seek prison practices that will provide rehabilitative environments to combat recidivism and foster stable lifestyles in the community. At the same time, these parties value practices that are cost effective, maintain facility safety and security, and benefit the widest range of IP.
Both visitation and prison programs offer within-prison opportunities to build social and human capital. Visits from family and friends are integral to the maintenance of networks to provide instrumental (i.e., monetary, housing) and expressive (i.e., emotional) resources after release (Pleggenkuhle et al., 2016). Further, visits can foster a sense of optimism and hope for reentry efforts among IP (Visher & O’Connell, 2012). Similarly, evidence-based prison education, vocation, and life skills programs administered with fidelity are associated with post-release gainful employment and reductions in subsequent justice contact (Bozick et al., 2018; Newton et al., 2018). Less studied is how visitation may itself act as a mechanism to promote program engagement and completion.
Recent renewed interest in breaking open the “black box” of incarceration provides evidence that the wide heterogeneity in prison experiences is impactful for within-prison behavior and, in turn, post-release outcomes (Mears et al., 2015). Importantly, these experiences are interconnected in complex patterns that can foster or preclude attitudes and efforts for reform (see Cochran, 2012; Western, 2021). The present study seeks to add to this nuanced literature by examining whether visitation acts to promote persistence in prison program participation to foster completion of a period of program enrollment. Findings suggest that visits experienced within the period of enrollment are beneficial for program completion, even beyond the general experience of receiving or not receiving visits during an incarceration. Results bolster evidence that visitation acts as a mechanism to promote prosocial behaviors, and support calls for the expansion of in-person visitation opportunities for incarcerated persons and their families.
Visitation and Prison Behavior
A growing body of literature suggests that IP who receive visits while incarcerated engage in fewer infractions and violent behaviors while behind bars (Cihan et al., 2020; Cochran, 2014; McNeeley & Duwe, 2020). The positive, restraining force of visits is most often viewed through the theoretical lens of informal social control mechanisms. Visitation is perceived as a proxy measure for social support and social bonds, including such elements as attachments to valued others and commitments to conventional institutions and behavior (Cochran & Mears, 2013; Hirschi, 1969). Further, social bonds are thought to strengthen or weaken in the face of intervening events that restructure routine activities and alter perceived opportunity structures (Laub & Sampson, 1993). Importantly, theorists suggest that social bonds must be actively maintained (Laub et al., 1998), a difficult task when incarcerated. Visits offer a mechanism of social bond maintenance, ensuring contact with valued others and fostering attachment (Cochran & Mears, 2013).
Evidence supports visitation as a conduit for social bond maintenance. Anderson et al. (2020) find proximal effects of visitation on misconduct, suggesting that the experience of visits is not enough compared to the “activation” of social bonds through recency of visits (see also Mowen & Visher, 2016). Receiving visits helps to maintain general social ties and relationships with those who are likely to provide tangible and emotional support during reentry efforts, making them an essential component of rehabilitative efforts (Dixey & Woodall, 2012; Meyers et al., 2017; Mowen & Visher, 2016).
Visits may also work to alleviate the harsh experience of incarceration. The deprivation model of prison behavior suggests that IP internalize the norms of institutional life (Clemmer, 1940). The oppressive nature of the prison facility deprives individuals of resources such as social relationships, liberty and autonomy, goods and services, and security; collectively, such deprivations are known as “pains of imprisonment” (Sykes, 1958). Such deprivations are likened to general strains which induce negative emotions and, under certain circumstances, induce maladaptive coping mechanisms (Blevins et al., 2010). Visits can both alleviate and exacerbate the pains of imprisonment. With respect to the former, visits alleviate pains of imprisonment such as removal from social support networks. IP report that receiving visits assists in breaking up the monotony of prison life by giving them something to look forward to (Dixey & Woodall, 2012; Turanovic & Tasca, 2019), and that seeing family, friends, and children sustains crucial support systems (Dixey & Woodall, 2012; Liu et al., 2016). Further, pre-posttests suggest that negative emotionality is substantially reduced after visits, supporting hypotheses that visits ease strain and reduce likelihood of antisocial behaviors (Bachar & Guetzkow, 2023).
Conversely, qualitative evidence suggests that the perception of visits is variable. IP report that visits can increase the strain of imprisonment, with some IP opting to discourage visits (Pleggenkuhle et al., 2018). Visits can increase the perceptions of deprivations due to repeated exposure of separation from loved ones (i.e., when visitors leave, IP cannot follow), knowledge that visitors experience financial and emotional difficulties associated with visits, and negative visit experiences such as in the case of children who do not wish to engage with incarcerated parents during visits (Dixey & Woodall, 2012; Pierce, 2015; Pleggenkuhle et al., 2018). Indeed, several studies find an increase in antisocial behavior in the weeks directly following visits (Berghuis et al., 2024; Siennick et al., 2013), suggesting potential increases in negative emotionality that trigger antisocial coping mechanisms. Though this evidence is contrary to hypotheses of positive visitation effects, authors find that the majority of IP who receive visits perceive them as positive events for the quality of their incarceration and positive emotional stimulus (Turanovic & Tasca, 2019).
Dimensions of Visitation: Quantity and Quality
Authors suggest that it is not simply the experience of visitation that impacts prison behavior, but also the quality and consistency of visits. Anderson et al. (2020) propose that visits act as a conduit through which social capital is activated, suggesting that receiving sporadic visits is not enough to promote prosocial behavior and post-visit benefits. Examinations of longitudinal patterns of visitation and prison behavior support this argument. Cihan et al. (2020) found that prison misconduct is predicted by trajectories of visitation, where those who receive visits early and more consistently are less likely to engage in misbehavior (see also Cochran, 2014). The importance of timing and chronological proximity for positive benefits of visitation may explain some lukewarm findings regarding visitation which suggest that visits are more indicative of social support rather than a beneficial mechanism on their own (Atkin-Plunk & Armstrong, 2018). The frequency and consistency of visits vary substantially across IP (Pleggenkuhle et al., 2018), but are associated with outcomes such as within-prison behavior, perceptions of visit quality, and optimism for post-release support and success (Baker et al., 2022; Duwe & McNeeley, 2021; Turanovic & Tasca, 2019).
