Mitigating the Effects of Parental Incarceration through Social Intervention

Abstract

This study addresses two central research questions: (1) are children with incarcerated parents (CIP) more deviant than nonimpacted peers and (2) is a regional mentoring social intervention program effective for CIP? Two sources of data were used, longitudinal data gathered from 173 children involved with a regional branch of Big Brothers Big Sisters (BBBS) and a sample of children from the Fragile Families (FF) and Child Wellbeing Study. Based on the BBBS data, results find that CIP were more deviant than nonimpacted peers. Unexpectedly, children involved with BBBS reported more deviance after a year of social intervention, compared with children from FF.

Keywords

parental incarceration deviance mentoring social intervention social control social capital

In the face of a growing need for evidence that social intervention can be effective for children with incarcerated parents (CIP; Eddy and Poehlmann 2010; Nation et al. 2003; Parke and Clarke-Stewart 2003), the Big Brothers Big Sisters (BBBS) mentoring program has begun identifying the unique challenges faced by these children (cf. Big Brothers Big Sisters of America [BBBSA] 2011). The one-on-one social intervention model used by BBBS has an empirically demonstrated ability to reduce negative outcomes for other at-risk populations of children, with positive outcomes including reduced drug and alcohol use, improved school attendance and performance, and improvements in a child’s interpersonal relationships (Tolan et al. 2014). The BBBS program achieves these outcomes by pairing children with an active, close, consistent, and long- lasting relationship with a positive adult role model.

The rationale behind this model of social intervention is to give children access to a unique source of social capital. Coleman (1988) argued that human social capital is in part reflected in the differing access people have to one another and the resources made available through those people. Access to disparate groups of people varies. Mentoring programs carefully screen volunteers in hopes of providing at-risk children with a novel source of pro-social human capital, ostensibly steering children to opportunities and influences otherwise unavailable. The term “pro-social” is borrowed from developmental psychology (cf. Baumeister and Bushman 2014) to emphasize the point that sources of social capital differ; not all sources of social capital direct children toward socially accepted norms and expectations. For example, a drug dealer with limited social capital for other job opportunities might have an abundance of social capital making the drug market accessible (Fagan 1989).

Despite a strong foundation in social capital theory, it is uncertain if social intervention strategies such as those found in community-based mentoring programs can successfully mitigate the intergenerational effects of parental incarceration, a situation of risk that has been shown to increase a child’s self-reported levels of sadness (Poehlmann 2005), academic misconduct including cheating (Travis and Waul 2003), and increased likelihood for deviant behavior (Turanovic, Rodriguez, and Pratt 2012).

Furthermore, contemporary findings are in disagreement about the total effect parental incarceration has as a predictor of criminal behavior for CIP, with some studies reporting as much as “60-80% of the intergenerational crime relationship” being accounted for by intergenerational transmission of human capital and behavior from parent to child (Hjalmarsson and Lindquist 2012:553; see also Reed and Reed 2004 who find that CIP are approximately five times more likely to be incarcerated than similar peers). The National Institute of Corrections recently followed up on contemporary claims, adjusting the estimate and reporting that CIP are approximately three times more likely to be involved with the justice system at some point in their own lives, compared with their peers (Conway and Jones 2015).

Debates about total effect are likely to continue; regardless, CIP face unique risks and are in need of empirically grounded social intervention strategies. The present study focused on the efficacy of additional social support/capital in the form of an adult role model for CIP. In particular, analyses examined the ability of a regional mentoring social intervention program to mitigate the effects of parental incarceration on a child’s self-reported levels of sadness, cheating in school, and engagement in deviant behaviors. Two data sources were used, primary data gathered from 173 children involved with a regional branch of the BBBS mentoring program and a survey given to children in the Fragile Families (FF) and Child Wellbeing Study (Reichman et al. 2001). FF data are a nationally representative sample of approximately 5,000 children living in urban settings. The FF sample focuses on at-risk populations of children, including CIP (for more on FF, see Reichman et al. 2001).

The analysis centered on two central hypotheses; the first replicates studies working to examine the negative effects experienced by CIP (cf. Farrington, Coid, and Murray 2009; Haskins 2014; Murray, Farrington, and Sekol 2012):

Hypotheses 1 (H1): Children at risk for intergenerational crime (i.e., CIP) will have an increased likelihood for the negative outcomes identified.

The second hypothesis tests the efficacy of a regional social intervention program for CIP living in the area:

Hypotheses 2 (H2): CIP who are a part of this social intervention program will report decreased levels of sadness, cheating in school, and deviant behavior compared with a nationally representative sample of similarly at-risk children.

Literature Review

Broadly defined, mentoring occurs when three things are in place: (1) a protégé (in this case, a child) is involved with a role model who has greater experience or wisdom, (2) the mentor offers guidance to facilitate the development of the mentee, and (3) there is an emotional bond characterized by a sense of trust (Baber 2005).

Sipe (2002) noted three conditions that are critically important for developing successful social intervention through mentoring. She recommended that mentors go through a rigorous process of (1) screening, (2) orientation and training, and (3) ongoing supervision and support (see also Baber 2005). BBBS is the oldest social intervention program in the United States adopting this model of mentoring, providing mentors to children for more than 100 years. Across the United States, in 2014, BBBS served “more than 170,000 children . . . in need of a positive role model” (BBBSA 2014). BBBS matches one child (called a “Little”) with one screened, trained, and supervised adult mentor (called a “Big”). Volunteer mentors commit to spending at least four hours per month with their mentee, typically spread over two to four visits, engaging in social activities such as going to museums, playing video games, or just “hanging out” together.

To provide evidence that social intervention is working, a program must define what “working” means, essentially asking, “What are the intended outcomes of the social intervention and how can they be measured?” Growing scholarly work focused on the efficacy of social intervention programs has catalyzed a BBBS emphasis on providing empirically tested best practices (Grossman and Tierney 1998; Rhodes 2008). In recent years, BBBS has made efforts to clearly define what a mentoring relationship is and what it provides to children being served by the program. With specific reference to children at risk for deviance, BBBS (2011) has stated,

JUVENILE JUSTICE—January of 2011, we launched a Juvenile Justice Initiative that focuses on . . . delinquency prevention through research and evaluation . . . We believe mentoring can prevent children who have come into contact with the juvenile justice system from having further involvement with the system [by matching] children with mentors that offer a positive influence and help children achieve their goals by instilling higher aspirations, greater confidence and modeling positive adult-child relationships. We hold ourselves accountable for positive measurable youth outcomes. (P. 8)

However, notwithstanding the good intentions, empirical work supporting positive outcomes for CIP is lacking (Nakkula and Harris 2005; Tolan et al. 2013); at present, reviews of the literature are full of statements that read, “little is known about the impact that [parental incarceration] may have on children, families, and society at large” (cf. Bongers et al. 2004; Loeber and Burke 2011; Naudeau 2010:47). More work is needed to determine if social intervention programs can mitigate the risk factors that children face in these homes, and in particular if social intervention can successfully mitigate the likelihood of deviance across the life-course (Thornberry 2009).

Despite a lack of empirical evidence for this subpopulation of at-risk children, the overall support for mentoring programs has grown steadily (Britner et al. 2006) with more than two million adults serving as a volunteer mentor to children in the United States, in BBBS and similar mentoring programs (DuBois et al. 2011). The popularity of these programs stems from increasing knowledge that a child’s relationships with nonparental adults plays a significant role in promoting overall positive adolescent development (Coakley, Shears, and Randolph 2014; DuBois and Karcher 2005). The literature on attachment, social learning, as well as self- and social control, all highlight the importance of connectedness to nonparental adults for the positive development and health of children (DuBois and Silverthorn 2005). Children who find adults other than their parents to act as a mentor and guide them through the adolescent years have better outcomes (Grossman and Bulle 2006).

Jekielek, Moore, and Hair (2002) as well as DuBois and colleagues (2011) conducted reviews of the literature on child mentoring. They found that children involved with mentoring experience many positive outcomes including a reduction in school absences, more positive school attitudes and behavior, higher college participation, less drug and alcohol use (especially among at-risk minority children), a decreased likelihood for violent behavior, and a decreased likelihood of criminal sentencing. What is not clear, across the literature (cf. Farruggia et al. 2011), is the strength of the effect mentoring has on the negative outcomes facing the children in these programs (see Tolan et al. 2013 for a comprehensive review of the empirical literature).

An additional challenge faced by social intervention is attrition. More than half of all BBBS mentoring relationships end in the first few months (Rhodes 2002; Serido, Borden, and Wiggs 2014; Spencer 2006). Research finds that, when mentoring falls short, social intervention can do more harm than good (Grossman et al. 2012). Evidence-based best practices suggest that children benefit most with at minimum six months of social intervention and, for greatest impact, at least a full year or more of the mentoring relationship (Grossman 2005). For social intervention through mentoring to have a lasting impact on the life of a child, programs need to support relationships of trust lasting at least six months and ideally one year or more (Miller et al. 2013; Roberts et al. 2004).

Mentoring Children at Risk for Intergenerational Crime

Establishing a trusting relationship with CIP is particularly challenging (Ellis et al. 2012). CIP live in a daily reality where trusting relationships are scarce, absentee parents are common, and crime and incarceration are often normative (Besemer and Farrington 2012; Blumstein and Cohen 1987; Campbell et al. 2006; Wakefield and Wildeman 2011). Even after a parent is released from jail or prison, a child continues to navigate the frequent presence of criminal justice personnel, electronic monitoring, and the stigma involved with criminal behavior (Braman 2004; Braman and Wood 2003; Turney, Schnittker, and Wildeman 2012).

