Unhealthy Diet of High Socioeconomic Children of Color: Marginalization-Related Diminished Returns

Research Article

Int J Nutr Sci. 2022; 7(1): 1061.

Unhealthy Diet of High Socioeconomic Children of Color: Marginalization-Related Diminished Returns

Assari S1,2,3*, Najand B1 and Bazargan M2,3

1Marginalization-Related Diminished Returns (MDRs) Research Center, Charles R Drew University of Medicine and Science, Los Angeles, California, USA

2Department of Family Medicine, Charles R Drew University of Medicine and Science, Los Angeles, California, USA

3Department of Urban Public Health, Charles R Drew University of Medicine and Science, Los Angeles, California, USA

*Corresponding author: Shervin Assari, Marginalization-Related Diminished Returns (MDRs) Research Center, Charles R Drew University of Medicine and Science, 1731 E 120th St, Los Angeles, California, USA

Received: January 05, 2022; Accepted: February 08, 2022; Published: February 15, 2022


Background: While socioeconomic status (SES) indicators such as parental educational attainment and household income are among the primary drivers of individual health, the effects of household SES indicators (for example, parental educational attainment and family income) on health behaviors such as a healthy diet may differ by ethnicity, as discussed by the Marginalization related Diminished Returns (MDRs) phenomenon.

Objectives: Built on the MDRs, this study had two aims: first, to test the associations between family SES indicators (parental educational attainment and household income) and diet quality, and second, to test ethnic variation in these associations.

Materials and Methods: This longitudinal study used the Adolescent Brain Cognitive Development (ABCD) baseline and year 2 data. Participants included 5,856 individuals who were either Black, White, Latino, or non-Latino. Age, sex, family structure (parental marital status), parental education, and family income were studied. The outcomes were the amount and frequency of consuming fish, soup, vegetables, fruits, hot dogs, French fries, ketchup, soda, and sugary beverages. Linear regression was used for performing the main data analysis.

Results: Overall, high educational attainment and family income showed a positive association with fruit consumption and a negative association with the consumption of hot dogs, fries, soda, and sugary beverages in the overall population. We documented statistically significant interactions between ethnicity and educational attainment and household income on our dietary habits of interest, indicating weaker associations between family SES and diet in Black and Latino than non-Latino White individuals.

Conclusion: We observe that household SES differently improves the dietary quality of diverse ethnic groups. Due to MDRs of education and income in ethnic minorities, children from highly educated and high-income families eat less healthily than their non-Latino White counterparts. This finding is in line with the MDRs framework that ethnic health disparities sustain across class lines.

Keywords: Educational attainment; Diet; Socioeconomic status; Population groups


While Marmot [1,2], Hayward [3-5], Link and Phelan [6], Ross and Miroswky [7-9], and others [10] have shown that socioeconomic status (SES) indicators, such as educational attainment, promote population and individual health outcomes, recent research has documented weaker SES effects for ethnic minorities than non-Latino White populations [11,12]. Although one of the mechanisms that explains the impact of education and income on health is improved health behaviors such as diet [13-20], more recent studies have shown that the effects of parental and one’s own education and income on dietary habits and food options are weaker for ethnic minorities than non-Latino Whites [21,22]. This is in part because ethnic minorities are still segregated, and their high SES does not improve access to healthy foods to the same level as high SES non-Latino Whites. This is important because a healthy diet is associated with a lower risk of obesity, diabetes, heart, and metabolic diseases [13,23,24].

The Marginalization-related Diminished Returns (MDRs) phenomenon [11,12] refers to the general observation that the effects of educational attainment and household income on generating outcomes are weaker for ethnic minorities than their non-Latino White counterparts [25-31]. Similar MDRs are shown for diet [22], obesity [32,33], heart disease [34], disability [35], chronic disease [36], hospitalization [37], and mortality [38-41]. While stress, labor market discrimination, segregation, food access, neighborhood quality, and various aspects of the social environment are all potential mechanisms, one of the proximal mediators of the MDRs might be dietary behaviors. We expect that pro-health diets are less influenced by educational attainment and income in ethnic minority families because they face many barriers in their lives given racism and social stratification [42-44]. In addition, because ethnic minorities live in marginalized communities and are more likely to be under the influence of parents who had spent their childhoods in poverty, their family SES may not have a large effect on diet and eating for ethnic minority families. As an adaptation to poverty, food culture may also contribute to the diminished returns of SES on the dietary quality of Black and Latino people in the US.

The Marginalization-related Diminished Returns (MDRs) phenomenon [11,12] also refers to the weaker economic and health effects of SES indicators, such as educational attainment, for the members of marginalized groups (particularly ethnic minorities) than US-born heterosexual non-Latino Whites [11,12]. Assari [11,12], Ferarro [45], Thorpe [46-48], Hudson [49-51], Kaufman [52], Braveman [53], Shapiro [54,55], Williams [56,57], Ceci [58], Navarro [59-61], and others [62] have reported weaker effects of SES indicators for ethnic minorities than non-Latino Whites. Kaufman has discussed the poor overlap between SES across ethnic groups that result in residual and unmeasured confounding as well as not-comparability of SES across ethnic groups [52]. Navarro has described this as “ethnicity and SES,” rather than “ethnicity or SES” effects due to the complex interplay between ethnicity and SES [59- 61]. Ceci has mentioned that the Have-Nots (ethnic minorities) may gain access to health less than the Haves (non-Latino Whites) from the same resources (SES indicators) due to their lower readiness to uptake and navigate the complex social systems [58]. However, most of the existing literature is on Black-White comparison rather than Latino-non-Latino comparison. Thus, there is a need to study a wider range of ethnic groups that include Black groups and compare Latino and non-Latino groups.