Visits also vary in multiple dimensions that impact perceptions of visit quality. Cochran and Mears (2013) call for better research on the heterogeneity of the visitation experience, as visits vary along such dimensions as timing, rate and consistency, type of visit, and closeness to visitors. Evidence suggests, for instance, that many IP feel that visit options, both in-person as well as phone and video-calls, are too short to provide meaningful interactions with family and friends (Dixey & Woodall, 2012; Pierce, 2015). Other IP report that in-person visits which go on for too long are fatiguing and stressful (Dixey & Woodall, 2012). A number of studies also suggest that certain visitors produce negative visit experiences. Family and friends can generate stress for IP through visits that lead to arguments, shame, or guilt for the individual’s incarceration status (Baker et al., 2022; Dixey & Woodall, 2012; La Vigne et al., 2005; Pleggenkuhle et al., 2016; Turanovic & Tasca, 2019).
Visits that generate negative emotions or strain already weakened family ties may thus have an undesirable effect on IP behavior. Indeed, some evidence finds an increase in antisocial behavior after receiving visitors that are perceived as negative or stress-inducing (Atkin-Plunk & Armstrong, 2018; La Vigne et al., 2005). Further examinations of the dimensions of visitation as they relate to IP behavior are thus warranted.
Visitation as a Catalyst for Prosocial Behavior
The bulk of studies examining the impact of visitation on in-prison behavior examine how visitation acts as a restraining mechanism for antisocial behavior such as misconduct. Fewer studies seek to determine whether visitation can promote engagement in prosocial activities, including program engagement, beyond the absence of antisocial behaviors.
There is reason to test whether visitation can spur prosocial behavior rather than reduce the incidence of unwanted behavior. Theories of informal social control and identity change suggest that social ties and capital are mechanisms of short- and long-term behavioral change. Visits foster commitment to conventional values through long-term goal setting. IP report that visits remind them of the collateral costs of their imprisonment, including the impact of their removal from networks and the burden placed on visitors (Liu et al., 2016). The internalization of informal control mechanisms through the social bond can assist in the formation of long-term commitment to conventional goals, leading IP to develop prosocial habits that will contribute to desistance and reestablishment of familial roles (Maruna, 2001; Mills & Codd, 2008).
Visits may also act to facilitate movement toward new identities for incarcerated individuals. Identity change is considered to be a longitudinal process that is instigated by openness to change (Giordano et al., 2002). Several theories of desistance suggest that an individual must commit to identity change before they take steps to achieve a new positive identity (Paternoster & Bushway, 2009). This process occurs when the cost of current antisocial or offending behavior comes to outweigh benefits gleaned from offending (Paternoster & Bushway, 2009). Incarceration itself may act as the catalyst to impose sufficient “costs” of offending lifestyles (Hickert et al., 2021). Enrollment in prison programs is thought to signal those who are ready for substantive lifestyle changes, providing initial evidence that this group is in the beginning stages of prosocial identity transformation (Bushway & Apel, 2012). Importantly, once the individual is open to change and prepared to take tangible steps toward a desired prosocial identity, social relationships become integral to the positive affirmation of this path and to the facilitation of opportunities to achieve this identity (Giordano et al., 2002; Paternoster & Bushway, 2009). Visits thus offer a mechanism through which identity change is encouraged and dedication to conventional opportunities, like prison programs, is fostered.
These issues are particularly salient to prison programs, as many IP perceive programs to provide useful skill and knowledge accumulation that will be of use during reentry (Richmond, 2014). Importantly, completing a program is shown to provide substantial benefits beyond participation alone. These include such elements as longer and more complete program dosage, signals of desistance to community members, and better skill and knowledge acquisition (Bushway & Apel, 2012; Duwe, 2018; Pompoco et al., 2017). Unfortunately, completing a period of enrollment is not assured. Studies reveal that internal factors, such as motivation and ability (Prendergast et al., 2009; Silver et al., 2020), and external factors, such as program quality and rehabilitative environment (Gendreau et al., 2016), are associated with program persistence and completion. However, far less work examines how in-prison experiences like visitation may act on program completion, though the evidence above suggests that visits are likely to provide a catalyst for prosocial identity change and tangible steps toward conventional lifestyles, as in program participation.
There is some initial evidence that visits can spur participation in prosocial activities in prison. Brosens et al. (2016) found that receiving visits increases the likelihood of participation in prison work and programs. Similarly, Rose and Rose (2014) found that receiving visits from children is associated with increased likelihood of enrollment in college courses while incarcerated. While these studies offer initial evidence that visits are associated with prosocial behaviors, they examine the positive impact of visits through coarse measures, such as whether the individual ever received a visit rather than the timing or chronology of visit and program events. Further, these studies do not account for various dimensions of visit quality, such as length of visits or commonality of visits.
The Current Study
The current study adds to literature examining the impact of visits on IP behavior through program completion. Two hypotheses are tested based on theory and empirical evidence from prior literature. Drawing from literature suggesting that visitation restrains antisocial behavior and can promote prosocial activity engagement:
Hypothesis 1: Incarcerated persons who receive a visit during a period of program enrollment will be more likely to complete the semester of enrollment.