In most cases, incarceration adds to the burden of struggling families (Hagan and Foster 2006; Jones, Cauffman, and Piquero 2007; Muller and Wildeman 2013). Wildeman and Western (2010) pointed out families that are impacted by incarceration often deal with difficult life circumstances beyond familial criminality, calling these families “fragile.” CIP often deal with issues of poverty (D. J. Smith 2005; Wacquant 2001), uncertain and irregular housing arrangements (Harper and McLanahan 2004; Johnston 2006), family histories of substance abuse (Hagan, McCarthy, and Foster 2002; Rivers and Anwyl 2000), single parented homes (Harper and McLanahan 2004; Martone 2005; McLanahan and Schwartz 2002), and family histories of mental, physical, and sexual abuse (Ammerman et al. 1999).

It is important to note, even in situations of extreme risk, the probability that parental incarceration will increase the likelihood of intergenerational crime and incarceration is not certain. Children can be surprisingly resilient and the removal of an abusive or negligent parent can be a net positive in a child’s life (Sampson 2011). The negative effects a child experiences as a result of incarceration are related to a number of complicating conditions such as the parent-child separation, the trauma of witnessing the crime and arrest, proceeding periods of incarceration/separation, general instability, poverty, and the adequacy of care in the home. Furthermore, the degree to which a child is affected by parental incarceration may be influenced by a number of variables, including the age of the child, level of maturity and understanding of the crime and incarceration, the length of separation, the level of disruption, the frequency of separation experiences, and the availability of family or community support. Taking all of the risk-inducing factors into consideration, the potential effects of parental incarceration are based on the following four factors: (1) the child’s developmental level (Moffitt 1993; Vazsonyi and Huang 2010), (2) the child’s awareness of the crime and incarceration (Skogan 1988), (3) the cognitive and emotional state of the child (de Haan and Loader 2002; Giordano, Schroeder, and Cernkovich 2007; Hipp et al. 2004), and (4) available social support (Travis and Waul 2003).

In a comprehensive meta-analysis focused only on this population of children, children with an increased risk for deviant behavior, Tolan and colleagues (2013) found only small amounts of empirical evidence supporting positive outcomes through social intervention. They argue that the current state of empirical work on mentoring provides “little understanding” of how social interventions produce positive results (Tolan et al. 2013:21). The point is, despite indications that mentoring can have a positive impact on other at-risk populations of children, when it comes to the population of CIP, little is known about how social intervention might mitigate a child’s likelihood for deviant behavior (for more on social intervention for juvenile deviance, see Jarjoura et al. 2013; E. I. Johnson and Easterling 2012; Kim, Merlo, and Benekos 2013).

In criminological terms, the positive outcomes anticipated from social intervention are rooted in sociogenic processes. Laub and Sampson’s (1993) theory of social control focused on the social conditions that mentoring programs hope to create. Laub and Sampson (2001) presented the concept of the “turning point” as the foundational expression of how social influence can either steer people away from or toward deviant behaviors (Laub and Sampson 2003). Life events transition people into distinct settings. For example, turning points transition people from family care into schooling, from changes in friend groups, to the role obligations of getting married. Events occur in the context of larger structural and cultural conditions that present new sets of norms and expectations to negotiate. As people learn to navigate and interact with new groups of people, opportunities for, and outcomes related to, behavior also change (Stryker and Craft 1982; Turner 1962). This is not a new idea. The notion of a turning point taps into the processes of socialization (Gottfredson and Hirschi 1990), social learning (Akers and Jensen 2006; Sutherland 1947), and role-based theories of criminal behavior (Matsueda and Heimer 1997).

A turning point at once recognizes the interactive nature of mentoring as well as the human agency involved in such a pairing. Placing a child in a deliberately constructed relationship with the “Little” role to occupy recognizes the power of socialization experiences to influence behavioral outcomes. Combining the individual-level interactions of socialization with the macro/cultural norms and expectations providing context for interaction, social-control theory and turning points argue that pro-social behavior is directly proportional to a person having access to a pro-social environment (Laub and Sampson 2002). A mentor provides a child with novel access to social roles and social capital that might otherwise be out of reach (Coleman 1988; Rose and Clear 1998).

Accordingly, current cross-cultural research on the effects of social capital finds that access to pro-social capital in the form of norms and interaction networks predicts not only perceptions of crime (Ferguson and Mindel 2007), but the commission of property (Paolo, Montolio, and Vanin 2009) as well as violent crimes (Kennedy et al. 1998). The hope is that, over time, as children in the BBBS mentoring program develop a sense of identity around the Little role, they will begin to internalize the pro-social capital they have access to through the relationship with their mentor (Mead 1934). Increased exposure to pro-social capital (i.e., norms, expectations, identities) increases pro-social attitudes, including self-control (Ray et al. 2013), thereby decreasing deviant behaviors. BBBS posits that the positive social influence of a mentor will provide children with a pro-social role to model, alternate expectations to consider, and will potentially represent a “turning point” to a life-course trajectory outside of the life conditions that might otherwise predict a life of crime and other antisocial outcomes (Fusco 2012).

Hypotheses

The present study focused explicitly on the efficacy of additional social support/capital in the form of an adult role model for children with an incarcerated parent. Another contribution of the current study is the source of data. Data in the current study were gathered directly from impacted children. Few studies examine intergenerational risk from the perspective of the affected children (T. Johnson 2012). Two hypotheses were tested:

Hypothesis 1 (H1): Children impacted by parental incarceration will report more deviant behavior than nonimpacted peers.

Hypothesis 2 (H2): After one full year of mentoring, children impacted by parental incarceration in BBBS will report lower negative outcomes, compared with a national sample of children impacted by parental incarceration, children not being mentored.

Outcomes tested in this study were (1) deviance, (2) cheating in school, and (3) self-reported levels of sadness.

Analytic Strategy

Two data sources were used in this study to determine if the social support services provided through mentoring result in the anticipated turning point. The first data source came from primary data gathered from 173 children in the BBBS mentoring program in Central Indiana (BBBSCI). BBBSCI serves the Marion county region; a 99 percent urban environment of approximately one million people surrounding Indianapolis, Indiana. Panel data were collected in three waves (i.e., T1 = baseline collected prior to treatment, T2 = six months of social intervention, and T3 = one year of social intervention). The second data source came from the FF and Child Wellbeing Study. FF data are a nationally representative sample of at-risk children living in urban settings. FF data are ideal given the focus of the data collection on FF, including study measures relating to familial crime and incarceration (Wildeman and Western 2010:158). FF data include a cohort of approximately 5,000 children who were born between 1998 and 2000. It should be emphasized that FF data come from a nine-year follow-up survey given to children and are a cross-sectional assessment of children at this point in time. During analyses that did not use the probability matching found in propensity scores, FF data were treated as a representative national “average” of at-risk children in urban settings across the United States. This average applies to three FF dependent variables not modeled longitudinally.

Several limitations exist in the primary data. It should be noted that the results of this study are an analysis of a particular BBBS implementation with limited information about the programming beyond the assurance that all national BBBS standards were being met. Indeed, there may be unmeasured mentoring relationship factors that explain why this particular BBBS achieves the outcomes reported. For example, it is possible that hours spent or quality of pro-social capital of mentoring at BBBSCI vary from the BBBS national average. Too, based on limitations in the FF data, parental incarceration in this study collapses the mother and/or father into one category, with no distinction made. Studies have demonstrated differing effects of maternal versus paternal incarceration for CIP (Geller and Franklin 2014; Huebner and Gustafson 2007; Western and McLanahan 2000).

Furthermore, attrition in the primary data from baseline to the third wave of data resulted in about 50 percent of the original BBBSCI sample leaving the study. Attrition challenges most panel data analyses and, in particular, studies that examine at-risk children involved with mentoring (Bell and Bradley 2013; DuBois et al. 1992). Based on missingness, data were analyzed using regression models under maximum likelihood estimation in a mixed-model framework, rather than ordinary least squares; this approach has a demonstrated robustness to missingness when data are monotone, or missing at random (Rabe-Hesketh and Skrondal 2012:278–82). A missing data analysis was run finding that missingness was monotone; that is, children left the program and did not return to provide data at later waves. To further account for any bias that may have come from a misspecified model, the sandwich estimator for standard errors with the command vce(robust) was used in Stata (Rogers 1993).

Measures of Dependent and Independent Variables

Each dataset contains a scaled measure of deviant behavior, an identical measure of self-reported cheating in school recorded as Yes/No, and an identical measure of self-reported sadness recorded as a frequency. The deviance scales represent different underlying dimensions based on each survey (see tables below for more); for this reason, the scales were standardized using Bartlett Factor Scale Scores with a mean of 0 and a standard deviation of 1 (for more on factor scores, see DiStefano, Zhu, and Mîndrilă 2009; Reise, Waller, and Comrey 2000; Russell 2002). The deviant behavior items in the FF data were all coded Yes/No and required cleaning prior to component analysis. Responses coded “Not in wave,” “Missing,” “Don’t know,” and “Refuse” were all coded as missing. Restructuring of the FF data resulted in the principal components analysis (PCA) statistics displayed in Table 1 (using Promax rotation with a coefficient of alienation set to .25).¹ The Kaiser-Meyer-Olkin (KMO) & Bartlett test of sampling adequacy was .825 (p = .001) indicating adequate sampling after recoding.

Table 1.

Principal Components Analysis of Deviant Behavior Scale in FF Data.

