Examining Patterns of Family Resilience and Neighborhood Cohesion as Moderators of the Effects of Adverse Childhood Experiences on the Mental Health of Black Adolescents

Abstract

Adverse childhood experiences (ACEs) confer risk to the mental health of Black youth, but few studies have examined how youth gender, family, and neighborhood factors jointly influence the psychological impact of adversity. This study investigates if family resilience and neighborhood cohesion jointly moderate the link between latent ACE profiles and mental health among Black girls and boys. This study uses data from the National Survey of Children’s Health, combined across the years 2016 through 2021, and includes a nationally representative sample of 5,493 Black youth (48% female) between the ages of 12 and 17. Two patterns of ACEs were identified using latent class analysis characterized by no-to-minimal ACE exposure and moderate-to-high ACE exposure. Membership in the high-ACEs class increased the risk for internalizing problems among Black boys (b = 0.56, p < .001) and girls (b = 0.42, p <.01). Only boys in the high-ACEs class who also reported low levels of family resilience and low neighborhood cohesion evidenced an increased risk for externalizing concerns (b = 0.70, p < .001). Conversely, only girls in the high-ACEs class who reported high levels of family resilience and low levels of neighborhood cohesion evidenced an increased risk for externalizing problems (b = 0.69, p < .01). Findings suggest that the impact of ACEs on mental health is not uniform across Black boys and girls, and that family and neighborhood-level factors may collectively shape the impact of ACEs on the mental health among Black youth in unique ways.

General Scientific Summary

This study suggests that the absence of family resilience and neighborhood cohesion increases the risk for externalizing problems among Black boys with a chronic history of adverse childhood experiences. Further, results suggest that high levels of family resilience coupled with low levels of neighborhood cohesion increase the risk for internalizing problems among Black girls with a chronic history of adverse childhood experiences.

Exposure to adverse childhood experiences (ACEs), or potentially traumatic experiences that occur prior to the age of 18 (e.g., experiencing abuse or neglect, witnessing violence, or experiencing discrimination), pose a significant public health concern and problem to the mental health and well-being of youth. Building from the seminal work of Felitti et al. (1998), decades of scholarship have produced clear evidence demonstrating that ACEs exhibit a dose-dependent effect on health, with greater ACE exposure being associated with negative psychological, physiological, and behavioral functioning across the life course (Hughes et al., 2017). Alarmingly, recent research has found that by the age of 9, nearly 95% of American youth have experienced at least one ACE (X. Zhang & Monnat, 2022), signaling that childhood exposure to potentially traumatic experiences is unsettlingly common. While all youth regardless of sociodemographic characteristics are at risk of being exposed to ACEs, chronic ACE exposure is not uniform across racial and ethnic groups. Indeed, prior research has documented that structural inequities within the United States position Black youth to be more likely to be exposed to multiple ACEs relative to their White peers, and subsequently at greater risk for diagnosable mental health concerns (Bernard et al., 2021).

Despite the prevalence and nefarious implications of ACE exposure, it is important to note that not all Black youth who experience adversity exhibit poor psychological outcomes (Volpe et al., 2022). As such, there is a critical need for research to elucidate factors that may buffer or mitigate the impact of ACEs on the mental health of Black youth. To this end, literature has highlighted that Black boys and girls may respond differently to ACEs (Jones et al., 2022), and that developmental health responses to ACEs may vary as a function of family resilience and neighborhood cohesion (McGowan et al., 2023; Uddin et al., 2020). However, questions remain regarding how the gendered impacts of ACEs might be shaped by family and neighborhood-level dynamics, which undermines the development of targeted prevention and intervention strategies that can mitigate the effects of ACEs on health. Thus, the purpose of the current study is to examine the degree to which patterns of family resilience and neighborhood cohesion moderate the association between ACEs and mental health within a large sample of Black girls and boys.

ACEs and Mental Health Among Black Youth

The deleterious psychological effects of ACEs among youth have been replicated across a range of populations, particularly in regard to internalizing and externalizing concerns (Petruccelli et al., 2019). The association between ACEs and mental health is arguably most pronounced among Black youth, who are more likely to be exposed to multiple ACEs, including unique forms of adversity associated with their historically marginalized background including discrimination, policing, and exposure to community violence (Hampton-Anderson et al., 2021). Extant literature has produced clear evidence highlighting the gradient relationship between ACEs and elevations in internalizing and externalizing symptomatology among Black youth. As an example, when using an expanded ACE assessment that included unique forms of adversity that marginalized youth may be more likely to be exposed to (e.g., racism, poverty, and neighborhood violence), one study found that 51% of Black youth reported exposure to four or more ACEs (Freeny et al., 2021). In this same study, the authors also found that individuals reporting four and above ACEs were at greater risk for depressive symptoms relative to those who reported fewer ACE exposures. In a different study, Black adolescents reporting five ACEs were found to be 2.5 times more likely to report substance use problems, compared to White adolescents reporting the same number of ACEs (Johnson et al., 2023). Notably, the link between ACEs and negative mental health outcomes among Black youth has been documented as early as age nine (Bernard et al., 2022), and has been found to persist over time across sensitive developmental periods (Hicks et al., 2020).