Given evidence that visits vary along many dimensions that, in turn, impact the quality of visits and subsequent behavior, the following hypothesis is examined:
Hypothesis 2: Incarcerated persons who receive visits that are longer and involve multiple visitors will be more likely to complete the semester of enrollment.
The current study seeks to improve on facets of prior work examining visitation and prison behavior in several ways. First, the current study examines visitation as an inducer of prosocial behavior rather than a restraint mechanism for antisocial behavior. The majority of prior work assessing the impact of visitation on behavior of IP examines misconduct and post-release recidivism (e.g., Cochran, 2012; Mears et al., 2012). While these studies reveal a great deal about the positive impacts of visitation, they do not suggest that visits act to strengthen engagement in conventional activities, such as prison programs. Second, the current study uses a multilevel modeling method to adjust for within-person propensity to successfully persist in programs until the end of an enrollment period. Multilevel models are useful to examine within- and between-unit variance in outcomes of interest, and have been used previously in the examination of both visits and prison programs (Pelissier et al., 2003; Turanovic & Tasca, 2019). Third, dated administrative files allow for the examination of chronological proximity of visitation experiences to the primary outcome of interest, an issue that is understudied. Finally, the current study employs two measures of visitation quality to examine the dimensions of prison visitation rather than solely the presence or absence of visits.
Data
The data represent a random sample drawn from the 2016, 2017, and 2018 release cohorts of men incarcerated in a large Northeastern state prison system. 1 The data are administrative files provided by the state’s Department of Corrections (DOC) and were input by correctional staff responsible for overseeing the experiences of IP (i.e., staff who detected and processed misconduct tickets input data on misconduct events, staff responsible for intake interviews input demographic and risk information, etc.). Data include detailed information on IP characteristics and activities during their stay, including intake classification, convicting offense and sentence, programming and prison work, misconduct events and processing, and visitation. All events are dated to create a time-ordered series of events that transpire from intake to release.
Data are structured at the person-enrollment level, where an enrollment period is defined as the program start and ends dates which are loosely comparable to semester periods (though programs within this dataset range in duration). During an enrollment period, the individual may either complete the program in its entirety or make sufficient progress to enroll in the next enrollment period. At any time during an enrollment period, the individual may voluntarily dropout, be dropped administratively, fail to meet program expectations, or experience another program interruption such as facility transfer or release; the latter are excluded from analyses, as these experiences are outside the individual’s control (see Pelissier et al., 2003 for similar methodology). Visits from attorneys are also excluded from the analyses, as they do not represent social support systems. 2
The original data transferred by DOC included 12,000 unique person-stays. It was found that 369 files were duplicates. Further, 504 were mistakenly transferred without intake records, 207 individuals had been reincarcerated within the same year and thus historical records did not represent their current incarceration, and 74 files were for persons of a race/ethnicity other than White, Black, or Hispanic/Latino.
3
The current sample is further restricted to those who participated in a program at some point during their incarceration to create the most comparable standards among observations. Of the remaining sample, 3,239 individuals were not documented as participating in a program during their stay and 273 were enrolled in ongoing programs (i.e., specialty programs with no natural termination date, such as outpatient mental health care). Finally, 1,084 individuals displayed missingness on visitation, intake/legal, and/or programming data. The nature of administrative files produces some data missingness that is unavoidable. A comparison of the eligible and ineligible samples can be found in the Appendix A. The final sample of analysis yielded 6,250 unique persons with a combined 35,702 enrollment periods (
Analytic Method
Analyses progress in several stages. First, data are examined descriptively and Chi-square analyses are employed to assess the relationship between visitation and program completion. Second, I estimate two-level logistic regression models including random intercepts for each individual. Level 1 is characterized as the enrollment period (where the outcome is whether the individual completed the current enrollment period; discussed further below), while Level 2 is the individual (who may enroll in multiple program periods during their incarceration). 4
Use of two-level logistic regression models is preferred to account for the nested nature of the data (Raudenbush & Bryk, 2002). Success in prison programming is known to be related to observable and unobservable individual-level characteristics, including prior education, personal motivation, and mental ability (Bosma et al., 2016; Silver et al., 2020). Failure to correct for higher-level unit clusters masks relationships between independent variables and outcomes, leading to misestimation of critical measures such as standard errors and statistical significance (Raudenbush & Bryk, 2002). Level 1 variables are person-mean centered (i.e., group mean centered) to examine the within-person variance in program completion given a visit during a period of enrollment. Following the guidance of Curran and Bauer (2011), I include both the person-mean centered values and the person means for the primary independent variables in estimated models. This method allows for the disaggregation of within-person effects (derived from the person-mean centered values; entered as a Level 1 variable) and between-person effects (derived from the calculated person means; entered as a Level 2 variable).
To assess whether multilevel models provide improved model fit compared to one-level logistic regression models, I estimate a likelihood-ratio (LR) test using unconstrained (i.e., models without the primary independent variable or control variables) two-level and single-level models. Results from the LR test suggest that the two-level model is preferred for these data (Chi-value[1] = 2,597.01, p < .001). Further, estimation of the intraclass correlation coefficient (ICC), defined as a measure of the proportion of outcome variance described by a cluster (Luke, 2020), suggests that between-person variance accounts for 30% of the total variance in the outcome in the unconstrained model. Level 1 variance is assumed to be heteroskedastic and a function of the intercept in multilevel logistic regression models (Luke, 2020; Raudenbush & Bryk, 2002), and is not estimated here. Together, these values suggest a strong preference for multilevel corrections using random intercepts for the individual.
Measures
Dependent Variable
The dependent variable is a dichotomous indicator of enrollment period completion status. The variable collapses two forms of program completion: program is completed in its entirety or the individual made sufficient progress during the enrollment period and may continue with the program in the following enrollment period.