Deviant Behavior Items	Factor analysis components
Deviant Behavior Items	1	2	3	4
Purposely damaged property	.386			.428
Taken or stolen something				.709
Taken money at home				.755
Been in a fight	.767
Hurt animal on purpose			.492	.263
Trespassed			.496
Run away from home			.569
Skipped school			.711
Drank alcohol		.391
Smoked marijuana		.613
Used tobacco		.785
Kicked out of school	.797
Graffiti	.339		.369
Arson		.540
Avoid paying for things	.268
Thrown rocks or bottles

Note. Rotation converged in 6 iterations. The highest correlation between components was 1 and 4 at .330. FF = Fragile Families.

Restructuring of data reduced component fit based on the premise that there was nothing meaningful to interpret in the missing values; and for the purposes of this study, missing data were not analyzed beyond their missingness (see Allison 2002; Schafer and Graham 2002 for more on the treatment of missing data values).

Four items from Table 1 dealing with property damage and stealing were used in a Confirmatory Factor Analysis (CFA) and create the deviant behavior scale in the FF data. CFA showed that these measures fit the data well (e.g., Root Mean Square Error of Approximation [RMSEA] = .035) despite moderate loadings (average CFA loading of .45) and Cronbach’s alpha (α = .547). These items were used during analysis for three reasons: (1) they loaded together during component analysis, (2) they partially overlapped with the BBBSCI items, and (3) they represent actions that are criminal, and because the likelihood of intergenerational crime is the focus of this research, they proved most sensible as a comparison of deviant behavior bent toward criminality (R. C. Johnson 2009).

The deviant behavior scale in the BBBSCI data came from four measures. The first item asked the frequency children had stolen things (loading .744), done things against the law (loading .796), been stopped and questioned by the police (loading .816), and been arrested (loading .909). All items were coded 1–8 with 1 = Never and 8 = Every Day. Component analysis of the deviant behavior items in the BBBSCI data returned a KMO-Bartlett statistic of .763 (p = .001) and a Cronbach’s alpha of .876. BBBSCI deviant behavior items hang together very well as a scaled measure of deviance.² Table 2 presents the mean and standard deviation for each of the outcome variables analyzed and also includes information organized by children who are impacted by parental incarceration. In both sample populations, CIP had higher negative outcomes, in some cases, double the standard deviation. The differences were smaller, or reversed, in the case of BBBSCI data for academic dishonesty (i.e., at BBBSCI baseline, children not impacted by parental incarceration self-reported more cheating).

Table 2.

Means and Standard Deviations of Dependent Variables by CIP and Treatment.

Variable name	BBBSCI children						FF children
	Baseline		6 months		1 year		National “average”
	No Impact	Yes Impact	No Impact	Yes Impact	No Impact	Yes Impact	No Impact	Yes Impact
Deviance	−.27 (.61)	.16 (1.42)	−.24 (.06)	.13 (1.27)	−.17 (.39)	.13 (1.37)	−.11 (1.21)	.11 (1.48)
Cheating	.31 (.47)	.24 (.43)	.14 (.36)	.26 (.47)	.29 (.46)	.19 (.37)	.05 (.22)	.08 (.27)
Sadness	1.31 (1.07)	1.45 (1.15)	.57 (.51)	.71 (.46)	.81 (.40)	.70 (.47)	.61 (.88)	.79 (1.0)
N	62	108	21	38	21	27	1,606	1,662

Note. BBBSCI = treatment; FF = control. CIP = children with incarcerated parent; BBBSCI = BBBS mentoring program in Central Indiana; BBBS = Big Brothers Big Sisters; FF = Fragile Families.

Descriptive statistics found in Table 2 suggest that CIP are more deviant and sad than a comparison group of peers, and that while negative outcomes decrease overall, the BBBSCI children consistently experience more negative outcomes than FF children. The causal nature of these relationships was explored in greater detail during subsequent analyses.

The measure for CIP was constructed as a binary variable 0 = No Impact and 1 = Yes Impact. In both samples, coding came from a combination of measures asking if either the child’s mother or father (or both) had ever been incarcerated during a child’s life. In the BBBSCI data, coding sources included agency data records, parent reports, mentor reports, or a child’s self-reported awareness of parental incarceration at each wave of data collection. Any response that confirmed parental incarceration was coded a 1 at that wave and later waves of data collection. In the FF data, both mother’s reports and father’s reports were used to construct the parental incarceration Impact variable as well as “constructed” variables created by the FF researchers regarding parental incarceration Impact. Any response that confirmed parental incarceration during a child’s life was coded 1.

The measure of parental incarceration in the FF data was drawn from multiple sources, none of which include a measure of child awareness related to the crime and incarceration of the parent, as was present in the BBBSCI data. As the FF data lack this level of sensitivity to the effect of parental incarceration, the analyses presented here likely missed some of the child awareness effects of parental incarceration in the FF data; and because it is impossible to know the level of awareness of the children in the FF data, it is uncertain if the bias was positive or negative. This is a known sensitivity, or measurement error, limitation of the FF data.

Controls, including controls for multivariate matching purposes, included age as a continuous variable of actual age; age has a strong positive relationship to deviance (Farrington 1986). Based on a demonstrated linkage between socioeconomic status and the likelihood for deviant behavior (see Giordano et al. 2011; Nagin and Tremblay 1999; 2001), a child receiving free school lunch (binary: yes/no), the primary caregiver’s employment status (categorical: employed, unemployed, and other), as well as household income (categorical: arranged in standard categories) were all controlled for. Finally, because both gender (see Lauritsen, Heimer, and Lynch 2009) and race (see Western 2006) are known to be related to deviant behavior, both of these controls were also included. Table 3 displays the controls used.

Table 3.

Means and Standard Deviations of Independent Variables by BBBSCI Baseline and CIP.

Variable name	BBBSCI children		FF children
	No CIP	Yes CIP	No CIP	Yes CIP
	$\bar{x}$ (SD)	$\bar{x}$ (SD)	$\bar{x}$ (SD)	$\bar{x}$ (SD)
Age	12.41 (2.08)	12.48 (1.90)	9.28 (0.563)	9.27 (0.697)
Gender
Female	.55	.55	.47	.46
Male	.45	.45	.53	.53
Income	3.19 (2.24)	3.50 (2.22)	5.07 (2.54)	3.67 (2.49)
Employed
Yes	.41	.46	.21	.18
No	.56	.54	.12	.21
Other	.3	.0	.67	.62
Race
Black	.52	.66	.40	.59
White	.36	.21	.27	.16
Other	.13	.13	.33	.24
School Lunch	0.88 (0.33)	0.84 (0.36)	0.58 (0.49)	0.84 (0.37)
N	64	109	1,861	1,874

Note. $\bar{x}$ = Mean or Proportion. BBBSCI = BBBS mentoring program in Central Indiana; CIP = children with incarcerated parent; FF = Fragile Families; BBBS = Big Brothers Big Sisters.

Method

Three methods were used to examine the relationship between parental incarceration and the effectiveness of social intervention for reducing negative outcomes. First, before determining if the BBBS social intervention reduced deviant behaviors, the a priori relationship that social intervention programs assume represented in H1 was modeled, that is, CIP are more likely to be deviant than nonimpacted peers. Using only the BBBSCI data—because it is longitudinal—two-level random intercept and random coefficient models for each outcome were estimated. This analysis provided regression estimates for the child as well as estimates for grouped cases by parental incarceration accounting for any variation occurring in the outcome level as a result of correlated error. Second, using both BBBSCI and FF data, I ran fixed-effects repeated measures analysis of variance (ANOVA) models treating FF as a national average comparison indicating how children in BBBSCI fare across a year of social intervention compared with the nationally representative “average” of FF children. Finally, to determine if social intervention is working, I compared the FF data to the BBBSCI data using propensity score matching (PSM; Rosenbaum and Rubin 1983, 1984). PSM was used for three reasons: (1) PSM has become a popular method for making quasi-experiential comparisons (Becker and Ichino 2002; Dehejia and Wahba 2002; Richey and Ikeda 2009), (2) PSM methods of balancing data are attractive especially to account for the age differences in these data, and (3) PSM was chosen because it replicates a recent report issued by BBBS finding positive outcomes of mentoring as a social intervention comparing children currently being mentored with children on a waiting list for a mentor (Valentino and Wheeler 2013):

To draw conclusive evidence . . . we used a technique known as Propensity Score Matching to match each treatment youth with a youth from the comparison group who was similar on these potentially influential characteristics. (P. 20)

PSM analysis was run using a logit probability model. Scores were created using the controls mentioned above and were based on prior theoretical and empirical work dealing with at-risk children (Heckman, Ichimura, and Todd 1998; Latimer 2001; Rosenbaum and Rubin 1985). PSM was used to make the final comparison between the BBBSCI children and the FF children.

The limited data available on CIP and the difficulty of establishing a traditional experiment along with the ethical concerns of delaying children’s access to social services (for more on this point, see Blechman and Bopp 2005:461) make PSM an attractive substitute to randomization. Propensity scores were used to balance data so that the distribution of pretreatment covariates was similar but independent (unconfoundedness) of the treatment effect (Caliendo and Kopeinig 2008; Dehejia and Wahba 1999, 2002). In other words, a child’s place in the treatment group was determined by carefully selected observed variables, and based on these conditions, assignment to treatment was effectively randomized. Once balancing was satisfied, the propensity score had the same distribution properties independent of the treatment. During analysis, balancing was achieved for each of the PSM analyses run. The parameter estimate called the “average treatment effect on the treated” (ATT) was then used for comparison between cases.