ACEs, Mental Health, and Gender

There is increasing, albeit equivocal evidence, that demonstrates that ACEs may differentially impact youth mental health as a function of gender. To date, scholarship has largely emphasized gender differences in the context of externalizing symptoms, yet findings in this area have been conflicting. While some scholarship using large-scale, predominantly Black youth samples have found that ACEs tend to be associated with more pronounced elevations in externalizing symptoms among boys (Leban, 2021), other work has found the inverse to be true with the ACE-externalizing symptom link being only significant for girls (Hunt et al., 2017). This general lack of agreement in findings pertaining to gender differences in psychological responses to ACEs also extends to internalizing symptoms. As an example, Foster et al. (2004) found that the impact of ACEs related to community violence exposure on anxiety and depression was more pronounced for Black girls relative to Black boys. Similarly, in examining the impact of ACEs on health in a community sample of youth, Meeker et al. (2021) found that compared to male adolescents, female adolescents reported significantly higher depressed mood, self-injury, suicidal ideation, and suicide attempts in relation to ACEs. Yet, subsequent work among diverse samples of children and adolescents has found that ACEs are not only more strongly associated with elevations in internalizing concerns for boys but also a greater breadth of internalizing symptoms (Gajos et al., 2022; Jones et al., 2022). Collectively, these equivocal findings underscore the need for additional work to clarify the significance of gender in the context of ACEs and mental health among Black youth.

Conceptualizing ACEs Through Latent Variable Approaches

Although cumulative risk conceptualizations of ACEs have produced unequivocal evidence linking adversity to poor health outcomes, an overreliance on conventional count scores may oversimplify and minimize the effects of adversity on health. Scholars who warrant caution when using count score approaches highlight that such conceptualizations of adversity obfuscate data which signals that ACEs evidence differential impacts on health and may group together in distinct ways to impact health (Briggs et al., 2021; Ford et al., 2014). Awareness of this issue has resulted in a proliferation of research conceptualizing ACEs through latent class analysis (LCA) approaches in order to delineate how heterogeneous patterns of ACEs uniquely shape risk for developmental outcomes (Lanier et al., 2018; Maguire-Jack et al., 2019; Wang et al., 2022).

To date, we are unaware of any within-group investigations that explicitly illuminate how different patterns of ACEs are uniquely associated with the mental health of Black youth. That said, there are a number of cross-race comparative studies that underscore the need for this work, which demonstrate that patterns of ACEs and their associations with mental health outcomes are considerably different among Black youth relative to youth from other racial and ethnic backgrounds. For example, Liu et al. (2018) identified a two-class solution for ACE exposure among Black youth, whereas a three-class solution best fit the data for White and Latinx youth. Similarly, Maguire-Jack et al. (2019) identified a three-class solution for ACE exposure among Black youth, whereas a six-class and five-class solution best fit the data for White and Latinx youth, respectively. Most recently, N. Zhang et al. (2022) found a three-factor model of ACE exposure among Black youth, while a four-class model best described the data for White and Latinx youth. Notably, each of these studies identified two specific profiles that differentiated youth who had been exposed to a low number of ACEs from youth who had been exposed to a high number of ACEs, regardless of racial or ethnic background. Despite the generalizability of these profiles, each of these studies also highlighted that Black youth evidenced greater odds of being classified into the high ACE exposure category, and also the greatest odds of reporting poor mental health outcomes relative to White and Latinx youth (Liu et al., 2018; Maguire-Jack et al., 2019; N. Zhang et al., 2022).

Significance of Family Resilience and Neighborhood Cohesion in the Context of ACEs

The significance of the family system as a proximal influence on the development and well-being of youth has been well documented (Bronfenbrenner, 1977). One specific family characteristic that is important to consider in the context of ACEs and mental health is family resilience, broadly defined as a family’s capacity to adaptively cope, process, and maintain positive functioning in the face of adversity (Walsh, 2003). Among children and adolescents, family resilience has been associated with a number of positive outcomes among youth (e.g., flourishing, self-regulation, and family connection), which can collectively offset the deleterious impact of ACEs (Bethell et al., 2019; Westphaln et al., 2022). Thus, it follows that family resilience has been argued to be among the most critical sources of protection that facilitate the successful negotiation of adversity among Black families (Hollingsworth, 2013). In support of this notion, prior work among Black youth has shown that positive, supportive, and cohesive family environments can mitigate the deleterious effects of traumatic stress reactions (Kiser et al., 2010). Such findings support prior theories highlighting the significance of Black families as determinants of youth psychological adjustment in the aftermath of stress and trauma (Murry et al., 2018), and also echo a larger body of literature that has shown that the presence of family resilience can buffer the degree to which ACEs are associated with internalizing and externalizing symptoms among youth (Uddin et al., 2020).

Beyond proximal familial characteristics, community factors have also been found to modulate the impact of ACEs on youth health. In particular, higher levels of neighborhood cohesion, defined as a sense of trust and feelings of kinship among community members (Sampson et al., 2002), have been shown to promote positive mental health outcomes among Black youth in the face of adversity. As an example, DiClemente et al. (2018) found that Black adolescents who reported higher levels of neighborhood cohesion were more likely to report elevations in psychosocial well-being (e.g., self-esteem and positive affect) following exposure to violence, relative to those who reported lower levels of neighborhood cohesion. Further, in line with recent expansions of the ACE framework to include culturally relevant manifestations of adversity (Bernard et al., 2021; Cronholm et al., 2015), scholarship has also documented that greater perceptions of neighborhood cohesion can buffer the impact of racial discrimination on depressive symptoms among Black adolescents (Saleem et al., 2018). Notably, though not examined among Black youth, work by Kingsbury et al. (2020) found that greater perceptions of neighborhood cohesion attenuated the association between ACEs and internalizing and externalizing symptoms and suicidality among youth over a 2-year period.

Although the above-mentioned scholarship clearly positions family resilience and neighborhood cohesion as unique protective factors in the face of adversity among youth, there remains a paucity of literature examining how the interplay between these processes may mitigate the impact of ACEs on health.

