Pathways from Childhood Trauma and Social Instability to Father Involvement Among Rural, Unmarried, African American Fathers: Methylation of the Oxytocin Receptor Gene as a Mediating Mechanism

Abstract

Father involvement contributes uniquely to children’s developmental outcomes. The antecedents of father involvement among unmarried, African American fathers from rural areas, however, have been largely overlooked. The present study tested a conceptual model linking retrospective reports of childhood trauma and early adulthood social instability to father involvement among unmarried, African American men living in resource-poor, rural communities in the southeastern United States. We hypothesized these factors would influence father involvement indirectly, via DNA methylation of the oxytocin receptor gene (OXTR). A sample of 192 fathers participated in three waves of data collection in early adulthood. Fathers reported on social instability at Wave 1, OXTR methylation was assessed via saliva samples at Wave 2, and measures of father involvement, retrospective childhood trauma, and quality of the fathers’ relationships with their children’s mothers were collected at Wave 3. Structural equation modeling indicated that childhood trauma was related directly to reduced levels of father involvement and to increased social instability. Social instability was associated with elevated levels of OXTR methylation, which in turn predicted decreased father involvement. The indirect effect from social instability to father involvement via OXTR methylation was significant. These associations did not operate through fathers’ relationship with the child’s mother, and remained significant even accounting for associations between inter-parental relationship quality and father involvement. Findings suggest that OXTR methylation may be a biological mechanism linking social instability to father involvement among unmarried, African American fathers in vulnerable contexts and underscore the detrimental influence of childhood trauma on father involvement.

Father involvement beginning in infancy and early childhood is related to a host of beneficial cognitive and socio-emotional outcomes for children (Sarkadi, Kristiansson, Oberklaid, & Bremberg, 2008). Children whose fathers are highly involved in their lives exhibit more advanced language development (Tamis-LeMonda, Baumwell, & Cabrera, 2013), better emotion regulation (Downer & Mendez, 2005), and fewer behavior problems (Yoon, Bellamy, Kim, & Yoon, 2018) than those whose fathers are less involved. Father involvement affects children independent of mothers’ contributions. This occurs indirectly via its influence on inter-parental relationship quality and directly via the impact of fathers’ socialization practices and father-child relationship quality on children (Cabrera, Volling, & Barr, 2018; Lamb, 2010). Evidence for fathers’ unique contributions to child development underscores the need for research on the personal, biological, and socio-contextual correlates of involved fathering. The current study aims to meet this need by examining whether life stress in childhood (i.e., childhood trauma) and adulthood (i.e., social instability) are related to father involvement via an epigenetic modification (i.e., methylation of the oxytocin receptor gene) among unmarried, African American fathers in rural environments.

African American fathers have often been considered “invisible men” in the scientific literature (Coley, 2001) in part due to common limitations with study designs and sampling procedures in family research. First, despite substantial gains in recent decades, fathers remain vastly underrepresented in parenting research relative to mothers (e.g., Phares, Lopez, Fields, Kamboukos, & Duhig, 2005; Volling & Cabrera, 2019). As such, the plurality of research has opted to sample from middle-class, European American, and largely homogenous target populations of fathers that fail to represent the lives, experiences, and contributions of African American men to family life (e.g., Coley, 2001). Second, stereotypes of unmarried, African American fathers as “absent” from the lives of their children remain pervasive (Coles, Coles, & Green, 2010; Harris, 2002) despite considerable evidence to the contrary (e.g., Downer & Mendez, 2005; Jones & Mosher, 2013). These stereotypes are perpetuated by study designs that either do not include African American fathers, or rely on mothers’ reports of fathering behavior (Coley & Morris, 2002; Hernandez & Coley, 2007). Finally, African American fathers are disproportionately likely to be unmarried and not living in the same home as their children (Castillo, Welch, & Sarver, 2011), whereas the majority of research on the contributions of fathering to child development has focused on married and co-residing couples (Carlson & McClanahan, 2010). As a result of these limitations, the personal and contextual factors that shape how unmarried, African American fathers engage with their children outside of the mother-father relationship are not well-understood.

Socio-Historical Context and the Lives of Rural, African American Fathers

The lived experiences of rural, unmarried, African American fathers have been shaped by a confluence of social, cultural, and historical risk factors that can serve as barriers to enacting their parenting roles. Historically, African American men have experienced disproportionately high levels of unemployment and underemployment (Chetty, Hendren, Joes, & Porter, 2018), as well as prejudicial and discriminatory hiring practices that take toll on young, African American men’s ability to find and maintain steady work (Mong & Roscigno, 2010). These disparities are exacerbated when accounting for the impact of mass incarceration (e.g., Western & Pettit, 2005) and child support policies that hinder or discourage fathers’ involvement and economic support for their children (e.g., Roy, 1999).

The impact of economic hardship for African American fathers goes beyond the provision of financial resources to adversely affect the stability of relationships and fathers’ consistent presence in African American family life (McLoyd, 1990). Life stress experienced by fathers is also likely to be transmitted intergenerationally, such that the conditions that place African American fathers at risk may in turn affect their sons’ abilities to maintain positive engagement with their own children (e.g., Cooper, Ross, Dues, Golden, & Burnett, 2019). As such, the deleterious impact of economic disadvantage, racism, and problematic social policies on African American fathers has persisted across generations.

