Abstract
One in fourteen children experience the incarceration of a parent only to be reunited upon release where risk for recidivism remains high—the implications of which for children’s well-being is not well understood. To fill this gap, the study uses data from the Multi-site Family Study on Incarceration, Parenting and Partnering to follow 543 justice-involved fathers 18-months post-release. Results suggest that stable reentry trajectories are associated with children’s decreased behavior problems, particularly for those who are non-residential. Boys and older children may be more negatively influenced by frequency of recidivism whereas girls and younger children may benefit more from longer durations of community time. Findings inform strategies that ease reintegration processes and improve well-being for children in justice-involved families.
Introduction
A series of policy shifts fueled by “tough on crime” rhetoric launched the U.S. into an era of mass incarceration, resulting in unprecedented rates of imprisonment (National Research Council, 2014). At its peak in 2009, 1.6 million individuals were incarcerated in prison with millions more detained in jails and on community supervision (Carson, 2020). Even with a decrease in population of 11.4% between then and 2019, approximately 1.4 million people remain imprisoned annually—nearly 600,000 of which are new admissions (Carson, 2020). The majority of those imprisoned are men, half of whom are parents, resulting in 2% of all children with a currently incarcerated parent (Glaze & Maruschak, 2010; Murphey & Cooper, 2015).
That said, 95% of people in state prisons are eventually released, equating to nearly 567,000 in 2019 (Carson, 2020; Hughes & Wilson, 2003). However, reentry is rarely linear with a significant proportion of individuals “churning” back through the system, or recidivating. The National Institute of Justice (n.d) defines recidivism as “acts that resulted in re-arrest, reconviction or return to prison with or without a new sentence.” Estimates suggest that 68% recidivate within 3 years of release, 77% within 5 years, and 83% within 9 years; one-third of recidivisms occur within the first 6-months of release and more than half by the end of that first year (Alper et al., 2018). However, administrative, policy, and practice responses, as opposed to new crimes or safety concern, largely contribute to risk of recidivism. Indeed, 45% of prison re-admissions are caused by probation or parole revocations, 26% due to technical violations, and many others resulting from unpaid fines and fees (Pettus-Davis & Kennedy, 2020).
This churning means that many of the 5 to 8 million children who experience parental incarceration have post-release reunification processes halted as they endure additional cycles of loss and separation (Murphey & Cooper, 2015; Pettus-Davis & Kennedy, 2020). Despite evidence of family support as pivotal to returning men’s community reintegration, and children as motivators for fathers’ post-release success (e.g., Charles et al., 2019; Naser & La Vigne, 2006), little research documents how paternal recidivism contributes to children’s well-being. The current study sets out to examine how fathers’ reentry trajectories are associated with children’s behavior following prison; the findings have implications for reentry strategies that stabilize fathers post-release so as to support the well-being of their children.
Background
Reentry and Recidivism
Examining recidivism patterns is necessary for understanding trajectories of individuals released from prison. According to Alper et al. (2018), a measure of recidivism has three components: (1) a starting point, such as prison discharge; (2) a “failure,” such as a re-arrest or re-incarceration; and (3) an observed period (e.g., 6-months, 1-year, 3-years). With the threat of recidivism particularly high in the first months of release (Alper et al., 2018), the field has set out to identify contributing risk factors for measurements of “failure.” In a recent meta-analysis, personal demographics, antisocial behaviors, criminogenic needs, and family factors were all pointed to as influential in individuals’ risk for churning back in the system (Katsiyannis et al., 2018). For instance, men and younger individuals are more likely to recidivate with both criminogenic (antisocial tendencies and peers) and non-criminogenic factors (mental health, community disadvantage, housing instability) contributing additional risk for carceral churn (Olson et al., 2016; Taxman & Caudy, 2015). For those that are parents, increased involvement and positive parent-child relationships after prison have been found to potentially reduce criminality (Visher, 2013), though post-release structural challenges may still derail success.
For instance, upon release, individuals are often met with competing demands that obstruct reintegration. The formerly incarcerated face challenges in obtaining and maintaining employment (Visher et al., 2011). Some may be barred from public housing, while others face hurdles in application processes, credit requirements, and deposit or rent payments—contributors to housing insecurity (Geller & Curtis, 2011). These challenges come coupled with barriers to accessing healthcare, treatment, and social welfare benefits (e.g., Western, 2018). On top of this, many are released with high legal fees and those on extended supervision face strict rules of release (Link, 2021). Each structural barrier has the potential to destabilize reentry experiences, thus increasing risk for recidivism with potential detriments to children’s adjustments.