Current Study

Three limitations constrain our understanding of the association between ACEs and mental health among Black youth. First, research that has used LCA approaches to identify different classes of ACEs among Black youth has done so primarily in efforts to show that ACE exposure, and subsequent risk for mental health problems, is not uniform across racial and ethnic groups. Given the comparative nature of these studies, the within-group variability that may exist in unique ACE groupings has been missed, making it difficult to understand what unique profiles of ACEs may exist among Black youth and how such profiles are associated with Black youth mental health. Second, and perhaps relatedly, examinations as to the gendered effects of ACEs on the mental health of Black youth remain relatively unclear, in part, because of the lack of consistency in how ACEs are operationalized and measured. As such, further research is needed to clarify the unique associations between ACEs and mental health among Black boys and girls, to inform tailored intervention and prevention efforts. Third, although family resilience and neighborhood cohesion have been identified as unique protective factors that can offset the deleterious effects of ACEs, we are unaware of any studies that have looked at these factors in tandem among Black youth. Yet, sociological perspectives of resilience call for the joint examination of protective factors across neighborhood and family level processes, as their synergistic influence may confer greater benefits in the context of ACEs compared to when examined in isolation (Liu et al., 2020).

In light of these limitations, the present study aimed to understand how familial and community-level factors jointly influence the association between ACEs and mental health within a nationally representative of Black girls and boys. Consistent with efforts to move beyond cumulative risk conceptualizations of ACEs, the first aim of the study was to elucidate distinct patterns of ACE items using parent-reported data on youth experiences. Based on conflicting reports in the literature, it is difficult to hypothesize a priori the number of classes that will emerge. However, we expected to find at least two classes that are characterized by “high adversity” and “low adversity,” which has been consistently replicated within prior work. The second aim of the study was to examine how co-occurring patterns of family resilience and neighborhood cohesion jointly moderated the association between ACEs and mental health among Black boys and girls. In line with prior work suggesting that the combined presence of positive family and community-level resources are associated with reductions in ACE-related mental health concerns among youth (Liu et al., 2020), we hypothesize that the joint presence of high levels of family resilience and neighborhood cohesion would significantly attenuate the association between ACEs and mental health among Black boys and girls.

Method

Data Source and Sample

The current study leveraged public-use data, combined across years 2016 through 2021, from the National Survey of Children’s Health (NSCH; Child and Adolescent Health Measurement Initiative, 2016–2021). The NSCH offers nationally representative, annual cross-sectional samples of noninstitutionalized children ages 0–17 in the United States. The data source offers rich, caregiver-report data related to various domains of children’s health, access to quality health care, and aspects of the child’s social environment (e.g., family, neighborhood, school, and larger social contexts). To address our research questions, we focused on children who were identified as non-Hispanic Black or African American. We also limited the analytic sample to include only focal children between the ages of 12 and 17 (i.e., adolescents) who had sufficient information regarding ACEs (i.e., were not missing values across all ACE items) and other key model variables. This study was not preregistered.

The final analytic sample included 5,493 respondents. In terms of unweighted descriptive information, the focal children had an average age of 14.66 years (SD = 1.69 years), and about 48% identified as female. Roughly 46% of caregiver respondents indicated having a college degree or more education; about 30% indicated having completed some college or an associate degree; 19% indicated completing high school; the remaining 5% indicated having less than a full high school education. About 82% of caregiver respondents indicated being the biological or adoptive parent of the focal child, whereas 11% indicated being a grandparent, 2% indicated being a stepparent, and the remaining 5% indicated being either a foster parent, relative, or nonrelative. See Table 1 for more details regarding the demographic composition of the sample.

Table 1.

Sample Description

Variable	% or M	SD	Min.	Max.
Youth age	14.66	1.69	12	17
Youth is female	0.48
Youth physical healtha	4.34	0.84	1	5
Caregiver mental healtha	4.04	0.97	1	5
Caregiver physical healtha	3.62	0.97	1	5
Youth insurance coverage	0.94	0.24	0	1
Caregiver educational attainment
Less than high school	0.05
High school	0.19
Some college or an associate degree	0.30
College degree or more	0.46
Youth internalizing problems	0.11
Youth externalizing problem	0.17
ACE latent classes
High ACEs	0.18
Low ACEs	0.82
Family resilience/connection and neighborhood cohesion groupings
HF–HC	0.29
HF–LC	0.22
LF–HC	0.17
LF–LC	0.32

Note. SD, minimum values, and maximum values are only reported for continuous variables. ACE = adverse childhood experiences; Min. = minimum; Max. = maximum; HF = high family resilience and connection index value (LF = low); HC = high neighborhood cohesion (LC = low).

^aValues of 1 = poor, values of 5 = excellent.

Measures

ACEs

Nine caregiver-reported ACE items were used as observed indicators within the present analyses (to be described in more detail below) that align with commonly used measures of ACEs (Bethell et al., 2017). The first ACE item asked caregivers the following: “since this child was born, how often has it been very hard to cover the basics, like food or housing, on your family’s income?” Response options included never, rarely, somewhat often, and very often. We dichotomized this item, such that responses of somewhat often or very often were coded as 1, and responses of never or rarely were coded as 0. For the remaining eight ACE items, caregivers were asked to indicate, to the best of their knowledge, whether the focal child ever experienced any of the following: parent or guardian divorced or separate; parent or guardian died; parent or guardian served time in jail; saw or heard parents or adults slap, hit, kick, punch one another in the home; was a victim of violence or witnessed violence in their neighborhood; lived with anyone who was mentally ill, suicidal, or severely depressed; lived with anyone who had a problem with alcohol or drugs; and treated or judged unfairly because of their race or ethnic group. These eight items were dichotomous, coded such that a value of 1 indicated an affirmative response (i.e., yes) and 0 indicated a nonaffirmative response (i.e., no).