The unmarried fathers who are the focus of the present study live in small towns and rural communities in the southeastern United States in which poverty and unemployment rates for African Americans are among the highest in the nation (Carl Vinson Institute of Government, 2003). These communities are part of a geographic concentration of rural poverty that coincides with some of the country’s worst economic and health disparities by race (Wimberley, 2010; Chi, Shapley, Yang, & Wang, 2019). For fathers with little discretionary income, residing in rural areas can be more challenging than living in urban areas (Burton, Lichter, Baker, & Eason, 2013). Within these contexts, the transition to fatherhood can be a demoralizing process that can take a toll on their involvement with their infants and young children (Coley & Chase-Lansdale, 1999; Roy, 2005). Thus, African American fathers who reside in rural environments are exposed to a host of cumulative risk factors that can make early father involvement challenging.

Theoretical and Conceptual Framework

The present study is informed by recent efforts to develop ecological models of father involvement and father-child relationships (see Cabrera, Fitzgerald, Bradley, & Roggman, 2014; Volling & Cabrera, 2019). An ecological approach to fatherhood focuses on the experiences of fathers in the contexts they inhabit and the impact of cultural, social, and physical environments on fathers’ parenting. Further, the ecological framework incorporates paternal history and life course influences by including biological, cultural, and rearing histories (Cabrera et al., 2014), as well as life transitions (Volling & Cabrera, 2019), as critical determinants of fathering. A strengthened focus on contextual influences and developmental history is important for understanding fathers of all types, but especially so when considering the unique life experiences and contextual circumstances faced by rural, unmarried, African American fathers.

Biopsychosocial models of development (e.g., Gottlieb, 2007) emphasize the importance of bidirectional, biological mechanisms for all levels of individual development. As applied to an ecological understanding of fathers, risk factors in the rearing history and social experiences of vulnerable boys and young men can influence the underlying neurobiology of these men as they become parents (e.g., Rilling & Mascaro, 2017), which in turn can affect their engagement with young children (e.g., Abraham & Feldman, 2018). DNA methylation is one epigenetic mechanism that appears to play an important role in linking life stress to social and behavioral phenotypes (Szyf & Bick, 2013). Given emerging epigenetic evidence and the importance of the oxytocin system for fathers more generally (e.g., Gordon, Zagoory-Sharon, Leckman, & Feldman, 2010), DNA methylation of the oxytocin receptor gene (OXTR) is a particularly promising biological mechanism (Maud, Ryan, McIntosh, & Olsson, 2018) that is considered here. Specifically, we examine the possibility that life stress in both childhood and adulthood could alter the expression of the OXTR receptor gene, which may then affect fathers’ ability to remain positively engaged with their children.

As presented in the conceptual model in Figure 1, we hypothesize that rural African American men’s exposure to life stress in childhood and adulthood plays a central role in unmarried fathers’ ability to maintain involvement with their young children. We focus on life stress that disproportionately affect unmarried, African American fathers in resource-poor, rural areas: childhood trauma and social instability. Informed by recent models of father involvement (Cabrera, et al., 2014; Volling & Cabrera, 2019), and calls for the integration of biological mediation in fathering designs (Abraham & Feldman, 2018; Kuo et al., 2018), we considered the potential for these life stressors to induce biological changes that affect father involvement. Specifically, we investigate the mediating influence of DNA methylation of the oxytocin receptor gene (OXTR).

Figure 1.

Conceptual Model Linking Life Stress to Father Involvement via OXTR Methylation

Childhood Trauma, Social Instability, and Father Involvement

The transition to parenthood is a powerful and life-changing challenge at any age; for young African American men in rural communities who are exposed to persistent economic and race-related stressors, this challenge is amplified (McLoyd, 1990). African American fathers in impoverished environments are tasked with fulfilling their fathering roles while facing a host of economic, employment, and educational barriers (McAdoo, 1993; Threlfall, Seay, & Kohl, 2013). A multitude of factors contribute to life stress among these men. The present study considers two such factors that are disproportionately likely to affect this population: experiences of trauma in childhood (Health Resources and Services Administration, 2011) and social instability in emerging adulthood (Cho & Kogan, 2016).

Childhood trauma refers to the experience of maltreatment prior to age 18. Childhood trauma – particularly in the context of individuals facing an already stressful set of environmental circumstances – can have far-reaching consequences for parents’ mental health and ability to form close relationships (Horwitz, Widom, McLaughlin, & White, 2001; Murphy et al., 2014). The majority of African American children are likely to have experienced at least one episode of abuse or neglect (Flaherty et al., 2006). African American men in resource-poor environments are more likely than other men to experience trauma and less likely to access mental health services to support their recovery (Motley & Banks, 2018). African American men are also more likely to report stronger and more persistent childhood trauma than men of other ethnicities, which may account for some health disparities experienced by African American men (Umberson, Williams, Thomas, Liu, & Thomeer, 2014). Further, the effect of childhood trauma on poor health, problematic alcohol and substance abuse, and impaired social relationships is stronger among African American men than among African American women or men of other ethnicities (Umberson et al., 2014; Cross, Crowe, Powers, & Bradley, 2015). Although its association with parenting among African American fathers specifically has not been examined, considerable evidence links childhood trauma to difficulties in relationship functioning in general and parenting in particular in predominantly European American populations (e.g., Danese & McEwen, 2012; Ehrensaft, Knous-Westfall, Cohen, & Chen, 2015). As such, we hypothesize that young, African American fathers who have experienced trauma are more likely to be disengaged from their own children than those without a trauma history.

