Intersectional Analysis of Adverse Childhood Experiences By Poverty, Race/Ethnicity, and Gender: Findings from the Panel Study of Income Dynamics

Background and Purpose: Studies show that the risk of experiencing adverse childhood experiences (ACEs) is greater the general US population among low-income and marginalized racial/ethnic groups, though recent evidence suggests that racial/ethnic differences are confounded by economic status (Mersky et al., 2018, 2021; Slopen et al., 2016). Extending this line of research, the current study considers whether differences in the prevalence of ACEs across racial/ethnic groups is moderated by poverty and gender. We also apply advanced quantitative approaches to examine variation in ACEs across subgroups defined by intersections of different racial/ethnic, economic, and gender statuses.

Methods: We sampled 5,196 adults from the Panel Study of Income Dynamics (PSID) and its supplemental Childhood Retrospective Circumstances Study (CRCS). The eligible sample consisted of adults born after 1950 who completed both PSID and CRCS surveys in 2013. We randomly selected one adult per household/mother to prevent clustering effects and ensure that participants had valid household economic data. Ten ACEs were measured: physical abuse, emotional neglect, physical neglect, emotional abuse, crime victimization, being bullied at school, maternal mental health problems, household substance abuse, domestic violence, and parental divorce/separation. Childhood poverty was defined as living below the federal poverty threshold for at least one year before turning 18. Participants’ race/ethnicity was categorized as Hispanic (i.e., Latinx), non-Hispanic Black, non-Hispanic White, or Other. Gender was coded dichotomously (man; woman). Bivariate analyses were used to describe group differences in ACE scores. We performed ordinary least squares regressions with fixed effects to explore pairwise interaction effects. Regression analyses were repeated after replacing interaction terms with composite variables representing distinct combinations of racial/ethnic, economic, and gender categories.

Results: In the full sample (N=5,196), 78.7% reported at least one ACE and 53.9% reported multiple ACEs. The mean cumulative ACE score was highest among Hispanic participants (M=2.38), followed by White (M=2.15), Black (M=1.96), and Other (M=1.84) racial/ethnic subgroups. Participants who lived below the poverty threshold in childhood reported more ACEs on average than did non-poor participants (M= 2.18 vs. 2.02; p< .001). Compared to men, women reported more ACEs (M= 2.20 vs. 1.96; p< 0.01). Results from pairwise interaction tests showed that variation in ACE scores by race/ethnicity was not significantly moderated by poverty status or gender. However, intersectional analyses identified between-group contrasts indicating that racial/ethnic differences were confounded by poverty and gender. For instance, compared to non-poor White men, higher ACE scores were observed among non-poor and poor White and Hispanic women as well as poor Black women (p< 0.05). Remarkably, compared to non-poor white men, lower ACE scores were observed among poor Black men (p< 0.05)

Conclusions and Implications: Our study adds to recent advances in analytic intersectionality and extends evidence suggesting that variation in ACEs by race/ethnicity is confounded by economic status. Results will be discussed while considering the strengths and limitations of the PSID sample. These limitations notwithstanding, our findings have implications for disparities research, including ongoing debate about why children and families from marginalized racial/ethnic groups are overrepresented in the child welfare system.