Youth Mental Health

Two core measures of youth mental health, one reflecting youth internalizing problems and another reflecting youth externalizing problems were drawn from NSCH data. In terms of internalizing problems, caregivers were asked whether a doctor or other health care provider ever told them that the focal child had depression or anxiety, and if so, whether the focal child had the condition at the time of survey completion. Respondents were coded as 1 if they indicated yes to the focal child having current depression or current anxiety; respondents were coded as 0 otherwise. About 11.2% (n = 616) of youth in the analytic sample had internalizing problems.

With respect to externalizing problems, caregivers were asked whether a doctor or other health care provider ever told them that the focal child had attention-deficit/hyperactivity disorder (ADHD) or behavioral/conduct problems, and if so, whether the focal child had the condition at the time of survey completion. Although ADHD represents a neurodevelopmental disorder that can precede ACE exposure, chronic ACE exposure can also significantly elevate the risk for an ADHD diagnosis and the severity and ADHD-related behavioral problems, even after accounting for other individual level and familial ADHD risk factors (Brown et al., 2017; Walker et al., 2021). There is also longitudinal evidence to support the idea that ACEs can shape ADHD risk, with one study finding that ACEs experienced before age 5 and during middle childhood, were associated with the presence of an ADHD diagnosis at age 9 (Jimenez et al., 2017). These findings strongly align with results from a meta-analysis of 70 studies examining associations between ACEs and ADHD, wherein N. Zhang et al. (2022) found that exposure to one, two, and three or more ACEs increased the risk for ADHD symptoms by 1.51, 1.99, and 2.87 times. Considering this evidence, we conceptualize the presence of ADHD as an externalizing problem that could be associated with chronic ACE exposure. Respondents were coded as 1 if they indicated yes to the focal child having current ADHD or current behavior/conduct problems; respondents were coded as 0 otherwise. About 16.8% (n = 923) of youth in the analytic sample had externalizing problems. Although we could have focused on whether any of these internalizing or externalizing conditions were present for the focal child at any point in the past, it was important to ensure temporal precedence between the ACE measures and measures of youth mental health—an advantageous feature that will become clearer after we describe our analyses.

Grouping Variable (Moderator)

We constructed a grouping variable drawing from information about youths’ family environment (i.e., family resilience and connection index; Bethell et al., 2019; Song et al., 2021) and neighborhood environment (i.e., neighborhood cohesion). Turning first to the family environment, we constructed the family resilience and connection index (congruent with the approach used by Song et al., 2021), which incorporated information from six items. The first four items measured family resilience as follows: “family members talk together about what to do when the family faces problems”; “family members work together to solve the problem when the family faces problems”; “family members know we have strengths to draw on when the family faces problems responses”; and “family members stay hopeful even in difficult times when the family faces problems.” Response options for these items included “all of the time,” “most of the time,” “some of the time,” and “none of the time.” The remaining two items associated with the family resilience and connection index captured information about caregiver–child connection and caregiver coping as follows: “How well can you and this child share ideas or talk about things that really matter?” and “How well do you think you are handling the day-to-day demands of raising children?” Response options for these items included “very well,” “somewhat well,” “not very well,” and “not well at all.” To create the index, we assigned one point for each time a respondent indicated “all of the time” to the four family resilience items and for each time a respondent indicated “very well” to the remaining two items. We then created a count variable using these scores (ranging from a possible score of zero to six for each respondent), following which we categorized respondents into either low-scoring (scores of 0–3; coded as 0) or high-scoring (scores of 4–6; coded as 1) groups.

Turning to neighborhood cohesion, we incorporated the following three items: “People in this neighborhood help each other out,” “We watch out for each other’s children in this neighborhood,” and “When we encounter difficulties, we know where to go for help in our community.” Response options for these items included “definitely agree,” “somewhat agree,” “somewhat disagree,” and “definitely disagree.” Consistent with how neighborhood cohesion is constructed and reported by the Data Resource Center for Child and Adolescent Health (Child and Adolescent Health Measurement Initiative, 2023), focal children were coded as having high levels of neighborhood cohesion (coded as 1) if caregiver respondents indicated “definitely agree” to at least one of three neighborhood cohesion items while also indicating “somewhat agree” or “definitely agree” to the other two items. Otherwise, respondents were coded as 0, reflecting relatively low levels of community cohesion.

Finally, we created the final grouping variable by combining the low-high dichotomous variables for family resilience/connection and neighborhood cohesion. This resulted in four specific and distinct groups: (a) one with high levels of both family resilience/connection and neighborhood cohesion (hereafter referred to as HF–HC; n = 1,585), (b) one with high levels of family resilience/connection but low levels of neighborhood cohesion (hereafter referred to as HF–LC; n = 1,239), (c) one with low levels of family resilience/connection but high levels of neighborhood cohesion (hereafter referred to as LF–HC; n = 923), and (d) one with low levels of both family resilience/connection and neighborhood cohesion (hereafter referred to as LF–LC; n = 1,746).

Covariates

When assessing associations between ACEs and youth mental health, we accounted for several relevant covariates that have been historically associated with ACE exposure and mental health outcomes among youth. Specifically, we incorporated measures of youth age (continuous in year units), youth physical health (reverse coded such that 5 = excellent and 1 = poor), caregiver physical health (reverse coded such that 5 = excellent and 1 = poor), caregiver mental health (reverse coded such that 5 = excellent and 1 = poor), youth insurance coverage (1 = youth has any kind of health insurance, 0 = otherwise), and caregiver educational attainment (1 = college degree or more,0 = otherwise). See Table 1 for descriptive information related to all model variables.

