An Exploratory Investigation of Parental Incarceration, Emotional Independence, and Adult Children’s Criminal Activity

Abstract

Although research suggests that parental incarceration is associated with intergenerational continuity in crime, the mechanisms underlying this association remain unclear. Using multi-population structural equation modeling and data from the Toledo Adolescent Relationships Study (n = 1207), the current study explored specific experiences associated with labeling as well as internalizing labels, including experiencing corporal punishment during childhood, criminal arrests during adolescence, and identifying as a troublemaker/partier in young adulthood (measured with reflected appraisals), as potential mechanisms linking parental incarceration and young adults’ offending. We assessed whether this association differed by young adults’ level of emotional independence, that is, freedom from the need for parental approval. We found that parental incarceration indirectly influenced criminal activity particularly through identifying as a troublemaker/partier during young adulthood but only for those who sought parental approval. Overall, we concluded that high emotional independence, or not seeking parental approval, may be a protective factor that facilitates intergenerational discontinuities in crime.

In the USA, most individuals who are incarcerated are also parents of children under 18 years of age (Glaze & Maruschak, 2008). During the past four decades as incarceration rates increased (Bronson & Carson, 2019), so did the number of incarcerated parents and children experiencing parental incarceration (Glaze & Maruschak, 2008). In 2012 alone, over 2.6 million minors experienced parental incarceration (Sykes & Pettit, 2014). Moreover, Murphey and Cooper (2015) estimated that over 5 million children have ever experienced the incarceration of a resident parent. Given the number of children exposed to parental incarceration, concerns surrounding the collateral consequences of incarceration have amplified.

Researchers (e.g., Davis & Shlafer, 2017; Lee et al., 2013; Young et al., 2020) have found that parental incarceration has deleterious, wide-ranging consequences for children’s outcomes throughout the life course including economic disadvantage and poorer mental and physical health. Most relevant to the current study, parental incarceration is associated with the intergenerational transmission of crime and contact with the criminal justice system across the life course (Dallaire et al., 2015; Mears & Siennick, 2016; Wildeman & Muller, 2012). However, several scholars (e.g., Dallaire et al., 2015; Foster & Hagan, 2015; Giordano et al., 2019) have noted that the mechanisms underlying the consequences of parental incarceration on children need further exploration.

Using population-based survey data from a contemporary sample of men and women surveyed across the transition from adolescence to young adulthood (Toledo Adolescent Relationships Study, n = 1207), we explored experiencing corporal punishment during childhood, criminal arrests during adolescence, and identifying as a troublemaker/partier in young adulthood (measured with reflected appraisals), as possible mediating mechanisms underlying the association between parental incarceration and young adults’ offending. Additionally, this exploratory study examined whether emotional independence, or freedom from the need for parental approval, facilitated intergenerational discontinuities in offending.

Mechanisms Underlying Parental Incarceration and Children’s Criminal Activity

Parental incarceration is associated with problem behavior, criminal activity, and criminal justice system contact during adolescence and young adulthood (e.g., Muftić & Smith, 2018; Wakefield & Wildeman, 2014; Wildeman & Muller, 2012). For example, Roettger and Swisher (2011) found that male adolescents who experienced paternal incarceration had an increased propensity for delinquency that persisted into young adulthood and an increased risk of adult arrest before age 25. Mears and Siennick’s (2016) research illustrated that the association between parental incarceration, offending, and marijuana use extended into late young adulthood. Such patterns are also found in studies addressing issues of selection. For example, Murray, Farrington et al. (2012), Murray et al. (2012), conducting propensity score matching and fixed effects analyses, found that parental incarceration was associated positively with youth (ages 7–22) theft.

Moreover, recent studies have made strides in empirically examining the mechanisms that underlie the association between parental incarceration and children’s antisocial behavior. For instance, Dallaire and colleagues’ (2015) model, which analyzed data from a sample of 151 children (ages 6–13) of incarcerated mothers in Virginia, illustrated maternal incarceration predicted children’s problematic internalizing and externalizing behaviors through incarceration-specific risk experiences, such as witnessing their mother’s criminal behavior, arrest, or sentencing. Additionally, Johnson et al. (2019), examining data from a predominantly Black, adolescent sample from Alabama, found that household members’ contact with the justice system (as measured by anyone in the household experiencing an arrest) influenced adolescents’ substance use indirectly through poor supervision, fewer rules, and no curfew restrictions.

Analyzing data from the Linking Interests of Families and Teachers (LIFT) study, which included a sample of 655 fifth graders from 12 local schools in Oregon in 1991, Kjellstrand and Eddy (2011) examined the influence of parental incarceration (prior to 10 years of age) on children’s externalizing behaviors in the 5th, 8th, and 10th grade and delinquency in the 10th grade. Overall, using structural equation modeling, the authors found that the influence of parental incarceration largely operated in an indirect fashion. That is, parental incarceration was associated negatively with parental advantage (a latent construct that assessed household income, weekly hours worked, and socioeconomic status), which was associated positively with parental health (a latent construct that assessed parents’ physical health and depression). Further, parental health was associated positively with parenting (a latent construct that assessed parents’ monitoring, involvement, praise, appropriate and consistent use of discipline, as well as the quality of the parent/child relationship) and negatively associated with children’s externalizing behaviors and delinquency. Notably, the LIFT study included an intervention, so Kjellstrand and Eddy (2011) controlled for intervention group membership. However, the authors did not include any other relevant control variables, such as parents’ antisocial behavior or children’s race, age, or gender.

Although these studies have made valuable contributions to our understanding of the mechanisms of parental incarceration on children’s involvement in antisocial behavior, existing research has not investigated whether informal or formal labeling or the internalization of labels underlie the influence of parental incarceration on young adults’ offending. Moreover, studies examining the mechanisms underlying the influence of parental incarceration have not examined children’s outcomes later in life (e.g., in adulthood). Importantly, efforts to mitigate the consequences of parental incarceration will be more feasible once scholars explore potential mechanisms of the consequences of parental incarceration throughout the life course.

