Deterioration of Postincarceration Social Support for Emerging Adults

Abstract

More than 2.5 million emerging adults (ages 18-25) are incarcerated annually and most do poorly after release. Social support after an individual’s release from incarceration is a critical protective factor against recidivism for emerging adults. However, little is known about the stability of support for emerging adults post incarceration. This study uses hierarchical linear modeling (HLM) to examine whether social support declines over time after incarceration and how change in support may vary by incarceration length. Our findings show that, while a nonincarcerated group of justice-involved emerging adults experience relatively stable social support, there are negative and volatile effects of social support among their incarcerated counterparts. Moreover, longer incarceration stays are related to greater deterioration of support over time after community reentry for emerging adults. Study findings advance the field of postincarceration intervention development by responding to the challenge of determining the appropriate targets and length of interventions designed for emerging adults.

Keywords

reentry social support intervention family emerging adult youthful offender incarceration

Introduction

More than 2.5 million emerging adults (i.e., individuals aged 18 to 25 years) have contact with the criminal justice system each year, and approximately 30% are arrested by age 23 (Brame, Turner, Paternoster, & Bushway, 2012). Once incarcerated, emerging adults have the highest recidivism rate of any age group with 76% rearrested for a new crime within 3 years post release (Durose, Cooper, & Snyder, 2014). Thus, incarceration during this critical time increases the likelihood of further involvement with jail and prison into adulthood.

During emerging adulthood, most individuals are navigating new social roles through employment and education, developing adult relationships outside the family unit, and exploring long-term intimate partnerships. In contrast, incarcerated emerging adults spend this time confined in a potentially volatile environment. The experience of imprisonment leave these individuals facing substantially reduced social, occupational, marriage, and civic opportunities, further limiting the chance for a healthy transition into adult social roles and related social support (Bonnie, Stroud, & Breiner, 2015; Steinberg, Chung, & Little, 2004; M. L. Sullivan, 2004).

Based on Cullen’s (1994) social support theory, positive relationships are a key factor for formerly incarcerated emerging adults to navigate desistance from criminal behaviors during reentry (Cullen, Wright, & Chamlin, 1999). In fact, research shows social support is “positively and causally related to mental health, physical health, and longevity” and “buffers the harmful physical and mental health impacts of stress exposure” (Thoits, 2011a, p. 145). Underscoring the salience of social support to criminal involvement, Andrews, Bonta, and Wormith (2006) highlight low or poor social support as one of the “big four” risk factors for general recidivism. Thus, rehabilitative interventions that help formerly incarcerated emerging adults receive and maintain social support may reduce recidivism (Colvin, Cullen, & Vander Ven, 2002; Wright & Cesar, 2013).

Although ample evidence suggests that social support is critical to community reentry outcomes (Colvin et al., 2002), significant knowledge gaps remain regarding incarceration’s influence on the trajectory of an individual’s social support post release. Yet, we know that incarceration disrupts social support networks, and the longer an individual is incarcerated, the more social support atrophies (Martinez & Abrams, 2013). Therefore, prisoner reentry researchers and program developers are increasingly designing postincarceration interventions with strategies that enhance naturally occurring (i.e., informal) social support (Byrne & Taxman, 2004; Fontaine, Taxy, Peterson, Breaux, & Rossman, 2015; Hanson & Harris, 2000; Petersilia, 2003; Pettus-Davis et al., 2015; Willis & Grace, 2009; Wright & Cesar, 2013).

Many of these social support–focused interventions concentrate on the period immediately after release—during the first 6 months after incarceration—when people are most likely to recidivate (Altschuler & Armstrong, 2001; Chung, Schubert, & Mulvey, 2007). However, evaluation results regarding the impact of such short-term social support interventions have been mixed (Fontaine et al., 2015; E. Sullivan, Mino, Nelson, & Pope, 2002; Wilson & Davis, 2006). The findings from this body of intervention research suggest that initial spikes in social support upon reentry to the community are followed by slow but steady declines after the first 6 months post incarceration (Fontaine, Gilchrist-Scott, Denver, & Rossman, 2012; Pettus-Davis et al., 2017; Seal, Eldrige, Kacanek, Binson, & MacGowan, 2007). In short, if social support deteriorates over time during the first year of reentry, short-term interventions are likely not sufficient. Yet, systematic empirical investigation of the changes in social support over these longer periods of time post incarceration is lacking. This study aims to advance understanding of how social support from caring adults changes over the first 2 years post incarceration among formerly incarcerated emerging adults. Filling this knowledge gap can improve postincarceration interventions by providing guideposts for the timing and amount of social support needed beyond the initial reentry period.

Background

Social Support

Social support is commonly defined as a process of interpersonal interaction that involves the provision or exchange of social resources that persons perceive to be available or are actually provided by others (Cohen, Underwood, & Gottlieb, 2000; House, 1981). Social support occurs because of integration into a social network, which operates as a vital resource for well-being (Cohen et al., 2000; Hobfoll, 2009). A positive social network exposes individuals to family and nonfamily social controls that influence adaptive (normative) behaviors, thus providing individuals with predictability, purpose, and a sense of stability and belonging.

Individuals may have access to different dimensions of social support, such as “formal” and “informal” resources, upon reentry (Vaux, 1988). Local social service offices, hospitals, or correctional agencies provide formal support (McCamish-Svensson, Samuelsson, Hagberg, Svensson, & Dehlin, 1999; Walker, 2010). In contrast, informal support comes from naturally occurring relationships, or individuals who do not receive pay for their efforts (McCamish-Svensson et al., 1999) such as family, intimate partners, mentors, or neighbors. This study focuses on informal support.

Emerging Adulthood and Social Support

Emerging adulthood is a distinct developmental stage between adolescence and adulthood when individuals establish key interpersonal relationships (Arnett, 2000). Social support during this phase plays an important role in the transition to adulthood as individuals are developing their independence (Arnett, 2007; Bonnie et al., 2015), and the primary social influence on behavior moves from the family to peers and nonfamily caring adults (Farrington, 2003). For example, in a study of 3,334 emerging adults, more than half of respondents (56%) reported having informal mentors who were nonfamily adults (Hurd, Stoddard, Bauermeister, & Zimmerman, 2014).

Social support from caring adults, both family and nonfamily, is predictive of reduced violence (Ullrich & Coid, 2011), decreased psychiatric symptoms (Draine & Solomon, 2000), lower sexual risk behaviors (Hurd & Zimmerman, 2010), increased coping skills (Hurd et al., 2014), improved quality of life (Draine & Solomon, 2000), employment (Hochstetler, DeLisi, & Pratt, 2010; McDonald & Lambert, 2014), stable housing (La Vigne, Visher, & Castro, 2004), and positive substance abuse treatment access, retention, and recovery (Dobkin, De Civita, Paraherakis, & Gill, 2002; Thoits, 2011b; White, 2009). Supportive relationships from caring adults also prevent and alleviate stress (Gardner & Cutrona, 2004; Garrity et al., 2006; Taylor, Doane, & Eisenberg, 2014). Even disadvantaged and transient youth report the importance of relationships with caring adults—74% of homeless youth indicate the relationships have positive impacts on perceived social support, satisfaction with social support, and engaging with prosocial peers (Dang, Conger, Breslau, & Miller, 2014).

Social Support and Community Reentry

By influencing the cognitions, beliefs, bonds, and behaviors of individual actors, social relationships play a role in criminal behaviors (Andrews et al., 2006; Gendreau, Little, & Goggin, 1996; Hawkins & Weis, 1985; Hirschi, 1969; Laub & Sampson, 2003; Welch, Tittle, Yonkoski, Meidinger, & Grasmick, 2008). For instance, Hirschi’s (1969) social bond theory describes that deviancy occurs when social bonds to conventional society are weakened, which often occurs during or after incarceration (Martinez & Abrams, 2013). In contrast, investing in conventional activities and cultivating relationships can inhibit deviant behavior (Chriss, 2007; Hirschi, 1969).