Independent Variables
Three independent variables are examined. First, a dichotomous indicator of whether the individual received a visit during the period of enrollment (1) or did not receive a visit during the period of enrollment (0). Second, among those who received a visit, whether the individual received at least one group visit during the semester (1) or received only single-person visits during the semester (0). Though an imperfect proxy, group visits are indicative of larger family and peer group visits, potentially including children, siblings, or both parents. Such visits indicate high levels of social support. Finally, length of visits is examined, operationalized as the average number of hours per week an individual was visited over the enrollment period. This measure is calculated as the total number of hours the individual was visited during the enrollment period divided by the number of weeks elapsed between the start and end date of the enrollment period. This metric standardizes visit hours across individuals, as enrollment periods differ between programs. I examine this metric in the context of prior evidence that suggests IP prefer longer visits as a method to exchange information and resources, and express frustration at visitation policies that restrict visit time, especially with children (Dixey & Woodall, 2012; Pierce, 2015).
The number of visits the individual received prior to the enrollment period is used as a proxy for existing social support networks, as some evidence suggests that the experience of visitation is less important than the knowledge that an individual has pre-existing support networks to assist them tangibly and emotionally during reentry (Atkin-Plunk & Armstrong, 2018). 5 Models are estimated with and without this variable to test whether this social support hypothesis is true, or whether the chronological proximity of visits is central to behavior as suggested by activation models of social support (Anderson et al., 2020).
Control Variables
A number of Level 2 control variables are included in models. Four variables assess demographic characteristics. Race/ethnicity is characterized in three distinct groups: non-Hispanic White, non-Hispanic Black, and Hispanic/Latino. Age is a continuous variable representing age at entry to DOC. Highest education completed prior to facility entry is a dichotomous variable representing 11th grade or less (0) or 12th grade or higher (1; including high school equivalency degrees). Marital status is measured as a three-category variable: Single (0; reference), married (1), and other (2; including widowed, divorced, and separated).
Seven variables assess characteristics defined by sentence or DOC. Minimum sentence years is a continuous variable. Longer sentence length indicates greater exposure time to participate in programs and receive visits. The individual’s governing offense is a three-category variable: Violent (0; reference), non-violent (1), or drug-related (2). A dichotomous variable is used to assess whether the individual has ever been convicted as a parole violator in the state. Dichotomous variables indicate whether the individual was flagged at intake as a substance user, in need of mental health services, or an assault risk. 6
Three Level 1 control variables are included. These variables vary by the period of enrollment. The number of program hours offered is included, as programs with longer enrollment periods may be more difficult to complete. Facility security level is measured as a four-category variable indicating whether the individual resided in a minimum (0; reference), medium (1), close (2), or maximum (3) security facility during program enrollment, as defined under the DOC’s categorization of each facility. The proportion of the individual’s sentence served at the time of program enrollment is considered. Some evidence suggests that the timing of program enrollment is related to success, that individuals may experience program fatigue across their stay, and that persons incarcerated longer seek novel activities to fill their time (Zamble & Porporino, 1990). Finally, program type is a five-category variable indicating the general type of program for the current period of enrollment: Education (0; including Adult Basic Education, General Education Development, English as a Second Language, and post-secondary; reference), Vocation (1; including finance courses, construction and building maintenance, Heating-Ventilation-Air Conditioning Certification, trade apprenticeships, etc.), Substance (2; including cognitive-behavioral interventions, therapeutic communities, 12 Step Facilitation Group etc.), Violence (3; including batterer programs, aggression management, sexual offender management, etc.), and “Other” (4; including life skills, hygiene, parenting, etc.). 7 Visits may be uniquely beneficial in persistence in certain programs as compared to others. For example, high perceived emotional support among women in a high-intensity prison substance dependence program was associated with improved 12-month outcomes (Andersen, 2018). Conversely, adult learners in prison education program report strong personal motivation to receive their degree to evoke personal betterment (Manger et al., 2010). Thus, the benefit of visits may not extend equally to all program enrollees.
Within the context of this state, IP experience program enrollment through two methods. First, the individual may approach “education staff” to voluntarily enroll in most educational and vocational programs. Counseling, therapy, and specialized behavior programs are also available for voluntary enrollment, though priority is based on risk-needs assessment. Second, individuals can be recommended or mandated to attend programs. Recommendations and mandates are determined through actuarial risk tools that consider an individual’s risk-needs. Mandates are primarily used for education (where persons with less than a high school degree must take at least 360 hours of education programming; less if they complete their degree within this timeframe) and substance use treatment (where mandated treatment is focused on withdrawal symptom maintenance but may be a function of court requirements). Referred programming is not required, but encouraged. In both voluntary and mandatory contexts, an individual may either complete the semester or fail to complete the semester. Among mandatory enrollments, non-completion can be a factor of administrative discharge for disruptive or dangerous behavior, failure to meet semester success metrics, and even voluntary withdrawal from the semester to manage other needs when staff deem it appropriate. Though prior scholarship points to mixed evidence regarding the relationship between mandatory programming and retention (Parhar et al., 2008), mandatory versus voluntary enrollment could not be reliably determined within the data provided and thus their assessment is outside the scope of this study.
Findings
The sample is primarily Black (46%) and White (42%). A slight majority of the sample report a high school degree or higher (59%) and most report that they are single (77%). Most individuals were flagged as a substance user (85%) and an assault risk (63%), but fewer were flagged as in need of mental health services (22%). Most individuals were incarcerated for a violent offense (41%) and about one-quarter are known to have violated parole in the past. The sample was, on average, 33 years old at the time of intake and were sentenced to slightly longer than 3 years.