Currently and Formerly Incarcerated Fathers and Their Children
As alluded, research suggests that family relationships both during and after incarceration play a role in parents’ risk of recidivism. For instance, more frequent and positive family contact during prison has been found to contribute to lower rates of recidivism and greater likelihood of desistance for parents (Bales & Mears, 2008; Thomas et al., 2021). These visitation experiences are also known to bolster other post-release outcomes that may indirectly improve odds of successful reentry, such as improved family support, child attachment, and identity formation (DiZerega & Verdone, 2011; La Vigne et al., 2005; Naser & La Vigne, 2006). Indeed, families play a key role in supporting fathers in prison thereby bolstering successes post-release.
At the same time, many formerly incarcerated fathers (like all those released) face challenges in employment, housing, and family relationships which come coupled with complexities in maintaining—or even rebuilding—relationships with children (Charles et al., 2019). Yet, upwards of 86% of released men rely on family for housing post-release, with two-thirds leaning on family for financial needs (Naser & La Vigne, 2006). These stronger post-release family ties have been associated with lower risk of recidivism (Visher, 2013), potentially by positively influencing mechanisms such as employment, substance use treatment, and mental health care (Berg & Huebner, 2011; La Vigne et al., 2005). Across the board, family relationships are reported as a leading factor in keeping people out of prison, with renewed dedication to children being one specific motivator (Charles et al., 2019; La Vigne et al., 2005).
Understanding what contributes to stable reentry trajectories for fathers matters given the challenges that their incarceration sparks for children. Though these minors face significant adversity beyond parental incarceration alone, the imprisonment, itself, is directly associated with a range of poor developmental outcomes (e.g., Murphey & Cooper, 2015; Wildeman et al., 2018). That said, limited work has tied paternal carceral churn to child well-being despite recidivism’s inherent exposure to law enforcement being found traumatic for children and parental instability (including that of multiple incarcerations) being associated with increased child behavior problems (Muentner et al., 2019, 2021). In the case of maternal incarceration, recidivism has been found to sever parent-child relationships and, more generally, lead to children viewing law enforcement as unjust, taking away control, and destabilizing family life (Few-Demo & Arditti, 2014; Smith & Young, 2017). However, more work is needed to contextualize how fathers’ recidivism may complicate child development. As noted by Wildeman et al. (2016), the majority of work on parental recidivism to date focuses on outcomes for parents, whereas little work has examined links between churning and children’s outcomes—a significant gap that this study begins to fill.
The utility in examining links between parental reentry trajectories and child outcomes is emboldened by evidence around the intergenerational transmission of crime. For decades, scholars have posited that aspects of crime and punishment are aggregated within households, as youth with family members involved in the criminal legal system may be more likely to become systems-impacted themselves, at rates between 20% and 125% higher than their peers (e.g., Farrington et al., 2001; Hjalmarsson & Lindquist, 2012; Laub & Sampson, 1993; Wildeman, 2020). While there have been some studies that identify direct links between parental incarceration and children’s criminality (e.g., Billings, 2019; Wildeman & Andersen, 2017), a number of key mechanisms are likely at play. For instance, parental incarceration may increase material hardship, family instability, and stigma, which in turn increase children’s risk of antisocial behaviors (Dallaire et al., 2010; Geller et al., 2011; Massoglia et al., 2011). Other work discusses how parenting behaviors can mediate this relationship (Thornberry et al., 2009a, 2009b), calling into question how aspects of parents’ recidivism, such as churning, may be a risk for youth adjustment whereas time spent together in the community may buffer poor outcomes.
In general, much remains unknown as to how youth fare when parents are released. Evidence suggests that those with formerly incarcerated parents, on average, exhibit poorer mental health and more frequent substance use (Davis & Shlafer, 2017a, 2017b). When reentry is operationalized more concretely, this adversity becomes more evident as children see continuing, and even worsening, concern—with the largest associations for boys and older youth (Craigie et al., 2018; Yaros et al., 2018). Though rates of parent-child co-residence are lower post-release than pre-incarceration (McKay et al., 2018), some evidence suggests that stronger ties between formerly incarcerated parents and their child may help buffer adverse associations of parental incarceration and child well-being but that negative father-child co-residential experiences may induce additional challenge (Yaros et al., 2018). Finally, children who experience parental incarceration are at risk for adverse outcomes across the life course, including lower healthcare utilization and labor force participation (Heard-Garris et al., 2018; Mears & Siennick, 2016). Even still, more work is needed to understand ongoing consequences of paternal imprisonment that extend for children beyond an initial instance of confinement.