During community reentry, up to 92% of formerly incarcerated persons rely on family to provide emotional (e.g., trust and empathy), instrumental (e.g., money and housing), informational (e.g., advice), and appraisal support (e.g., information for self-evaluation; Berg & Huebner, 2010; Naser & La Vigne, 2006; Visher & Courtney, 2006). Each type of social support is associated with positive outcomes in the transition from prison to the community. First, the emotional support that stable partnerships and families provide increases access to other positive support providers (Giordano, Cernkovich, & Holland, 2003), adherence to postrelease supervision requirements, and reduces use of illicit substances (Cullen & Gendreau, 2000; Mallik-Kane & Visher, 2008; Rossman & Fontaine, 2015; Skeem, Louden, Manchak, Vidal, & Haddad, 2009). Similarly, instrumental support is negatively correlated with recidivism (Gendreau et al., 1996; Makarios, Steiner, & Travis, 2010). Last, informational and appraisal support is related to formerly incarcerated individuals’ success by providing encouragement and advice about resources in the community (Giordano et al., 2003; Mallik-Kane & Visher, 2008; Seal et al., 2007; E. Sullivan et al., 2002).

Emerging adults with a criminal history receive, on average, more assistance from their parents than either their nonjustice-involved siblings or peers (Siennick, 2011). However, many of these families provide support under considerably strained circumstances. Families of formerly incarcerated individuals frequently report low incomes, little education, mental and physical health concerns, and competing caregiving responsibilities (Arditti, Lamber-Shute, & Joest, 2003; Fontaine et al., 2012; Shollenberger, 2009; Wildeman & Western, 2010). Some families experience substantial frustration and resentment with their loved one’s incarceration leading to conflict (diZerega, 2010; Hairston & Oliver, 2006). Family conflict may be especially problematic for formerly incarcerated emerging adults because it can lead to negative effects on social and emotional health as well as increase the likelihood for drug use and crime upon reentry (Fosco, Caruthers, & Dishion, 2012; Mowen & Visher, 2013). Therefore, though emerging adults are largely dependent on family after incarceration, these same relationships could be impeding the natural developmental process of establishing stable and positive relationships with nonfamily. Arditti and Parkman (2011) described formerly incarcerated emerging adults as being in a “developmental paradox” (p. 211) because emerging adulthood is a phase in which individuals establish independence from family rather than increased reliance.

Recognizing the important role of social support in reentry success (Colvin et al., 2002; Cullen et al., 1999), researchers are interested in improved access to informal social support as a key outcome of interventions for formerly incarcerated individuals (Wright & Cesar, 2013). Interventions that focus on social support have been defined as “systematic activities designed to change the existing quality, level, or function of an individual’s social network or to create new networks and relationships [to mobilize social support to achieve specific outcomes]” (Budde & Schene, 2004, p. 342). Prior work on social support interventions suggests that “within social networks, it is the appropriate match of social support needs and provisions that reinforces an individual’s emotions, beliefs, and behaviors that buffer internal and external stress . . . to promote positive outcomes” (Pettus-Davis, Howard, Roberts-Lewis, & Scheyett, 2011, p. 484). Understanding whether and how social support changes over time for emerging adults is a critical first step in assessing the “appropriate match” of support needs and support resources.

Present Study

In this study, we examine whether social support declines over the 2 years after incarceration and how changes in social support may vary by incarceration length. We begin with a descriptive question: What are the patterns of family and nonfamily social support among incarcerated and nonincarcerated emerging adults? We then pose two main research questions:

Research Question 1: Are there differential trends of social support over time for those who have and have not been incarcerated?

Research Question 2: Among those incarcerated, does incarceration length moderate the relationship between social support and time?

In relation to the first research question, we hypothesize that formerly incarcerated emerging adults will experience declines in social support after initial release compared with their nonincarcerated counterparts, who will not experience similar declines. Regarding the second research question, we hypothesize that, among those incarcerated, experiencing longer incarceration stays will be associated with greater deterioration of support over time compared with experiencing shorter incarceration stays.

Method

Sampling

This study uses data from the Pathways to Desistance Project, a longitudinal multisite investigation following criminal offenders from adolescence to young adulthood in Maricopa County, Arizona, and Philadelphia County, Pennsylvania (Inter-University Consortium for Political and Social Research, 2012). A total of 1,354 youth aged 14 to 17 were enrolled during the recruitment period (November 2000 through January 2003). Youth were selected for potential enrollment if they had been adjudicated (i.e., found guilty) of a serious offense (i.e., predominantly felonies, with a few exceptions for some misdemeanor property offenses, sexual assault, or weapon offenses).¹ Youth who agreed to enroll in the study completed the baseline interview between 75 and 90 days of their adjudication hearing. Data were collected using computer-assisted interviews that took place in the participants’ homes, in libraries (or other public places), or in facilities. The first six waves of follow-up interviews correspond to 6-month recalls, while Waves 7 to 11 correspond to 12-month recalls. For details on the theoretical foundation, participant recruitment, full sample description, and methodology of the Pathways to Desistance Project, please refer to the following publications (see Mulvey et al., 2004; Schubert et al., 2004).

The current study uses data from the first six waves of the Pathways to Desistance data (referred to here as Waves 1 through 6) spanning 3 years to ensure equal time between waves (i.e., 6 months). Because we are interested in emerging adulthood and the development of social support beyond 1 year, individuals who were aged at least 17 years and had at least four consecutive waves, or 2 years, of complete data (hereafter referred to as Time 1 through Time 4) were included in the sample. Each of these four consecutive waves had to meet one of two criteria: (a) The wave occurred after an incarceration stay of greater than or equal to 30 days or (b) the individual did not experience any incarceration stay greater than or equal to 30 days. Incarceration was defined as confined in a residential setting without community access (e.g., detention, jail, prison, secure residential facility). Thus, all individuals in the sample had four consecutive waves of data, which are labeled Time 1 through 4, yet these four time periods could come from Pathways Waves 1 through 4, 2 through 5, or 3 through 5. Because we examine change in social support after incarceration, we chose the 30-day cutoff to establish someone was “incarcerated” during that wave. This period of time is long enough to allow one’s environment to alter and, in turn, social support to change as a result of the incarceration. Although this decision may artificially differentiate between those incarcerated for 30 days and those incarcerated for fewer than 30 days, it should be noted that 88% of those labeled nonincarcerated experienced zero days incarcerated.

As depicted in Figure 1, the sample was reduced from the original 1,354 for several reasons. First, we removed any individuals younger than 17 in the first three waves (n = 283) to focus on emerging adulthood. We also removed any individuals who did not have at least four consecutive waves of data and were missing valid data on time spent in a secure facility without community access (n = 164) because they needed ample time to be able to “experience” social support for us to assess changes. Collectively, these exclusion criteria reduced the sample from 1,354 to 907. The final criterion was to establish those who were either “nonincarcerated” for at least four consecutive waves or “incarcerated” during Pathways Waves 1, 2, or 3 followed by three consecutive waves of “nonincarceration.” After these criteria were imposed, an additional 428 participants were excluded largely due to reincarceration in subsequent waves, resulting in a final sample of 479. Of these participants, 283 (59.1%) met criteria for the “nonincarcerated” group and 196 (40.9%) met criteria for the “incarcerated” group.

Figure 1:

Participant Flow Chart

Measures²

Dependent Variable: Social Support

Social support is measured using the Contact With Caring Adults Inventory (Center for Research on Health Care Data Center, n.d.). This inventory measures perceived amount and type of social support provided by caring adults. The measure is a compilation of commonly used items with adolescents and emerging adults to identify available social support from within and outside of the family (Institute of Behavioral Science 1990; Nakkula, Way, Stauber, & London, 1990; Phillips & Springer, 1992). This measure of social support is consistently available at each wave for the study sample, and it addresses both family and nonfamily social support providers.

Respondents provided the relationship type of the person to which they are referring when responding to items for both family (i.e., mother/stepmother, father/stepfather, older sister, older brother, other female relative, or other male relative) and nonfamily (i.e., other female adult in the community, other male adult in the community, or other specified). The interview questions cover the four types of social support of most interest for this study: emotional, instrumental, informational, and appraisal. The question asks respondents to identify, if possible, at least one caring adult for the following eight domains: (a) someone you admire and want to be like, (b) someone you could talk to if you needed advice, (c) someone you could talk to about trouble at home, (d) someone you could tell about your accomplishments, (e) someone you could talk to about important decisions, (f) someone you could depend on for help, (g) someone you could talk to about problems, and (h) someone who cares about your feelings.