Data Analysis

To begin, we conducted LCA using the nine ACE items as observed indicators (Collins & Lanza, 2009). Our aim was to assess whether there were distinct patterns, or latent classes, of ACE-item responses in the sample, and if so, ascertain how many such patterns were evident in the data. Identified patterns would then be used as the focal predictor of youth mental health in subsequent analytic models. Class enumeration, or the process of determining the optimal number of latent classes, was carried out by contrasting model output from solutions with an increasing number of latent classes specified. The following pieces of information were examined jointly to evaluate absolute and relative fit across possible solutions: Akaike information criterion, Bayesian information criterion, sample-size adjusted Bayesian information criterion, several likelihood ratio tests (compares relative fit between k − 1 and k number of classes), the size of the smallest latent class (a strategy for assessing whether a latent class solution represents an overextraction; we favored solutions that did not yield latent classes possessing less than 5% of the total sample), mean posterior probabilities (the average probability of cases being assigned to each latent class, conditional on item-response patterns; we favored solutions that yielded values equal to or greater than 0.85), entropy (a coarse summary of overall classification uncertainty, where higher values [ranging from 0 to 1] indicate greater precision), and substantive interpretability and utility (Weller et al., 2020). Sampling weights were incorporated into the LCA to generate representative parameter estimates. At this stage of the analysis, 170 cases from the focal sampling frame (i.e., non-Hispanic Black-identifying youth) possessed missing values across all ACE items, resulting in a reduction of usable cases from 5,922 to 5,752. An exploration of data missingness yielded evidence of covariate-dependent missingness (with youth age, youth physical health, caregiver mental health, caregiver physical health, youth insurance coverage, and caregiver educational attainment as covariates)—a special case of missing at random—for the ACE items (chi-square distance = 1,237.42, p = 1.00). That, coupled with the relatively small percentage of impacted cases (i.e., 2.9%), suggested that the omission of cases was likely to exert negligible influence on the parameter estimation process.

Following the selection of an optimal latent-class solution, we applied multiple group comparison analysis in a path analysis framework to specify latent-class membership (representing distinct portfolios of ACEs) as a predictor of youth mental health, both in terms of internalizing problems and externalizing problems (two outcome variables with correlated residuals). Youth mental health outcomes were also regressed on predetermined covariates. Modeling at this stage was partitioned on the basis of youth gender, with one multiple group comparison analysis conducted for male youth only and another multiple group comparison analysis conducted for female youth only. In each case, the four combinations of low-versus-high levels of family resilience/connection and neighborhood cohesion were used as the grouping variable. This grouping variable can be conceptualized as a categorical moderator of the path analysis detailed above. Analytic models were evaluated as having acceptable fit if they met most or all of the following criteria: nonsignificant chi-square test of model fit; comparative fit index (CFI) and Tucker–Lewis index (TLI) values greater than 0.95; and root-mean-square error of approximation (RMSEA) values (and upper 90% confidence interval [CI] values) less than 0.06 (Bowen & Guo, 2011). Given the binary form of the two mental health outcome variables, we used a means- and variance-adjusted weighted least squares estimator and probit link function.

Starting with a model in which all structural parameters were allowed to vary across the four combinations of low-versus-high levels of family resilience/connection and neighborhood cohesion, we assessed whether each structural parameter could be constrained to equality across groups without worsening model fit. Wald tests were used to conduct these assessments. In the event a Wald test for a specific parameter constraint was nonsignificant, that parameter was constrained to equality across the four groups; otherwise, the parameter was allowed to vary freely across the four groups (an indication that the grouping variable significantly moderated that parameter estimate). We assessed one structural parameter at a time, which included all regression coefficients and the correlation between the residuals of the two outcome variables. Thus, for male and female youth distinctly, we arrived at final models that featured all invariant and noninvariant structural parameters with respect to the four groupings of low-versus-high family resilience/connection and neighborhood support. Similar to the LCA, sampling weights were incorporated to generate representative parameter estimates in these models. At this stage of the analysis, 259 cases from the LCA sample possessed missing values related to exogenous variables in the model, resulting in a reduction of usable cases from 5,752 to 5,493 (the final analytic sample reported earlier). An exploration of data missingness yielded evidence of covariate-dependent missingness (with the ACE grouping, child age, and child gender as covariates)—a special case of missing at random—for the exogenous variables (chi-square distance = 149.34, p = .63). That, coupled with the relatively small percentage of impacted cases (i.e., 4.5%), suggested that the omission of cases was likely to exert negligible influence on the parameter estimation process. All data management was conducted in Stata 17.0, whereas LCA and multiple group comparison analyses were conducted in Mplus 8.6.

Results

Latent-Class Enumeration

Table 2 features results from the latent-class enumeration process. Upon reaching a three-class solution, mean posterior probabilities began to drop below our pre-specified cutoff of 0.85, and two out of the three likelihood ratio tests were nonsignificant. In addition, latent-class solutions beyond a two-class solution yielded at least one latent class possessing less than 5% of the total sample. Taken together, we selected the two-class solution as optimal.

Table 2.