Current Study
Strain theory is a leading framework to discuss stress as it relates to criminology (Agnew, 1985, 1992). The theory posits that, despite whether individuals strive for positive outcomes, negative stimuli (e.g., re-arrest or re-incarceration) intersect paths toward success and induce strain that extends beyond the individual. In the context of recidivism, the stressor can affect family members and lead to problems in family functioning and behavior. As such, strain associated with churning may cause stress (provoked by additional loss and separation) for parents and children that, in turn, prompts negative emotions for children which can cause adversity. For example, Davis and Shlafer (2017b) describe strain as compromising children’s post-release adjustment and Mears and Siennick (2016) discuss its lasting consequences. In short, parental justice-involvement brings a host of negative stimuli that can lead to diminished well-being. As such, recidivism patterns may play a substantial role in these processes.
This study examines paternal recidivism as a strain for children by first describing fathers’ trajectories of community residence and recidivism after release from prison. Second, it examines associations of these trajectories with child behavior problems, considering stable (permanent) re-entry, number of re-arrests and re-incarcerations, and months of community time, as well as whether father-child co-residence moderates these associations. Finally, it examines potential heterogeneity in these associations by child gender and age.
Method
Procedures and Participants
Data for this study comes from the Multi-site Family Study on Incarceration, Parenting and Partnering (MFS-IPP), a project that evaluated federally funded relationship-building programs for men in prison (for more information, see: McKay et al., 2019). Program enrollment began in 2008 with varied recruitment efforts in five prison sites in Indiana, New Jersey, Ohio, New York, and Minnesota. Across locales, the MFS-IPP employed either a matched-comparison or wait-list design for treatment assignment; site and treatment are included as controls in this study’s analyses. To be eligible, men needed to be in an intimate or co-parenting relationship, over the age of 18, English speaking, and able to provide contact information for partners.
After completing the baseline interview, the study team contacted partners to be consented and interviewed. All participants were interviewed again after 9- and 18-months, and a subset at 34-months (because the final wave interviewed only a subsample, this study uses only the first three waves). At baseline, incarcerated men identified a target child to answer specific questions about across the subsequent waves. Target children were chosen based on being co-parented by both members of the couple and being as close to 8-years-old as possible. In the interviews, men were asked questions that pertained to their ongoing criminal involvement and both respondents reported on aspects of family stability and children’s well-being. Those in the community (men who had been released and their partners) received a cash incentive of $35 to $75 per interview. The current study is limited to male respondents that were a parent to a minor child between the ages of 3- and 17-years-old and had been released at some point before the 18-month interview, yielding an analytic sample of 543 currently and formerly incarcerated fathers.
Table 1 displays demographics of the fathers and their children. Fathers ranged in age from 20- to 53-years-old, averaging 34-years (SD = 7.11), and were majority Black (n = 310, 57.09%). About two-thirds had a high school degree or more (n = 362, 66.67%) and over half were employed in the months before prison (n = 326, 60.26%). At baseline, fathers had been incarcerated for an average of 2.5 years (SD = 2.55), approximately two-thirds were incarcerated for non-violent offenses (n = 349, 64.27%). Prior to imprisonment, fathers had been arrested an average of 14 previous times (SD = 13.61) and incarcerated approximately 7 (SD = 9.76). Target children were an average of 8-years old (SD = 3.91), half were boys (n = 268, 50.28%), and half were girls (n = 265, 49.72%). Nearly two-thirds of fathers lived with the target child in the months before their current incarceration (n = 355, 66.60%).
Participant Demographics and Bivariate Descriptive Statistics.
Note. aThose with community time by 9-months (n = 317); bThose with community time by 18-months (n = 482).