To calculate the social support received from family, we summed the total number of domains the respondents identified with a caring adult from his or her family. We report nonfamily social support as the sum of the number of domains for which the respondent identified at least one caring nonfamily adult. Respondents who reported having no caring adults in their lives were coded as zero for both measures. Although respondents could have included individual service providers as nonfamily caring adults, the instructions provided enough orientation toward informal social support that most respondents would not report on formal support relationships.

Time-Varying Control Variables

Prior research indicates several factors influence social support among those involved in the criminal justice system, such as peer delinquency (Haynie, 2002; Haynie & Osgood, 2005), exposure to violence (Wolff & Shi, 2012), criminal involvement (Andrews & Bonta, 2010), and substance abuse (Haynie & Osgood, 2005; Skeem et al., 2009). All statistical models control for these conceptually and empirically relevant constructs at each wave as described below.

Peer delinquency is measured as the proportion of friends (up to four) who were arrested in each 6-month recall period. Those who reported no friends received a value of zero. The measure of exposure to violence assesses both experienced and witnessed acts of violence; this measure ranges from zero, indicating no exposure to violence, to 13, indicating exposure to six different forms of victimization and seven different forms of witnessed violence in each 6-month recall period. Criminal involvement is a dichotomous measure of self-reported offending of any of 22 types of offenses committed in each 6-month recall period. Although a majority of individuals reported zero to one offenses in a recall period, a few extreme cases indicated upward of 500 separate offenses; because of this wide range in the number of offenses reported, the measure is dichotomized as either an offense was reported or not. Finally, the substance abuse measure draws from a substance use consequences question, which asked respondents to indicate either social problems or dependency attributable to alcohol and/or drug use across 31 items coded yes (1) or no (0). These items were summed into a total score. The time-varying controls are significantly correlated with one another within each time period; however, multicollinearity is not a major issue as the correlations range from .09 to .42 with the strongest correlations between self-reported offending and substance use consequences in each time period.

Time-Stable Control Variables

We include basic demographic variables as time-stable covariates to control for between-group differences in support experiences. These variables include the age of the respondent at the first of the four consecutive time periods (Sweeten, Piquero, & Steinberg, 2013); race/ethnicity, which includes categories White, Black, Hispanic, or other, with White treated as the reference group (Giordano, Cernkovich, & Rudolph, 2002); and gender, a dichotomous variable coded as either male or female (Lipsey & Derzon, 1998). We also include socioeconomic status and family structure collected at the first interview as additional time-stable controls (Lipsey & Derzon, 1998). Socioeconomic status is a mean score of the mother’s and father’s level of education ranging from one (some graduate or professional school) to six (grade school or less) such that higher numbers indicate lower socioeconomic status. If data for only one parent were available, the score reflects that individual’s level of education. We use mean imputation for the 10 cases with respondents who were missing both parents’ education level. Family structure was originally computed from the household roster indicating the combination of caregivers residing in the subject’s home. This variable is categorized into two-parent families (e.g., two biological, two adoptive parents, stepparent combinations), single-parent families (e.g., single mother or father), and an “other” category (e.g., other adult, no adult).

As mentioned previously, examination of changes in social support across time requires at least four consecutive waves of data. However, to maximize sample size, respondents could meet this criterion for inclusion by having “Time 1” of the four starting with Pathways to Desistance Wave 1, Wave 2, or Wave 3. Consequently, we control for which of the interview waves correspond to each respondent’s Time 1. Furthermore, both the incarcerated and nonincarcerated groups could have been incarcerated in the waves prior to Time 1. Because chronic incarcerations negatively correlate with social support (Andrews & Bonta, 2010), prior incarceration is included as a dichotomous variable to indicate whether the respondent was incarcerated in any of the waves prior to entering Time 1.

Moderators

Because incarceration socially isolates individuals from positive social relationships in the community and the length of incarceration can affect the level of deterioration in social relationships (Fontaine et al., 2012; Hairston & Oliver, 2006; Martinez & Abrams, 2013), we examine the potential moderating effects of incarceration status and length of incarceration at Time 1 on the relationship between time and social support. The first moderator is incarceration status at Time 1, which is a dichotomous measure of whether the respondent was incarcerated for more than 30 days at Time 1. Among those incarcerated at Time 1, we use the length of incarceration, which is a dichotomous measure based on the proportion of time incarcerated at Time 1. The proportion of time incarcerated ranges from .15 to 1.00 (i.e., 30 days to 6 months based on a 6-month recall); the mean proportion of time incarcerated is .65 (SD = .30), which translates to approximately 4 months incarcerated within a 6-month time frame. We cut the sample at the mean so that those with a proportion of time of .65 or less were coded “0” to represent a shorter incarceration stay and those greater than .65 were coded “1” to represent a longer incarceration stay.

Analysis

We begin the analysis by calculating the means and standard deviations for the time-stable and the time-varying measures. We then plot the mean number of domains of family and nonfamily support separately to address our first descriptive research question.

To address our main research questions, we use hierarchical linear modeling (HLM; Raudenbush & Bryk, 2002) to model individual change in support over the four time periods (i.e., domains of family and nonfamily support). To study individual change within the HLM framework, social support for each individual by wave is viewed as nested within that individual. Thus, there are two levels of analysis. Level 1 estimates each person’s development with a unique individual growth trajectory that depends on a set of parameters. The growth parameters from Level 1 become the outcome variables in Level 2 of the model and are determined by person-level characteristics. Therefore, Level 1 in the model estimates the within-individual change and Level 2 represents the between-individual analysis (Raudenbush & Bryk, 2002).

Specifically, at the within-individual level (i.e., Level 1), the outcome of social support is modeled as a function of time, which is operationalized as a dummy variable representing each of the four time periods, using a hierarchical overdispersed Poisson model (Raudenbush, Bryk, & Congdon, 2004). Time 2, as opposed to Time 1, serves as the reference time period as this period corresponds to the first 6 months post incarceration (for the incarcerated group). Although Time 1 includes some proportion of time incarcerated, Time 2 is the time period hypothesized to have the highest level of social support for this group, allowing us to compare the peak time of social support post incarceration (Time 2) with social support during incarceration (Time 1) and social support after the peak (Times 3 and 4). Thus, the Level 1 (i.e., within-individual) equation is as follows:

\begin{matrix} {Social Support}_{i t} = π_{0 i} + π_{1 i} Time 1_{i t} + π_{2 i} Time 3_{i t} + π_{3 i} Time 4_{i t} + π_{4 i} Peer Delinquency + \\ π_{5 i} Exposure to Violence + π_{6 i} Offender + π_{7 i} Substance Use Consequences + e_{i t}, \end{matrix}

where i is the index for individuals and t stands for the longitudinal observations. The π_1i through π_3i parameters estimate the average rate of change in social support between that time period and Time 2 (i.e., the first 6 months post incarceration for the incarcerated sample; Raudenbush & Bryk, 2002). The time-varying controls are group-mean centered.

At the between-individual level, the parameters from Level 1 are estimated from the Level 2 equations. The basic elements of the between-individual model (i.e., Level 2) are as follows:

\begin{matrix} π_{0 i} = β_{00} + β_{01} {Age}_{i} + β_{02} {Gender}_{i} + β_{03} {Black}_{i} + β_{04} {White}_{i} + β_{05} {Hispanic}_{i} + β_{06} \\ Incarceration Status at Time 1_{i} + β_{07} Mean Exposure to Violence + β_{08} Mean Peer \\ Delinquency + β_{09} Mean Offender + β_{10} Mean Substance Use Consequences + Prior \\ Incarceration + Corresponding Pathways Wave for Time 1 + r_{0 i} \end{matrix}

π_{1 i} = β_{10} + β_{11} Incarceration Status at Time 1_{i},

π_{2 i} = β_{20} + β_{21} Incarceration Status at Time 1_{i},

π_{3 i} = β_{30} + β_{31} Incarceration Status at Time 1_{i},

π_{4 i} = β_{40} \dots β_{7 i} = β_{70},

where age, gender, race, incarceration status at Time 1, prior incarceration, the Pathways wave that corresponds to Time 1, and the mean of each of the time-varying controls act as time-stable control variables that are presumed to predict social support (Raudenbush & Bryk, 2002). Cross-level interactions are then modeled where incarceration is allowed to moderate the impact of time (e.g., each dummy variable of time on the outcome of social support).