Latent-Class Enumeration

Classes	AIC	BIC	aBIC	p			Entropy	Smallest class n	Smallest class % of total sample	M posterior probabilities
				LRT	aLRT	BS LRT				1	2	3	4	5
1	42,212	42,272	42,244
2	38,730	38,856	38,796	.00	.00	.00	0.78	1,106	19.2	.89	.95
3	38,567	38,760	38,668	.74	.74	.00	0.72	259	4.5	.91	.83	.80
4	38,481	38,741	38,617	.33	.33	.00	0.77	68	1.2	.90	.83	.81	.79
5	38,413	38,739	38,584	.54	.54	.00	0.79	69	1.2	.78	.84	.73	.81	.90

Note. AIC = Akaike information criterion; BIC = Bayesian information criterion; aBIC = adjusted BIC; LRT = Lo–Mendell–Rubin likelihood ratio test; aLRT = adjusted LRT; BS LRT = Bootstrap likelihood ratio test.

Figure 1 offers a visualization of the two-class solution. The first latent class was marked by low or near-zero conditional probabilities of endorsing any of the nine ACE items. One potential exception was the ACE related to a parent or guardian divorcing or separating, which yielded a conditional probability of .31. In the interest of parsimony, we opted to label this latent class “Low-ACEs,” which possessed 81% of the analytic sample. The second latent class was marked by moderate to high conditional probabilities of endorsing each of the nine ACE items, with one exception—the conditional probability of endorsing the ACE item related to a parent or guardian dying was .18. Although a relatively low probability in absolute terms, .18 is a notable probability for the death of a parent or guardian—a generally rare event for youth to experience (an estimated 3% of all children in the United States, and 6% of all Black children, experienced the death of a caregiver per 2021–2022 NSCH weighted data). Again, for the sake of parsimony, we opted to label this latent class “high-ACEs,” which possessed the remaining 19% of the analytic sample. To ensure we were not missing meaningful potential contributions from the three-class solution, we examined the output from the three-class solution and noticed that the third class (which, again, possessed less than 5% of the total sample) appeared to draw cases from the low and high classes, essentially forming a mid-range class with respect to conditional probabilities of ACE-item endorsement. From this perspective, the third class would not necessarily add value beyond the two-class solution to subsequent analyses intended to contrast portfolios of ACEs with respect to youth mental health.

Figure 1. Visualization of Latent-Class Solution.

Note. ACE = adverse childhood experience; See the online article for the color version of this figure.

Multiple Group Comparison Analysis

Figure 2 displays results from the final multiple group comparison analysis for male youth (Panel A), which yielded acceptable model fit as indicated by the following model fit indices: χ²(30) = 29.68, p = .48; CFI = 1.00, TLI = 1.00, RMSEA = 0.00, upper 90% CI [0.03]. For male youth, membership in the high-ACEs class (vs. membership in the low-ACEs class) was associated with a significant increase in the probability of youth possessing current internalizing problems (b = 0.56, p < .001). The estimated coefficient did not differ significantly across the four groupings of low-versus-high family resilience/connection and neighborhood cohesion (Wald test p = .11); that is, family resilience/connection and neighborhood cohesion did not significantly moderate the association between ACE-class membership and internalizing problems for male youth. In contrast, family resilience/connection and neighborhood cohesion did significantly moderate the association between ACE-class membership and externalizing. Specifically, membership in the high-ACEs class (vs. the low-ACEs class) was associated with a significant increase in the probability of youth possessing current externalizing problems, but only in the LF–LC group (b = 0.70, p < .001). For the other three combinations of low-versus-high levels of family resilience/connection and neighborhood cohesion, the association was nonsignificant. Also notable was the relatively lower R² values associated with the two outcome variables among members of the HF–HC group (0.15 for internalizing problems and 0.06 for externalizing problems, respectively). Although the residual correlation between outcome variables differed significantly across the four low-versus-high family resilience/connection and neighborhood cohesion groups, the general magnitude of the correlation was similar (and positive) across groups, ranging between 0.56 and 0.74). See Figure 2 for more details.

Figure 2. Multiple Group Comparison Analysis of Associations Between Latent Classes of ACEs and Youth Mental Health, by Youth Gender.

Note. Both endogenous variables are regressed on the following covariates: youth age, youth physical health, caregiver mental health, caregiver physical health, youth insurance coverage, and caregiver educational attainment. Males only (Panel A): Model fit: χ²(30) = 29.68, p = .48; CFI = 1.00, TLI = 1.00, RMSEA = 0.00, upper 90% CI [0.03]; HF–HC n = 841; HF–LC n = 618; LF–HC n = 506; LF–LC n = 914. Females only (Panel B): Model fit: χ²(30) = 25.39, p = .65; CFI = 1.00, TLI = 1.00, RMSEA = 0.00, upper 90% CI [0.02]; HF–HC n = 744; HF–LC n = 621; LF–HC n = 417; LF–LC n = 832. ACE = adverse childhood experience; HF = high family resilience and connection index value (LF = low); HC = high neighborhood cohesion (LC = low); CFI = comparative fit index; TLI = Tucker–Lewis index; RMSEA = root-mean-square error of approximation; CI = confidence interval.