Social instability is defined by a constellation of life uncertainties that include economic distress, numerous housing transitions, difficulty finding employment, and a lack of education and vocational engagement (Cho & Kogan, 2016). Economic disparities between African American and European American males have been longstanding and well-documented, including lower wages, less wealth accumulation, and substantially lower rates of upward mobility for African American men relative to European American men (e.g., Chetty et al., 2018). Historically, young African American men are more likely to be excluded from the labor market, less likely to be employed or enrolled in school (Holzer & Offner, 2006; Williams, 2008), and more likely to experience housing instability (Kushel, Gupta, Gee, & Haas, 2006; Vijayaraghavan et al., 2013) than their White or Hispanic counterparts.

Thus, many unmarried and rural African American fathers must navigate the transition to parenthood while facing significant social instability in work, education, and living arrangements (Roy, 2005, 2006; Cho & Kogan, 2016). The life stress associated with this instability is especially likely to affect the involvement of fathers who are not in stable, cohabiting relationships with the mothers of their children (Fagan & Palkovitz, 2007; Fagan, Palkovitz, Roy, & Farrie, 2009). Young, unmarried fathers who are struggling to make ends meet while navigating transitory housing and employment situations may in turn struggle to establish fathering identities or consistent, predictable patterns of involvement with their children (Shirani, Henwood, & Coltart, 2012; Woldoff & Cina, 2007). Thus, social instability is hypothesized to undermine African American fathers’ ability to maintain involvement with their children.

Life course perspectives on close relationships emphasize how life stress in childhood exhibits continuity into adulthood (Umberson, Crosnoe, & Reczek, 2010). For example, there is evidence that the psychological challenges associated with childhood trauma could lead to instability in relationships and educational attainment in late adolescence and emerging adulthood (McMahon, 2014), as well as employment-related struggles among young adults (Henry, Fulco, & Merrick, 2018). Notably, family of origin characteristics associated with increased risk for childhood trauma - such as intimate partner violence and substance abuse - are also risk factors for social and relationship problems in the next generation (e.g., Rossow, Felix, Keating, & McCambridge, 2015; Fredland et al., 2015). Although the present study is unable to disentangle the direction of effects among these variables, childhood trauma should be associated with emerging adults’ engagement (or lack thereof) with school, work, or other institutions that support economic and social stability. As such, we hypothesize that reports of childhood trauma will be linked to elevated social instability during the emerging adult transition.

The Role of OXTR Methylation

Epigenetic regulation of gene expression is one mechanism whereby life experiences can become biologically embedded and can exert influences on behavioral and relational outcomes. Epigenetic regulation occurs when biochemical mechanisms are triggered that influence the genome to express (upregulate or downregulate) particular genes (Champagne, 2010). DNA methylation is the addition of a methyl group to a DNA base, downregulating gene expression; it is the most well-known and extensively studied example of an epigenetic modification (Meloni, 2014). An emerging body of research suggests that methylation of the oxytocin receptor gene (OXTR) could be a biological mechanism whereby life stress affects fathering behavior (Unternaherer et al., 2012). Oxytocin is a hormone and neurotransmitter that influences a variety of social and affective processes in human adults, and is linked to the development of fathering behaviors (Feldman et al., 2010; Gordon et al., 2010). Although numerous biological and environmental factors can regulate (up or down) OXTR expression, hypermethylation of OXTR may contribute to reduced OXTR gene expression with implications for close relationships (Harony-Nicolas et al., 2014; Kumsta, Hummel, Chen, & Heinrichs, 2013).

DNA methylation of specific OXTR CpG sites blocks the binding of proteins responsible for transcription, thus decreasing gene expression (Dadds et al., 2014). In mouse models, OXTR methylation has been linked to reduced OXTR expression in both body (Mamrut et al., 2013) and brain tissues (Harony-Nicolas et al., 2014). In humans, hypermethylation of OXTR is associated with inhibited gene expression in liver tissue (Kusui et al., 2001), as well as in peripheral blood cells and the temporal cortex tissue of individuals with autism (Gregory et al., 2009). Several studies have also investigated consequences of OXTR methylation in humans that may be relevant for parenting. Elevated levels of OXTR methylation have been linked to deficits in social processing (Haas et al., 2016) and affective functioning (Puglia, Lillard, Morris, & Connelly, 2015), social-cognitive deficits and problems with emotion recognition (Maud et al., 2018), depression and maladaptive relationship schemas (Simons, Lei, Beach, Cutrona, & Philibert, 2017), greater attachment anxiety (Ebner et al., 2019), and the expression of callous unemotional traits (Cecil et al., 2014). Collectively, this work suggests that OXTR methylation may be a biological mechanism that affects emotional functioning, social-information processing, and affiliative processes in the context of close relationships. Although associations with fathering have not yet been examined, based on the available evidence we anticipate that elevated levels of OXTR methylation will be associated with reduced father involvement.