Measures
Paternal reentry and recidivism
At the 9- and 18-month follow-up interviews, fathers provided information on whether, since the prior interview, they had been released and if they had been re-arrested or re-incarcerated. Those who were released and experienced any community time by either the 9- or 18-month community waves were coded as experiencing either (1) stable reentry or (0) recidivism. This is measure is used in the regression analyses to examine average differences in child behavior problems for children whose fathers ever and never recidivated. However, descriptive statistics are also provided for more fine-grained categories, including: long and stable reentry (released before 9-month wave and did not recidivate by 18-month interview); a short but stable reentry (released between 9- and 18-month wave and did not recidivate); and unstable reentry (released and recidivated between waves at either time point).
In addition to general recidivism, the cumulative numbers of re-arrests and re-incarcerations between waves were self-reported by fathers. Counts of these variables were summed together to total the number of any form of recidivism experienced between waves.
Finally, fathers reported the number of days they were in the community or incarcerated between each wave. These variables were used to compute the number of days the father spent in the community, which was then divided by 30 to create a measure of months of community time.
Child well-being
Though both mothers and fathers reported on children’s behaviors, these analyses use mothers’ reports because they would have had more involvement with the children (and thus be better positioned to report on their behavior) during the incarceration (Arditti & McGregor, 2019). The study uses a scale of eight items from the National Survey of Children’s Health based on the Problem Behavior Index (Zill, 1990). Factor analyses in a previous study using the scale in this dataset found that responses fit into two constructs: internalizing and externalizing problems (Yaros et al., 2018). Internalizing problems consisted of whether the child was stubborn, or irritable; felt worthless or inferior; was unhappy or depressed; or was withdrawn. Externalizing problems consisted of whether the child bullied or was mean to others; showed respect for adults; got along with other children; or was disobedient. When necessary, responses were reverse coded such that higher values meant more behavior problems (on a scale of 0–3). Each set was summed to account for total behavior problems in the domain, values were then standardized to a mean of 0 and a standard deviation of 1 for ease of interpretation.
Covariates
The study controls for time-variant indicators that may be correlated with children’s well-being or risk of paternal recidivism. Father-child co-residence accounts for whether the dyad lived together at the time of the interview. Father employment provides an indicator of whether the father had a job while in the community. Community supervision accounts for whether the father was released on probation or parole after prison. The models also hold constant a number of time-invariant variables, such as parent age, race, and education; previous recidivism and offense degree; as well as child gender and age.
Analytic Strategy
To handle missing data, a multiple imputation procedure was conducted in STATA statistical software to create 10 imputed complete case datasets across which all analyses were performed. Descriptive statistics are presented at the father level (e.g., one observation per father). However, regression analyses are performed using father-wave level data such that each father is observed at baseline and at each follow-up. To begin, regression models were estimated for each combination of outcomes (internalizing and externalizing behavior problems) and key predictors (stable reentry vs. recidivism; cumulative rearrests or reincarceration; months in the community). The first models consist of OLS regressions estimating average between-child differences in child behavior problems at 18-month follow up as a function of father reentry trajectory. The second are OLS regression with father-specific fixed effects, which estimate the average within-father (child) differences in child behavior problems from baseline (during father incarceration) to 18-month follow up as a function of father reentry pattern. Fixed effects are particularly advantageous in addressing the study’s questions because they control for unmeasured time-invariant characteristics that may contribute to both child well-being and paternal recidivism. However, the models continue to be subject to bias due to unobserved time-variant variables or time-invariant variables that have time-varying effects on child well-being. The standard OLS and fixed-effects regressions are then estimated with the inclusion of an interaction term between father-child co-residence and the key predictor to determine whether results vary by whether a father and child are co-resident. Finally, the fixed effects models are estimated separately for boys, girls, and those older and younger than 8-years-old to examine potential heterogeneity in the associations by child gender and age.
Results
Reentry Patterns
Over the study period, roughly 25% of released fathers experienced recidivism. Of those with stable reentries, more than half (57%) were classified as experiencing a long and stable reentry and just under half (43%) had short but stable reentry. Total counts of recidivism across the whole sample ranged from 0 to 12 (M = 0.59; SD = 1.35); however, instances of recidivism were concentrated among those who had unstable reentry trajectories, experiencing either re-arrest(s), re-incarceration(s), or both. For this portion, the number of re-arrests ranged from 0 to 9 (M = 1.22; SD = 1.18) and re-incarcerations from 0 to 5 (M = 1.05; SD = 0.92), resulting in an average of 2.28 (SD = 1.77) total counts of recidivism ranging from 1 to 12. In terms of total community time, fathers were released for an average of 11 months (SD = 3.83) with durations that ranged from under a week in total to nearly the entire 18-month observation period.