We first conduct main models for the total sample (n = 479; that is, without incarceration status at Time 1 as the moderator; Model 1). We then add incarceration status at Time 1 as a cross-level interaction (Model 2) to address our first main research question of whether incarceration status moderates the relationship between time and social support. To address our second main research question of whether length of incarceration differentially affects social support among the incarcerated sample (n = 196), we repeat these analyses for both Models 1 and 2. In the incarcerated only analyses, we replace the moderator of incarceration status at Time 1 with the dichotomous measure of longer versus shorter incarceration stays at Time 1.

Results

Descriptive Analyses

Table 1 displays the descriptive statistics for the time-stable and time-varying measures. The sample is largely male (78.9%), predominately Black or Hispanic (26.5% White, 37.6% Black, 31.5% Hispanic, 4.4% other), and 17.48 years old, on average (SD = 0.63), at Time 1. Although each of the original Pathways to Desistance sample members was adjudicated for a serious offense, not all of them spent time in secure settings. Out of the 479 youth used in this study, 34.4% were incarcerated prior to the first of four consecutive 6-month time periods (i.e., Time 1; 72.5% of the 196 incarcerated group and 8.1% of the 283 nonincarcerated group).

Table 1:

Description of Sample: Time-Stable and Time-Varying Measures

Time-stable measures	Percent or M (SD)	Range
Incarcerated at Time 1	40.9%	0-1
Age at Time 1	17.48 (0.63)	17-20
Male	78.9%	0-1
Ethnicity		1-4
White	26.5%
Black	37.6%
Hispanic	31.5%
Other	4.4%
Family type
Two parent	41.1%
Single parent	45.5%
Other parent combination	13.4%
Socioeconomic status	4.22 (0.93)	1-6
Pathways wave of entrance		1-3
Pathways Wave 1	45.7%
Pathways Wave 2	33.0%
Pathways Wave 3	21.3%
Incarcerated prior to Time 1	34.4%	0-1
Time-varying measures	T1	T2	T3	T4
Caring Adult–Family	4.95	4.99	4.63	4.59
Caring Adult–Nonfamily	1.03	0.93	1.02	0.97
Peer delinquency	0.40	0.41	0.36	0.35
Exposure to violence	1.15	0.98	0.94	0.80
Self-reported offender	44.9%	42.0%	39.7%	37.0%
Substance use consequences	0.92	0.96	1.02	1.00

On average, the sample reported having at least one caring adult family member in close to five of the domains at Time 1, decreasing to 4.59 domains on average by Time 4. The number of domains reported for a nonfamily caring adult is much lower with close to one domain reported across the four time periods. These patterns differ for those who were incarcerated at Time 1 and by length of incarceration, among those incarcerated at Time 1. Figure 2 shows the mean levels of family and nonfamily social support for each time period by incarceration status (Figures 2a and 2c, respectively) and by incarceration length (Figures 2b and 2d, respectively).³ Figures 2a and 2b, show that while all emerging adults decline in their family social support over time, those who were incarcerated and were incarcerated for longer experience a steeper decline from Time 2 to Time 3. With respect to nonfamily social support, there is instability across time for those who were incarcerated (Figure 2c), but no apparent change in this type of support for those who were not incarcerated. Interestingly, those with longer incarceration time not only have higher levels of nonfamily social support but also have a greater decline in that support over time (Figure 2d). Overall, these descriptive analyses suggest that there are differences in the patterns of social support based on incarceration. However, these patterns do not take into account the multiple influences on social support in emerging adulthood or assess whether these differences are statistically significant. Next, we conduct a series of HLM models to control for additional observed variables that may influence these variations in support.

Figure 2:

Changes in Domains in Social Support in Emerging Adults

Main Analyses

Domains of Family Support Over Time

Table 2 displays the HLM results for the domains of family support for the total sample (n = 479) and incarcerated subsample (n = 196) separately. Models 1 and 3 present the main models with Models 2 and 4 introducing the cross-level interactions. In both the total sample models and incarcerated only models, the respondents report significantly fewer domains of family support as time progresses, as indicated by the significant coefficients in Model 1 for Time 3 compared with Time 2 (b = −0.076, p = .001) and for Time 4 compared with Time 2 for the total sample (b = −0.082, p ≤ .001); and in Model 4 for the incarcerated sample for Time 3 compared with Time 2 (b = −0.095, p = .026) and for Time 4 compared with Time 2 (b = −0.099, p = .012). However, there are no significant differences in these trends based on incarceration status (Model 2). In contrast, the reported domains of family support for Time 3 compared with Time 2 do differ by length of incarceration with a large reduction in family support for those with longer incarceration stays and a slight increase for those with shorter incarceration stays (see Model 4: b = −0.224, p = .014).

Table 2:

Hierarchical Nonlinear Models of Domains of Family Social Support Post Incarceration: Comparison of Incarcerated and Nonincarcerated in First 6 Months (Total Sample) and Comparison of Time Incarcerated in First 6 Months (Incarcerated Sample)

	Total sample (n = 479)						Incarcerated sample (n = 196)
	Model 1			Model 2			Model 3			Model 4
	Coefficient	Significance	SE	Coefficient	Significance	SE	Coefficient	Significance	SE	Coefficient	Significance	SE
Intercept	3.106	***	0.625	3.102	***	0.625	4.093	***	0.837	4.058	***	0.837
Time 1	−0.008		0.022	−0.013		0.027	−0.001		0.039	0.022		0.050
Time 3	−0.076	**	0.024	−0.064	*	0.027	−0.095	*	0.043	0.008		0.053
Time 4	−0.082	***	0.025	−0.073	*	0.031	−0.099	*	0.039	−0.089		0.051
Time 1 × Incarceration				0.010		0.047				−0.048		0.076
Time 3 × Incarceration				−0.030		0.051				−0.224	*	0.091
Time 4 × Incarceration				−0.024		0.051				−0.020		0.083
Within-individual controls
Friends arrested	0.025		0.045	0.023		0.042	0.029		0.072	0.033		0.073
Exposure to violence	0.008		0.009	0.008		0.008	0.000		0.015	0.001		0.015
Self-reported offending	−0.040		0.032	−0.039		0.028	−0.061		0.048	−0.058		0.050
Substance use	0.002		0.008	0.002		0.008	0.008		0.017	0.009		0.017
Between-individual controls
Incarceration (status/time)	−0.045		0.077	−0.035		0.063	−0.081		0.073	−0.011		0.081
Gender	0.047		0.068	0.047		0.051	−0.197	**	0.067	−0.197	**	0.067
Age	−0.092	*	0.046	−0.092	*	0.035	−0.146	**	0.047	−0.145	**	0.047
Black	−0.017		0.075	−0.017		0.052	−0.008		0.077	−0.008		0.077
Hispanic	−0.076		0.075	−0.076		0.055	−0.188		0.099	−0.188		0.099
Other	−0.110		0.138	−0.110		0.103	−0.434	*	0.186	−0.434	*	0.186
Socioeconomic status	0.008		0.031	0.008		0.023	0.010		0.038	0.010		0.038
Single parent	−0.017		0.058	−0.017		0.044	0.189	*	0.078	0.189	*	0.078
Other parent	−0.069		0.086	−0.069		0.070	0.017		0.116	0.017		0.116
Prior incarceration	0.007		0.087	0.007		0.068
Friends arrested (mean)	0.093		0.096	0.093		0.074	0.010		0.122	0.010		0.122
Exposure to violence (mean)	0.003		0.027	0.003		0.020	0.030		0.037	0.030		0.037
Self-reported offending (Mean)	−0.056		0.097	−0.055		0.074	−0.105		0.129	−0.105		0.129
Substance use	0.005		0.020	0.005		0.014	−0.007		0.030	−0.007		0.029

Note. Models presented utilized Penalized Quasi-Likelihood (PQL) estimation. We report the population-average with robust standard error results. All models include the wave the respondent entered the study.

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

Figure 3 graphically displays the change in the domains of family social support between each time period and Time 2 (the reference category) for the total sample and the incarcerated subsample to better depict the relationships reported in Table 2. This figure shows that among the incarcerated group, the length of time incarcerated at Time 1 affects one’s family support. The critical finding is that the decline in family social support is only apparent for those who are incarcerated for a long time (i.e., 4 months or longer assuming a 6 month recall). There is a 19% reduction in family support between Time 2 and Time 3 for those who are incarcerated longer {[(exp^{(0.008 − 0.224)}) − 1] × 100} while those who are incarcerated for a shorter time show a small (1%) increase in the domains of family social support {[(exp^0.008) − 1] × 100}, controlling for several related covariates.