While most periods of enrollment in this sample ended in program completion (78%), only 36% of the sample received a visit during the period of enrollment. Of the visited group, 68% received at least one group visit during the period of enrollment and individuals experienced an average of 1.65 hours of visits per week during enrollment. Most enrollments were for education programs (38%) or vocation programs (24%). Data are primarily drawn from medium security facilities (53%) which are the most abundant in the state. The average number of hours offered for an enrollment period was 75 hours. Finally, individuals had served an average 47% of their sentence at the time of program enrollment (Table 1).
Descriptive Statistics.
Note. Standard deviations not presented for categorical variables.
Among the 12,872 observations in which an individual received a visit during the semester.
Chi-square analyses (not displayed) suggest that enrollment period completion and receiving a visit during the enrollment period are statistically related (Chi-value[1] = 127.75, p < .001). While 31% of individuals who did not complete the enrollment period received a visit, 38% of individuals who completed the enrollment period received a visit. Initial results suggest a positive relationship between visits and enrollment period success.
Visit Presence, Prior Visits, and Program Completion
Table 2 displays results from three models assessing likelihood of program completion given visitation, a test of Hypothesis 1. I present only results from the primary independent variables for space and relevance though all models include the control variables described above; I present full models in Appendix B. Variables notated with (w) represent the person-mean centered values, while variables notated with (b) represent the person-means. Interpretations of these coefficients are thus within-person effects and between-person effects, respectively.
Odds of Enrollment Completion with Visits, Two-Level Random Effects Logistic Regression.
Note. Coefficients denoted with (w) represent the deviation from the person mean and thus are interpreted within-person. Coefficients denoted with (b) represent the calculated person mean and thus are interpreted between-person. Models based on 6,250 unique incarcerated persons, 35,702 program enrollments. Models clustered at the individual level (L2). All control variables described in the text are included in all models. Reference categories: White; Single; Violent Convicting Offense; Education Program.
p < .001.
Results from Model 1 suggest that an individual who receives a visit during enrollment display 21% greater odds of completing the enrollment period compared to enrollment periods when they do not receive a visit (p < .001). To test whether such results are a function of the general presence of visits (i.e., the presence of social support at any time during an incarceration) or whether the proximity of visits is crucial to positive outcomes, I examine whether the number of prior visits is related to program completion (Model 2). The coefficient fails to reach statistical significance, suggesting that when an individual experiences a greater number of prior visits they are no more or less likely to complete a period of enrollment compared to enrollment periods in which they experienced fewer prior visits. When visits received during enrollment and the number of prior visits are included together in Model 3, the number of prior visits again fails to reach statistical significance while persons who receive a visit during enrollment continue to display 21% higher odds of enrollment period completion as compared to enrollment periods in which they did not receive a visit (p < .001). Results speak to the import of the chronological proximity of visits to positive outcomes, as suggested by “activation” models of behavior.
To examine the between-person variation in enrollment period completion in the presence of visits, person mean values for both variables are considered in Model 4. Results suggest that persons who receive a visit during enrollment display 72% greater odds of completing the enrollment period as compared to persons who do not receive visits during the enrollment period (p < .001). Again, the number of visits prior to enrollment failed to reach statistical significance, suggesting that persons with a greater number of prior visits were no more or less likely to complete the program enrollment period as compared to persons with fewer prior visits. Results for within-person estimates are similar to estimates presented in Models 1 to 3. Findings suggest that both the within-person and between-person experience of visitation during an enrollment period offer substantively greater odds of completing an enrollment period, lending support to Hypothesis 1 and illustrating the importance of both the presence and proximity of visits as social support mechanisms that prompt persistence in rehabilitative programming.
Visit Qualities and Program Completion
To examine whether certain qualities of visits are differentially important for persistence in program semesters (Hypothesis 2), I analyze the presence of group visits and average hours per week of visits during the enrollment period. Models include only those persons who received a visit during enrollment to differentiate between visit qualities and the presence of visits. To simplify models, only within-person estimates are considered to examine how such visit qualities are impactful. Variables are entered into the model alone (Models 1 and 3) and with the number of prior visits to consider past support (Models 2 and 4; Table 3).
Visit Qualities as Related to Program Completion, Two-Level Random Effects Logistic Regression.
Note. Models include only those individuals who received a visit during program enrollment. Models based on 3,567 unique incarcerated persons, 12,872 program enrollments. Models clustered at the individual level (L2). All control variables described in the text are included in all models. Reference categories: White; Single; Violent Convicting Offense; Education Program.
p < .05. ***p < .001.
Among those who received a visit during enrollment, individuals who received at least one group visit display 37% greater odds of completing their enrollment period compared to enrollment periods in which they received only single-person visits (p < .001). Further, this relationship remains significant when number of prior visits is added to the model (OR = 1.38, p < .001; Model 2). Additionally, within the visited sample, number of prior visits reaches statistical significance as a predictor of enrollment period completion. Results suggest a modest positive effect, wherein persons who experienced a greater number of prior visits display higher odds of completing an enrollment period compared to periods of enrollment in which they received fewer prior visits (OR = 1.004, p < .05). 8 Findings suggest that receiving visits in which multiple persons are present provides additional positive effects of visitation. It is likely that such visits indicate visits from parents or partners with children, both of which represent important support systems during and after incarceration. Further, among the visited sample, more persistent past social support provides a benefit beyond the present indicator of social support.