Bivariate tests to examine differences in characteristics by post-release stability (stable reentry vs. ever recidivism) are displayed in Table 1. These findings suggest that fathers who recidivated were younger and more likely to be unemployed prior to baseline. Fathers with stable reentries averaged longer prison sentences than those who recidivated (2.6 years compared to 1.91). However, those with unstable reentries experienced more previous adult arrests (M = 17.55) and incarcerations (M = 8.97) compared to those who had stable reentry trajectories.
Full-Sample Regression Results
Table 2 presents the full set of standard OLS and fixed effects regressions for each combination of key predictors and outcomes. The standard OLS results, reported in the first four columns of Table 2, reveal no statistically significant associations of stable reentry (Panel A), frequency of recidivism (Panel B), or duration of community time (Panel C) with children’s behavior problems. This lack of association holds regardless of whether the child and father were coresident, as indicated by the lack of statistical significance of both the main effect in all models and of the resident father × stable reentry interaction in all models in which it is included.
Model Results for Children’s Well-being as Predicted by Recidivism Measures.
Notes. †p < .1. *p < .05. **p < .01. ***p < .001; aOLS Regression Results include a full set of covariates not displayed.
The final 4 columns of Table 2 present results from regressions that include individual-specific fixed effects. Results are, again, presented for models with and without the inclusion of an interaction between father co-residence and the primary predictor. Consistent with the standard OLS results, none of the key predictors is associated with child behavior problems in models that exclude the interaction term. However, this appears to reflect heterogeneity in the within-child association of these variables with child behavior problems based on father-child co-residence. For example, the fixed-effects results for stable reentry (Panel A) illustrate that stable reentry for children who are not coresident with their father is associated with 0.28 standard deviations (SDs) fewer externalizing behavior problems. However, stable reentry is associated with only 0.15 SDs fewer externalizing behavior problems among children who are co-resident. No association is found between stable reentry and internalizing behavior problems.
The fixed-effects models also highlight a direct association between frequency of recidivism and children’s behavior problems (Panel B) such that every re-arrest or re-incarceration is associated with increases in externalizing and internalizing problems of 0.05 and 0.04 SDs, respectively, and this association does not statistically significantly vary by co-residence with the father. Note, however, that the marginally significant (p < .10) interaction effect for internalizing behavior problems suggests that this association does not hold for children who are living with their father for internalizing behavior problems. Results from the duration of community time models (Panel C) show that every month that a father remains in the community is associated with 0.02 SDs fewer internalizing behavior problems and, again, this effect does not statistically significantly vary by co-residence status with the father.
In summary, children experienced, on average, fewer behavior problems if fathers did not recidivate, though the strength of this association may be less so when children are co-resident children with their father. These trends are mirrored in models measuring frequency and duration, such that additional recidivism counts are generally associated with greater problem behaviors but every month in the community is associated with fewer internalizing behavior problems. Moreover, while the interaction terms fail to attain statistical significance in these models, the direction of the estimates suggests that the association is likely larger when children are non-co-resident children than co-resident with their father.
Heterogeneity by Child Gender and Age
Table 3 presents results from fixed-effects models that include the key predictor × father-child co-residence interaction, but are estimated separately for boys, girls, children older than 8, and children 8 and younger to examine potential heterogeneity by child gender and age. For boys who are not living with their father, stable reentry (Panel A) is associated with 0.33 SDs fewer externalizing problems, whereas stable reentry is associated with only 0.12 SDs fewer externalizing problems for boys who live with their father. For girls, having a father with stable reentry is associated with 0.24 and 0.19 SDs fewer externalizing and internalizing behaviors, respectively. Unlike boys, there is a significant association for internalizing behaviors; the results of which do not significantly differ as a function of co-residence. The results in Panel B suggest that boys may be more vulnerable to churn than girls. For boys, each additional count of recidivism is associated 0.07 SDs more externalizing and 0.06 SDs more internalizing behavior problems. However, residence moderates this relation for externalizing behaviors (significant at p < .10), suggesting that every instance that a co-resident father recidivates is associated with a 0.14 SD increase in behavior problems. In contrast, frequency of recidivism is non-significant for girls’ internalizing and externalizing behavior problems, regardless of co-residence status. Finally, fathers’ time in the community (Panel C) has no association with boys’ behaviors despite each additional month the father is in the community being associated with 0.04 fewer internalizing behavior problems for girls, regardless of co-residence status.