To better understand whether MDRs observed in the ABCD data for brain outcomes can be, in part, explained by MDRs in diet, we conducted a secondary analysis of the ABCD data to determine the association between two SES indicators namely, household income and parental educational attainment and dietary indicators and variation in these effects by ethnicity. We hypothesize a positive association between educational attainment and household income, and healthy dietary habits of individuals. We also expected a negative association between family SES and unhealthy dietary practices such as hot dogs, fries, and sugary beverage consumption. Built on the MDRs framework, we hypothesize that these positive and inverse associations would be weaker for Latino and Black children than non-Latino White children. As a result, we expect more healthy diets in non-Latino White families with high SES than ethnic minority families with similar SES, indicating a diminished effect of household income and educational attainment on healthy diet in ethnic minority families (due to access and their need to allocate resources to other necessities). These MDRs in diet [21,22] then would partially explain the MDRs on the effects of SES on brain development [63-69].

Materials and Methods

Design and setting

This study is a secondary analysis of the first two years (waves 1 and 2) of data (2016/2018 to 2018/2020) of the Adolescent Brain Cognitive Development (ABCD) study [70-74]. The ABCD is a stateof- the-art and national longitudinal study of children’s development in the United States [70,75].

Sampling and participants

In the ABCD study, participating children were 9-10-years-old at the time of recruitment, which took place between 2016 and 2018. Recruitment occurred across 21 study sites in multiple cities across 15 U.S states. The primary recruitment strategy was through the school systems [76]. The original/overall study included 10,875 children at baseline.

Eligibility and analytical sample

From the 10,875 participants, we included children at wave 1 and wave 2 who were either Latino or non-Latino White or Black. This was based on parents’ report of ethnicity (see below). This study did not include other ethnic groups such as Asian, Native American, Mixed, Other, or unknown ethnic groups. Participants were only included if they had data on SES, ethnicity, diet, and covariates (n=5856).

Study variables

Primary outcomes: The outcomes were the amount and frequency of consuming fish, soup, vegetables, fruits, hot dogs, French fries, ketchup, soda, and sugary beverages. These food indicators were calculated based on Block Kids Food Screener (BKFS). The BKFS is a 41-item food frequency questionnaire which was developed by NutritionQuest (Berkeley, CA, USA). This instrument is ideal for the measurement of dietary intake of nutrients and food groups for children and youth aged 2-17. The questionnaire can be selfadministered or asked using the parents’ report. BKFS asks the participant to report the frequency and quantity of food and beverage they consumed during the past seven days. The response items for frequency ranges from ‘none’ to ‘every day’. For each food type, three or four food items are measured. The BKFS provides estimates for the amount of intake of fruit, vegetables, dairy, whole grains, protein sources (meat, poultry, and fish in ounce equivalents), fast food, sugary beverages, soda, and other food types. This measure also details calorie intake, saturated fat, and consumed sugars. For this study, BKFS was administered at wave 2 when the child was 11 or 12 years old. This measure was collected two years after measuring all other variables such as household income, parental education, family structure, and other covariates. This measure is validated and shows reliability and accuracy in this age group.

Independent variable

Socioeconomic status: This study used two indicators of household SES. These indicators were household income and parental education, both treated as continuous measures. Parents reported their years of schooling. This variable ranged between 0 (for no formal education) and 21 (doctoral degree). Annual family income had a range between 1 and 10 that referred to the following income levels: 1 = less than $5000; 2 = $5000; 3 = $12,000; 4 = $16,000; 5 = $25,000; 6 = $35,000; 7 = $50,000; 8 = $75,000; 9 = $100,000; 10 = $200,000+.


Ethnicity: Ethnicity was composed of two categorical variables which were identified by the parents. All participants were non- Hispanic or Hispanic White or Black.


Demographic factors. Age, sex, and family structure were measured as covariates. Parents reported the children’s ages, the child’s age acting as a continuous variable, and measured in months. The sex of the child was a dichotomous variable with 1 representing males and 0 representing females. Family structure was married or unmarried.

Data analysis

We used the SPSS 25 for data analysis. First, we ruled out multi-collinearity between our study variables and confirmed that our outcomes had near to normal distributions. Next, we applied multivariable linear regression models with parental educational attainment and household income as the independent variables, a diet indicator as an outcome, and ethnicity as moderators; all models were performed in the pooled sample. Model 1 did not include an interaction term but included all the confounders. Model 2, however, did include interaction terms between parental education and household income and ethnicity, in addition to all the confounders. We performed similar models for outcomes. We reported the values b, SE, 95% CI, and p from our regression models.

Ethical considerations

This analysis was exempt from a full IRB review by Charles R Drew University of Medicine. The study of origin (ABCD) was approved by the Institutional Review Board (IRB) at the University of California, San Diego (UCSD). Assent and consent were received from children and their parents, respectively [75].


Overall, 5,856 individuals entered our analysis. Our participants were either White, Black, non-Latino, or Latino. Table 1 shows the summary of descriptive statistics overall.