Average hours per week of visits during enrollment did not reach statistical significance as a predictor of enrollment period completion in either Model 3 or Model 4. Results align with mixed results from prior literature suggesting that longer visits are reported as both beneficial to promote togetherness and provide quality time to exchange information, as well as fatiguing and potentially frustrating when they go on for too long. It is possible that different IP perceive the benefits and drawbacks of lengthy visits differently. However, the number of prior visits again displays a modest positive relationship with enrollment period completion (OR = 1.004, p < .05)
Discussion
Prison visits have long been conceptualized as positive and prosocial events in the lives of incarcerated individuals. Prior work focuses on visitation as a method of informal social control which restrains antisocial behavior (e.g., Cochran, 2012). The present study provides new evidence that visitation does not only restrain antisocial behavior, but promotes prosocial behavior through prison program perseverance during a period of enrollment. Benefits of program completion are numerous, including positive evaluations for parole and signals to community representatives that the individual is prepared to desist (Bushway & Apel, 2012; Ruhland, 2020). The examination of methods to promote program retention are thus warranted, and this study aimed to determine whether visitation can act in such a manner. Several conclusions are drawn from the results presented above.
First, experiencing a visit is associated with increased odds of successfully completing the program enrollment period, where results hold both within-person (i.e., a person who receives a visit during one enrollment period but receives no visits during another enrollment period) and between-person (i.e., a person who receives a visit compared to a person who does not receive a visit during enrollment). Results support Hypothesis 1 and lend credence to theoretical models that suggest social support is a mechanism that prompts prosocial behavior. Prior work examining the importance of visits in the context of the most restrictive environment suggest that visits act to bolster or reaffirm social bonds, especially attachment to valued others and commitment to conventional lines of behavior (Cochran & Mears, 2013; Laub et al., 1998). For incarcerated persons, visits and active maintenance of family and peer social ties may also bolster active identity change, in which incarcerated persons seek to attain a more prosocial identity signaled by participation in conventional activities such as programs (Giordano et al., 2002). Further, visits may protect against harmful “prisonization” effects by alleviating the many deprivations and subsequent negative emotionality that are prompted by the prison environment (Blevins et al., 2010; Sykes, 1958). Novel to this study, results suggest that visits work to influence prosocial behavior within prisons through the completion of program enrollment periods. Such findings align with prior studies which find important concurrent development of behavior during a prison stay (Cochran, 2012). Though direct tests of the mechanisms at work in this relationship are beyond the scope of the present analyses, results lend support to the concept of visits and contact with valued others as important for prosocial identity transformation wherein incarcerated persons who experience visits during enrollment may lean on visitors to affirm steps toward conventional lifestyles and desistance.
Further, models suggest that the conditions of the visit impact completion. Specifically, that receiving group visits is more beneficial than receiving only solo visits. While the relation of visitors and incarcerated persons was unable to be determined from the data at hand, group visits are indicative of family or peer group visitors. Prior visitation literature suggests that receiving visits from children and parents can assist in the maintenance of identities such as father, partner, or child, all of which provide evidence to the individual that they possess an identity outside of the prison walls and their offending behavior (Turanovic & Tasca, 2019). While there are numerous practical and emotional barriers to group visits, especially when children are involved, results suggest that receiving group visits can further assist in program persistence, potentially by reifying prosocial identities among persons who are committed to taking positive steps toward conventional lifestyles.
Second, results suggest that the chronological proximity of visits is more important to the promotion of prosocial behavior than the general experience of visitation. Models assessing prior visitation did not support general positive effects of the number of visits prior to the current enrollment period, though there is some evidence that prior visits are impactful for those who actively receive visits. Rather, findings aligned with activation models of social support that suggest timing and close chronological proximity of an event are more relevant for outcomes (Anderson et al., 2020). Though social support is a crucial tool for behavior and reentry, prior theoretical scholarship contends that these support networks are weakened if not actively maintained (Laub et al., 1998). This issue is especially salient for incarcerated persons and their families who experience substantial financial, social, and administrative barriers to visitation. Prosocial benefits from visitation thus appear to be neutralized if new and recent visitation experiences are absent. Such findings are integral to our understanding of how social support operates within the prison context and provides direct policy implications for the expansion of in-person visiting hours and modalities, which are discussed further below.
Policy Implications
Though visitation is a correlate of prosocial behavior and social network maintenance, the availability and ease of access to in-person visitation is under consistent scrutiny. This phenomenon has become increasingly true in the wake of the COVID-19 pandemic, where visitation was halted at both the state and federal levels (Muñiz et al., 2023). Additionally, technological advancements in video-chat methodologies and the intrusion of third-party private industry into visitation modalities has expanded virtual visitation methods while inadvertently discouraging the expansion of in-person visitation policies (Kukorowski et al., 2013). Though heralded as a positive and easier method by which to offer contact with families, both incarcerated persons and visitors report strong preferences for in-person visitation compared to virtual visitation methods (Tartaro & Levy, 2017).
The current study provides evidence that in-person visits offer something substantial to the prosocial behavior of incarcerated individuals. While some administrators suggest that expansion of these opportunities would pose significant monetary and safety costs to the institutions, it is possible to alter existing infrastructure to accommodate more accessible in-person visitation policies. This includes shifting visitation hours to accommodate post-work and weekend schedules. In this way, facilities do not need to offer more visitation hours but simply more accommodating visitation hours. Facilities are also increasingly at the mercy of online visitation scheduling tools (i.e., “reserving” visitation hours through online portals). For families and peers of incarcerated persons who are not computer literate or who do not possess regular access to computers, smart devices, or internet, these online tools are a burden rather than a technological advancement. Encouraging facilities to offer a wider range of drop-in hours or alternative scheduling modalities would facilitate ease of visitation experiences.