Fixed Effects Model Results of Children’s Well-being, by Child Gender and Age.
Turning to child age, stable reentry (Panel A) is significantly associated with 0.22 SDs fewer externalizing behavior problems for children who older than eight and 0.34 SDs fewer externalizing behaviors for children eight or younger. However, the latter association is larger among children who do not live with their father; younger co-resident children exhibit only 0.23 SDs fewer externalizing problems. Each re-arrest or re-incarceration (Panel B) is associated with 0.05 SDs more internalizing behavior problems for children older than eight, regardless of co-resident status, and with .08 SDs more externalizing behavior problems (marginally significant at p < .10) for children eight or younger, regardless of co-residence status. Finally, each additional month the father is in the community (Panel C) is associated with 0.03 SDs fewer internalizing problems for children older than eight, regardless of co-residence status, and with 0.02 SDs fewer externalizing problems for children eight or younger, regardless of co-residence status.
Discussion
With unprecedented rates of mass incarceration in the U.S., the country remains the world’s leader in imprisonment. Upon release, significant challenges await the formerly incarcerated that may elevate their risk for cycling back through the carceral system. Though considerable adversity has been documented for children while parents are incarcerated, less work examines children’s adjustments after release, including potential associations of recidivism with their well-being. Strain theory posits that father recidivism can induce stress that dysregulates the family system, which may adversely affect children. This study set out to examine associations of fathers’ reentry trajectories with child behavior problems, contributing to the literature on the well-being of children with currently and formerly incarcerated parents.
In general, those released from prison have high likelihoods of recidivism with estimates of more than 4 of 5 individuals churning back through the system within a decade of their release, the majority of which occurs within the first year (Alper et al., 2018). Though reentry timelines for fathers in this sample were not standardized (meaning some had been released more or less recently than others), participants had been in the community for an average of 11-months and only about 25% had recidivated—rates significantly lower than national averages. While no data follows reentry trajectories of parents, specifically, it is possible that fathers have lower rates of recidivism than the general population of releases. This may, at least in part, reflect fathers’ (renewed) dedication to parenting and their attribution of children as motivators for post-release success (Charles et al., 2019) or the ways in which family relationships and contact with children can be associated with lower levels of criminal behavior and risk for recidivism (Visher, 2013). That said, more work that tracks release statistics for parents and uses causal identification strategies are needed to fully answer these questions.
The study begins to provide a profile of fathers who may be particularly likely to experience churn after release. Descriptively, those who are younger and were unemployed prior to imprisonment were more likely to have unstable reentries, as were those with shorter sentences and those who had more adult arrests and incarcerations. These findings are largely in line with other work that studies recidivism in the general population. Indeed, in their meta-analysis, Katsiyannis et al. (2018) discuss the role of gender, age, poverty, anti-social tendencies, and criminogenic histories as being significant predictors of recidivism.
Though these correlates of churn are well-documented in the field, questions around what makes for stable reentry remain harder to grapple with. Western (2018) points to technological advancements while incarcerated; issues with poverty, unemployment, housing instability, and welfare program eligibility; community violence and historical trauma; mental, emotional, and physical health and substance use issues; racism and discrimination; and depleted social ties as only some potential barriers ahead of the formerly incarcerated in the year after prison. Indeed, the reentry period can be intense, challenging, and demoralizing as the same concerns known to await people back in the community are also contributors to risk of recidivism, which can result in a cyclical nature of incarceration. Because of this, fathers must navigate competing demands meanwhile attending to issues around maintaining, or potentially rebuilding, relationships with children—the consequences of which for the children, themselves, remain largely unknown.
