Diminished Protective Effects of Household Income on Internalizing Symptoms among African American than European American Pre-Adolescents

Aim: To investigate the differential role of race on the effect of household income on pre-adolescents’ internalizing symptoms in a national sample of U.S. pre-adolescents. Methods: This is a cross-sectional study that used data from the Adolescent Brain Cognitive Development (ABCD) study. Wave 1 ABCD data included 5,913 adolescents between ages 9 and 10 years old. The independent variable was household income. The primary outcome was internalizing symptoms measured by the teacher report of the Brief Problem Monitor (BPM) scale. Results: Overall, high household income was associated with lower levels of pre-adolescents internalizing symptoms. Race showed statistically significant interaction with household income on pre-adolescents’ internalizing symptoms, controlling for all confounders, indicating weaker protective effect of high household income on internalizing symptoms for African American than European pre-adolescents. Conclusion: High household income is a more salient protective factor against internalizing symptoms of socially privileged European American pre-adolescents than of historically marginalized African Americans pre-adolescents. Elimination of internalizing behavioral gaps across racial groups requires more than equalizing socioeconomic status. Future research should study the moderating role of institutional and structural racism experienced by African American families across all income levels. Such research may explain why pre-adolescent African Americans with high household income remain at high risk of internalizing symptoms. www.scholink.org/ojs/index.php/jetmm Journal of Economics, Trade and Marketing Management Vol. 2, No. 4, 2020 39 Published by SCHOLINK INC.


Introduction
Race and socioeconomic position (SEP) are two overlapping social determinants of adolescents' behaviors. Compared to non-Hispanic European American, ethnic minority adolescents, particularly African Americans, are at an increased risk of emotional and behavioral problems such as depression, anxiety, substance use, aggression, conduct disorder, and bullying (McLaughlin, Hilt, & Nolen-Hoeksema, 2007). African American adolescents report higher internalizing symptoms compared to European American adolescents (McLaughlin et al., 2007). Some examples of the behavioral risk of African American compared to European American adolescents include increased risk of substance use (Wallace et al., 2009), aggressive behaviors (McLaughlin et al., 2007), conduct disorders (McLaughlin et al., 2007), and poor academic achievement and school drop-out (Bumpus, Umeh, & Harris).
As emotional and behavioral problems operate as a gateway to future economic and health problems later in life (Burchinal et al., 2011;Cohen & Sherman, 2005;Gorey, 2009;Hair, Hanson, Wolfe, & Pollak, 2015), there has been an interest in understanding the joint effects of race and SEP on shaping inequalities in adolescents' internalizing symptoms. Such information on the complexities of race and SEP's effects may enable us to more effectively eliminate or at least reduce subsequent inequalities later in life by reducing adolescents' behavioral inequalities early in life (Burchinal et al., 2011;Cohen & Sherman, 2005;Gorey, 2009;Hair et al., 2015).
In fact, household income may partially explain some of the existing social inequalities and gaps in adolescents' behavioral outcomes between marginalized and privileged groups (Bell, Sacks, Thomas www.scholink.org/ojs/index.php/jetmm Journal of Economics, Trade and Marketing Management Vol. 2, No. 4, 2020 40 Published by SCHOLINK INC. Tobin, & Thorpe, 2020;Fuentes, Hart-Johnson, & Green, 2007;Kaufman, Cooper, & McGee, 1997;Samuel, Roth, Schwartz, Thorpe, & Glass, 2018). If household income is partially responsible for the existing gaps, then minimum wage policies and earned income tax credit should be regarded as a primary strategy for addressing the social inequalities that we observe between social groups (D. R Williams, 1999;D. R. Williams, Costa, Odunlami, & Mohammed, 2008).

Aims
In this investigation, we compared African American and European American pre-adolescents for the effects of household income on internalizing symptoms. While high household income may be associated with fewer internalizing symptoms, this protection is expected to be weaker for African American than European American pre-adolescents. The second hypothesis is based on an extensive literature generated by Assari on various SEP resources and behavioral and health outcomes across age groups (Assari, 2017;Assari, 2018a).

Materials and Methods
This cross-sectional study was a secondary analysis of existing data. We borrowed data from the Adolescent Brain Cognitive Development (

Analytical Sample
This study included 5913, 9-10 years-old American pre-adolescents who had data on the Brief Problem Monitor (BPM) scores (teacher report), rage, sex, race, ethnicity, parental education, household income, and marital status of the family. Pre-adolescents from any racial or ethnic background were included in our analysis.

Independent Variable
Household income. Household income was a nominal variable: Parents were asked, "What is your total combined highest income for the past 12 months? This should include income (before taxes and deductions) from all sources, wages, rent from properties, social security, disability and veteran's benefits, unemployment benefits, workman". Responses included less than $50,000, $50,000-$100,000; and $100,000 +.

Moderator
Race. Race, a self-identified variable, was a categorical variable: African Americans vs. European Americans (reference group).

Confounders
Child sex. Child sex was a categorical variable with 1 for males and 0 for females.
Age. Age was a dichotomous measure in months and was reported by parents.