All PSM estimates were obtained in Stata 13 using pscore (Becker and Ichino 2002). ATT was based on a range of propensity scores determined by common support. Counterpart observations are differentiated from controls using two matching methods. The first is nearest neighbor matching with replacement, comparing children from each sample who have the closest propensity score and allowing untreated FF children to be used more than once as a match. Running a PSM with common support and replacement adjusts for the difference in control conditions in each sample, for example, age. The second matching method, kernel matching, is a nonparametric matching based on a weighted average of all of the FF children (for more on matching methods, see J. A. Smith and Todd 2001).

Results

Results from the Random Intercept and Random Coefficient Models

Table 4 presents the results of multivariate examination using a random intercept model of deviant behavior, and Table 5 presents a similar analysis using a random coefficient model testing nesting of random effects.

Table 4.

Multilevel Regression of Deviant Behavior, Random Intercept Model Estimates.

Variables Modeled	Model 1A		Model 1B		Model 1C
Variables Modeled	β	SE	β	SE	β	SE
Children with Incarcerated Parent Impact	.34*	.14	.38***	.12	.39***	.11
Age	.10^†	.05	.08^†	.04	.08^†	.04
Gender	−.14	.15
Employment
Unemployed	.18	.18
Other	−.15	.16
Income	−.03	.05
Race
Black	.24*	.12	.09	.09
Other	.54	.51	.42	.39
School Lunch	−.46	.36
Intercept	−1.00	.70	−1.31**	.52	−1.20**	.11
Random Effect	1.13	.80	1.10	21.34	1.11	.211

†

p ≤ .10. *p ≤ .05. **p ≤ .01. ***p ≤ .001.

Table 5.

Multilevel Regression of Deviant Behavior, Random Coefficient Model Estimates.

Variables Modeled	Model 2A		Model 2B		Model 2C
Variables Modeled	β	SE	β	SE	β	SE
Children with Incarcerated Parent Impact	.41*	.19	.43***	.14	.44***	.14
Age	.08^†	.04	.07^†	.04
Gender	−.03	.18
Employment
Unemployed	.22	.20
Other	−.11	.33
Income	−.04	.05
Race
Black	.12	.20
Other	.22	.30
School Lunch	−.38	.32
Intercept	−.81	.68	−1.05**	.42	−2.45***	.05
Random Effect	.86	.00	.79	.18	.79	0.0

†

p ≤ .10. *p ≤ .05. **p ≤ .01. ***p ≤ .001.

Model 1A included all of the controls. Model 1B dropped nonsignificant results after the first run. Model 1C included only the variables that were significant after the second run.

Results suggest that the best predictor variable, over the year of social intervention, for a child’s involvement in deviant behavior was whether or not they were impacted by an incarcerated parent. Across each model tested, CIP were .3 to .4 standard deviations higher on the deviance scale than children who are not impacted by parental incarceration.

Results of the random coefficient model support those found in the random intercept model. The likelihood-ratio test after estimation of the nested coefficient model within the random intercept model revealed that the mixed models were not nested, and because children’s rate of deviant behavior is essentially the same across mixed models, either model may be accepted.

Models of the outcome for cheating and sadness did not find significant effects for parental incarceration and are not reported here. There were, however, significant effects for controls finding that children from Black (β = 1.59, p = .01) and Other (β = 2.60, p = .001) racial categories were more likely to cheat. This finding is consistent with the literature on academic dishonesty (McCabe and Trevino 1997; Morris 2012).

Results of the mixed-effects regression models found that children who have an incarcerated parent are significantly more likely to commit deviant acts, supporting H1. The predictor for parental incarceration Impact was predictive across a year of social intervention even when controlling for variables such as age, gender, and race, variables that have been used successfully to predict criminal activity in numerous empirical settings.

Results from Fixed-effects ANOVA Models

Table 6 displays the results of the fixed-effects repeated measures ANOVA with BBBSCI children compared with the FF “average.” Following the results of the mixed models reported in Tables 4 and 5 above, I only report the results for children from both samples that were impacted by parental incarceration (modeled with controls). In each of the fixed-effects models, the assumption of sphericity was violated (the assumption that the variance between all predictors is equal, Mauchly’s Test = .406[2df], p = .001) and the test of within-subjects change over time was based on the Greenhouse-Geisser correction. The Greenhouse-Geisser correction of the mean scores for each outcome was statistically significantly different between time points by parental incarceration risk (see within-subjects statistics reported in Table 6). The BBBS program had a statistically significant effect on the children who were impacted by parental incarceration over time. However, how did this effect compare with the FF average?

Table 6.

Fixed-effects ANOVA of FF and BBBS Data across Three Waves of BBBS Data.

ANOVA Models	Sum of squares	df	M ²	F	Significance
Deviance
Model 3A
Between Groups	.948	1	.948	.19	.668
Within Subjects	.168	1.05	.16	65.37	.001
Cheat at school
Model 3B
Between Groups	1.48	1	1.48	8.28	.004
Within Subjects	.024	1.43	.016	4.83	.017
Sadness
Model 3C
Between Groups	27.54	1	27.54	10.35	.001
Within Subjects	9.80	1.26	7.81	1,167.36	.001
Figure 1. Deviance^a	Figure 2. Cheating^a		Figure 3. Sadness^a

Note. ANOVA = analysis of variance; FF = Fragile Families; BBBS = Big Brothers Big Sisters; BBBSCI = BBBS mentoring program in Central Indiana.

Control = FF Children – Black Line/Constant, Treatment = BBBSCI Children – Red Line/Changing.

Looking at Figure 1 in Table 6, I found a bouncing trajectory for the BBBSCI children’s self-reported deviance over the year of social intervention. This resulted in a nonsignificant amount of change compared with the FF average. Over a year of social intervention, there was enough negative and then positive change to eliminate a significant effect between subject groups for self-reported deviance. The net amount of change equaled out over the year of social intervention, with the mean value for BBBSCI Deviance at baseline nearly identical to the FF national “average.” Children start the program with the same amount of self-reported deviant behavior (see Figure 1 in Table 6). After six months of social intervention, the likelihood for deviant behavior drops significantly, but at one year of social intervention, Deviance climbs above the FF data (time by treatment within-subjects effect F = 65.37, p = .001), and based on this, there was no statistically significant difference between groups (F = .19, p = .668). Further analysis of these differences at six months and one year were analyzed in PSM models.

Both Cheating and Sadness had statistically significant within- and between-subject effects. The self-reported instance of cheating went down at the six-month interviews, but at one year, rose slightly. Even after the increase at one year, the level of reported cheating stayed well below the amount reported at the baseline measurement. However, as Figure 2 in Table 6 indicates, at each time point, self-reported cheating among BBBSCI children was higher than the amount reported by FF children.

In terms of Sadness, BBBSCI children experience a consistent decline in the frequency of self-reported sadness over the year of social intervention. However, when treating the model as fixed, at each time point, the levels of sadness experienced by BBBSCI children were higher than those reported by FF children (see Figure 3 in Table 6). It appears that social intervention is working over the year of social intervention when it comes to cheating in school and the frequency of sadness these children experience. However, the Deviance outcome cautions against a singularly positive interpretation of the findings. For this reason, BBBSCI data were balanced against FF data and compared in a quasi-experimental PSM framework. The PSM allows for a more conclusive argument addressing H2. Did BBBSCI produce positive outcomes after a year of social intervention had taken place?

Results from Propensity Score Models

Table 7 displays the results of the first PSM analysis. After six months of social intervention, the ATT resulted in BBBSCI children reporting less deviant behavior compared with FF children. However, after a full year of social intervention, the ATT was positive and like the fixed-effects model was not significant.

Table 7.

Mentoring Effect on Deviance for Children Impacted by Incarceration.

Panel Waves	N _T	N _C	ATT	SE	t Value
Time 1 = Baseline, child at time of match assignment
See Table 6 ANOVA results.
Time 2 = Six month anniversary of match
Nearest Neighbor	64	273	−.395	.257	−1.536
Kernel	64	324	−.427	.463	−.921
Time 3 = One year anniversary of match
Nearest Neighbor	64	273	.097	.181	.536
Kernel	64	324	.066	.150	.438

Note. N_T = sample size of treatment; N_C = sample size of control; ATT = average treatment effect on the treated; ANOVA = analysis of variance.

Small t values for Deviance were consistent with the ANOVA results. Taken as a whole, both sets of analyses indicate that social intervention did not produce the expected positive outcomes, with the mentoring effect decreasing deviance at six months of social intervention, but after one year, rising to the point that there was no significant effect from mentoring.

Table 8 displays the results of the PSM analysis comparing BBBSCI children with FF children on Cheating. Nearest neighbor matching for the ATT supported the results already reported. Using a nonparametric comparison based on a weighted average through kernel matching also found an increased likelihood for self-reported cheating among BBBSCI children, finding that over the year of social intervention, the ATT increased. Taken with the fixed-effects results, these findings indicate that BBBSCI children were consistently more likely to report involvement in academic dishonesty compared with FF children.

Table 8.

Mentoring Effect on Cheating for Children Impacted by Incarceration.

Panel Waves	N _T	N _C	ATT	SE	t Value
Time 1 = Baseline, child at time of match assignment
See Table 6 ANOVA results.
Time 2 = Six month anniversary of match
Nearest Neighbor	64	275	.089	.053	1.665
Kernel	64	324	.102	.098	1.046
Time 3 = One year anniversary of match
Nearest Neighbor	64	275	.148	.058	2.526
Kernel	64	324	.161	.107	1.509

Note. N_T = sample size of treatment; N_C = sample size of control; ATT = average treatment effect on the treated; ANOVA = analysis of variance.