Labeling and Parental Incarceration

To describe the potential underlying mechanisms linking parental incarceration and children’s involvement in crime, we drew insights from labeling theory (Lemert, 1951, 1967), which contends that the informal and formal labels that society attaches to an individual can influence subsequent behavior through the internalization of labels. Individuals adjust their self-conceptions and identities to reflect the labels and expectations that others have imposed on them (e.g., Asencio & Burke, 2011; Brownfield & Thompson, 2005). That is, identities, the “content” component of the self-concept (e.g., Gecas & Burke, 1995; Rosenberg, 1981; Stryker & Burke, 2000), are based, in part, on reflected appraisals (Matsueda, 1992; Rosenberg, 1981), which refer to individuals’ perceptions of how others label them. Moreover, individuals’ subsequent behavior tends to conform to internalized labels (Gecas, 1982; Heimer & Matsueda, 1994; Matsueda, 1992), including the “troublemaker/partier” label (Copp et al., 2020; Johnson et al., 2016).

Studies examining informal labels as measured by actual and perceived parental labeling have found that informal labels influence youths’ reflected appraisals, which predicted delinquency (Bartusch & Matsueda, 1996; Liu, 2000). Moreover, some scholars have proposed that parents’ use of corporal punishment influences children’s self-identities and behavior by informally labeling children as “bad” (Straus, 2017; Vissing et al., 1991). Further, research on labeling through official sanctions, such as police contact, arrest, or incarceration as a juvenile, has found that formal, punitive labels influence antisocial identity formation and subsequent behavior (Saunders, 2018; Wiley, 2015; Wiley et al., 2013).

Importantly, entire family units can be deemed criminal when others label them due to their connection to a “criminal” family member (Shaw, 2016). This is supported by the notion of linked lives (Elder, 1985, 1994, 1997), a core principle of the life course perspective, which is based on the premise that individuals’ lives influence and are influenced by the individuals in their immediate and extended social network, including family members. Also paralleling the idea that experiences are consequential across generations, Goffman (1963) used the phrase “courtesy stigma” to refer to stigma that is due to being associated with an individual, such as a parent, who has a discredited label and used the term “spoiled identity” to refer to an identity originating from a negative label that results in transitioning through the life course with an attribute that is extremely stigmatizing. Indeed, plausibly due to linked lives (Elder, 1985, 1994, 1997), courtesy stigma (Goffman, 1963), and the process of labeling (Lemert, 1951, 1967), children of incarcerated parents experience social, psychological, and financial costs throughout the life course (Dallaire et al., 2010; Phillips & Gates, 2011; Wildeman et al., 2017). Children of incarcerated parents may therefore internalize negative labels and develop spoiled identities through the courtesy stigma that accompanies parental contact with the justice system (Augustyn et al., 2019; Finkeldey et al., 2020; Saunders, 2018).

Given that courtesy stigmas are associated with social exclusion, hostility from others, and a lack of respect for individuals whose parents have been incarcerated, such stigma may result in limited access to conventional identities and perhaps greater acceptance of antisocial identities and behavior. Johnston and Sullivan’s (2016) description of adult children’s (ages 18–59) personal experiences supported how parental incarceration and courtesy stigmas influenced their self-identities. For example, Tony expressed, “As a child of an incarcerated parent it is expected that you will be immersed in the lifestyle of prison, that there is no future or success in the years ahead of you” (Johnston & Sullivan 2016:125). Similarly, Jeremy, whose father was incarcerated for most of his childhood, stated: “From a young age, I was convinced that my purpose in life was to follow in my dad’s footsteps; everyone seems to expect it …with people’s expectations, it created my image of myself as well” (Johnston & Sullivan 2016:97). Furthermore, Giordano (2010) found that adolescents were aware of being stigmatized because of their parents’ antisocial behavior. One adolescent, Jason, stated, “When my mom was ten years old … she was smoking marijuana and drinking alcohol. And, like, everybody expects me to be like just like my mom” (Giordano, 2010:168). Thus, children may develop antisocial reflected appraisals from the courtesy stigma that they experience due to their connection to an incarcerated or formerly incarcerated parent (Finkeldey et al., 2020). Given that antisocial identities and reflected appraisals are associated with antisocial behavior across the life course (e.g., Crank, 2018; Giordano et al., 1999; Rocque et al., 2016), internalizing negative labels that accompany parental incarceration may be relevant for understanding intergenerational patterns in crime among families exposed to incarceration.

Emotional Independence

Since children often look to parents as models for identity formation, the salience of incarcerated parents’ inmate or criminal identities (Arditti et al., 2005; Dyer, 2005) may, understandably, increase the odds of children adopting a deviant identity. However, an important developmental task that occurs during adolescence and young adulthood is separation-individuation, wherein individuals actively craft a sense of self separate from their parents (Mahler et al., 2015; Steinberg & Cauffman, 1996; Zupančič & Kavčič, 2014). One dimension of separation-individuation involves developing emotional independence, or freedom from the excessive need for parental approval, closeness, togetherness, and emotional support (Hoffman, 1984), which may be relevant to combating the internalization of labels associated with the courtesy stigma that accompanies parental incarceration.

Whereas low emotional independence may involve child-like dependence on parents for approval, high levels may involve unhealthy detachment from and distrust of parents (Koepke & Denissen, 2012). In contrast, moderate levels of emotional independence involve a healthy balance between complete emotional enmeshment and detached isolation (Koepke & Denissen, 2012; Lapsley et al., 1989; Zupančič & Kavčič, 2014). Although both low and high emotional independence are associated with internalizing and externalizing problem behaviors (Allen et al., 2002; Ingoglia et al., 2011; Pace & Zappulla, 2010), moderate levels are associated with psychosocial maturity (Koepke & Denissen, 2012). However, studies examining the consequences of low and high emotional independence generally have assumed that children have parents who hold conventional prosocial values (i.e., law-abiding, non-incarcerated parents). For children of currently or formerly incarcerated parents, higher levels of emotional and identity distancing may be an important dimension of positive identity formation processes (Finkeldey et al., 2020; Johnson & Easterling, 2015; Luther, 2016). In other words, it is plausible that high emotional independence may be a protective psychological factor against the internalization of antisocial labels among those with a history of parental incarceration. Consequently, those who emotionally separate from their incarcerated parent(s) may therefore be less likely to adopt an antisocial identity or engage in crime (Finkeldey et al., 2020; Johnson & Easterling, 2015; Luther, 2016).