Figure 3:

Percent Change in Rate of Domains of Family Support

Domains of Nonfamily Support Over Time

With respect to nonfamily support, Table 3 shows that, though there are no significant main effects in the domains of nonfamily social support for the total sample (Model 1), there is a significant interaction with incarceration status. Specifically, the domains of nonfamily social support are higher in Time 1 than Time 2 for the incarcerated group compared with the nonincarcerated group (Model 2: b = 0.317, p = .034). For ease of interpretation, Figure 4 graphically depicts the relationships between incarceration and nonfamily social support over time. The coefficients in Table 3 translate to a 35% higher mean of domains of nonfamily social support in Time 1 when compared with Time 2 for those who were incarcerated. In contrast, those not incarcerated have a 1% lower mean of domains of nonfamily support in Time 1 when compared with Time 2 (see Figure 3).

Table 3:

Hierarchical Nonlinear Models of Domains of Nonfamily Social Support Post Incarceration: Comparison of Incarcerated and Nonincarcerated in First 6 Months (Total Sample) and Comparison of Time Incarcerated in First 6 Months (Incarcerated Sample)

	Total sample (n = 479)						Incarcerated sample (n = 196)
	Model 1			Model 2			Model 3			Model 4
	Coefficient	Significance	SE	Coefficient	Significance	SE	Coefficient	Significance	SE	Coefficient	Significance	SE
Intercept	−4.705	**	1.706	−4.665	***	1.301	−6.252	***	1.753	−6.436	***	1.758
Time 1	0.099		0.068	−0.013		0.094	0.295	*	0.115	0.259		0.194
Time 3	0.108		0.068	0.023		0.084	0.275		0.149	0.457	*	0.225
Time 4	0.057		0.070	0.062		0.092	0.051		0.154	0.470	*	0.211
Time 1 × Incarceration				0.317	*	0.149				0.055		0.248
Time 3 × Incarceration				0.244		0.175				−0.299		0.302
Time 4 × Incarceration				−0.018		0.184				−0.707	*	0.305
Within-individual controls
Friends arrested	0.222		0.123	0.215		0.122	0.298		0.231	0.350		0.233
Exposure to violence	0.006		0.028	0.008		0.028	0.003		0.042	0.024		0.041
Self-reported offending	0.264	**	0.085	0.258	**	0.084	0.326	*	0.150	0.360	*	0.142
Substance use	−0.028		0.021	−0.028		0.021	−0.073		0.052	−0.067		0.051
Between-individual controls
Incarceration (status/time)	0.239		0.153	−0.384	*	0.177	0.230		0.184
Gender	−0.795	***	0.120	−0.796	***	0.120	−0.426	*	0.209	−0.420	*	0.208
Age	0.334	***	0.076	0.335	***	0.076	0.418	***	0.100	0.419	***	0.100
Black	−0.506	**	0.153	−0.507	**	0.153	−0.753	**	0.238	−0.759	**	0.237
Hispanic	−0.127		0.137	−0.127		0.136	−0.195		0.266	−0.197		0.267
Other	−0.007		0.244	−0.007		0.244	0.046		0.365	0.029		0.364
Socioeconomic status	−0.061		0.073	−0.062		0.073	−0.079		0.090	−0.079		0.091
Single parent	−0.078		0.130	−0.079		0.130	−0.485	*	0.201	−0.483	*	0.202
Other parent	0.033		0.164	0.032		0.164	0.059		0.213	0.063		0.214
Prior incarceration	0.111		0.221	0.112		0.221
Friends arrested (mean)	−0.037		0.196	−0.037		0.196	0.168		0.263	0.169		0.262
Exposure to violence (mean)	0.077		0.051	0.077		0.051	0.117		0.090	0.118		0.090
Self-reported offending (mean)	0.157		0.201	0.157		0.201	0.207		0.333	0.205		0.332
Substance use	−0.077	*	0.039	−0.077	*	0.039	−0.089		0.067	−0.088		0.067

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

Figure 4:

Percent Change in the Rate of Domains of Nonfamily Support

For the incarcerated only sample, there is also a significant interaction between time and length of incarceration for the time periods post incarceration between Time 4 and Time 2 (see Table 3, Model 4: b = −0.707, p = .021). Specifically, for those incarcerated for longer periods of time during Time 1, there is a 21% reduction in nonfamily social support between Time 4 and Time 2. This is drastically different from those who are incarcerated for shorter stays as their domains of nonfamily support increase to close to 60% between Time 4 and Time 2 (see Figure 4), controlling for key time-varying and time-stable controls.

Discussion

Social support from caring adults is important because it can mediate criminogenic risk factors (e.g., unemployment, substance abuse) for reengaging in crime and being reincarcerated (Hurd & Zimmerman, 2010; Hurd et al., 2014; McDonald & Lambert, 2014). This study builds on existing research by bringing attention to the potential disruption of social support as a result of incarceration and by investigating variations in social support upon reentry for emerging adults. Results show that though the nonincarcerated group experienced relatively stable social support throughout the study period, there are not only negative effects of incarceration on social support, but those negative effects are exacerbated by longer incarceration stays.

Impact of Incarceration Status on Declines in Social Support

Family Support

Existing research among emerging adult samples from the general population reports a fairly even distribution of support from family and nonfamily caring adults (Hurd et al., 2014). In contrast, the justice-involved emerging adults from this study demonstrate much higher support from family than from nonfamily adults. These findings suggest that this group of serious young offenders either have difficulty establishing nonfamily relationships that can provide social support throughout their adolescence and/or experience disruption in these nonfamily supports during emerging adulthood because of incarceration. Indeed, prior work suggests that formerly incarcerated emerging adults are more dependent on family than their nonincarcerated counterparts (Arditti & Parkman, 2011; Arnett, 2007; Martinez & Abrams, 2013; Naser & La Vigne, 2006).

Using the current study findings, we cannot determine which factors contribute to declines in family support over time for the study respondents. However, prior literature suggests some possible influencing characteristics. When youth become involved with the justice system, family members are almost always engaged in the justice process in some capacity. Research shows that justice involvement and the related increase in needed support from family members can create a sense of overwhelming obligation, conflict, or distress (Martinez & Abrams, 2013; Seal et al., 2007). Respondents in these studies indicate drifting away from family support because of family “being controlling” or relationships becoming strained. Future research should examine ways to mitigate family strain and enhance the sustainability of positive support from family.

Nonfamily Support

The patterns of support from nonfamily caring adults confirms our hypothesis that formerly incarcerated emerging adults experience declines in nonfamily social support that their nonincarcerated counterparts do not. The finding that nonfamily social support slightly increases for nonincarcerated individuals is consistent with emerging adulthood patterns (Arnett, 2004, 2007). In contrast, nonfamily social support is rather volatile for those with incarceration histories, showing declines in the initial period after incarceration and then a temporary increase only to be followed by more declines. Although the initial drop in nonfamily support might be the removal of caring adults in the secure setting, the continuation of instability indicates a difficulty in establishing and maintaining nonfamily caring adult relationships after incarceration.

Assuming that it is appropriate to expect and encourage formerly incarcerated emerging adults to expand or stabilize their nonfamily positive social support, the current study findings suggest incarceration stays stunt developmentally appropriate nonfamily relationships. Longer incarceration stays further compound the disruption in both family and nonfamily support for emerging adults. Future social support intervention research should assess how to facilitate the development of social support relationships with nonfamily caring adults.

Length of Incarceration and Deterioration of Support

Overall, those with longer incarceration stays demonstrated substantial instability in their support from both family and nonfamily caring adults. Results support our hypotheses that those participants with longer incarceration stays experience greater deterioration of support than those with shorter incarceration stays. The initial large reduction in family support as incarceration length increases might be related to family strain. Longer periods of disconnection may elevate strain and the relationships between emerging adults and their family members may become further fractured, which makes it more difficult for a successful reconnection following the individual’s return to the community (Martinez & Abrams, 2013). In contrast, more stable patterns of family and nonfamily support for emerging adults who are released after shorter incarceration stays indicate an easier process of resuming the key developmental tasks of emerging adulthood. These experiences may be due to a rallying effect by family members to help the individual with community reintegration.