We also hypothesize that OXTR methylation will be elevated among fathers who have experienced childhood trauma or social instability. OXTR methylation increases in response to childhood maltreatment (Cecil et al., 2014; Unternaehrer et al., 2012), less positive maternal care (Unternaehrer et al., 2015), stressful environments in non-human primates (Baker et al., 2017), and generalized adult adversity in humans (Simons, Lei, Beach, Cutrona, & Philibert, 2017). Recent prospective research with adults has linked life stress and OXTR methylation to cognitive schemas associated with attachment style (Simons et al., 2017) and internalizing problems (Gouin et al., 2017). Notably, there are some conflicting results, with some studies failing to find an association between early life adversity and OXTR methylation (Dadds et al., 2014; Kogan, Cho, Beach, Smith, & Nishitani, 2018). Consistent with the bulk of available evidence, however, we expect that childhood trauma and social instability will be associated with elevated OXTR methylation. Findings linking OXTR methylation to both stressful environments and problems in social relationship functioning also suggest the possibility that it may act as a biological mediator of these associations. We will thus test indirect effect hypotheses in which OXTR hypermethylation connects childhood trauma and social instability to father involvement.

In considering pathways from life stress to OXTR methylation to father involvement, it is also important to disentangle the effects of OXTR methylation on father involvement while accounting for the influence of fathers’ relationships with children’s mothers. OXTR methylation has implications for biological, emotional, and cognitive systems that overlap with parent-child and romantic or co-parental relationships (Maud et al., 2018), which can impact both fathers’ interactions with their children as well as their interactions with their children’s mothers. The behavioral outcomes that have been linked to OXTR methylation thus far (see Maud et al., 2018 for a review) seem likely to affect relationships broadly, including both interparental and parent-child functioning. Thus, the present study also examined whether OXTR methylation was related to fathering indirectly via the quality of fathers’ relationships with their children’s mothers.

The Present Study

The present study builds upon ecological models of fatherhood (e.g., Cabrera et al., 2014; Volling & Cabrera, 2019) by incorporating a biopsychosocial perspective on development to better understand the social and biological determinants of parenting in a hidden population of parents. Specifically, a conceptual model (Figure 1) examining pathways from childhood trauma and emerging adult social instability to father involvement among unmarried, rural, African American fathers via OXTR methylation was tested. We hypothesized that retrospectively reported childhood trauma would be related to elevated social instability, both of which will be linked to increased OXTR methylation. In turn, elevated OXTR methylation will be associated with reduced levels of father involvement and less optimal relationships with their children’s mothers, and will mediate associations from childhood trauma and social instability to father involvement. The test of study hypotheses has two important controls. First is the consideration of mother-father relationships. We will test whether OXTR methylation affects fathering directly and/or indirectly via the quality of fathers’ relationships with the mothers of their children. Second, studies suggest that DNA methylation can be affected by genetic influences. Drawing on recent research linking genetic variation in OXTR to OXTR methylation (Smearman et al., 2016), we control for genetic variation in the OXTR gene on methylation.

Method

Participants and Sampling Procedure

Participants were 192 fathers from the African American Men’s Project (AMP), a longitudinal study of the health risk behavior of young, African American men in the rural South. Men were 19 to 22 years of age (M = 20.18; SD = 1.08) at baseline, and participated in three waves of data collection, with approximately 18 months between waves. All men who reported being fathers at the third wave of data collection were included in the present study. Children’s ages ranged from 1 month to 7 years (M = 3.17 years; SD = 2.31); 52.1% of children were girls and 47.9% were boys. The analytical sub-sample of fathers did not differ significantly from non-fathers in AMP on any of the primary study variables. Fathers in this sample reported lower levels of educational attainment (t = 3.75, p < .01) but higher levels of monthly income (t = −2.42, p = .02) than non-fathers in the overall study. Because participants were selected for inclusion in the sub-sample based on being fathers at wave 3, all participants included in this sub-sample had complete survey data at this wave as well as the prior two waves. A subgroup of 41 fathers (21.4%) did not provide saliva samples (N = 151 total saliva samples; see description below) at wave 2 of the study. There were no significant differences in demographic or study variables for fathers who provided saliva sample vs. those who did not. To ensure adequate power to detect hypothesized effects in our model, we considered statistical power estimates at the model level. Using MacCallum, Browne, and Sugawara’s (1996) model-based power estimate protocols, power was .82 to detect a close-fitting model (RMSEA = .05). This study was approved by the University of Georgia Institutional Review Board, Protocol #2011100958, “The African American Men’s Project”.

Participants were sampled from seven geographically contiguous counties in central Georgia. Targeted counties were non-metropolitan counties in which block group areas were at least 50% rural as defined by the U.S. Census Bureau, and in which at least 25% of the population was African American. Participants were recruited using respondent-driven sampling (RDS), which combines a prescribed chain-referral recruitment method designed to reduce biases commonly associated with network-based samples and to improve approximation of a random sample in hidden populations (Heckathorn, 1997). Community liaisons who were African American and living in one of the target communities oversaw recruitment of initial seed participants. Each participant was then asked to identify three other men in his community from his personal network who met the criteria for inclusion in the study.