Indeed, children with high emotional independence may construct identities that are distinct from their incarcerated or formerly incarcerated parent(s). Luther’s (2016) qualitative research with college students (ages 18–39) who experienced parental incarceration during childhood found that physical and/or emotional separation from an incarcerated parent helped children to manage courtesy stigma and ultimately develop prosocial identities. For instance, illustrating high emotional independence, Hannah admitted that her incarcerated father was “kind of like a stranger” (Luther, 2016:6), and Julian admitted that he “pretty much discarded” his incarcerated father (Luther, 2016:7). Exhibiting the ways in which emotional independence facilitated an identity distinct from an incarcerated parent, Natalie stated “because I was able to separate myself from… [my incarcerated stepfather] … I just realized I’m my own person” (Luther, 2016:7). Moreover, Giordano’s (2010) interviews found that adolescents (ages 10–19) who developed an identity in contrast to that of antisocial parents were more likely to exhibit intergenerational discontinuities in crime. Thus, although it is assumed that high emotional independence, or not looking to parents for approval or identifying with parents, is detrimental for development, in instances where children face adverse experiences, such as parental incarceration, high emotional independence may actually facilitate healthier development.

Overall then, since emerging research has revealed that parental incarceration influences identities and reflected appraisals (e.g., Finkeldey et al., 2020; Johnson & Easterling, 2015; Luther, 2016), and self-identities are known to affect behavior (Adams et al., 2003; Giordano et al., 2009; Matsueda & Heimer, 1997), we explore whether labeling and internalizing antisocial labels might help to elucidate the intergenerational transmission of crime among young adult children with a history of parental incarceration. We also assess whether young adult children who develop high emotional independence will exhibit intergenerational discontinuities in offending despite experiencing parental incarceration.

Present Study

Based on a prospective population-based survey from a contemporary sample of men and women, the Toledo Adolescent Relationships Study (TARS), we used structural equation modeling to explore the direct and indirect influence of parental incarceration on young adults’ offending through events associated with labeling and internalizing labels. We chose to utilize structural equation modeling over other analytic methods (such as ordinary least squares regression) since it enabled us to more squarely explore our conceptualization of the sequential mechanisms underlying the association between parental incarceration and criminal involvement. Specifically, we investigated whether childhood experiences of corporal punishment, arrests during adolescence, and internalizing the troublemaker/partier identity during young adulthood (measured with reflected appraisals) acted as pathways underlying the association between parental incarceration and individuals’ involvement in crime. Multi-population analyses enabled us to explore whether the associations between parental incarceration, corporal punishment, arrests, and identity operated the same for young adults with low versus high emotional independence from parents. To further isolate these pathways, we included additional covariates of identity and offending, including parent’s own risk behavior during their adolescence (Boduszek et al., 2016; Giordano, 2010; Thornberry et al., 2009). We also included mother’s educational attainment, mother’s income, household receipt of public assistance, and disadvantaged neighborhood context as proxies for socioeconomic background (Archer, 2000; Derzon, 2010; Harding et al., 2015). Further, we accounted for family structure (Demuth & Brown, 2004), gender (Kruttschnitt, 2013; Rocque et al., 2016), race and ethnicity (Tonry, 2011), and age (Hirschi & Gottfredson, 1983; Rocque et al., 2016).

Since our study is more exploratory than explanatory, we assessed three research aims as opposed to testing formal hypotheses. The first research aim was to assess whether parental incarceration had a direct influence on (a) experiencing corporal punishment during adolescence, (b) criminal arrests as a juvenile, (c) identifying as a troublemaker/partier as an adult, and (d) criminal activity in adulthood. Given that the association between parental incarceration and involvement in crime may operate indirectly through labeling and the internalization of labels, our second research aim was to assess the association between parental incarceration and childhood corporal punishment, juvenile arrests, and identifying as a troublemaker/partier as a young adult. Finally, our third research aim was to assess whether these associations were stronger for those with low emotional independence. Based on our review of the literature, our working hypothesis was that young adults with high emotional independence—who do not need/want parental approval—would exhibit intergenerational discontinuities in crime.

Methodology

Data

We used data from the TARS, a contemporary sample of men and women interviewed five times over a period of 10 years as they transitioned from adolescence to young adulthood. The initial stratified, random sample of adolescents (and a parent/guardian interviewed separately at the first interview) were drawn from enrollment records of all students registered for the seventh, ninth, and eleventh grades in Lucas County, Ohio, in 2000 (school attendance was not required for inclusion in the sample). Developed by the National Opinion Research Center, the sampling frame was comprised of 62 schools across seven school districts and included over-samples of Black and Hispanic adolescents. These data were collected in the years 2001, 2002, 2004, 2006, and 2011. The initial sample included 1321 adolescents, the second interview included 1177, the third interview included 1114, the fourth interview included 1092, and the fifth interview included 1021 respondents (77% of the original respondents and 94% of the respondents from the fourth interview). Respondents were 12–19 years of age at the first interview and 22–29 years at the fifth interview. Additionally, these data were supplemented with official incarceration records of respondents’ biological parents, which were collected with publicly available online data. Although a regional survey, TARS’ respondents have sociodemographic background characteristics similar to young adults nationally. That is, a comparison with data from the American Community Survey (U.S. Census Bureau, 2011) suggested that the TARS respondents have similar backgrounds to young adults, ages 23–28, in terms of gender (50.7% men in the USA versus 50.9%), race (65% White in the USA versus 58%), education (28.8% college degree or more in the USA versus 27.9%), employment status (71.1% are employed in the USA versus 76.2%), and union status (49.7% are married or cohabiting in the USA versus 51.4%).

In the analyses, we used data from the respondent survey, parent/guardian survey, and official parental incarceration records search. Our sample was restricted to respondents who had valid data on the dependent variable (i.e., criminal activity), focal independent variables (i.e., parental incarceration and troublemaker/partier identity), and moderator (i.e., emotional independence), yielding an analytic sample of 1207.