A within-group comparison of patterns of support across the family and nonfamily domains for the incarcerated groups revealed interesting results. It appears that as support increased from one group (e.g., family), support declined for the other group (e.g., nonfamily). It is beyond the scope of this current study, but future research should examine the relationships in these patterns to try to better understand what factors contribute most to the seemingly symbiotic variations in support for these emerging adults as these patterns likely have ripple effects into issues of social capital, such as bridging networks to employment or other civic and social opportunities.

The broader social support literature demonstrates that recipients of social support have individual characteristics that influence their willingness and ability to attract and access social support and to increase diversity in their social networks. Providers of social support have a range of skills and abilities to provide various forms of support. Some providers do not have the capability to provide all forms of support or the form of support needed by the recipient. Richman, Rosenfeld, and Hardy (1993) contend this range of abilities is important because the benefits of social support will only occur to the extent that the support received is appropriately matched to the needs activated by the stress event such as incarceration (Sarason, Sarason, & Pierce, 1990). Social support may not be actualized if the recipient is unwilling or unskilled at seeking or accepting support, or if the provider is unwilling or incapable of providing support. Mismatches of support exchanges that occur in this way may help to explain changes in support over time for formerly incarcerated emerging adults. In addition, the recipient’s perception of the relevance of the support plays an important role in determining from whom an individual might seek help. As emerging adults navigate relationships, they may cycle through the types of support they feel are most relevant which may contribute to ebbs and flows in network makeup of family versus nonfamily social support providers. Future social support intervention research should seek to identify tools for matching formerly incarcerated clients’ support needs with the support resources in their social network or prospective social networks.

Limitations

This study has many strengths, including the relative racial/ethnic and regional diversity in the sample and the rigor in data collection for the original sample, but some limitations should be noted. First, the use of longitudinal data, while providing temporal order, introduces methodological issues such as panel attrition and respondent recall. Panel attrition was kept to an impressive minimum during the original data collection, ranging from 91% to 93% retention. Respondent recall tends to be more problematic as the time interval between interviews lengthens; thus, while recall biases may be present, these issues are kept to a minimum with the criterion of only using the waves with a 6-month recall period.

Second, the incarcerated sample includes people who were released from a secure setting in the first three waves of the study after an initial incarceration stay at Time 1. This selection criterion may have resulted in a biased sample by including only those who were doing “better” after incarceration while also excluding those who may have attrited from the study or who remained incarcerated after Time 1. Another limitation of the study is the lack of data to identify the nature of the relationship of the nonfamily caring adult to the participant. Therefore, we cannot speculate on whether changes in support have to do with changes in contact with social services systems that may breed supportive nonfamily relationships nor why these changes are occurring. Moreover, combining the type of support provided (e.g., emotional, instrumental, appraisal, or informational) precludes a greater understanding about the importance of each type of support for emerging adults, which is an important avenue for future research. Finally, although consistent with prevalence data, the small sample size of females did not allow us to conduct gender-based subgroup analysis. Despite the study limitations, the findings advance current understanding of the postrelease experiences of formerly incarcerated emerging adults.

Additional avenues for future research include an examination of how the deterioration of postincarceration social support affects recidivism as well as the identification of the malleable factors that contribute to the deterioration of social support over time to incorporate these into intervention development. Finally, a better understanding is needed of whether formerly incarcerated emerging adults are getting their support needs met if they are experiencing declines in support across both family and nonfamily caring adults.

Conclusion

Overall, this study suggests that the emerging adults who likely have high needs for positive support—those with incarceration histories—are less likely to receive or experience stability in this critically needed social support. The instability of social support may occur because well-intending support providers may not have the necessary skills to provide the needed support or for the necessary length of time; the formerly incarcerated emerging adult may not be receptive to available support or may not have the skills to access support; or there may be a mismatch between the formerly incarcerated individual’s support needs and the support resources available to the caring adult. Each of these scenarios suggests that social support is dynamic and a spectrum of support needs and resources may exist for formerly incarcerated emerging adults that are not yet being accessed effectively.

We identify several implications from the study that will contribute to advancing the field of postincarceration social support intervention development. First, existing social support intervention approaches that focus on garnering support immediately after release and then taper off intervention efforts within a few months may be insufficient. Declines in social support do not start to surface until much later in the reentry process. New social support intervention approaches may try a model of staggered intervention wherein intervention sessions are provided in blocks over a 2-year period. Those individuals who have been incarcerated for longer periods of time may need social support intervention “blocked” approaches to span more than 2 years to garner and stabilize positive social support. Second, social support interventions should be examined for gender responsiveness. Do males and females have different social support needs over time? Do males and females (and their support providers) respond differently to social support interventions? Third, interventionists should explore whether prioritizing the stabilization of nonfamily versus family support should be considered, or whether different approaches should be used to stabilize family versus nonfamily supports. Fourth, it is known that during incarceration, the presence of social support via consistent visitation with family members and other informal mentors increases psychological well-being and reduces the likelihood for recidivism after release (Cochran, 2014; Listwan, Colvin, Hanley, & Flannery, 2010). Therefore, what can be done prerelease to reduce the disruption of existing positive social relationships, especially for those sentenced for longer incarceration stays? Finally, given that longer incarceration stays likely contribute to worse social support outcomes, policy makers should consider whether lengthier incarcerations are necessary in all cases and whether lengthier incarcerations may in fact pose greater risks to public safety because they begin to strip away protective factors such as social support, particularly among emerging adults.

Footnotes

Notes

Carrie Pettus-Davis, PhD, MSW, is founding director of the Institute for Advancing Justice Research and Innovation and co-founding director of the Smart Decarceration Initiative. She is an assistant professor at Washington University in St. Louis. Her research focuses on the development and dissemination of policy and practice interventions for those involved in the criminal justice system and their loved ones.

Elaine Eggleston Doherty, PhD, is associate professor in the Department of Criminology and Criminal Justice at the University of Missouri–St. Louis. Her research interests focus on life course criminology and the onset, continuation, and desistance from crime. She has published in top-tier journals such as Criminology, Journal of Research in Crime & Delinquency, Journal of Quantitative Criminology, and Justice Quarterly.

Christopher Veeh, PhD, MSW, is research assistant professor in the Institute for Advancing Justice Research and Innovation at Washington University in St. Louis. His research interests include traumatic brain injury and well-being development within adolescents and emerging adults in the criminal justice system.

Christina Drymon, MA, is a PhD student at the George Warren Brown School of Social Work at Washington University in St. Louis. Her research focuses on a recovery-oriented systems approach to substance abuse and addiction. Her goal is to identify ways to enhance natural recovery from illicit drug addiction by harnessing her training in qualitative data collection, system dynamics, and her NIDA TranSTAR fellowship.

References

Altschuler

D. M.

Armstrong

T. L.

(2001). Reintegrating high-risk juvenile offenders into communities: Experiences and prospects. Corrections Management Quarterly, 5, 72-88.

Andrews

D. A.

Bonta

(2010). The psychology of criminal conduct (5th ed.). New Providence, NJ: Matthew Bender.

Andrews

D. A.

Bonta

Wormith

J. S.

(2006). The recent past and near future of risk and/or need assessment. Crime & Delinquency, 52, 7-27.

Arditti

J. A.

Lambert-Shute

Joest

(2003). Saturday morning at the jail: Implications of incarceration for families and children. Family Relations, 52, 195-204.

Arditti

J. A.

Parkman

(2011). The developmental paradox of young men’s reentry after incarceration. Family Relations, 60, 205-220.

Arnett

J. J.

(2000). Emerging adulthood: A theory of development from the late teens through the twenties. American Psychologist, 55, 469-480.

Arnett

J. J.

(2004). Emerging adulthood: The winding road from the late teens through the twenties. New York, NY: Oxford University Press.

Arnett

J. J.

(2007). Emerging adulthood: What is it, and what is it good for? Child Development Perspectives, 1, 68-73.

Berg

Huebner

B. M.

(2010). Reentry and the ties that bind: An examination of social ties, employment and recidivism. Justice Quarterly, 28, 382-410.

10.

Bonnie

R. J.

Stroud

Breiner

(Eds.). (2015). Investing in the health and well-being of young adults. Washington, DC: National Academies Press.

11.