Data Collection Procedures

African American community research associates residing in target communities visited participants at their home or a convenient community location. Participants completed an audio computer-assisted self-interview on a laptop computer to eliminate literacy concerns. At the baseline assessment (Wave 1; W1) participants reported on social instability. Approximately 18 months later, when men’s mean age was 22.05 years (SD = 1.30), a follow-up data collection visit was conducted (Wave 2; W2). At W2, participants provided DNA specimens using Oragene Discover OGR-500 kits (DNA Genotek Inc., Ottawa, ON, Canada). Participants rinsed their mouths with tap water and then deposited 2 ml of saliva in the Oragene sample vial. The vial was sealed, inverted, and shipped via courier to a central laboratory in Iowa City, Iowa, where samples were prepared according to the manufacturer’s specifications. A third visit (Wave 3; W3) occurred approximately 18 months later when men’s mean age was 23.30 years (SD = 1.26) during which men reported on father involvement, retrospective childhood trauma, and their relationship with the child’s mother. At W3 the average education completion was a high school diploma (53.2% of the sample), whereas 15.8% had not completed high school, 28.9% had completed some college or an associate’s degree, and 2.1% had completed a bachelor’s degree. Average total monthly income from all sources was $1506 per month, with values ranging from $495 to $4000.

Measures

Childhood Trauma.

At W3, men completed the Childhood Trauma Questionnaire-Short Form (CTQ-SF; Bernstein, Ahluvalia, Pogge, & Handelsman, 1997). The CTQ-SF contains 28 items consisting of 5 subscales that assess retrospective reports of sexual abuse (e.g., “Someone tried to touch me in a sexual way, or tried to make me touch them”), physical abuse (e.g., “People in my family hit me so hard that it left me with bruises or marks”), emotional abuse (e.g., “I thought that my parents wished I had never been born”), physical neglect (e.g., “I didn’t have enough to eat”), and emotional neglect (e.g., “I felt loved,” reverse coded) prior to the age of 16. The response set ranged from 0 (never true) to 4 (very often true). The five subscales were standardized and summed to form a childhood trauma index. Cronbach’s alpha for the total score was .78. Retrospective reports of the CTQ appear to be stable across adulthood and show good predictive validity with a host of psychological and intrapersonal outcomes (e.g., Cammack et al., 2016; Reuben et al., 2016).

Social Instability.

Consistent with research suggesting that the number of risk factors present has a more robust influence on development than does any particular risk factor (Epstein, Botvin, Griffin, & Diaz, 2001; Sullivan & Farrell, 1999), we developed a risk factor index to assess the social instability construct using 5 dichotomous indicators collected at W1. Men reported on the number of times they had moved in the past 6 months (0 = none or once, 1= more than once) and their school enrollment or employment status (0 = either enrolled in school or full time/part-time employed, 1= neither enrolled in school nor employed). Men reported their vocational engagement on a 6-item scale (e.g.: “I have trouble keeping jobs,” “I am a dependable employee”; α =.80; 0 = above the median and 1 = below the median), and economic distress on a 5-point scale (0 = below the median and 1 = above the median), and whether or not they had completed high school (0 = yes, 1 = no).

OXTR Methylation.

Previous research has evaluated methylation at individual CpG sites as well as mean methylation levels across sites in this region (Dadds et al., 2014; Szyf & Bick, 2013). The present study focused on 13 consecutive CpG sites across Chr3: 8810648 to 8810890. These sites are located in Intron 1; a region with demonstrated associations to social behavior and cognition (Gregory et al., 2009; Kumsta et al., 2013). All 13 consecutive CpG sites resulted in reasonable convergence to a single, continuous scale (α = .58), which exceeds convergence in prior research (Dadds et al., 2014).

Accumulating evidence indicates that high-quality methylation profiles can be generated from saliva for both genome-wide (Langie et al., 2017) and candidate gene (Nishitani, Parets, Haas, & Smith, 2018) analyses. In recent years, the use of saliva in studies of methylation has become more common, in part because of technical improvements in saliva sampling and storage (Wren, Shirtcliff, & Drury, 2015). A number of studies confirm that genome-wide DNA methylation profiles from saliva are more than 90% comparable to those from blood (Langie et al., 2017). Other evidence indicates that DNA extracted from saliva is more similar to the methylation patterns observed in brain tissues than is DNA from blood (Smith et al., 2015). Thus, saliva appears to be a good proxy tissue for methylation studies of human behavior.

OXTR Genetic Polymorphisms.

Genotyping at OXTR SNPs was determined using TaqMan SNP Genotyping technology (Life Technologies; Grand Island, NY). Detailed information about extraction and genotyping procedures is available from the authors. Eight candidate SNPs on the OXTR gene were selected based on evidence of association with behavioral outcomes in the literature. These SNPs were: rs1042778, rs237885, rs2268493, rs2254298, rs53576, rs237889, rs2301261, rs2268498. Preliminary analyses examined group differences in methylation for each variation of all OXTR SNPs. These analyses indicated differences in methylation by genetic polymorphisms on two SNPs: rs2301261 and rs2268498. In both cases C-carriers (i.e, CC/CT) showed significantly more methylation than did individuals who were homozygous for the T-allele (i.e., TT), t = −3.30, p < .01 for rs2301261 and t = −5.02, p < .01 for rs2268498. Thus, both of these polymorphisms were dichotomized (CC/CT = 1; TT = 0) and included as covariate predictors of OXTR methylation in the final structural equation model.

Relationship with Child’s Mother.