Measures

Endogenous Variables

We measured parental incarceration by combining data from parents’ reports and official records. In the parent interview, parents/guardians reported whether their child’s (core respondent) living situation changed due to a parent going to prison (maternal versus paternal imprisonment was not specified). That is, the parent-reported measure reflected a change in living situation due to the imprisonment of a parent up until the time of the first TARS interview (when some respondents were as young as 12 years old). To enhance the parent-reported measure, TARS gathered publicly available records for respondents’ biological parents to determine whether official records indicated a stay in jail or prison. The online public records supplemented parents’ reports by providing incarceration data even for respondents who did not reside with their biological parents, which is important because non-residential parental incarceration is consequential for children’s outcomes (Geller et al., 2012). The official data also captured cases of parental incarceration for a longer period of time (up until the child’s age of 18 years). Moreover, the publicly available official data supplemented parents’ reports by enabling us to capture even short stays in jail. Thus, combining data from parents’ reports and official records offered the most comprehensive and accurate counts of parental incarceration. We coded the final measure of parental incarceration as “1” when the parent report or official records indicated parental incarceration and “0” otherwise.

Corporal punishment, measured using two items from the Revised Conflict Tactics Scale, asked parents/guardians how often in the month preceding the interview they had threatened to physically hurt their child or pushed, grabbed, slapped, or hit their child (Straus et al., 1996). Responses reflected whether parents/guardians reported (0) never engaging in either of these behaviors, (1) engaging in either of these behaviors once a month, (2) once a week, (3) almost every day, or (4) every day.

We measured juvenile arrest (ranging from 0 to 4) with respondents’ retrospective responses at the fourth and fifth interviews. Respondents were asked how many times they had been arrested before the age of 18. Responses included (0) never, (1) one time, (2) two times, (3) three or four times, and (4) five or more times.

Troublemaker/partier identity was measured using reflected appraisals (e.g., Asencio, 2013; Giordano et al., 2009; Matsueda, 1992) at the first respective interview when respondents were 18 years of age or older.¹ Similar to existing research (e.g., Copp et al., 2020; Finkeldey et al., 2020; Johnson et al., 2016; Reslan et al., 2011), respondents reported the extent to which they agreed that other individuals would describe them as a “troublemaker” and “partier.” Responses ranged from (0) strongly agree to (4) strongly disagree. We dichotomously coded the troublemaker/partier identity as “1” when respondents strongly agreed or agreed with both descriptions (i.e., troublemaker and partier) and “0” otherwise. We created this variable with a more “stringent” criteria (i.e., reporting both troublemaker and partier reflected appraisals as opposed to having only one reflected appraisal) since holding one deviant identity (i.e., troublemaker or partier) is arguably not unusual during young adulthood, a life course stage in which identity experimentation and exploration are normative (Boduszek et al., 2016).

We measured self-reported criminal activity at the first interview when respondents were 18 years of age or older. Respondents reported how often they engaged in the following behaviors in the preceding year: (1) “stolen (or tried to steal) things worth $5 or less”; (2) “damaged or destroyed property on purpose”; (3) “carried a hidden weapon other than a plain pocket knife”; (4) “stolen (or tried to steal) something worth more than $50”; (5) “attacked someone with the idea of seriously hurting him/her”; (6) “sold drugs”; and (7) “broken into a building or vehicle to steal something or just to look around” (Elliott & Ageton, 1980) (Cronbach’s α = 0.87, 0.82, 0.78, 0.72, and 0.78 at the first through fifth interview, respectively). Responses ranged from (0) never to (8) to more than once a day. To account for both the frequency and seriousness of offending, we created a weighted offending scale (see Giordano et al., 2008; Kazenrian & Le Blanc, 2007; King et al., 2007). Each offense was assigned a crime seriousness score derived from the National Survey of Crime Severity (Wolfgang, 1985), which gives less weight to less serious crimes and more weight to more serious crimes. Then, we multiplied each self-reported offense frequency by its corresponding seriousness score and summed the resulting products to create a cumulative weighted criminal activity scale.

Exogenous Variables

We accounted for the following factors known to influence individuals’ self-identities as well as criminal behavior: parent’s own risk behavior as a teenager; mother’s education; mother’s income; household receipt of public assistance; disadvantaged neighborhood context; family structure; and respondent’s gender, race/ethnicity, and age. Parents/guardians reported whether they engaged in the following behaviors during adolescence: (1) “I was involved in a lot of extracurricular activities” (reverse coded); (2) “I was suspended or expelled from school”; (3) “I got pregnant/got someone pregnant”; (4) “I was arrested by the police”; (5) “I drank alcohol”; and (6) “I used drugs.” Responses for each item were (0) no and (1) yes. We summed the six items to create the parent’s teen risk behavior measure (ranging from 0 to 6).

Mother’s education, mother’s income, household receipt of public assistance, and disadvantaged neighborhood context were included as proxies for socioeconomic status while growing up (Harding et al., 2015; Krieger et al., 1997). Mother’s education asked parents/guardians how far their child’s (focal respondent) mother went in school. Responses ranged from (1) 1st–8th grade to (7) beyond a 4-year college degree. We created a dichotomous variable indicating whether respondents’ mothers obtained a high school education or higher (1 = high school graduate or higher; 0 = less than high school). Ranging from (1) less than $10,000 to (9) $75,000 and over, mother’s income was also based on reports from the child’s (focal respondent) parent/guardian. Household receipt of public assistance, also reported by parents/guardians, was a dichotomous indicator for whether the parent/guardian or any member of their household received public assistance in the year preceding the survey (1 = household public assistance; 0 = no). Following Giordano et al. (2016), disadvantaged neighborhood context was measured as a 10-item scale. Parents/guardians were asked about 10 potential problems in their neighborhoods (e.g., rundown buildings, fights, unemployment). Responses for each item were (0) no and (1) yes. We summed the 10 items to create the disadvantaged neighborhood measure (ranging from 0 to 10).

Respondents reported whom they were living with most of the time in the year preceding the first interview. We used this information to create a dichotomous variable that indicated whether respondents lived with both biological parents (1 = yes; 0 = no). Also self-reported at the first interview, gender was coded 1 = male and 0 = female, and race and ethnicity was coded 1 = non-Hispanic white and 0 = otherwise. Respondents’ age in years was measured at the first interview in which respondents were ages 18 years or older.