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

Figure 2 also displays results from the multiple group comparison analysis for female youth (Panel B), which yielded acceptable model fit as indicated by the following model fit indices: χ²(30) = 25.39, p = .65; CFI = 1.00, TLI = 1.00, RMSEA = 0.00, upper 90% CI [0.02]. For female youth, membership in the high-ACEs class (vs. membership in the low-ACEs class) was associated with a significant increase in the probability of youth possessing current internalizing problems (b = 0.42, p < .01). The estimated coefficient did not differ significantly across the four groupings of low-versus-high family resilience/connection and neighborhood cohesion (Wald test p = .09); that is, family resilience/connection and neighborhood cohesion did not significantly moderate the association between ACE-class membership and internalizing problems for female youth (similar to male youth). Family resilience/connection and neighborhood cohesion, however, did significantly moderate the association between ACE-class membership and externalizing problems. Specifically, membership in the high-ACEs class (vs. the low-ACEs class) was associated with a significant increase in the probability of youth possessing current externalizing, but only in the HF–LC group (b = 0.69, p < .01). For the other three combinations of low-versus-high levels of family resilience/connection and neighborhood cohesion, the association was nonsignificant. Across groups, the R² value associated with internalizing problems was especially high (0.38) relative to the other groups (0.13–0.27). Although the residual correlation between outcome variables differed significantly across the four low-versus-high family resilience/connection and neighborhood cohesion groups, the general magnitude of the correlation was similar (and positive) across groups, ranging between 0.52 and 0.74).

Discussion

This study is among the first to examine if joint patterns of family resilience and neighborhood cohesion moderate the association between ACE exposure and internalizing and externalizing problems among a nationally representative sample of Black male and female youth. Notably, we identified two unique profiles of ACE exposure characterized by high and low exposure which is consistent with prior work using the NSCH and other population-based samples (Lanier et al., 2018; N. Zhang et al., 2022). Further, although family resilience and neighborhood cohesion did not jointly inform the link between ACE exposure and internalizing problems, the combination of these factors did inform the association between ACE exposure and externalizing problems among Black girls and boys in unique ways; a pattern also observed in previous work (Leban & Delacruz, 2023; Meeker et al., 2021). Collectively, the findings from this study underscore the complexity of identifying proximal and contextual processes that may augment risk associated with ACEs exposure, and suggest that efforts to produce policy and practice interventions likely require a deeper understanding of the role of family and community-level factors.

Consistent with the generalizable patterns of ACE exposure identified within cross-race-comparative studies, we found two distinct patterns of ACE exposure within our sample of Black youth characterized by high and low exposure to adversity, respectively. Critically, over 80% of this population-based sample of Black youth evidenced a low probability of having their parent endorse that they had experienced any of the nine ACEs assessed in the current study. Although studies have repeatedly highlighted that Black youth are at increased risk for ACEs relative to other groups (Liu et al., 2020; X. Zhang & Monnat, 2022; N. Zhang et al., 2022), this unexpected, yet, important finding suggests that there is significant within-group heterogeneity with respect to adversity exposure among Black youth. To this end, prior scholarship suggests that the presence of cultural assets unmeasured in the current investigation (e.g., community-based organization involvement, positive schooling environment, and neighborhood capital) may protect Black youth from ACE exposure and mitigate the effects that ACEs have on their mental health (Hampton-Anderson et al., 2021; Woods-Jaeger et al., 2021). That said, in comparison to the low ACE group, youth within the high ACE exposure group were at elevated risk for internalizing problems and externalizing problems which is consistent with prior scholarship documenting the deleterious effects of chronic ACE exposure among Black youth (Bernard et al., 2022; Hicks et al., 2020).

Regarding our second prediction, family resilience and neighborhood cohesion were hypothesized to moderate the association between ACEs and mental health problems. Surprisingly, we found that family and community resilience factors did not moderate the association between ACEs and internalizing problems, in either male or female youth. That is, regardless of the level of family resilience or neighborhood cohesion, the association between ACEs and internalizing problems did not change significantly. This finding is a somewhat unexpected finding given the preponderance of guidance suggesting that intervening to improve neighborhood conditions and family resilience, including enhancing parent-child relationships, are effective strategies to prevent the development of internalizing conditions among children who have experienced trauma (McGowan et al., 2023; Traub & Boynton-Jarrett, 2017).

We point to the demographic composition of our sample to explain why previous recommendations differ from our current findings. Indeed, while enhancing the presence of family resilience and neighborhood cohesion has many benefits, it may not be enough to protect Black youth, in particular, from being exposed to ACEs or their associated consequences. As illustrated by Melton-Fant (2019) living in supportive neighborhoods does reduce the risk for some (e.g., having difficulty getting by on their family’s income) but not for all ACEs (e.g., living with an incarcerated parent or guardian) among Black youth. Furthermore, Liu et al. (2020) revealed that in predicting health outcomes, exposure to moderate-to-high levels of adversity outweighed the presence of protective factors at the family and community level—a trend that was discernible solely among Black youth and was not observed among their White or Latinx counterparts. Thus, it is possible that for Black youth family resilience and neighborhood cohesion are reflective of “protective-reactive factors” or constructs that are protective in the face of low levels of adversity, but whose promotive effects diminish at high levels of risk or adversity (S. T. Li et al., 2007; Luthar et al., 2000). Thus, current findings suggest that mitigating the effects of ACEs on internalizing concerns requires intervention approaches that go beyond enhancing family and neighborhood factors.

In contrast with our findings related to internalizing problems, we found that family resilience and neighborhood cohesion did moderate the relationship between ACEs and externalizing problems, and that this moderating effect differed slightly between male and female youth. Specifically, the association between adversity and externalizing problems (e.g., conduct disorder and ADHD) was only significant for boys who also reported low levels of both family and neighborhood-level factors. In contrast, the link between adversity and externalizing problems was only significant for females who also reported low community cohesion but higher family resilience. So, regardless of child gender, the link between ACE exposure and externalizing problems was only present when low neighborhood cohesion was identified, which appears to replicate findings from a recent study using the full NSCH sample from similar years (Khanijahani & Sualp, 2022).