Table 9 displays the results of the PSM analysis on Sadness that compared BBBSCI children with FF children. At six months and one year of social intervention, the ATT resulted in BBBSCI children reporting less sadness compared with FF children.

Table 9.

Mentoring Effect on Sadness for Children Impacted by Incarceration.

Panel Waves	N _T	N _C	ATT	SE	t Value
Time 1 = Baseline, child at time of match assignment
See Table 6 ANOVA results.
Time 2 = Six month anniversary of match
Nearest Neighbor	64	274	−.165	.158	−1.042
Kernel	64	324	−.173	.138	−1.248
Time 3 = One year anniversary of match
Nearest Neighbor	64	274	−.165	.163	−1.011
Kernel	64	324	−.173	.144	−1.201

Note. N_T = sample size of treatment; N_C = sample size of control; ATT = average treatment effect on the treated; ANOVA = analysis of variance.

PSM results for Sadness find a positive outcome from the effect of mentoring on children. The fixed-effect model reported in Table 6 found consistently declining levels of sadness among mentored children. Using the quasi-experimental methods of PSM, results find that BBBSCI children report less sadness than FF children as a result of social intervention. Despite this positive outcome, results of the PSM analyses were mixed.

Discussion

The findings reported in this study support previous research and the hypothesis that CIP report more deviant behavior (Murray et al. 2014). However, the hypothesized relationship between mentoring and positive outcomes was not fully supported. I hypothesized that after one full year of mentoring, CIP in BBBS would report less deviance, cheating, and sadness compared with a nationally representative sample of similar children. When impacted children from FF were compared with impacted children from BBBS, I found that mentoring did not result in all of the expected positive outcomes, that is, results dipping below the “average” represented by the FF children. To the contrary, children in BBBS experienced increases in both Deviance and Cheating after a year of social intervention.

The mixed models suggested that a significant amount of deviant behavior comes from the effect of parental incarceration, but why a jump in cheating specifically? The increased likelihood for academic dishonesty may be explained by the fact that BBBS spends a lot of time focusing on a child’s performance in school (Rhodes, Grossman, and Resch 2000). BBBS children are oriented toward academic pursuits, spending mentoring time on homework and thinking about how schoolwork impacts their future opportunities. This focus on academics potentially has the unintended effect of leading children who are struggling to innovate an avenue to success (Merton 1938), believing cheating is one way to achieve good grades and increase their life opportunities (Bernardi et al. 2004; Davis et al. 1992; for more on this possibility, see McCabe and Katz 2009; T. R. Smith 2004). This antecedent “cause” for the increased likelihood of academic dishonesty among the BBBS children is beyond the scope of this study but provides an avenue for further research on the long-term effects of social intervention and academic integrity.

The bouncing trajectory may also be explained by a “shock effect” experienced as children join the mentoring program. This would explain the drop from baseline to six months of social intervention. With respect to Deviance and Cheating, as the social intervention relationship nears a close, the initial effect diminishes. Mentoring programs will need to find ways to continue the initial positive gains found in the first six months of mentoring. One suggestion might be to support continued contact between the Big and Little. With the increasing ease and proliferation of electronic communication, one possibility is a pen-pal program, requiring Bigs to commit to continued and consistent contact with their Little for two to five years beyond mentoring.

Looking at the results of mentoring on Deviance and Cheating, it appears that compared with the FF children, social intervention is making things worse for CIP (increasing the likelihood for these outcomes). However, this interpretation of the findings should be treated with caution. One possible explanation for the findings that falls outside of the scope of this study relates to the BBBS baseline numbers. Both Cheating and Sadness were higher at baseline for BBBS children, with Deviance essentially the same. This alone may account for the overall “lack of success” of BBBS when compared with the FF children. This could indicate that the urban setting and conditions of this study are particularly challenging for CIP. The city where this study took place has relatively high crime rates (the city crime index for 2013 was 647.8 while the national average for that same year was 297.1³), which exposes more children to criminal environments and role models. In addition, children from the FF study are on average three years younger than children in the BBBS sample. This age difference likely affects the multilevel models, even when age is controlled for, as was done here.

Does social intervention act as a turning point for these children? Based on the final comparison using propensity matching with replacement (and the assumption of balancing satisfied) when CIP in the local BBBS program were compared with a nationally representative sample, on average, the BBBS children report more cheating and deviance, suggesting that social intervention is not working for this population of children. However, despite these negative outcomes, findings indicate that children receiving social intervention are happier than the comparison group of FF children. Is it working? In some ways, no; but, in others, there is evidence supporting social intervention as a turning point.

Limitations and Suggestions

The outcomes measured are not generalizable to the effect of mentoring across the BBBS social intervention network. These findings may reflect the efficacy of the program in the study setting. The effect of the BBBS program as a whole for CIP is largely unknown (Bruster and Foreman 2012; Jarjoura et al. 2013; Miller et al. 2013). Nonetheless, these findings are useful as an indicator of the efficacy of social intervention under the condition of the study setting. The expected positive outcomes of mentoring did not fully occur for CIP, indicating that programs serving this population of children may need to adjust how they meet their unique needs. If the BBBS social intervention is to provide a turning point for CIP, it may be necessary to restructure elements of the program. One change could be an increased length of mentoring commitment. More time in the program provides at-risk children with more opportunities to socialize with their mentor, thus, increasing the probability that this relationship will act as a turning point in the life of the child (Laub and Sampson 1993).

The most serious limitation is the lack of a generalizable mentoring effect. Research including a large representative sampling of BBBS would be able to make more precise claims. Future research would also benefit from mixed methods, augmenting the limitations of self-report measures relied upon in this study. Expanding the BBBS data is a worthy, but time-consuming and expensive undertaking, which made this level of generality impossible for the current project. Another limitation is the lack of longitudinal data from the FF study. Comparisons with a single cross-sectional wave of data, though nationally representative, should be interpreted with the knowledge that the BBBSCI data were being compared with a cross-section of FF data. The FF children provide a snapshot of children across the United States at a single point in time. A better measure of the changing likelihood for deviant behavior, cheating in school, and sadness would be to follow these children and survey them as they mature. As another researcher making use of these data notes, despite these limitations, the FF data are “the strongest data upon which to make causal claims about the relationship between [parental] incarceration and children’s” risk for intergenerational deviance (Wildeman 2010:308). The fixed-effect model provided a useful “average” comparison against my program data whereas the PSM models provided the quasi-experimental comparisons that determined social intervention at this site is not reducing deviance and cheating, but is increasing happiness for CIP.

In summary, the results of this study support the first hypothesis but do not fully support the second. In support of H1, results find that CIP report more involvement in deviant behavior compared with peers who are not impacted by parental incarceration. Mixed results related to H2 find that after a year of social intervention, mentoring does not reduce deviance and increases academic dishonesty. However, children involved with mentoring are happier than FF peers. Multiple modeling strategies and three different outcome variables were used demonstrating the robustness of these findings. Results suggest that the current implementation of BBBS social intervention is only partially successful at mitigating the negative effects of parental incarceration (see Phillips and O’Brien 2012 for more on the unique challenges and needs of children at risk for intergenerational criminal outcomes). It is recommended that the BBBS program evaluate its future plans for providing social intervention services addressing the unique needs of this population of at-risk children.

Additional Resources

For more information about Big Brothers Big Sisters visit the following Web site: http://www.bbbs.org/. For more information about the Fragile Families and Child Wellbeing study, visit the following Web site: http://www.fragilefamilies.princeton.edu/. Those interested in pursuing research investigating the collateral consequences of parental incarceration should visit the Web site of the Office of Juvenile Justice and Delinquency Prevention (OJJD): https://www.ojjdp.gov/funding/funding.html.

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.

Notes

Author Biography

R. C. Morris is an assistant professor of sociology at Weber State University. He received his PhD in sociology from Purdue University in 2014. He conducts research on social psychology and criminology as well as the scholarship of teaching and learning. He is currently working on two program evaluations for state-funded juvenile justice programs; each study investigates the social psychological impact intervention programs have on child development of a pro-social sense of self.

References

Akers

Ronald L.

Jensen

Gary F.

2006. “The Empirical Status of Social Learning Theory of Crime and Deviance: The Past, Present, and Future.” Pp. 1-37 in Taking Stock: The Status of Criminological Theory, edited by Cullen

Francis T.

Wright

J. P.

Blevins

K. R.

New Brunswick, NJ: Transaction Publishers.

Allison

Paul David

. 2002. Missing Data. Thousand Oaks, CA: Sage.

Ammerman

R. T.

Kolko

D. J.

Kirisci

Blackson

T. C.

Dawes

M. A.

1999. “Child Abuse Potential in Parents with Histories of Substance Use Disorder.” Child Abuse & Neglect 23(12):1225–38.

Baber

Kristine M.

2005. “Mentoring Youth.” Pp. 1-4 in UNH Charitable Foundation, edited by UNH Center on Adolescence. Madison, WI.

Baumeister

Roy F.

Bushman

Brad J.

2014. Social Psychology and Human Nature: Comprehensive Edition. Belmont, CA: Cengage Learning.

Becker

Sascha O.

Ichino

Andrea

. 2002. “Estimation of Average Treatment Effects Based on Propensity Scores.” The Stata Journal 2(4):358–77.

Bell

Stephen H.

Bradley

M. C.

2013. “Calculating Long-Run Impacts of Social Programs with Staggered Implementation.” Empirical Economics 44(1):163–87.