Moderator: Emotional Independence

We focused on one aspect of emotional independence: not wanting/needing parents’ approval (Hoffman, 1984). Respondents reported how important it was to gain their parents’ approval at the first interview in which they were 18 years or older. Responses ranged from (0) very important to (4) not at all important. Thus, lower values reflected lower levels of emotional independence (i.e., wanting/needing approval) and higher values reflected higher levels of emotional independence (i.e., not wanting/needing approval). The final dichotomous measure of high emotional independence was coded as “1” for those who reported not wanting/needing approval (i.e., parental approval was “somewhat,” “not too,” or “not at all” important) and “0” for those who wanted/needed approval (i.e., parental approval was “pretty” or “very” important).

Analytic Strategy

To explore the mediating mechanisms potentially underlying the associations between parental incarceration and adult children’s criminal activity, the current research employed structural equation modeling (SEM) in SAS 9.4 with the PROC CALIS procedure (Hatcher, 1996). SEM determines the unidirectional influences of exogenous variables, which influence other variables in the model, on endogenous variables, which are predicted by other variables in the model, in a path diagram, and decomposes the total influence of a variable into direct and indirect effects (Golob, 2003; Hatcher, 1996). Although normal-theory maximum likelihood estimation (MLE) functions generally require normally distributed data, MLE with SEM produces acceptable estimates even with normality violations (Bentler, 2010; Iacobucci, 2010; Lei & Lomax, 2005). However, some argue that using MLE with non-normal data raises concerns about test statistics and standard errors (Bentler, 2010). Thus, we used a robust form of MLE that relaxes the assumption of normality: the Satorra–Bentler (SB) scale correction (Satorra & Bentler, 1994). This estimation method adjusts model chi-squares, standard errors, and model fit indices using the proposed sandwich formula to produce parameter estimates that are unbiased asymptotically (Dardick & Tuckwiller, 2019; SAS, 2019; Xia et al., 2016). Further, given our working hypothesis that the influence of parental incarceration may be conditioned by young adults’ level of emotional independence, multi-population SEM analyses enabled us to examine the direct and indirect effects of parental incarceration separately for those with low and high emotional independence.

The indirect influence of parental incarceration on adults’ offending was examined through corporal punishment, juvenile arrests, and troublemaker/partier identity. Corporal punishment, juvenile arrests, and troublemaker/partier identity were regressed on parent’s teen risk behavior, mother’s education, mother’s income, household receipt of public assistance, disadvantaged neighborhood context, as well as respondents’ gender, race and ethnicity, and age. Additionally, parental incarceration was regressed on parent’s teen risk behavior and mother’s education. Ultimately, the order of our exploratory conceptual model was based on the theoretical sequence of these factors. For instance, based on labeling theory and the life course notion of linked lives, it is presumed that parental incarceration affects subsequent labeling (e.g., corporal punishment and juvenile arrests), labeling influences the internalization of labels (i.e., identifying as a troublemaker/partier), and internalized labels affect later criminal behavior. Models were estimated allowing the factor structures to differ between groups. We used the SB-scaled model chi-square (x²), standardized root mean square residual (SRMR), goodness of fit index (GFI), root mean square error of approximation estimate (RMSEA), and Bentler comparative fit index (CFI) to assess model fit (Golob, 2003; Hooper et al., 2008).

Results

The means and proportions (as well as bivariate test statistics) for all variables by parental incarceration are shown in Table 1. Approximately 26% of the sample experienced parental incarceration before age 18. Overall, 10% of young adults reported identifying as a troublemaker/partier; however, a higher proportion of respondents who experienced parental incarceration, compared with those who did not, reported a troublemaker/partier reflected appraisal (0.13 versus 0.09; p < 0.05). Moreover, a higher proportion of young adults who experienced parental incarceration, compared to those who did not, reported high emotional independence (i.e., not wanting/needing parental approval) (0.40 versus 0.33; p < 0.05). Further, those who experienced parental incarceration, compared to those who did not, reported engaging in more criminal activity (18.21 versus 9.09; p < 0.001).

Table 1.

Means and proportions for all variables by parental incarceration

	Parental Incarceration (N = 319)	No Parental Incarceration (N = 888)	Test Statistica	Total Sample (N = 1,207)
Endogenous Variables
Parental Incarcerationp	1.00	0.00	-	0.26
Corporal Punishmentm	0.24	0.16	2.39*	0.18
Juvenile Arrestsm	0.59	0.25	4.99***	0.34
Troublemaker/Partierp	0.13	0.09	1.98*	0.10
Criminal Activitym	18.21	9.09	3.47***	11.50
Exogenous Variables
Parent’s Teen Risk Behaviorm	1.74	1.41	3.53***	1.50
Mother’s High School Educationp	0.79	0.92	−5.19***	0.89
Mother’s Incomem	2.55	3.29	−6.21***	3.10
Household Public Assistancep	0.37	0.09	9.56***	0.16
Disadvantaged Neighborhoodm	4.69	1.71	9.01***	2.50
Lived with Both Biological Parentsp	0.28	0.61	−11.25***	0.52
Malep	0.48	0.48	−0.04	0.48
Whitep	0.39	0.72	−11.16***	0.63
Agem	18.76	18.66	1.05	18.68
Moderator
High Emotional Independencep	0.40	0.33	2.32*	0.35

^*p < .05;

^**p < .01;

^***p < .001

^aIndependent sample t-test for continuous variables and independent sample z-test for dichotomous variables

^mMean

^pProportion

As a preliminary investigation into the way in which the associations might vary by emotional independence, supplemental analyses (available upon request) examined labeling, internalization of labels, and criminal activity by both parental incarceration and emotional independence. Among young adults with low and high emotional independence, those who experienced parental incarceration reported more juvenile arrests and crime than those with no history of parental incarceration. Moreover, among those with low emotional independence, a higher proportion of respondents with parental incarceration experienced corporal punishment and identified as a troublemaker/partier than those without parental incarceration. In contrast, among those with high emotional independence, the proportion of respondents who experienced corporal punishment and identified as a troublemaker/partier did not differ by parental incarceration.

Table 2 presents all standardized estimates from the structural equation models examining criminal activity for those with low emotional independence, and Table 3 presents all standardized estimates for those with high emotional independence. The model for young adults who reported low and high emotional independence explained roughly 20 (R² = 0.2023) and 19% (R² = 0.1865), respectively, of the variation in offending. Assessing the overall fit summary using the SB-scaled model x²[212.7780, p < 0.0001], SRMR (0.0550), GFI (0.9748), RMSEA (0.1429), and CFI (0.8467) suggested the joint measurement model fit of these data adequately.