Ethnicity.
Ethnicity, a self-identified variable, was a categorical variable: Hispanics vs. non-Hispanics (reference category).
Marital status of the household. Family marital status was a dichotomous variable: married = 1 and not-married = 0.
Highest educational attainment of the parents. Parental education was asked using this item: "What is the highest grade or level of school you have completed or the highest degree you have received?" This variable was a five-level categorical variable: (1) less than high school (reference category), (2) high school / GED, (3) some college, (4) college degree, and (5) graduate+ school.

Data Analysis
The program Data Analysis and Exploration Portal (DEAP) was utilized for statistical analyses. DEAP is an online analytical tool and is based on the R program. To conduct our multivariable analysis, two mixed-effects model regressions were performed. We used mixed-effect models because observations were nested to individuals, which were nested to families, and then to sites. As a result, we needed to have a random effect and fixed terms to adjust for the sample's clustering and non-independence.
Internalizing symptom score was the outcome. Race was the moderator. Household income, as a categorical variable, was the predictor. All regression models were estimated in the overall/pooled sample. Model 1, the main effect model, was estimated in the absence of the household income by race interaction term. Model 2 (the interaction model) added an interaction term between race and household income. These models controlled for ethnicity, age, sex, highest educational attainment of the parents, and family marital status. Regression coefficient (b), and p-values were reported for each parameter. Table 1 summarized the model formulas. We also tested several assumptions before we ran our models that are shown in Figure 1.

Ethical Aspect
For this study, we used a fully de-identified data set, which is available to the public. As our paper was non-human subject research, this study was exempted from a full review of the Institutional Review Board (IRB). However, the protocol of the main study, the ABCD, was approved by the IRB at the University of California, San Diego (UCSD), and several other institutions. Participants signed consent or assent depending on their age (Auchter et al., 2018).  Table 3 summarizes the results of two linear regression models in the overall (pooled) sample. In

Overall Multivariate Analysis Without Interaction
Model 1 (Main Effect Model), we found a protective effect of high household income on our outcome (Table 3 and Figure 2).

Figure 2. Association between Household Income and Children's Internalizing Symptoms
Overall (n = 5913) Table 4 provides a summary of the results of a regression model in the overall (pooled) sample. Model 2 (Interaction Model) showed an interaction term between race and household income on our outcome, suggesting household income's protective effect against internalizing symptoms was weaker for African American than European American adolescents (Table 4 and Figure 3).

Discussion
Our first finding showed that high household income is associated with lower internalizing symptoms of American pre-adolescents. Our second finding was that among American pre-adolescents, race may alter the returns of household income, one of the most salient SEP indicators, in terms of reducing the internalizing symptoms of their children. We observe weaker effects of parental income on the behavioral profile of African American than European American pre-adolescents.
Our first results are in line with the literature on social determinants and fundamental cause theories.
Our second result can be interpreted in the light of the MDRs. These MDRs can contribute to the trans-generational effects of inequalities. Such MDRs are repeatedly shown across SEP sources, emotional and behavioral outcomes, age groups, and marginalization types (Assari, 2017;Assari, 2018a). MDRs are well documented within individuals as well as families. The results of this paper showed the transgenerational transmission of inequalities happen as a result of MDRs due to pre-adolescents' race.
Our results are similar to multiple studies on MDRs. A recent JAMA paper showed that MDRs of parental educational attainment is systematically weaker for African American and Hispanic adolescents than European American adolescents (Assari, Caldwell, et al., 2019). In that cross-sectional study, among 10,619 adolescents ages 12 to 17 years, parental educational attainment showed differential effects for African American and European American adolescents' psychological distress, aggression, tobacco dependence, school performance, and chronic disease (Assari, Caldwell, et al., 2019 Assari, Thomas, et al., 2018). Furthermore, other studies have shown that household income and parental education generate greater improvements in school performance for European American than African American adolescents (Assari & Caldwell, 2019b;Assari, Caldwell, et al., 2019) and young adults . Finally, some studies have shown that high income may even operate as a risk factor for high internalizing symptoms and depression for African American adolescents  and adults (Assari, Lankarani, & Caldwell, 2018).
While most studies have shown MDRs among African Americans (Assari, Thomas, et al., 2018), the same patterns may also exist for Hispanic (Assari, 2018e;Assari, Farokhnia, et al., 2019;Shervin & Ritesh, 2019), Asian American , Native American (Assari & Bazargan, 2019b), LGBTQ , immigrant , and even marginalized European individuals . In the US societies, all these marginalizing identities reflect lower access to the resources, worse treatment by the society, as European Americans occupy more prestigious jobs and show a relative privilege compared to all other racial groups, particularly African Americans (Feagin & O'brien, 2003;Franks, 2013). Any factor that can be a proxy of any type of social privilege may tend to be a source of differential marginal returns of SEP, such as income (Assari, 2018e;Assari & Bazargan, 2019a). That means, weaker effects of a wide range of SEP indicators on several outcomes have been documented for many marginalized communities compared to the socially privileged group.

Implications for Research
We argue that studies on behaviors or development should not merely control for race/ethnicity, class, or SEP. This is particularly true for studies that investigate behavioral, neural, and social correlates of