Bernardi

Richard A.

Metzger

Rene L.

Scofield Bruno

Ryann G.

Wade Hoogkamp

Marisa A.

Reyes

Lillian E.

Barnaby

Gary H.

2004. “Examining the Decision Process of Students’ Cheating Behavior: An Empirical Study.” Journal of Business Ethics 50:397–14.

Besemer

Sytske

Farrington

David P.

2012. “Intergenerational Transmission of Criminal Behaviour: Conviction Trajectories of Fathers and Their Children.” European Journal of Criminology 9(2):120–41.

10.

Big Brothers Big Sisters of America (BBBSA). 2011. “Big Brothers Big Sisters of America Annual Report 2011.” In Annual Report, edited by Big Brothers Big Sisters of America. Tampa, FL. Retrieved from http://www.bbbs.org/financial/

11.

Big Brothers Big Sisters of America (BBBSA). 2014. “Big Brothers Big Sisters of America Annual Report 2014.” In Annual Report, edited by Big Brothers Big Sisters of America. Tampa, FL. Retrieved from http://www.bbbs.org/financial/

12.

Blechman

E. A.

Bopp

Jedediah M.

2005. “Juvenile Offenders.” Pp. 454–66 in Handbook of Youth Mentoring, edited by DuBois

David L.

Karcher

Michael J.

Thousand Oaks, CA: Sage.

13.

Blumstein

Alfred

Cohen

Jacqueline

. 1987. “Characterizing Criminal Careers.” Science 237(4818):985–91.

14.

Bongers

Ilja L.

Koot

Hans M.

van der Ende

Jan

Verhulst

Frank C.

2004. “Developmental Trajectories of Externalizing Behaviors in Childhood and Adolescence.” Child Development 75(5):1523–37.

15.

Braman

Donald

. 2004. Doing Time on the Outside: Incarceration and Family Life in Urban America. Ann Arbor, MI: University of Michigan Press.

16.

Braman

Donald

Wood

Jennifer

, eds. 2003. From One Generation to the Next: How Criminal Sanctions Are Reshaping Family Life in Urban America. Washington, DC: The Urban Institute Press.

17.

Britner

P. A.

Balcazar

F. E.

Blechman

E. A.

Blinn-Pike

Larose

2006. “Mentoring Special Youth Populations.” Journal of Community Psychology 34(6):747–63.

18.

Bruster

B. E.

Foreman

2012. “Mentoring Children of Prisoners: Program Evaluation.” Social Work in Public Health 27(1–2):3–11.

19.

Caliendo

Marco

Kopeinig

Sabine

. 2008. “Some Practical Guidance for the Implementation of Propensity Score Matching.” Journal of Economic Surveys 22(1):31–72.

20.

Campbell

Susan B.

Spieker

Susan

Burchinal

Margaret

, Early Child Care Research Network. 2006. “Trajectories of Aggression from Toddlerhood to Age 9 Predict Academic and Social Functioning through Age 12.” The Journal of Child Psychology and Psychiatry, and Allied Disciplines 47(8):791–800.

21.

Coakley

Tanya M.

Shears

Jeffrey K.

Randolph

Schenita D.

2014. “Understanding Key Barriers to Fathers’ Involvement in Their Children’s Lives.” Child & Youth Services 35(4):343–64.

22.

Coleman

James S.

1988. “Social Capital in the Creation of Human Capital.” American Journal of Sociology 94:95–120.

23.

Conway

James M.

Jones

Edward T.

2015. Seven Out of Ten? Not Even Close: A Review of Research on the Likelihood of Children with Incarcerated Parents Becoming Justice-involved. Edited by Institute for Municipal and Regional Policy (IMRP). New Britain, CT: National Institute of Corrections, Central Connecticut State University.

24.

“Crime in the United States.” 2016. In Uniform Crime Report, edited by Federal Bureau of Investigation. Washington, DC: U.S. Department of Justice. Retrieved from https://ucr.fbi.gov/crime-in-the-u.s/2016/

25.

Davis

Stephen F.

Grover

Cathy A.

Becker

Angela H.

McGregor

Loretta N.

1992. “Academic Dishonesty: Prevalence, Determinants, Techniques, and Punishments.” Teaching of Psychology 19(1):16–20.

26.

de Haan

Willem

Loader

Ian

. 2002. “On the Emotions of Crime, Punishment and Social Control.” Theoretical Criminology 6(3):243–53.

27.

Dehejia

Rajeev H.

Wahba

Sadek

. 1999. “Causal Effects in Nonexperimental Studies: Reevaluating the Evaluation of Training Programs.” Journal of the American Statistical Association 94(448):1053–62.

28.

Dehejia

Rajeev H.

Wahba

Sadek

. 2002. “Propensity Score-matching Methods for Nonexperimental Causal Studies.” The Review of Economics and Statistics 84(1):151–61.

29.

DiStefano

Christine

Zhu

Min

Mîndrilă

Diana

. 2009. “Understanding and Using Factor Scores: Considerations for the Applied Researcher.” Practical Assessment, Research & Evaluation 14(20):1–11.

30.

DuBois

David L.

Felner

Robert D.

Brand

Stephen

Adan

Angela M.

Evans

Elizabeth G.

1992. “A Prospective Study of Life Stress, Social Support, and Adaptation in Early Adolescence.” Child Development 63(3):542–57.

31.

DuBois

David L.

Karcher

Michael J.

2005. Handbook of Youth Mentoring. Thousand Oaks, CA: Sage.

32.

DuBois

David L.

Portillo

Nelson

Rhodes

Jean E.

Silverthorn

Naida

Valentine

Jeffrey C.

2011. “How Effective Are Mentoring Programs for Youth? A Systematic Assessment of the Evidence.” Psychological Science in the Public Interest 12(2):57–91.

33.

DuBois

David L.

Silverthorn

Naida

. 2005. “Characteristics of Natural Mentoring Relationships and Adolescent Adjustment: Evidence from a National Study.” Journal of Primary Prevention 26(2):69–92.

34.

Eddy

J. M.

Poehlmann

2010. Children of Incarcerated Parents: A Handbook for Researchers and Practitioners. Washington, DC: Urban Institute Press.

35.

Ellis

Bruce J.

Giudice

Marco Del

Dishion

Thomas J.

Figueredo

Aurelio José

Gray

Peter

Griskevicius

Vladas

Hawley

Patricia H.

Jacobs

W. Jake

James

Jenée

Volk

Anthony A.

2012. “The Evolutionary Basis of Risky Adolescent Behavior: Implications for Science, Policy, and Practice.” Developmental Psychology 48(3):598–623.

36.

Fagan

Jeffrey

. 1989. “The Social Organization of Drug Use and Drug Dealing among Urban Gangs.” Criminology 27(4):633–70.

37.

Farrington

David P.

1986. “Age and Crime.” Crime and Justice 7:189–250.

38.

Farrington

David P.

Coid

Jeremy W.

Murray

Joseph

. 2009. “Family Factors in the Intergenerational Transmission of Offending.” Criminal Behaviour and Mental Health 19(2):109–24.

39.

Farruggia

Susan P.

Bullen

Pat

Davidson

Joy

Dunphy

Ann

Solomon

Frank

Collins

Efeso

. 2011. “The Effectiveness of Youth Mentoring Programmes in New Zealand.” New Zealand Journal of Psychology 40(3):52–70.

40.

Ferguson

Kristin M.

Mindel

Charles H.

2007. “Modeling Fear of Crime in Dallas Neighborhoods: A Test of Social Capital Theory.” Crime & Delinquency 53(2):322–49.

41.

Fusco

Dana

. 2012. “Use of Self in the Context of Youth Work.” Child & Youth Services 33(1):33–45.

42.

Geller

Amanda

Franklin

Allyson Walker

. 2014. “Paternal Incarceration and the Housing Security of Urban Mothers.” Journal of Marriage and Family 76(2):411–27.

43.

Giordano

Peggy C.

Schroeder

Ryan D.

Cernkovich

Stephen A.

2007. “Emotions and Crime over the Life Course: A Neo-Meadian Perspective on Criminal Continuity and Change.” American Journal of Sociology 112(6):1603–61.

44.

Giordano

Peggy C.

Seffrin

Patrick M.

Manning

Wendy D.

Longmore

Monica A.

2011. “Parenthood and Crime: The Role of Wantedness, Relationships with Partners, and SES.” Journal of Criminal Justice 39(5):405–16.

45.

Gottfredson

Michael R.

Hirschi

1990. A General Theory of Crime. Stanford, CA: Stanford University Press.

46.

Grossman

Jean Baldwin

. 2005. “Evaluating Mentoring Programs.” Pp. 251–65 in Handbook of Youth Mentoring, edited by DuBois

David L.

Karcher

Michael J.

Thousand Oaks, CA: Sage.

47.

Grossman

Jean Baldwin

Bulle

Meridel J.

2006. “Review of What Youth Programs Do to Increase the Connectedness of Youth with Adults.” Journal of Adolescent Health 39(6):788–99.

48.

Grossman

Jean Baldwin

Chan

Christian S.

Schwartz

Sarah E. O.

Rhodes

Jean E.

2012. “The Test of Time in School-based Mentoring: The Role of Relationship Duration and Re-matching on Academic Outcomes.” American Journal of Community Psychology 49(1–2):43–54.

49.

Grossman

Jean Baldwin

Tierney

Joseph P.

1998. “Does Mentoring Work? An Impact Study of the Big Brothers Big Sisters Program.” Evaluation Review 22:403–26.