Table 2.

Standardized estimates for structural equation models of parental incarceration on criminal activity in adulthood for young adults with low emotional independence

	Endogenous Variables (Outcomes)
	Parental Incarceration		Corporal Punishment		Juvenile Arrests		Troublemaker/Partier Identity		Criminal Activity
Predictors	b	se	b	se	b	se	b	se	b	se
Parental Incarcerationa	--	--	<0.01	0.04	0.07	0.04	0.09	0.04*	0.04	0.04
Corporal Punishment	--	--	--	--	--	--	0.11	0.05*	−0.02	0.04
Juvenile Arrest	--	--	--	--	--	--	0.14	0.05**	0.29	0.06***
Troublemaker/Partier Identity	--	--	--	--	--	--	--	--	0.21	0.06***
Parent’s Teen Risk Behavior	0.09	0.04*	0.08	0.04*	0.11	0.04*	0.04	0.04	0.03	0.04
Mother’s High School Education	−0.21	0.04***	0.08	0.04	−0.09	0.05	0.08	0.04*	−0.01	0.05
Mother’s Income	--	--	−0.05	0.03	−0.03	0.03	0.00	0.04	<0.01	0.03
Household Public Assistance	--	--	0.09	0.05	0.05	0.05	−0.10	0.04**	−0.04	0.05
Disadvantaged Neighborhood	--	--	0.07	0.06	0.08	0.05	<0.01	0.04	−0.07	0.04
Lived with Both Biological Parents	--	--	−0.07	0.04	−0.12	0.04**	−0.08	0.04*	−0.06	0.04
Male	--	--	−0.03	0.03	0.16	0.03***	0.08	0.04*	0.11	0.03***
White	--	--	−0.10	0.04*	<0.01	0.04	−0.01	0.04	−0.08	0.03*
Age	--	--	−0.02	0.02	−0.02	0.03	−0.06	0.02**	0.03	0.03

N = 783 Adults with Low Emotional Independence

R² = .2023

^aStandardized Total Effect of Parental Incarceration on Offending was .08^* and the Standardized Indirect Effect of Parental Incarceration on Offending was .04^*

^*p < .05;

^**p < .01;

^***p < .001

Table 3.

Standardized estimates for structural equation models of parental incarceration on criminal activity in adulthood for young adults with high emotional independence

	Endogenous Variables (Outcomes)
	Parental Incarceration		Corporal Punishment		Juvenile Arrests		Troublemaker/Partier Identity		Criminal Activity
Predictors	b	se	b	se	b	se	b	se	b	se
Parental Incarcerationa	--	--	<0.01	0.05	0.05	0.05	−0.03	0.05	0.08	0.05
Corporal Punishment	--	--	--	--	--	--	0.05	0.06	0.02	0.04
Juvenile Arrest	--	--	--	--	--	--	0.13	0.07*	0.22	0.07***
Troublemaker/Partier Identity	--	--	--	--	--	--	--	--	0.17	0.05**
Parent’s Teen Risk Behavior	0.08	0.05	0.10	0.05	0.09	0.05	0.12	0.05*	0.03	0.05
Mother’s High School Education	−0.10	0.05	0.02	0.05	−0.02	0.06	0.11	0.02***	ȡ0.07	0.05
Mother’s Income	--	--	−0.10	0.03**	−0.07	0.04	0.06	0.05	0.05	0.05
Household Public Assistance	--	--	−0.06	0.05	0.08	0.06	0.02	0.05	−0.10	0.05*
Disadvantaged Neighborhood	--	--	0.09	0.06	<0.01	0.06	−0.02	0.05	0.03	0.04
Lived with Both Biological Parents	--	--	−0.06	0.05	−0.15	0.04***	−0.02	0.05	−0.02	0.05
Male	--	--	0.04	0.05	0.12	0.05*	0.07	0.05	0.24	0.03***
White	--	--	0.03	0.05	−0.06	0.06	0.03	0.05	−0.06	0.05
Age	--	--	0.01	0.04	−0.04	0.03	−0.01	0.04	−0.04	0.03

N = 424 Adults with High Emotional Independence

R² = .1865

^aStandardized Total Effect of Parental Incarceration on Offending was .09 and the Standardized Indirect Effect of Parental Incarceration on Offending was .01

^*p < .05;

^**p < . 01;

^***p < .001

Beginning with those who reported low emotional independence (i.e., those wanting/needing parental approval; Table 2), parental incarceration was not associated with corporal punishment. However, corporal punishment positively influenced identifying as a troublemaker/partier among this group (b = 0.11; p < 0.05). A direct association between corporal punishment during adolescence and criminal activity during adulthood was not observed. Parental incarceration also was not associated with the number of juvenile arrests, although arrests were associated positively with the troublemaker/partier reflected appraisal for those with low emotional independence (b = 0.14; p < 0.01). Additionally, juvenile arrest exhibited a direct positive association with subsequent criminal activity (b = 0.29; p < 0.001). However, parental incarceration exhibited a direct positive association with having a troublemaker/partier identity (b = 0.09; p < 0.05), and identifying as a troublemaker/partier was positively associated with involvement in crime (b = 0.21; p < 0.001) for this group.

When accounting for the aforementioned mechanisms of parental incarceration, the direct association between parental incarceration and adult children’s involvement in crime was nonsignificant (albeit positive). However, the standardized total effect of parental incarceration on involvement in crime was significant for young adults with low emotional independence (0.08, p < 0.05). To calculate the portion of the total effect of parental incarceration attributed to each mediating mechanism, the direct effect from parental incarceration to each mediator was multiplied by the direct effect from the respective mediator to offending, and then the product term was divided by the standardized total effect for the respective group. Among those with low emotional independence, the portion of the total effect of parental incarceration on involvement in crime in adulthood through the troublemaker/partier identity was 0.21 [(0.09*0.21)/0.08], indicating that approximately 21% of the total effect of parental incarceration on offending among young adults with low emotional independence was due to internalizing the troublemaker/partier label.