In this study, fathers’ stable reentry was associated with decreased behavior problems for children in regression models that include individual fixed effects, but not in standard OLS regression. Whereas the standard OLS models estimate differences in behavior problems between children whose fathers experienced stable reentry and those who recidivated, the fixed-effects models estimate within child change in behavior problems between the initial (during father incarceration) and 18-month waves (by which point each fathers had reentered the community at least once) as a function of whether the father experienced stable reentry or some recidivism. Thus, differences between the two models indicate that declines in behavior problems over time are associated with stable father reentry, whereas there is no average difference in behavior problems for children who experience stable father reentry and those who experience recidivism. The fixed effects results represent the more rigorous of the two estimates as they are not subject to bias due to omitted time-invariant factors. These findings contribute to the literature that documents the interplay of family relationships and post-release stability (Naser & La Vigne, 2006; Visher, 2013), extending the focus now to include children. Because instances of parental residential instability (including spells of additional incarcerations) has been found to increase children’s behavior problems (Muentner et al., 2019), it is reasonable that fathers’ post-release stability would, in turn, be associated with improved outcomes seen herein.
The study findings point to the potentially challenging role that co-residence may present, despite a smaller proportion of children living with fathers post-release (McKay et al., 2018). Pointedly, while stable reentries are associated with decreased behavior problems compared to recidivism, the effect sizes of this relationship are smaller for co-resident children. These children may be forced to renegotiate norms, alter expectations, cope with changes in guardianship or allocation of resources, and manage stigma, all while fathers navigate the complexities that the first few months of reentry bring. Yaros et al. (2018) similarly discuss co-residence as being a complex phenomenon, finding that strained relationships increase children’s behavior problems when they live together post-release, commenting that “how much a child is exposed to his or her father helps to determine the influence the father-child relationship has on the child’s well-being” (p. 156). This trend is similarly found in the literature on the intergenerational transmission of crime, finding the strongest associations of children’s antisocial behaviors concentrated among parent-child dyads who have frequent contact (Thornberry et al., 2009a, 2009b). Because reentry is most smooth for children when it does not disrupt their regular routines (Smith & Young, 2017), it may be that witnessing shifts in household functioning makes it so that co-resident children see smaller improvements than non-co-resident children who go about this transition without such exposure.
Though indicators of stable reentry help to describe fathers’ general trajectories post-release, recidivism is often more than a “yes” or “no” experience. This lack of nuance may limit how frequency of recidivism as well as duration of community time may differentially influence children. First when considering frequency of recidivism, this study finds associations between additional re-arrests or re-incarcerations and increased behavior problems. Children may feel sadness, anger, resentment, and hurt when parents are re-incarcerated, describing parental recidivism as “a repeated cycle that kept going” (Smith & Young, 2017, p.17); such aspects of instability are associated with increased internalizing and externalizing behaviors for children (Muentner et al., 2019). Taken together, not only may recidivism contribute to child behavior problems but, more frequent post-release criminal justice contact may be compounding in nature.
Length of time in the community has received little attention in the literature on child father reentry and child well-being, despite documented concern for up to 9 years or longer after release (Craigie et al., 2018). This study’s results suggest that children may fare better when fathers remain in the community for longer periods of time, particularly with respect to internalizing behavior problems. Extended community time may increase opportunity for involvement—a driver for fathers to remain out of prison (Charles et al., 2019)—and ultimately lead to stronger relationships that improve children’s behaviors (Yaros et al., 2018). Taken in tandem, while frequency of recidivism is associated with increased behavior problems, the longer duration that a father remains in the community is related to decreases—findings that push the field beyond ever/never measures of incarceration-related experiences.
The results point to significant heterogeneity in these associations across gender and age. While stable reentry is associated with decreased problem behaviors for most children, frequency of recidivism may be most challenging for boys and older children, while community duration may be most meaningful for girls and younger children. These findings compliment other work that documents elevated post-release behaviors problems for boys and older children (Craigie et al., 2018; Yaros et al., 2018). Gendered social expectations and differences in children’s age-specific development likely play into this, such that boys may externalize socioemotional challenges and older children may exercise more agency in decision making in the extent to which they welcome fathers back into their lives. This study suggests that fathers’ community duration after release may be particularly beneficial for girls and younger youth as they employ different coping mechanisms than their male and older counterparts.
Associations of recidivism and children’s outcomes calls into question how to best promote stability, to which the Risk-Need-Responsivity (RNR) model has gained momentum among practitioners. This rehabilitative approach has been found to lower recidivism rates by more intensely treating those at heightened risk for re-offending, targeting their specific challenges, capitalizing on protective factors, and tailoring treatment plans to each person (Bonta & Andrews, 2007). These strategies have also proven effective in certain sub-populations, for instance those in community mental health programs (Campbell et al., 2015), thereby suggesting potential worth in improving reentry outcomes for parents and indirectly stabilizing children.