Limitations
The current study must be considered within its limitations. First, data are drawn from administrative files available from the state’s DOC database. Notably, these files do not include phone calls, mail, and video visits, which are stored by private third-party companies. While qualitative analyses reveal that visitors and IP prefer in-person visits, video visits and phone calls are perceived as convenient when in-person visits are not feasible (Pierce, 2015; Tartaro & Levy, 2017). Thus, the absence of in-person visits for these individuals does not indicate lack of social support. It is possible that these analyses underestimate the impact of social support modalities (i.e., in-person visits, phone calls, and video visits) on program completion, as work suggests positive effects of calls and video visits. Conversely, the impact of in-person visits may be overestimated if these alternative modalities are equally important for prosocial outcomes. Future work should expand on the examination of alternative communication modalities to determine whether they are equally successful at promoting prosocial behavior.
Second, the current data do not include indicators of actual and perceived social support, which may be crucial for individual motivation and conformity. Prior studies find that pre-incarceration networks are related to within-prison and post-release behavior (Atkin-Plunk & Armstrong, 2018). It is possible that individuals who possess strong support networks are simultaneously more likely to receive visits and successfully complete enrollment periods. Analyses employed proxy measures for persistent social support by examining prior visitation experiences. However, these measures are not perfect indicators of existing social support. Scholarship should strive to employ perceptual measures of support where possible.
Finally, the present sample is an all-male sample, thus the generalizability of findings to incarcerated women and gender non-binary persons is limited. Our understanding of social support within a gendered context may produce different conclusions regarding the effectiveness of visitation as a promoter of prosocial behavior. Extant work examining the general impact of social relationships on prosocial and antisocial behavior suggests gendered differences. For example, that marriage is almost always predictive of desistance among men, while the effect is more often null for women and may even be a catalyst for offending among women (Skardhamar et al., 2015). Indeed, some prior studies suggest that incarcerated women are more impacted by separation effects when visitors leave and report heightened distress after visits, especially within the context of parenting (De Claire & Dixon, 2017). The positive impact of visits may thus fail to extend to incarcerated female populations, though further research is needed in this area.
Conclusion
The current study adds to the mounting evidence that visits are not only a positive experience for many incarcerated persons, but also provide substantial benefits for outcomes within prison facilities. Expansion or alteration of visitation is a promising tool to promote prosocial activity engagement while in prison. In turn, skill and knowledge accumulation and movement toward prosocial identities will help those reentering into the community after a period of incarceration.
Footnotes
Appendix A
Comparison of Excluded and Included Sample.
| Excluded sample | Sample of analysis | Std. | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Freq. | M | SD | Min | Max | Freq. | M | SD | Min | Max | Diff. | |
| Race/ethnicity | |||||||||||
| White | 2,446 | 0.50 | — | 0 | 1 | 6,250 | 0.42 | — | 0 | 1 | 15.96 |
| Black | 1,872 | 0.38 | — | 0 | 1 | 6,250 | 0.46 | — | 0 | 1 | −14.93 |
| Hispanic/Latino | 486 | 0.10 | — | 0 | 1 | 6,250 | 0.12 | — | 0 | 1 | −6.65 |
| Other race/ethnicity | 73 | 0.02 | — | 0 | 1 | — | — | — | — | — | — |
| Completed 12th grade | 2,994 | 0.62 | — | 0 | 1 | 6,250 | 0.59 | — | 0 | 1 | 4.91 |
| Marital status | |||||||||||
| Single | 3,752 | 0.77 | — | 0 | 1 | 6,250 | 0.77 | — | 0 | 1 | 0.69 |
| Married | 596 | 0.12 | — | 0 | 1 | 6,250 | 0.14 | — | 0 | 1 | −4.29 |
| Other | 529 | 0.11 | — | 0 | 1 | 6,250 | 0.10 | — | 0 | 1 | 3.79 |
| Substance use flag | 4,354 | 0.89 | — | 0 | 1 | 6,250 | 0.85 | — | 0 | 1 | 11.74 |
| Assault risk flag | 2,924 | 0.60 | — | 0 | 1 | 6,250 | 0.63 | — | 0 | 1 | −5.81 |
| Mental health flag | 1,123 | 0.23 | — | 0 | 1 | 6,250 | 0.22 | — | 0 | 1 | 2.66 |
| Governing offense | |||||||||||
| Violent | 1,756 | 0.36 | — | 0 | 1 | 6,250 | 0.41 | — | 0 | 1 | −10.54 |
| Non-violent | 1,779 | 0.37 | — | 0 | 1 | 6,250 | 0.31 | — | 0 | 1 | 12.02 |
| Drug | 1,312 | 0.27 | — | 0 | 1 | 6,250 | 0.28 | — | 0 | 1 | −1.23 |
| Parole violation | 3,079 | 0.87 | — | 0 | 1 | 6,250 | 0.24 | — | 0 | 1 |
|
| Age at intake | 4,877 | 36.20 | 10.45 | 16 | 79 | 6,250 | 33.55 | 10.61 | 15 | 82 |
|
| Min sentence (years) | 4,877 | 2.54 | 3.03 | 0 | 42 | 6,250 | 3.03 | 4.032 | 0 | 45 | −13.74 |
Note. Standard deviations are not presented for categorical variables. Total excluded sample categories may not add up to the full excluded sample due to data missingness. Standardized differences are calculated between the excluded and included sample. Bolded values represent standardized differences above the threshold of |20| suggesting substantial differences between groups. The included group is younger, on average. Results for parole violation should be interpreted with caution, as this is the largest missing variable. Parole violation rates for the included sample align with current estimates of parole violation in the state.