When considering how combinations of family resilience and community cohesion influence the link between exposure to adversity in youth and externalizing outcomes, it is easier to explain the abovementioned findings for males. Given the hypothesized protective functions attributed to family resilience and community cohesion, we would expect that in instances of their diminished presence, a buffering effect may not be observed. In this way, low support from family and the broader community may leave Black boys more prone to maladaptive coping strategies (e.g., delinquency) and behavioral problems as they seek to process and cope with adversity exposure. Prior work suggests that Black youth residing in less supportive homes and/or less cohesive neighborhoods may have access to fewer adaptive coping resources and perceive fewer adults within or outside the home as trusted parties who can protect them from future ACE exposure (Riina et al., 2014).

More research is needed to replicate our partly counterintuitive finding for female Black youth. One potential explanation is that there are other influential social factors beyond that of the family (e.g., peers) that uniquely shape externalizing problems in gender-specific ways (see Hanish et al., 2005). It is also possible that the greater family resilience found among females with externalizing problems is an outcome of the family’s coping response to the problematic behaviors associated with externalizing problems. But why this effect would be observed for females and not for males in this analysis is unclear. What is clear is that we do not find a strong or consistent buffering effect of family resilience between ACEs and externalizing problems among Black female youth.

Implications

Results of this study suggest that family and community protective factors may have a direct effect on child mental health problems, but do not have a strong buffering effect in the wake of exposure to ACEs. There might be hope for the buffering effects of neighborhood support on externalizing problems, but family resilience appears to have no buffering effect on either internalizing or externalizing problems among Black youth who have experienced high levels of ACEs. This is not to say that family and community factors are not important, but among Black youth already exposed to ACEs, preventing internalizing problems likely requires understanding the influence of factors at lower (i.e., epigenetic processes) and higher (i.e., public policy) levels of youths’ social ecology. This finding also supports the prioritization of primary prevention of ACEs in childhood as the clearest strategy for reducing the increasing burden of internalizing problems such as depression and anxiety. For externalizing problems, such as conduct disorder and ADHD, our findings echo prior calls for “large-scale neighborhood support changes” to disrupt the harmful effects of ACEs (Khanijahani & Sualp, 2022, p. 176). Given the high co-occurrence of ACEs related to family poverty and neighborhood violence, particularly among Black youth (Cronholm et al., 2015), community interventions likely require substantial public investment to overcome the ongoing legacy of community disinvestment resulting from racist planning policies that drive contemporary racial disparities in ACE exposure (Strompolis et al., 2019).

Limitations and Future Directions

Findings from the current study should be interpreted in the context of its limitations. First, given the cross-sectional nature of these data, we are unable to establish causality in our findings or understand how the developmental timing of ACEs may uniquely inform the presence of mental health symptoms. That said, longitudinal work supports that ACE exposure precedes psychological problems among youth (e.g., S. M. Li et al., 2023). In any case, future research would benefit from longitudinal examinations that illuminate how the timing and frequency of ACEs can alter the psychosocial development of children, as well as the prospective moderating roles of family resilience and neighborhood cohesion on the link between ACEs and mental health among Black youth. Second, as the NSCH primarily relies on caregiver reports of youth experiences, including ACEs, the prevalence and impact of ACEs among Black youth may be underestimated. This possibility is supported by research that has found that there is generally poor agreement between youth and caregivers regarding youths’ exposures to ACEs (Oransky et al., 2013). As such, future research would benefit from youth reports of ACEs, mental health, and neighborhood/community level processes to further interrogate or corroborate the findings of the current study. Third, conceptualizations and subsequent measurements of ACEs within the NSCH and the broader literature typically capture sources of person-level sources of adversity that occur predominantly in or in close to proximity to the home. Yet, emerging research suggests that structural ACEs, which are associated with systemic social determinants of health and health disparity (e.g., segregation, reduced access to mental health services) not only enhance the risk for ACEs but can also serve as unique ACEs themselves that increase the risk for negative health outcomes among ethnoracially minoritized youth (Jewett et al., 2024). Thus, a promising area of future research would be to explore factors that augment psychological risk in the presence of both household and structural forms of ACEs. Fourth, examinations of family resilience and neighborhood cohesion relied on an abbreviated set of items and dichotomized categories that may not necessarily capture the richness or multidimensionality of family and neighborhood-level protective factors. Scholars are encouraged to further explore how family and neighborhood-level sources of protection may buffer the link between ACEs and mental health using more precise measures. Fifth, as a neurodevelopmental disorder, it is difficult to discern the degree to which the presence of ADHD symptoms, which was included within our measurement of externalizing symptoms, was informed by genetic versus adversity-related influences. Although prior research demonstrates that ACEs can enhance the risk for ADHD diagnoses and the severity of ADHD symptoms (Brown et al., 2017), it is also plausible for such symptoms to precede the presence of adversity exposure. Thus, future research is encouraged to replicate these findings to further clarify if and how ADHD may be related to or inform psychological responses to ACEs.

Conclusion

The purpose of this study was to examine the nuanced relationship between childhood adversity and mental health disorders in a nationally representative sample of Black youth. Findings align with and build upon prior scholarship demonstrating that high ACE exposure is a salient risk factor for internalizing concerns among Black youth regardless of gender, or familial and neighborhood-levels of support. Further, results illustrate that the link between ACEs and externalizing problems may be dependent on the levels of family and neighborhood support that are available to them. Collectively, study results indicate that family and neighborhood-level processes may be necessary, but not sufficient mechanisms to reduce the impact of ACEs among Black youth. Thus, there is a critical need for future work to investigate the interplay between individual, community, and systems-level factors that may offset the impact of ACEs, to identify potentially malleable mechanisms that can be targeted to enhance positive developmental outcomes in the face of adversity.