Brame

Turner

M. G.

Paternoster

Bushway

S. D.

(2012). Cumulative prevalence of arrest from ages 8 to 23 in a national sample. Pediatrics, 129, 21-27.

12.

Budde

Schene

(2004). Informal social support interventions and their role in violence prevention. Journal of Interpersonal Violence, 19, 341-355.

13.

Byrne

J. M.

Taxman

T. F.

(2004). Targeting offender reentry. Federal Probation, 68, 58-61.

14.

Center for Research on Health Care Data Center. (n.d.). Pathways to desistance: Caring adult—Subject follow-up. Retrieved from http://www.pathwaysstudy.pitt.edu/codebook/caring-adult-sf.html

15.

Chriss

J. J.

(2007). The functions of the social bond. The Sociological Quarterly, 48, 689-712.

16.

Chung

H. L.

Schubert

C. A.

Mulvey

E. P.

(2007). An empirical portrait of community reentry among serious juvenile offenders in two metropolitan cities. Criminal Justice and Behavior, 34, 1402-1426.

17.

Cochran

J. C.

(2014). Breaches in the wall: Imprisonment, social support, and recidivism. Journal of Research in Crime & Delinquency, 51, 200-229.

18.

Cohen

Underwood

L. G.

Gottlieb

B. H.

(Eds.). (2000). Social support measurement and intervention: A guide for health and social scientists. New York, NY: Oxford University Press.

19.

Colvin

Cullen

F. T.

Vander Ven

(2002). Coercion, social support, and crime: An emerging theoretical consensus. Criminology, 40, 19-42.

20.

Cullen

F. T.

(1994). Social support as an organizing concept for criminology: Presidential address to the Academy of Criminal Justice Sciences. Justice Quarterly, 11, 527-559.

21.

Cullen

F. T.

Gendreau

(2000). Assessing correctional rehabilitation: Policy, practice, and prospects. In Horney

(Ed.), Policies, processes, and decisions of the criminal justice system (pp. 110-175). Washington, DC: U.S. Department of Justice, National Institute of Justice.

22.

Cullen

F. T.

Wright

J. P.

Chamlin

M. B.

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

23.

Dang

M. T.

Conger

K. J.

Breslau

Miller

(2014). Exploring protective factors among homeless youth: The role of natural mentors. Journal of Health Care for the Poor and Underserved, 25, 1121-1138.

24.

diZerega

(2010). Engaging offenders’ families in reentry: Coaching packet. Washington, DC: Center for Effective Public Policy.

25.

Dobkin

P. L.

De Civita

Paraherakis

Gill

(2002). The role of functional social support in treatment retention and outcomes among outpatient adult substance abusers. Addiction, 97, 347-356.

26.

Draine

Solomon

(2000). Anxiety and depression symptoms and quality of life among clients of a psychiatric probation and parole service. Psychiatric Rehabilitation, 24, 38-45.

27.

Durose

M. R.

Cooper

A. D.

Snyder

H. N.

(2014). Recidivism of prisoners released in 30 states in 2005: Patterns from 2005 to 2010 (NCJ 244205). Washington, DC: Bureau of Justice Statistics.

28.

Farrington

D. P.

(2003). Developmental and life-course criminology: Key theoretical and empirical issues—The 2002 Sutherland Award address. Criminology, 41, 221-225.

29.

Fontaine

Gilchrist-Scott

Denver

Rossman

S. R.

(2012). Families and reentry: Unpacking how social support matters. Washington, DC: Urban Institute.

30.

Fontaine

Taxy

Peterson

Breaux

Rossman

(2015). Safer return demonstration: Impact findings from a research-based community reentry initiative. Washington, DC: Urban Institute.

31.

Fosco

G. M.

Caruthers

A. S.

Dishion

T. J.

(2012). A six-year predictive test of adolescent family relationship quality and effortful control pathways to emerging adulthood and emotional health. Journal of Family Psychology, 26, 565-575.

32.

Gardner

K. A.

Cutrona

C. E.

(2004). Social support communication in families. In Vangelisti

A. L.

(Ed.), Handbook of family communication (pp. 495-512). Mahwah, NJ: Lawrence Erlbaum.

33.

Garrity

T. F.

Prewitt

S. H.

Joosen

Tindall

M. S.

Hiller

M. L.

Leukefeld

C. G.

(2006). Correlates of subjective stress among drug court clients. Journal of Offender Therapy and Comparative Criminology, 50, 269-279.

34.

Gendreau

Little

Goggin

(1996). A meta-analysis of the predictors of adult offender recidivism: What works! Criminology, 34, 575-608.

35.

Giordano

P. C.

Cernkovich

S. A.

Holland

D. D.

(2003). Changes in friendship relations over the life course: Implications for desistance from crime. Criminology, 41, 293-327.

36.

Giordano

P. C.

Cernkovich

S. A.

Rudolph

J. L.

(2002). Gender, crime, and desistance: Toward a theory of cognitive transformation. American Journal of Sociology, 107, 990-1064.

37.

Hairston

Oliver

(2006). Women’s experiences with men’s incarceration and reentry. Women, Girls, & Criminal Justice, 7, 65-80.

38.

Hanson

K. R.

Harris

A. J. R.

(2000). Where should we intervene? Dynamic predictors of sexual offense recidivism. Criminal Justice and Behavior, 27, 6-35.

39.

Hawkins

J. D.

Weis

J. G.

(1985). The social development model: An integrated approach to delinquency prevention. Journal of Primary Prevention, 6, 73-97.

40.

Haynie

D. L.

(2002). Friendship networks and delinquency: The relative nature of peer delinquency. Journal of Quantitative Criminology, 18, 99-134.

41.

Haynie

D. L.

Osgood

D. W.

(2005). Reconsidering peers and delinquency: How do peers matter? Social Forces, 84, 1109-1130.

42.

Hirschi

(1969). Causes of delinquency. Berkeley: University of California Press.

43.

Hobfoll

S. E.

(2009). Social support: The movie. Journal of Social and Personal Relationships, 26, 93-100.

44.

Hochstetler

DeLisi

Pratt

T. C.

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

45.

House

J. S.

(1981). Work, stress, and social support. Reading, MA: Addison-Wesley.

46.

Hurd

N. M.

Stoddard

S. A.

Bauermeister

J. A.

Zimmerman

M. A.

(2014). Natural mentors, mental health, and substance use: Exploring pathways via coping and purpose. American Journal of Orthopsychiatry, 84, 190-200.

47.

Hurd

N. M.

Zimmerman

(2010). Natural mentors, mental health, and risk behaviors: A longitudinal analysis of African American adolescents transitioning into adulthood. American Journal of Community Psychology, 46, 36-48.

48.

Institute of Behavioral Science. (1990). Youth interview schedule: Denver Youth Study. Unpublished manuscript, University of Colorado, Boulder.

49.

Inter-University Consortium for Political and Social Research. (2012). Research on pathways to desistance [Maricopa County, AZ and Philadelphia County, PA]: Subject measures, 2000-2010 (ICPSR 29961). Ann Arbor, MI: Author. Retrieved from http://www.icpsr.umich.edu/icpsrweb/NAHDAP/studies/29961

50.

Laub

J. H.

Sampson

R. J.

(2003). Shared beginnings, divergent lives: Delinquent boys to age 70. Cambridge, MA: Harvard University Press.

51.

La Vigne

N. G.

Visher

Castro

. (2004). Chicago prisoners’ experience returning home. Washington, DC: Urban Institute.

52.

Lipsey

M. W.

Derzon

J. H.

(1998). Predictors of violent or serious delinquency in adolescence and early adulthood: A synthesis of longitudinal research. In Loeber

Farrington

D. P.

(Eds.), Serious & violent juvenile offenders: Risk factors and successful interventions (pp. 86-105). Thousand Oaks, CA: SAGE.

53.

Listwan

S. J.

Colvin

Hanley

Flannery

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

54.

Makarios

Steiner

B. M.

Travis

L. F.

(2010). Examining the predictors of recidivism among men and women released from prison in Ohio. Criminal Justice and Behavior, 37, 1377-1391.

55.

Mallik-Kane

Visher

C. A.

(2008). Health and prisoner reentry: How physical, mental, and substance abuse conditions shape the process of reintegration. Washington, DC: Urban Institute.

56.

Martinez

D. J.

Abrams

L. S.

(2013). Informal social support among returning young offenders: A metasynthesis of the literature. International Journal of Offender Therapy and Comparative Criminology, 57, 169-190.