Implications
This study’s findings have implications for policies that help to stabilize individuals upon release. Calls for justice reform range from shifts to rehabilitative responses, such as RNR models, to more progressive calls for prison abolition. Depending upon where one falls on this spectrum, a number of policy levers come into play that reduce risk of re-incarceration and improve odds of remaining in the community for longer. These include reducing sentence lengths, ending cash bail, reducing reliance on correctional institutions, amending rules of release so as to reduce risk of revocation, and reallocating funds to community programs and services. These policy reforms have the potential to improve outcomes for those involved in the system, thereby extending the amount of time that a parent may remain in the community with their child(ren) and reducing structural risks for re-arrest and re-incarceration.
The results also underscore the potential benefits of reductions in parental re-incarceration for children. While probation (community supervision in lieu of incarceration) is the most common form of this, its overreliance as a primary alternative comes coupled with inherent challenges as strict and overbearing rules of release can be detrimental for a released individual’s (and their family’s) well-being and elevate risk of revocation. Instead, increasing access to services such as community-based drug treatment or mental health programing have the potential to reduce future justice system contact which may benefit both parents and children. Some states (Washington) have even begun employing alternatives to incarceration specific for parents; for example, both the Family and Offender Sentencing Alternative and the Community Parenting Alternative make certain justice-involved parents eligible for community-based programing instead of incarceration, thus opening up the door for continued involvement with children. Expanding these services has the potential to mitigate risk for families by reducing challenges of separation and reunification while providing parents with treatment services.
On the practice front, reentry planning is a promising strategy that may help to stabilize post-release outcomes. While most institutions provide a form of this service to some incarcerated individuals, its utilization is not standardized. Preparation for reintegration would ideally begin at intake by connecting people with programs that fit self-determined goals and intervene upon challenges looming in the community. For fathers, this may include attention to parenting roles; for example, requesting support in preparing for in-prison contact with children or by sorting out challenges with child support. Not only does this strategy have the potential to improve outcomes for parents and children while in prison, it also begins reintegration processes early so as to situate fathers for a more stable release, ultimately reducing risk of recidivism.
Reentry planning can be coupled with a continuum of care that links parents with support upon release. In the best case, this would allow a continuation of programing received in an institution to extend to community settings. At a minimum, this would include a compilation of resources in one’s community that could be offered upon discharge. The greater extent to which parents can be offered service options that extend across multiple reentry challenges (i.e., residence, employment, social support, education, and health treatment), the better their odds of post-release stability. These services should also have a family focus, thereby allowing fathers to attend to challenges in parenting roles and continuing to mitigate risk for children’s well-being.
Limitations
Despite this study’s contributions, there are several limitations to consider. First, the study sample may reflect systematic self-selection bias in that participants selected into a relationship program, potentially resulting enrollment of those with more positive family relationships. The sample is also restricted to those released before the 18-month interview, potentially limiting the window of observation for recidivism. Taken together, these fathers are not representative of the general population of released men, thus follow-up studies at other notable desistance checkpoints (e.g., 3-, 5-, and 9-years; Alper et al., 2018) would further situate reentry patterns and study findings within a longer-term view of desistance. Additionally, the broad mother-reported behavioral measures may miss nuances in how children cope when fathers are released and recidivate. The analyses also include only a small range of child outcomes; future work would benefit by also studying health and learning. Additional attention to children’s well-being following maternal reentry is also warranted. Finally, it may be worthwhile for future research to examine variation in associations of parental reentry and child well-being by race, supervision status, and neighborhood characteristics, which were beyond the scope of this study.
Conclusion
As a whole, this study points to the need to promote strategies that reduce reintegrating fathers’ rates of reincarceration. Such efforts may be beneficial to children given the challenges that children experience during incarceration, the consequences of which that continue post-release are only beginning to be understood. This examination of children’s behavior problems when parents recidivate after prison begins to fill an important gap in the literature by providing evidence of how post-release stability may be associated with decreased problem behaviors, as well as heterogeneity seen herein by child gender, age, and residential status. Future work in this area has the potential to inform policy and practice strategies that promote post-release success for parents, thereby improving outcomes for their children.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