50.

Hagan

John

Foster

Holly

. 2006. “Profiles of Punishment and Privilege: Secret and Disputed Deviance during the Racialized Transition to American Adulthood.” Crime, Law and Social Change 46(1–2):65–85.

51.

Hagan

John

McCarthy

Bill

Foster

Holly

. 2002. “A Gendered Theory of Delinquency and Despair in the Life Course.” Acta Sociologica 45(1):37–46.

52.

Harper

Cynthia C.

McLanahan

Sara S.

2004. “Father Absence and Youth Incarceration.” Journal of Research on Adolescence 14(3):369–97.

53.

Haskins

Anna R.

2014. “Unintended Consequences: Effects of Paternal Incarceration on Child School Readiness and Later Special Education Placement.” Sociological Science 1(1):141–58.

54.

Heckman

James J.

Ichimura

Hidehiko

Todd

Petra

. 1998. “Matching as an Econometric Evaluation Estimator.” The Review of Economic Studies 65(2):261–94.

55.

Hipp

John R.

Bauer

Daniel J.

Curran

Patrick J.

Bollen

Kenneth A.

2004. “Crimes of Opportunity or Crimes of Emotion? Testing Two Explanations of Seasonal Change in Crime.” Social Forces 82(4):1333–72.

56.

Hjalmarsson

Randi

Lindquist

Matthew J.

2012. “Like Godfather, Like Son: Exploring the Intergenerational Nature of Crime.” Journal of Human Resources 47(2):550–82.

57.

Huebner

Beth M.

Gustafson

Regan

. 2007. “The Effect of Maternal Incarceration on Adult Offspring Involvement in the Criminal Justice System.” Journal of Criminal Justice 35(3):283–96.

58.

Jarjoura

G. Roger

DuBois

David L.

Shalfer

Rebecca J.

Haight

Konrad A.

2013. Mentoring Children of Incarcerated Parents: A Synthesis of Research and Input from the Listening Session Held by the Office of Juvenile Justice and Delinquency Prevention and the White House Domestic Policy Council and Office of Public Engagement. Edited by Office of Juvenile Justice and Delinquency Prevention. Washington, DC: U.S. Department of Justice.

59.

Jekielek

Susan

Moore

Kristin A.

Hair

Elizabeth C.

2002. Mentoring Programs and Youth Development: A Synthesis. Santa Monica, CA: RAND.

60.

Johnson

Elizabeth I.

Easterling

Beth

. 2012. “Understanding Unique Effects of Parental Incarceration on Children: Challenges, Progress, and Recommendations.” Journal of Marriage and Family 74(2):342–56.

61.

Johnson

Rucker C.

2009. “Ever-increasing Levels of Parental Incarceration and the Consequences for Children.” Pp. 177–206 in Do Prisons Make Us Safer? The Benefits and Costs of the Prison Boom, edited by Raphael

Steven

Stoll

Michale A.

New York: Russell Sage.

62.

Johnson

2012. “Mapping the Critical Service Needs of Adolescent Children of Prisoners.” Social Work in Public Health 27(1–2):45–68.

63.

Johnston

Denise

. 2006. “Children of Incarcerated Parents.” Journal of Children & Poverty 12(1):91–105.

64.

Jones

Shayne

Cauffman

Elizabeth

Piquero

Alex R.

2007. “The Influence of Parental Support among Incarcerated Adolescent Offenders: The Moderating Effects of Self-control.” Criminal Justice and Behavior 34(2):229–45.

65.

Kennedy, Bruce

Kawachi

Ichiro

Prothrow-Stith

Deborah

Lochner

Kimberly

Gupta

Vanita

. 1998. “Social Capital, Income Inequality, and Firearm Violent Crime.” Social Science & Medicine 47(1):7–17.

66.

Kim

Bitna

Merlo

Alida V.

Benekos

Peter J.

2013. “Effective Correctional Intervention Programmes for Juveniles: Review and Synthesis of Meta-analytic Evidence.” International Journal of Police Science & Management 15(3):169–89.

67.

Latimer

Jeff

. 2001. “A Meta-analytic Examination of Youth Delinquency, Family Treatment, and Recidivism.” Canadian Journal of Criminology/Revue Canadienne de Criminologie 43(2):237–53.

68.

Laub

John H.

Sampson

Robert J.

1993. “Turning Points in the Life Course: Why Change Matters to the Study of Crime.” Criminology 31(3):301–25.

69.

Laub

John H.

Sampson

Robert J.

2001. “Understanding Desistance from Crime.” Crime and Justice 28:1–69.

70.

Laub

John H.

Sampson

Robert J.

2002. “Sheldon and Eleanor Glueck’s Unraveling Juvenile Delinquency Study: The Lives of 1,000 Boston Men in the Twentieth Century.” Pp. 87–115 in Looking at Lives: American Longitudinal Studies of the Twentieth Century, edited by Phelps

Furstenberg

F. F.

Jr Colby

New York: Russell Sage.

71.

Laub

John H.

Sampson

Robert J.

2003. Shared Beginnings, Divergent Lives: Delinquent Boys to Age 70. Cambridge, MA: Harvard University Press.

72.

Lauritsen

Janet L.

Heimer

Karen

Lynch

James P.

2009. “Trends in the Gender GAP in Violent Offending: New Evidence from the National Crime Victimization Survey.” Criminology 47(2):361–99.

73.

Loeber

Rolf

Burke

Jeffrey D.

2011. “Developmental Pathways in Juvenile Externalizing and Internalizing Problems.” Journal of Research on Adolescence 21(1):34–46.

74.

Martone

Cynthia

. 2005. Loving through Bars: Children with Parents in Prison. Santa Monica, CA: Santa Monica Press.

75.

Matsueda

Ross L.

Heimer

Karen

. 1997. “A Symbolic Interactionist Theory of Role-transitions, Role-commitments and Delinquency.” Pp. 163-213 in Developmental Theories of Crime and Delinquency, edited by Thornberry

Terence P

. New Brunswick, NJ: Transaction Publishers.

76.

McCabe

Donald L.

Katz

Daniel

. 2009. “Curbing Cheating.” Education Digest 75(1):16–19.

77.

McCabe

Donald L.

Trevino

Linda Klebe

. 1997. “Individual and Contextual Influences on Academic Dishonesty: A Multicampus Investigation.” Research in Higher Education 38(3):379–96.

78.

McLanahan

Sara

Schwartz

Dona

. 2002. “Life without Father: What Happens to the Children?” Contexts 1(1):35–44.

79.

Mead

George Herbert

. 1934. Mind, Self & Society from the Standpoint of a Social Behaviorist. Chicago, IL: The University of Chicago Press.

80.

Merton

Robert K.

1938. “Social Structure and Anomie.” American Sociological Review 3 (October):672–82.

81.

Miller

J. Mitchell

Barnes

J. C.

Miller

Holly Ventura

McKinnon

Layla

. 2013. “Exploring the Link between Mentoring Program Structure & Success Rates: Results from a National Survey.” American Journal of Criminal Justice 38(3):439–56.

82.

Moffitt

Terrie E.

1993. “Adolescence-limited and Life-course-persistent Antisocial Behavior: A Developmental Taxonomy.” Psychological Review 100(4):674–701.

83.

Morris

R. C.

2012. “The Relative Influence of Values and Identities on Academic Dishonesty: A Quantitative Analysis.” Current Research in Social Psychology 19(2):1–20.

84.

Muller

Christopher

Wildeman

Christopher

. 2013. “Punishment and Inequality.” Pp. 169–86 in The SAGE Handbook of Punishment and Society, edited by Simon

Jonathan

Sparks

Richard

. Los Angeles, CA: Sage.

85.

Murray

Joseph

Bijleveld-Catrien

C. J. H.

Farrington

David

Loeber

Rolf

. 2014. Effects of Parental Incarceration on Children: Cross-national Comparative Studies. Washington, DC: American Psychological Association.

86.

Murray

Joseph

Farrington

David P.

Sekol

Ivana

. 2012. “Children’s Antisocial Behavior, Mental Health, Drug Use, and Educational Performance after Parental Incarceration: A Systematic Review and Meta-analysis.” Psychological Bulletin 138(2):175–210.

87.

Nagin

Daniel S.

Tremblay

Richard E.

1999. “Trajectories of Boys’ Physical Aggression, Opposition, and Hyperactivity on the Path to Physically Violent and Nonviolent Juvenile Delinquency.” Child Development 70(5):1181–96.

88.

Nagin

Daniel S.

Tremblay

Richard E.

2001. “Parental and Early Childhood Predictors of Persistent Physical Aggression in Boys from Kindergarten to High School.” Archives of General Psychiatry 58(4):389–94.

89.

Nakkula

M. J.

Harris

J. T.

2005. “Assessment of Mentoring Relationships.” Pp. 100–17 in Handbook of Youth Mentoring, edited by Dubious

D. L.

Karcher

M. J.

Thousand Oaks, CA: Sage.

90.

Nation

Maury

Crusto

Cindy

Wandersman

Abraham

Kumpfer

Karol L.

Seybolt

Diana

Morrissey-Kane

Erin

Davino

Katrina

. 2003. “What Works in Prevention: Principles of Effective Prevention Programs.” American Psychologist 58(6–7):449–56.

91.

Naudeau

Sophie

. 2010. “Children of Incarcerated Parents: Developmental Trajectories among School-aged Children.” Pp. 47–72 in Children of Incarcerated Parents: Theoretical, Developmental, and Clinical Issues, edited by Harris

Y. R.

Graham

J. A.

Carpenter

G. J. O.

New York: Springer.