To calculate the portion of the standardized indirect effect of parental incarceration attributed to each mediating mechanism, the direct effect from parental incarceration to each mediator was multiplied by the direct effect from the respective mediator to offending, and then the product term was divided by the standardized indirect effect for the respective group. For example, among young adult children with low emotional independence, the portion of the indirect effect of parental incarceration on adult crime through a troublemaker/partier identity was 0.47 [(0.09*0.21)/0.04]. That is, approximately 47% of the indirect effect of parental incarceration on involvement in crime was attributed to having a troublemaker/partier identity among this group.

Examining the standardized estimates for those with high emotional independence (i.e., those not wanting/needing parental approval; Table 3), parental incarceration was not associated with corporal punishment, and corporal punishment was not associated with identifying as a troublemaker/partier or criminal activity during adulthood. While parental incarceration was not associated with the number of juvenile arrests for this group, arrests were associated positively with the troublemaker/partier reflected appraisal (b = 0.13; p < 0.05) and criminal activity (b = 0.22; p < 0.001) in adulthood. The direct association between parental incarceration and identifying as a troublemaker/partier was not significant among those with high emotional independence, although identifying as a troublemaker/partier was positively associated with involvement in crime in young adulthood (b = 0.17; p < 0.01). Finally, the direct association between parental incarceration and involvement in crime was nonsignificant (albeit positive) for this group. In fact, the standardized total effect of parental incarceration on criminal activity was nonsignificant for those with high emotional independence (0.08, p > 0.05).

For ease of interpretation, Fig. 1 portrays the standardized effects reflecting our conceptualization of each model (the effects in bold font reflect estimates for the 783 adults with low emotional independence and the effects not bolded are estimates for the 424 adults with high emotional independence). Summarizing, our exploratory analyses revealed that, among young adults with low emotional independence (i.e., needing/wanting parental approval), parental incarceration indirectly influenced involvement in crime through internalizing the reflected appraisal of a troublemaker/partier. This association, however, was not observed among young adults with high emotional independence.

Fig. 1.

Completely standardized structural model for adult children’s criminal activity. Note: bold estimates are for the 783 young adults with low emotional independence and not bolded estimates are for the 424 adults with high emotional independence *p < 0.05, ** p < 0.01, *** p < 0.001

Discussion

This exploratory study investigated labeling (i.e., corporal punishment and juvenile arrests) and the internalization of negative labels (i.e., troublemaker/partier identity) as underlying mechanisms linking parental incarceration and involvement in crime in adulthood. Additionally, we explored whether these associations were conditional on young adults’ level of emotional independence, or desire for parental approval. We found that parental incarceration indirectly influenced criminal activity among young adults with low emotional independence (i.e., who wanted/needed parents’ approval). For this group, the influence of parental incarceration as a minor on involvement in crime during adulthood was, in part, due to the development of a troublemaker/partier identity during young adulthood. The larger implications of this research are pertinent to understanding the full effects of parental incarceration on adult children’s outcomes.

The current results parallel other research that has concluded that parental incarceration is consequential for adult children’s behavioral outcomes (e.g., Mears & Siennick, 2016; Murray et al., 2012; Murray et al., 2012; Wakefield & Wildeman, 2014). However, policies and programs aimed at minimizing the deleterious effects of parental incarceration are more viable when the mechanisms of its influence are understood (Murray et al., 2014). Although scholars have proposed why experiencing parental incarceration is detrimental for children, many (e.g., Dallaire et al., 2015; Foster & Hagan, 2015; Wildeman & Muller, 2012) have acknowledged that the mechanisms underlying the influence of parental incarceration remain unclear. Using labeling theory as a framework to understand the effects of parental incarceration, we followed the proposition (Cochran et al., 2018; Luther, 2016; Murray et al., 2012; Murray et al., 2012) that the labeling and stigma incarcerated parents experience likely spill over as a courtesy stigma (Goffman, 1963) to children based on the life course notion of linked lives (Elder, 1985, 1994, 1997).

Albeit exploratory in nature, our results paralleled earlier research illustrating the influence of self-conceptions on subsequent antisocial behavior (e.g., Lemert, 1951, 1967; Matsueda, 1992; Matsueda & Heimer, 1997) and research on parental incarceration and identity processes (e.g., Finkeldey et al., 2020; Johnston & Sullivan, 2016; Luther, 2016). Specifically, we found evidence that parental incarceration was associated with young adults’ offending indirectly through the internalization of labels during young adulthood, in the form of a troublemaker/partier reflected appraisal, but only for those with low emotional independence. Thus, we suggest that the influence of labeling, plausibly due to courtesy stigmas, on self-identities are mechanisms that, in part, explain the intergenerational transmission of crime among children exposed to parental incarceration, particularly among those who want/need parental approval.

Given that children who experience the incarceration of a parent are stigmatized and assumed to be similar to an incarcerated parent (Dallaire et al., 2010; Johnston & Sullivan, 2016; Luther, 2016), and are aware of such presumptions, it is understandable that some may develop identities consistent with criminal behavior. However, some children use their incarcerated parent as a model of what not to be and emotionally and physically distance themselves from their currently or formerly incarcerated parent (e.g., Johnston & Sullivan, 2016; Giordano, 2010; Luther, 2016). Thus, it is conceivable that some children who experience the incarceration of a parent may develop high emotional independence and ultimately display intergenerational discontinuities in crime.

Consistent with this notion, the current analyses, which partitioned the direct and indirect influence of parental incarceration on young adults’ behavioral outcomes by their level of emotional independence, found that the total influence of parental incarceration on involvement in crime was not significant among young adults with high emotional independence (who did not want/need parental approval). That is, we found high emotional independence was a protective factor for offending during young adulthood, supporting research (e.g., Finkeldey et al., 2020; Giordano, 2010; Johnson & Easterling, 2015; Luther, 2016) that has found that an “identity in contrast” to that of a criminal parent may protect children from problematic outcomes. Although few studies have been successful in elucidating factors that may be linked to intergenerational discontinuities in crime among families exposed to parental incarceration, our exploratory work suggests that high emotional independence among children exposed to parental incarceration is beneficial. Future research should continue to investigate whether the development of an identity in contrast to a criminal parent acts as a protective factor in the association between parental incarceration, labeling, reflected appraisals, and criminal behavior.