Three indices of the father’s relationship with their child’s mother at W3 were combined to create a latent variable assessing this construct. Fathers reported on contact frequency by responding to the question, “How often do you see your child’s mother?”. Responses ranged from 0 (Never) to 4 (Every Day). Fathers also reported on the nature of their relationship with the child’s mother on a 4-point scale (1 = We are not friendly; 4 = We are romantically involved on a steady bases). Fathers also completed the Dyadic Trust Scale (DTS; Larzelere & Huston, 1980) regarding their relationship with the child’s mother. The DTS consists of 8 items on a 4-point scale (0 = never; 3 = always) to assess interpersonal trust in the context of close relationships (e.g., “I feel that I can trust her completely”, “I feel that she can be counted on to help me”; α = .75).

Father Involvement.

At W3, father involvement was assessed using a 10-item measure adapted from the Fragile Families and Child Well-Being study (e.g., Carlson & McClanahan, 2010). Fathers were asked how often they participated in a series of caregiving (i.e., “How often did you help your child get dressed?”) and play (i.e., “How often did you play with your child?”) activities, as well as their contributions to financial and in-kind support (i.e., “How often did you buy clothes, diapers, toys, or presents for your child?”) Responses on all items ranged from 0 to 4 and were summed to create a composite father involvement score (α = .91).

Plan of Analysis.

Bivariate correlations among study variables were examined. Next, a measurement model specifying factor loadings on the relationship with child’s mother latent variable was assessed with a confirmatory factor analysis. Structural equation modeling using Mplus v. 7 (Muthen & Muthen, 2012) was used to test the hypothesized model, overall model fit, and direct and indirect pathways. Mplus uses full information maximum likelihood estimation (FIML) to account for missing data. Indirect effects using bias-corrected bootstrapping procedures were examined. Bootstrapping has distinct advantages as an estimation procedure for indirect effects, particularly with modest sample sizes (MacKinnon, Lockwood, & Williams, 2004). Indirect effects are considered significant if 95% confidence intervals do not contain zero.

Results

Means, standard deviations, and bivariate correlations among study variables are presented in Table 1. Correlations indicated few associations between demographic and primary study variables. Social instability was related to lower education and income, and fathers had more contact with their children’s mothers when the child was young. A number of significant correlations among primary study variables were observed. Childhood trauma and OXTR methylation were associated with reduced levels of father involvement, whereas all three indicators of relationship quality with the child’s mother were associated with higher levels of father involvement. Childhood trauma was associated with increased social instability and reduced dyadic trust. OXTR methylation was also associated with increased social instability and less frequent contact with the child’s mother.

Table 1.

Descriptive Statistics and Bivariate Correlations Among Study Variables

	1	2	3	4	5	6	7	8	9	10	11
Primary Model Variables
1. Childhood Trauma (W3)	-
2. Social Instability (W1)	.21**	-
3. OXTR Methylation (W2)	.00	.21*	-
4. Contact w/ Child’s Mother (W3)	−.05	.02	−.17*	-
5. Relationship w/ Child’s Mother (W3)	−.08	.00	−.06	.65**	-
6. Dyadic Trust (W3)	−.22**	−.04	.02	.35**	.45**	-
7. Father Involvement (W3)	−.23**	−.05	−.24**	.45**	.41**	.17*	-
Demographics
8. Father Age (W3)	.00	.06	.04	−.11	.04	−.08	−.06	-
9. Father Education Completed (W3)	−.14	−.47**	−.06	−.05	.06	.04	−.04	.12	-
10. Father Monthly Income (W3)	−.14	−.16*	−.01	−.07	.05	−.03	−.09	.09	.06	-
11. Child Age (W3)	.15	.07	−.05	−.22*	.11	−.06	−.19	.28**	−.03	−.03	-
Mean	.67	1.92	.25	2.94	2.37	20.79	16.89	23.30	4.51	1506.29	3.17
SD	4.32	1.25	.04	1.15	1.10	4.37	6.47	1.20	1.50	1239.99	2.31

Note:

^*p < .05

^**p < .01

A confirmatory factor analysis indicated that each of the observed indicators of relationship quality with mother loaded significantly onto a single latent construct. The standardized factor loadings for contact frequency, nature of relationship, and dyadic trust were .71, .50, and .91, respectively (all p < .001). Thus, analyses proceeded with the structural model including the relationship with child’s mother as a latent variable.

Structural equation modeling results are presented in Figure 2, with standardized path coefficients. The overall model fit was good, χ² = 24.92 (20), p = .20; CFI = .98; RMSEA = .04 and explained a substantial proportion of the variance in father involvement at W3 (R² = .29, p < .01). Childhood trauma was associated directly with low levels of father involvement and elevated levels of social instability. Social instability was associated with elevated levels of OXTR methylation, which in turn was associated with low levels of father involvement. The quality of the relationship with the child’s mother was associated with increased father involvement. Notably, the association between OXTR methylation and father involvement remained significant even after accounting for the robust association between relationship with child’s mother and father involvement. Covariate paths from OXTR SNPs (rs2301261 and rs2268498) to OXTR methylation were also significant, indicating that, for both SNPs, individuals with the CC/CT genotype showed more methylation than those with the TT genotype.

Figure 2.

Structural Equation Model Testing Study Hypotheses

Tests of indirect effects using bias-corrected bootstrapping procedures (10,000 samples) indicated a significant indirect effect from childhood trauma to OXTR methylation via social instability, indirect effect = .05, 95% CI: [.01, .11]. Further, there was a significant indirect effect of social instability on father involvement via OXTR methylation, indirect effect = −.05, 95% CI: [−.01, −.10], suggesting that the link between social instability and father involvement was mediated through OXTR methylation. In contrast, the direct link between OXTR methylation and relationship with the child’s mother was not significant, nor was the indirect effect linking OXTR methylation to father involvement via relationship with child’s mother, indirect effect = −.05, 95% CI: [−.15, .03].