92.

Paolo

Buonanno

Montolio

Daniel

Vanin

Paolo

. 2009. “Does Social Capital Reduce Crime?” Journal of Law & Economics 52(1):145–70.

93.

Parke

Ross D. and K.

Clarke-Stewart

Alison

. 2003. “The Effects of Parental Incarceration on Children: Perspectives, Promises, and Policies.” Pp. 189–232 in Prisoners Once Removed: The Impact of Incarceration and Reentry on Children, Families, and Communities, edited by Travis

Jeremy

Waul

Michelle

. Washington, DC: Urban Institute Press.

94.

Phillips

Susan D.

O’Brien

2012. “Learning from the Ground up: Responding to Children Affected by Parental Incarceration.” Social Work in Public Health 27(1–2):29–44.

95.

Poehlmann

Julie

. 2005. “Representations of Attachment Relationships in Children of Incarcerated Mothers.” Child Development 76(3):679–96.

96.

Rabe-Hesketh

Skrondal

2012. Multilevel and Longitudinal Modeling Using Stata (Vols. I & II, 3rd ed.). College Station, TX: Stata Press.

97.

Ray

James V.

Jones

Shayne

Loughran

Thomas A.

Jennings

Wesley G.

2013. “Testing the Stability of Self-control: Identifying Unique Developmental Patterns and Associated Risk Factors.” Criminal Justice and Behavior 40:588–607.

98.

Reed

D. F.

Reed

E. L.

2004. “Mothers in Prison and Their Children.” Pp. 261–74 in The Criminal Justice System and Women: Offenders, Victims, and Workers, edited by Price

Barbara R.

Sokoloff

Natalie J.

New York: McGraw-Hill.

99.

Reichman

Nancy E.

Teitler

Julien O.

Garfinkel

Irwin

McLanahan

Sara S.

2001. “Fragile Families: Sample and Design.” Children and Youth Services Review 23(4):303–26.

100.

Reise

Steven P.

Waller

Niels G.

Comrey

Andrew L.

2000. “Factor Analysis and Scale Revision.” Psychological Assessment 12(3):287–97.

101.

Rhodes

Jean E.

2002. Stand by Me: The Risks and Rewards of Mentoring Today’s Youth. Cambridge, MA: Harvard University Press.

102.

Rhodes

Jean E.

2008. “Improving Youth Mentoring Interventions through Research-based Practice.” American Journal of Community Psychology 41(1–2):35–42.

103.

Rhodes

Jean E.

Grossman

Jean B.

Resch

Nancy L.

2000. “Agents of Change: Pathways through Which Mentoring Relationships Influence Adolescents’ Academic Adjustment.” Child Development 71:1662–71.

104.

Richey

Sean

Ikeda

Ken’ichi

. 2009. “Institutional Incentives and Trust: Marginalized Groups and the Creation of Trust in Local Government.” Social Science Quarterly 90(4):911–26.

105.

Rivers

James E.

Anwyl

Robert S.

2000. “Juvenile Assessment Centers: Strengths, Weaknesses, and Potential.” Prison Journal 80(1):96–113.

106.

Roberts

Helen

Liabo

Kristin

Lucas

Patricia

DuBois

David L.

Sheldon

Trevor A.

2004. “Mentoring to Reduce Antisocial Behaviour in Childhood.” British Medical Journal 328(7438):512–14.

107.

Rogers

W. H.

1993. “Regression Standard Errors in Clustered Samples.” Stata Technical Bulletin 3(13):19–23.

108.

Rose

Dina R.

Clear

Todd R.

1998. “Incarceration, Social Capital, and Crime: Implications for Social Disorganization Theory.” Criminology 36(3):441–80.

109.

Rosenbaum

Paul R.

Rubin

Donald B.

1983. “The Central Role of the Propensity Score in Observational Studies for Causal Effects.” Biometrika 70(1):41–55.

110.

Rosenbaum

Paul R.

Rubin

Donald B.

1984. “Reducing Bias in Observational Studies Using Subclassification on the Propensity Score.” Journal of the American Statistical Association 79(387):516–24.

111.

Rosenbaum

Paul R.

Rubin

Donald B.

1985. “Constructing a Control Group Using Multivariate Matched Sampling Methods That Incorporate the Propensity Score.” The American Statistician 39(1):33–38.

112.

Russell

Daniel W.

2002. “In Search of Underlying Dimensions: The Use (and Abuse) of Factor Analysis in Personality and Social Psychology Bulletin.” Personality and Social Psychology Bulletin 28(12):1629–46.

113.

Sampson

Robert J.

2011. “The Incarceration Ledger.” Criminology & Public Policy 10(3):819–28.

114.

Schafer

Joseph L.

Graham

John W.

2002. “Missing Data: Our View of the State of the Art.” Psychological Methods 7:147–77.

115.

Serido

Joyce

Borden

Lynne M.

Wiggs

Christine B.

2014. “Breaking Down Potential Barriers to Continued Program Participation.” Youth & Society 46(1):51–69.

116.

Sipe

Cynthia L.

2002. “Mentoring Programs for Adolescents: A Research Summary.” Journal of Adolescent Health 31(6, Supp. 1):251–60.

117.

Skogan

Wesley G.

1988. “Community Organizations and Crime.” Crime and Justice 10:39–78.

118.

Smith

David J.

2005. “Ethnic Differences in Intergenerational Crime Patterns.” Crime and Justice 32:59–129.

119.

Smith

Jeffrey A.

Todd

Petra E.

2001. “Reconciling Conflicting Evidence on the Performance of Propensity-score Matching Methods.” American Economic Review 91(2):112–18.

120.

Smith

Tony R.

2004. “Low Self-control, Staged Opportunity, and Subsequent Fraudulent Behavior.” Criminal Justice and Behavior 31(5):542–63.

121.

Spencer

Renee

. 2006. “Understanding the Mentoring Process between Adolescents and Adults.” Youth & Society 37(3):287–315.

122.

Stryker

Sheldon

Craft

Elizabeth A.

1982. “Deviance, Selves and Others Revisited.” Youth & Society 14(2):159–83.

123.

Sutherland

Edwin Hardin

. 1947. Principles of Criminology. Chicago, IL: J. B. Lippincott.

124.

Thornberry

Terence P.

2009. “The Apple Doesn’t Fall Far from the Tree (or Does It?): Intergenerational Patterns of Antisocial Behavior—The American Society of Criminology 2008 Sutherland Address.” Criminology 47(2):297–325.

125.

Tolan

Patrick H.

Henry

David

Schoeny

Michael

Bass

Arin

. 2013. “Mentoring Interventions to Affect Juvenile Delinquency and Associated Problems: A Systematic Review.” Campbell Systematic Reviews 9(10):1-148.

126.

Tolan

Patrick H.

Henry

David B.

Schoeny

Michael S.

Lovegrove

Peter

Nichols

Emily

. 2014. “Mentoring Programs to Affect Delinquency and Associated Outcomes of Youth at Risk: A Comprehensive Meta-analytic Review.” Journal of Experimental Criminology 10(2):179–206.

127.

Travis

Jeremy

Waul

Michelle

, eds. 2003. Prisoners Once Removed: The Impact of Incarceration and Reentry on Children, Families, and Communities. Washington, DC: Urban Institute Press.

128.

Turanovic

Jillian J.

Rodriguez

Nancy

Pratt

Travis C.

2012. “The Collateral Consequences of Incarceration Revisited: A Qualitative Analysis of the Effects on Caregivers of Children of Incarcerated Parents.” Criminology 50(4):913–59.

129.

Turner

Ralph H.

1962. “Role Taking: Process versus Conformity.” Pp. 323–39 in Human Behavior and Social Processes: An Interactionist Approach, edited by Rose

Arnold M.

Boston, MA: Houghton Mifflin.

130.

Turney

Kristin

Schnittker

Jason

Wildeman

Christopher

. 2012. “Those They Leave Behind: Paternal Incarceration and Maternal Instrumental Support.” Journal of Marriage and Family 74(5):1149–65.

131.

Valentino

Salem

Wheeler

Marc

. 2013. “2013 Youth Outcomes Report.” In Big Brothers Big Sisters Report to America: Positive Outcomes for a Positive Future, edited by Big Brothers Big Sisters of America. Philadelphia, PA.

132.

Vazsonyi

Alexander T.

Huang

. 2010. Where Self-control Comes from: On the Development of Self-control and Its Relationship to Deviance over Time. Washington, DC: American Psychological Association.

133.

Wacquant

Loic

. 2001. “Deadly Symbiosis: When Ghetto and Prison Meet and Mesh.” Punishment & Society 3(1):95–134.

134.

Wakefield

Sara

Wildeman

Christopher

. 2011. “Mass Imprisonment and Racial Disparities in Childhood Behavioral Problems.” Criminology & Public Policy 10(3):793–817.

135.

Western

Bruce

. 2006. Punishment and Inequality in America. New York: Russell Sage.

136.

Western

Bruce

McLanahan

Sara

. 2000. “Fathers behind Bars: The Impact of Incarceration on Family Formation.” Contemporary Perspectives in Family Research 2:309–24.

137.

Wildeman

Christopher

. 2010. “Paternal Incarceration and Children’s Physically Aggressive Behaviors: Evidence from the Fragile Families and Child Wellbeing Study.” Social Forces 89(1):285–309.

138.

Wildeman

Christopher

Western

Bruce

. 2010. “Incarceration in Fragile Families.” The Future of Children 20(2):157–77.