57.

McCamish-Svensson

Samuelsson

Hagberg

Svensson

Dehlin

(1999). Informal and formal support from a multidisciplinary perspective: A Swedish follow-up between 80 and 82 years of age. Health and Social Care in the Community, 7, 163-176.

58.

McDonald

Lambert

(2014). The long arm of mentoring: A counterfactual analysis of natural youth mentoring and employment outcomes in early careers. American Journal of Community Psychology, 54, 262-273.

59.

Mowen

T. J.

Visher

C. A.

(2013). Drug use and crime after incarceration: The role of family support and family conflict. Justice Quarterly, 32, 337-359.

60.

Mulvey

E. P.

Steinberg

Fagan

Cauffman

Piquero

A. R.

Chaissin

. . . Losoya

S. H.

(2004). Theory and research on desistance from antisocial activity among serious adolescent offenders. Youth Violence and Juvenile Justice, 2, 213-236.

61.

Nakkula

M. J.

Way

Stauber

H. Y.

London

(1990). Teenage risk prevention questionnaire and interview: An integrative assessment of adolescent high-risk behavior. Piscataway, NJ: Rutgers University, Graduate School of Applied and Professional Psychology.

62.

Naser

R. L.

La Vigne

N. G.

(2006). Family support in the prisoner reentry process: Expectations and realities. Journal of Offender Rehabilitation, 43, 93-106.

63.

Petersilia

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

64.

Pettus-Davis

Dunnigan

Veeh

C. A.

Howard

M. O.

Scheyett

A. M.

Roberts-Lewis

(2017). Enhancing social support postincarceration: Results from a pilot randomized controlled trial. Journal of Clinical Psychology. Advance online publication. doi:10.1002/jclp.22442

65.

Pettus-Davis

Howard

M. O.

Murugan

Roberts-Lewis

Scheyett

A. M.

Botnick

Vance

(2015). Acceptability of a social support intervention for re-entering prisoners. Journal of the Society for Social Work and Research, 6, 51-89.

66.

Pettus-Davis

Howard

M. O.

Roberts-Lewis

Scheyett

A. M.

(2011). Naturally occurring social support in interventions for former prisoners with substance use disorders: Conceptual framework and program model. Journal of Criminal Justice, 39, 479-488.

67.

Phillips

Springer

(1992). Extended national youth sports program 1991-1992 evaluation highlights, part two: Individual Protective Factors Index (IPFI) and risk assessment study. Report prepared for the National Collegiate Athletic Association, EMT Associates, Sacramento, CA.

68.

Raudenbush

S. W.

Bryk

A. S.

(2002). Hierarchical linear models: Applications and data analysis methods (Vol. 1). Thousand Oaks, CA: SAGE.

69.

Raudenbush

S. W.

Bryk

A. S.

Congdon

(2004). HLM6: Hierarchical linear and nonlinear modeling. Lincolnwood, IL: Scientific Software International.

70.

Richman

J. M.

Rosenfeld

L. B.

Hardy

C. J.

(1993). The Social Support Survey: A validation study of a clinical measure of the social support process. Research on Social Work Practice, 3, 288-311.

71.

Rossman

S. B.

Fontaine

(2015). Safer return demonstration: Implementation findings from a research-based community reentry initiative. Washington, DC: Urban Institute.

72.

Sarason

I. G.

Sarason

B. R.

Pierce

(1990). Social support: The search for theory. Journal of Social and Clinical Psychology, 9, 133-147.

73.

Schubert

C. A.

Mulvey

E. P.

Steinberg

Cauffman

Losoya

S. H.

Hecker

. . . Knight

G. P.

(2004). Operational lessons from the Pathways to Desistance Project. Youth Violence and Juvenile Justice, 2, 237-255.

74.

Seal

D. W.

Eldrige

G. D.

Kacanek

Binson

MacGowan

R. J.

(2007). A longitudinal, qualitative analysis of the context of substance use and sexual behavior among 18- to 29-year-old men after their release from prison. Social Science & Medicine, 65, 2394-2406.

75.

Shollenberger

T. L.

(2009). When relatives return: Interviews with family members of returning prisoners in Houston, Texas. Washington, DC: Urban Institute.

76.

Siennick

S. E.

(2011). Tough love? Crime and parental assistance in young adulthood. Criminology, 49, 163-195.

77.

Skeem

Louden

J. E.

Manchak

Vidal

Haddad

(2009). Social networks and social control of probationers with co-occurring mental and substance abuse problems. Law and Human Behavior, 33, 122-135.

78.

Steinberg

Chung

H. L.

Little

(2004). Reentry of young offenders from the justice system: A developmental perspective. Youth Violence and Juvenile Justice, 2, 21-38.

79.

Sullivan

Mino

Nelson

Pope

(2002). Families as a resource in recovery from drug abuse: An evaluation of La Bodega de la Familia. New York, NY: Vera Institute of Justice.

80.

Sullivan

M. L.

(2004). Youth perspectives on the experience of reentry. Youth Violence and Juvenile Justice, 2, 56-71.

81.

Sweeten

Piquero

A. R.

Steinberg

(2013). Age and the explanation of crime, revisited. Journal of Youth and Adolescence, 42, 921-938.

82.

Taylor

Z. E.

Doane

L. D.

Eisenberg

(2014). Transitioning from high school to college: Relations of social support, resiliency, and adjustment during emerging adulthood. Emerging Adulthood, 2, 105-115.

83.

Thoits

P. A.

(2011a). Mechanisms for linking social ties and support to physical and mental health. Journal of Health and Social Behavior, 52, 145-161.

84.

Thoits

P. A.

(2011b). Perceived social support and the voluntary, mixed, or pressured use of mental health services. Society and Mental Health, 1, 4-19.

85.

Ullrich

Coid

(2011). Protective factors for violence among released prisoners—Effects over time and interactions with static risk. Journal of Consulting and Clinical Psychology, 79, 381-390.

86.

Vaux

(1988). Social support: Theory, research, and intervention. New York, NY: Greenwood Press.

87.

Visher

C. A.

Courtney

S. M

. (2006). Cleveland prisoners’ experiences returning home. Washington, DC: Urban Institute.

88.

Walker

(2010). “His mam, my dad, my girlfriend, loads of people used to bring him up”: The value of social support for (ex) offender fathers. Child & Family Social Work, 15, 238-247.

89.

Welch

M. R.

Tittle

C. R.

Yonkoski

Meidinger

Grasmick

H. G.

(2008). Social integration, self-control, and conformity. Journal of Quantitative Criminology, 24, 73-92.

90.

White

W. L.

(2009). The mobilization of community resources to support long-term addiction recovery. Journal of Substance Abuse Treatment, 36, 146-158.

91.

Wildeman

Western

(2010). Incarceration in fragile families. The Future of Children, 20, 157-177.

92.

Willis

G. M.

Grace

R. C.

(2009). Assessment of community reintegration planning for sex offenders: Poor planning predicts recidivism. Criminal Justice and Behavior, 36, 494-512.

93.

Wilson

J. A.

Davis

R. C.

(2006). Good intentions meet hard realities: An evaluation of the project greenlight reentry program. Criminology and Public Policy, 5, 303-338.

94.

Wolff

Shi

(2012). Childhood and adult trauma experiences of incarcerated persons and their relationship to adult behavioral health problems and treatment. International Journal of Environmental Research and Public Health, 9, 1908-1926.

95.

Wright

K. A.

Cesar

G. T.

(2013). Toward a more complete model of offender reintegration: Linking the individual-, community -, and system-level components of recidivism. Victims and Offenders, 8, 373-398.

Deterioration of Postincarceration Social Support for Emerging Adults

Abstract

Keywords

Introduction

Background

Social Support

Emerging Adulthood and Social Support

Social Support and Community Reentry

Present Study

Method

Sampling

Measures 2

Dependent Variable: Social Support

Time-Varying Control Variables

Time-Stable Control Variables

Moderators

Analysis

Results

Descriptive Analyses

Main Analyses

Domains of Family Support Over Time

Domains of Nonfamily Support Over Time

Discussion

Impact of Incarceration Status on Declines in Social Support

Family Support

Nonfamily Support

Length of Incarceration and Deterioration of Support

Limitations

Conclusion

Footnotes

Notes

References

Measures²