Discussion

Findings support the proposition that OXTR methylation may link social instability to father involvement among rural, unmarried, African American fathers. Childhood trauma was associated with low levels of father involvement, as well as increased social instability. Social instability was not associated directly with father involvement, but was linked to elevated OXTR methylation independent of genotype. OXTR methylation was associated with increased father involvement and mediated the link between social instability and father involvement, independent of the quality of the father’s relationship with the child’s mother. To our knowledge, results are the first to report associations between DNA methylation and fathering behavior in humans, and have important implications for the understanding of biopsychosocial mechanisms underlying father involvement in vulnerable populations.

The Influence of Childhood Trauma on Social Instability and Father Involvement

Study findings indicated that men who retrospectively reported more childhood trauma also reported (a) elevated levels of social instability, and (b) less father involvement. The influence of childhood trauma on father involvement may in part operate through its effects on social instability. The damage done by childhood trauma can affect relational and psychological functioning in many domains (e.g., work, school, home). Social instability is a particularly critical factor given the developmental timing and context of young men in the current study. Young, African American fathers in resource-poor communities face a myriad of risk factors in early emerging adulthood. Those that experience trauma in early childhood may find it especially difficult to develop adaptive relationships and coping abilities to surmount those challenges. The downstream consequences of childhood trauma could manifest in an inability to effectively engage with stabilizing social relationships and institutions (e.g., work, school) in addition to difficulties in committed relationships with one’s children. The long reach of childhood trauma among vulnerable, African American men warrants further attention in subsequent research.

That the direct effect of childhood trauma on father involvement remained even after accounting for other relational and biological factors is consistent with other work documenting the long-term effects of adverse childhood experiences on parenting (e.g., Roberts, O’Connor, Dunn, Golding, & Team, 2004), and the detrimental impact of early trauma among African American men specifically (e.g., Janusek, Tell, Gaylord-Harden, & Mathews, 2017). Findings are among the first to link childhood trauma to fathering directly, and suggest that abuse and neglect in childhood may have lasting, intergenerational effects on men’s involvement with their children.

OXTR Methylation and Father Involvement

This study is the first that we know of to link DNA methylation to fathering. Specifically, increased methylation in the promoter region of the OXTR gene was related to low levels of father involvement. These results provide supporting evidence for the oxytocin system as particularly salient for the development of father-child relationships (e.g., Feldman et al., 2010; Gordon et al., 2010). Beginning with early groundbreaking work in animal models, epigenetic mechanisms with lifecourse-relevant consequences for mothers’ parenting have been examined (Champagne, 2018). Our findings suggest that fathering also may be a behavioral outcome affected by epigenetic modification. Functional studies examining the interplay between OXTR methylation, gene expression, and the behavior of oxytocin in the body are rare (e.g. Maud et al., 2018), and caution is warranted in interpreting the present findings in the absence of measures of gene expression or endogenous oxytocin. Nonetheless, our results raise the possibility that when the OXTR gene is highly methylated men in challenging environments may have difficulties engaging with their young children.

Consistent with a body of research linking interparental relationship quality and fathering (e.g., Fagan & Cherson, 2017; Fagan, Levine, Kaufman, & Hammar, 2016), the quality of fathers’ relationships with their children’s mothers was related robustly to fathers’ involvement with children. Fathers’ access to children is determined to some extent by the ability of those fathers to engage in stable, trusting relationships with the mothers of their children. Nonetheless, the effects of OXTR methylation on father involvement were independent of the quality of the father’s relationship with the child’s mother. Thus, the present study suggests that OXTR methylation could impact men’s relationships with their children directly, regardless of whether they have frequent contact or close, trusting relationships with their children’s mothers.

Social Instability and OXTR Methylation

Further, social instability was related to increased OXTR methylation, which mediated the link between social instability and father involvement. Young fathers in this sample were making the transition to parenthood while navigating the demands of entering the workforce in rural communities with relatively few opportunities (Gore & Aseltine Jr, 2003). Men who experience volatility in their work and home lives, and uncertainty in their ability to make ends meet, may have few financial and/or psychological resources to devote to involved and consistent parenting. These stressors forecast increased DNA methylation in OXTR, potentially affecting the oxytocin system in general. Although some prior research has suggested that the association between life stress and OXTR methylation may be stronger in females than males (Gouin et al., 2017), results of the present study suggest that this link could extend to fathers.

For men who are dealing with social instability, the biological and behavioral consequences of OXTR methylation may be an additional difficulty that can impede father involvement. Future work should continue to explore the role of methylation in general - and OXTR methylation in particular - as a mediating mechanism from life stress to father involvement, particularly among vulnerable populations of fathers, whose contributions to child development are both critical and often overlooked.