Appendix B
Visit Qualities as Related to Program Completion, Two-Level Random Effects Logistic Regression.
| Model 1 | Model 2 | Model 3 | Model 4 | |||||
|---|---|---|---|---|---|---|---|---|
| OR | 95% CI [lower, upper] | OR | 95% CI [lower, upper] | OR | 95% CI [lower, upper] | OR | 95% CI [lower, upper] | |
| Group visit during program | 1.37*** | [1.16, 1.62] | 1.38*** | [1.17, 1.63] | — | — | — | — |
| Average visit duration | — | — | — | — | 0.99 | [0.96, 1.03] | 0.99 | [0.96, 1.03] |
| Number of visits before enrollment | — | — | 1.00* | [1.00, 1.01] | — | — | 1.00* | [1.00, 1.01] |
| Race/ethnicity | ||||||||
| Black | 0.78** | [0.66, 0.93] | 0.79** | [0.66, 0.93] | 0.78** | [0.66, 0.93] | 0.79** | [0.66, 0.93] |
| Hispanic/Latino | 0.80 | [0.63, 1.02] | 0.80 | [0.63, 1.02] | 0.80 | [0.63, 1.02] | 0.80 | [0.63, 1.02] |
| Completed 12th grade | 1.64*** | [1.41, 1.91] | 1.64*** | [1.41, 1.91] | 1.64*** | [1.41, 1.91] | 1.64*** | [1.41, 1.91] |
| Age at intake | 1.02*** | [1.01, 1.03] | 1.02*** | [1.01, 1.03] | 1.02*** | [1.01, 1.03] | 1.02*** | [1.01, 1.03] |
| Marriage | ||||||||
| Married | 0.92 | [0.73, 1.15] | 0.92 | [0.73, 1.16] | 0.92 | [0.73, 1.15] | 0.92 | [0.73, 1.16] |
| Other | 1.13 | [0.83, 1.54] | 1.13 | [0.83, 1.55] | 1.13 | [0.82, 1.54] | 1.13 | [0.83, 1.54] |
| Substance dependence flag | 0.86 | [0.69, 1.07] | 0.86 | [0.70, 1.07] | 0.86 | [0.69, 1.07] | 0.86 | [0.70, 1.07] |
| Assault risk flag | 0.77** | [0.64, 0.92] | 0.77** | [0.64, 0.92] | 0.77** | [0.64, 0.92] | 0.77** | [0.64, 0.92] |
| Mental health needs flag | 0.73** | [0.61, 0.89] | 0.74** | [0.61, 0.89] | 0.73** | [0.61, 0.89] | 0.73** | [0.61, 0.89] |
| Minimum sentence (years) | 1.01 | [0.99, 1.03] | 1.01 | [0.99, 1.03] | 1.01 | [0.99, 1.03] | 1.01 | [0.99, 1.03] |
| Governing offense | ||||||||
| Non-violent | 0.91 | [0.75, 1.12] | 0.91 | [0.74, 1.12] | 0.91 | [0.75, 1.12] | 0.91 | [0.74, 1.12] |
| Drug | 1.04 | [0.84, 1.30] | 1.04 | [0.84, 1.29] | 1.04 | [0.84, 1.29] | 1.04 | [0.84, 1.29] |
| Parole violator | 1.00 | [0.81, 1.23] | 1.00 | [0.81, 1.22] | 1.00 | [0.81, 1.23] | 1.00 | [0.81, 1.23] |
| Facility security | ||||||||
| Medium | 1.16 | [0.74, 1.82] | 1.15 | [0.73, 1.82] | 1.14 | [0.73, 1.80] | 1.14 | [0.72, 1.79] |
| Close | 1.17 | [0.61, 2.22] | 1.16 | [0.61, 2.20] | 1.17 | [0.62, 2.23] | 1.17 | [0.61, 2.22] |
| Maximum | 1.10 | [0.69, 1.74] | 1.10 | [0.69, 1.74] | 1.09 | [0.69, 1.73] | 1.09 | [0.69, 1.73] |
| Program type | ||||||||
| Vocation | 1.37** | [1.13, 1.67] | 1.37** | [1.12, 1.67] | 1.38** | [1.13, 1.68] | 1.37** | [1.13, 1.67] |
| Substance | 4.11*** | [2.72, 6.22] | 4.12*** | [2.73, 6.22] | 4.08*** | [2.70, 6.16] | 4.09*** | [2.71, 6.17] |
| Violence | 6.48*** | [5.15, 8.16] | 6.47*** | [5.14, 8.14] | 6.57*** | [5.22, 8.26] | 6.56*** | [5.22, 8.25] |
| Other | 6.00*** | [4.70, 7.68] | 6.03*** | [4.72, 7.71] | 6.06*** | [4.74, 7.74] | 6.08*** | [4.76, 7.78] |
| Program hours offered | 1.00 | [1.00, 1.00] | 1.00 | [1.00, 1.00] | 1.00 | [1.00, 1.00] | 1.00 | [1.00, 1.00] |
| Proportion of sentence served | 2.03*** | [1.54, 2.68] | 1.63** | [1.16, 2.29] | 1.98*** | [1.51, 2.61] | 1.61** | [1.15, 2.26] |
| Intercept | 5.37*** | [3.62, 7.98] | 5.38*** | [3.62, 7.98] | 5.37*** | [3.62, 7.97] | 5.37*** | [3.62, 7.97] |
| Level 2 variance component | 1.45 | [1.23, 1.72] | 1.45 | [1.22, 1.72] | 1.45 | [1.22, 1.72] | 1.45 | [1.22, 1.71] |
Note. Models include only those individuals who received a visit during program enrollment. Models based on 3,567 unique incarcerated persons, 12,872 program enrollments. Models clustered at the individual level (L2). Reference categories: White; Single; Violent Convicting Offense; Education Program.
p < .05. **p < .01. ***p < .001.
Acknowledgements
I would like to thank Holly Nguyen, Tara Emmers-Sommer, and Lee Bernick for their support of this project. I thank all reviewers for invaluable feedback on this manuscript.
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