The analyses, however, did not fully account for the relationship between parental incarceration and adult children’s involvement in crime. Future research should therefore further investigate identity as an underlying mechanism and continue to explore possible mechanisms in this relationship. Because peers and partners are especially influential in late adolescence and young adulthood (Copp et al., 2020; Haynie, 2002; Seffrin et al., 2009; Warr, 2002), it would be worthwhile to examine the association between parental incarceration and subsequent social relationships, identity, and behavioral outcomes.

This research was not without limitations. Given the stigma associated with criminal justice contact, parental incarceration is likely to be underreported in self-reported data. Although the TARS data provided a more comprehensive representation of the occurrence of parental incarceration by combining parent reports and official records of parental incarceration, the official records were not perfect. For instance, some parents may have been incarcerated in locations outside of the search radius. Additionally, some parent names were not reported, and it was not possible to officially determine whether those respondents experienced parental incarceration. Moreover, our measure of parental incarceration was limited in that it did not provide information on the gender of the incarcerated parent, the sentence length, or the specific timing of the incarceration. Additionally, we were unable to assess whether the parent was living with the child prior to incarceration or whether respondents of incarcerated parents had any contact (e.g., letters, phone calls, visits) with their incarcerated parent. Future research should aim to test the robustness of these results using official parental incarceration records and should strive to further disentangle the influence of parental incarceration. Additionally, future studies should investigate whether the mechanisms of the influence of parental incarceration differ by respondents’ gender and race/ethnicity.

Furthermore, we were unable to address issues of temporal and causal ordering and therefore do not interpret our results as causal. That is, parental incarceration was measured as occurring any time before respondents’ age of 18. Yet, the measure of corporal punishment reflected experiences of corporal punishment that occurred in the month before the first interview (when respondents in the sample were between the ages of 12 and 19), and the measure of juvenile arrests reflected the number of times respondents were arrested any time before the age of 18. Thus, while our conceptual model has parental incarceration predicting corporal punishment and juvenile arrests, these events may or may not have occurred after instances of parental incarceration. Additionally, the current study examined the influence of parental incarceration on children’s outcomes in adulthood utilizing contemporaneous measures of identity, emotional independence, and offending. That is, young adults’ troublemaker/partier identity, emotional independence, and offending were all measured at the first respective interview respondents who were 18 years of age or older. Moreover, although respondents’ troublemaker/partier identity and emotional independence were measured when respondents were young adults, these could have actually developed any time from early childhood to adulthood. Although the theoretical sequence of events guided the order of our conceptual model, it is also possible that the directionality of the relationships we explored could go in opposite directions. For instance, given the notion of a “courtesy stigma” (Goffman, 1963), parental incarceration alone may be associated with the development of a troublemaker/partier identity, which may in turn result in corporal punishment or juvenile arrests. Thus, although structural equation modeling enabled us to explore the theoretical sequencing of mechanisms underlying the association between parental incarceration and criminal activity, our pathways were limited due to our inability to address issues of temporal ordering and causality. Therefore, to better establish the causal order of these associations, future research should build on this exploratory study by investigating the mediating mechanisms underlying the relationship between parental incarceration and children’s outcomes with mediators that are measured between the occurrence of parental incarceration and when outcomes are measured.

Notably, the linked lives aspect of the life course perspective contends that parents’ behaviors and attainments are influential for children’s outcomes. Thus, we explored parent’s teen risk behavior and mother’s education as predictors of parental incarceration. Additionally, in examining children’s outcomes, we examined parent’s risk behaviors, mother’s education, mother’s income, household receipt of public assistance, and disadvantaged neighborhood context. However, the parental factors included in the current study were limited. For instance, the retrospective reports of parent’s risk behavior as a teenager were not available for both biological parents. Moreover, given that the predictors and observed consequences of parental incarceration may be due to other familial characteristics associated with parental incarceration (e.g., Giordano & Copp, 2015; Murray et al., 2014), future studies could delve deeper into other parental characteristics (for incarcerated parents and non-incarcerated primary caregivers), such as substance use/abuse, mental health, offending, and prosocial values, throughout their child’s upbringing that may explain the association between parental incarceration and criminal behavior.

Additionally, although the current study accounted for corporal punishment, this measure was also limited. Specifically, parents/guardians reported whether they inflicted corporal punishment in the month preceding the interview, but we were unable to determine whether the parent/guardian in the survey was the incarcerated parent or not. In addition to addressing the said measure limitations, it would be worthwhile to further examine other parenting strategies, such as parental control, known to influence children’s identity development and offending (e.g., Harris-McKoy, 2016; Hoeve et al., 2009; Schroeder et al., 2010).

Although our exploratory analyses suggested that high emotional independence from one’s incarcerated parent(s) was a protective factor for offending, the indicator of emotional dependence was limited. Specifically, Hoffman’s (1984) emotional independence scale included 17 items reflecting freedom from the excessive need of parents’ approval, closeness, togetherness, and emotional support. Yet, we were only able to assess one aspect of independence: not wanting/needing parental approval. Moreover, respondents may not have answered the question regarding the importance of gaining parental approval with their incarcerated parent in mind (i.e., respondents may have answered the question thinking about their non-incarcerated parent or primary caregiver). It is also plausible that young adults do not subjectively define emotional independence or having an identity in contrast to an incarcerated parent based on the importance they attribute to gaining parental approval. Moreover, low emotional independence may reflect emotional immaturity as opposed to reflecting respondents’ intention to model parents’ antisocial behavior. Future research should therefore aim to examine alternative operationalizations of emotional independence.

Conclusion

With these limitations in mind, the results of the present study illustrated that the association between parental incarceration and young adults’ offending was partially explained by the internalization of negative labels. That is, a substantial portion of the influence of parental incarceration on adult children’s involvement in crime was attributed to developing a reflected appraisal of a troublemaker/partier during young adulthood. However, this pattern was only observed among those with low emotional independence. The direct and indirect effects of parental incarceration on criminal behavior for adults with high emotional independence (who did not want/need parental approval) were not significant. Overall, our results suggested that the development of an identity in contrast to an incarcerated parent may combat the commonly observed intergenerational transmission of crime. However, given that our findings are exploratory rather than explanatory, it is hoped that future research investigates these potential mechanisms in more detail.