Implications for Unmarried, Rural, African American Fathers

These findings have implications for our understanding and perceptions of rural, African American fathers’ lived experiences and parenting contributions. Results are consistent with evidence suggesting that life stress experienced by young, African American men can get “under the skin” to affect numerous biological systems (Brody, Yu, Miller, & Chen, 2015). Paternal absence from children’s lives is not always a conscious choice. Low father involvement may instead be a product of lifecourse-persistent stress that is embedded within biological systems that affect men’s parenting. Future research examining father involvement from a biopsychosocial perspective (e.g., Gordon & Feldman, 2015) may continue to inform long-neglected research focused on father-child relationships and strengths-based programs targeting father-child relationships in rural, unmarried, African American populations.

More broadly, results support ecological and life course perspectives on father involvement in rural, African American families. The interplay among complex systems including family of origin history, engagement with employment/vocational contexts, educational opportunities, and epigenetic modifications combine to alter men’s ability and willingness to engage with young children. Vulnerabilities in one domain seem likely to lead to increased risks in other domains. Though this interconnectedness may (in some cases) lead to circumstances too challenging to overcome, it also suggests the possibility that preventive interventions targeting one domain or ecological context may have downstream consequences for others, including malleable and sometimes reversible epigenetic processes.

Limitations and Future Directions

This study provides new insight on the mechanisms underlying the development of father involvement in a hidden population. Nonetheless, there are a number of limitations to this study that should be addressed in future research. Key among these is further elucidating the direction of effects among study variables. Despite the longitudinal design employed in this study, it remains difficult to ascertain the exact timing of when methylation occurs, and hence the direction of effects among methylation, social instability, and father involvement. Evidence of methylation can be seen prenatally or shortly after birth (e.g., Rijlaarsdam et al., 2017), and is reversible (Ramchandani, Bhattacharya, Cervoni, & Szyf, 1999). These issues further obscure clarity in direction of effects. Moreover, although it is assumed that methylation affects father involvement, it may well be that aspects of the father-child relationship are in part responsible for OXTR methylation. This interpretation cannot be fully dismissed without prospective, longitudinal studies examining associations among these variables. Future studies incorporating repeated assessments of both fathering behavior and DNA methylation are essential for disentangling directions of effects among these variables.

Relatedly, the present study included a measure of OXTR methylation and several candidate SNPs, but did not collect data on gene expression or circulating oxytocin. This limitation prevents us from understanding biological pathways from epigenetic modification to behavioral outcomes. Further, there remains some debate regarding whether methylation from saliva samples is seen in brain tissue (Smith et al., 2015). Unlike genetic studies, epigenetic modifications are tissue-specific, prompting preclinical studies of behavior to focus on DNA from brain tissues. However, research with humans must rely on DNA extracted from proxy tissues. The extent to which methylation from saliva samples will be evident in human brain tissue is still not yet well-understood.

The oxytocin system appears to be an important biological system for understanding fathering. But recent research indicates other aspects of the neuroendocrine system, including cortisol and testosterone (e.g., Bos et al., 2018; Kuo et al., 2018) may play a role in fathering, and DNA methylation at other sites relevant to these systems could be critical. We hope that this may be the first of many studies integrating epigenetics and fathering, and more comprehensive assessments of neurobiological functioning in the future could continue to elucidate the mechanisms underlying father-child relationships in vulnerable contexts.

Other measurement limitations also render interpretations regarding causality speculative. The use of retrospective reports of childhood trauma at Wave 3 is one such limitation. Future studies that incorporate earlier assessments of childhood trauma could further elucidate the links between trauma and both social instability and OXTR methylation. Moreover, objective assessments such as public records of reported maltreatment may provide stronger evidence for these associations. In general, the present study relied on a single informant for survey data. As such, the risk of artificially inflated associations among variables assessed at Wave 3 due to shared-method variance must be noted. Future investigations incorporating multi-method approaches ideally in larger samples with more statistical power would be beneficial.

Finally, although this study is among the few to assess the parenting behavior of rural, unmarried, African American fathers directly, the measure of father involvement employed in this study was limited to quantitative assessments of time spent with children and financial contributions. More nuanced assessments of fathers’ parenting that tap into the quality of proximal interactions and relationships with children may further document the myriad ways in which these fathers contribute to children’s development, and the contextual and biological factors that are responsible for positive fathering. Further, given key differences in the correlates and developmental consequences of different types of fathering activities (e.g., caregiving vs. play) (Brown, Mangelsdorf, Shigeto, & Wong, 2018), more comprehensive and multidimensional assessments of father involvement would be fruitful for elucidating the biological and contextual correlates of different parenting domains for fathers.

Conclusion

The present study informs a deeper understanding of the hidden population of rural, unmarried, African American fathers by elucidating contextual factors that may foster or inhibit involvement with their young children. Results suggest that childhood trauma has implications for father involvement and social instability, and that the consequences of social instability for father involvement may in part be biologically mediated by epigenetic modification of the OXTR gene. Interventions targeting the oxytocin system via intranasal administration (e.g., Naber et al., 2010) could be useful, but so too might preventive interventions tailored to the unique contexts in which rural, African American fathers reside. The contributions of these men to their children’s development have been largely overlooked; research and practice should recognize that they can and do engage with children in ways that contribute to healthy development. Clearly, those contributions are dependent upon creating safe early environments that provide nurturant care in childhood and promote stability during critical life transitions in emerging adulthood. Those environments can equip fathers with the biological and psychosocial tools to successfully engage in rewarding and meaningful interactions with their children. Future research identifying the particular needs of rural, unmarried, African American fathers will continue to inform strengths-based programs serving this overlooked population of parents.