Parental Educational Attainment and Frequency of Marijuana Use in Youth: Hispanics’ Diminished Returns

Background: While socioeconomic status (SES) indicators such as parental educational attainment show robust associations with health behaviors such as substance use, the protective effects of these indicators may differ across racial groups. This phenomenon of weaker associations between SES indicators and health outcomes for marginalized and minoritized groups relative to non-Hispanic White people has been labeled “Marginalization-related Diminished Returns” (MDRs). Here, we test both whether parental educational attainment is associated with marijuana use frequency in youth as well as whether we observe racial and ethnic variation in this association consistent with MDRs. Methods: This study used data from the cross-sectional 2019 Monitoring the Future survey (MTF 2019). Participants included 29,230 youth who were either Hispanic (24.1%) , non-Hispanic Black (16.1%), or non-Hispanic White (59.9%). We used weighted logistic regression models to test for (1) associations between maternal educational attainment and youth cannabis use frequency as well as (2) moderation of this association by race/ethnicity, while adjusting for the complex sample design of the MTF 2019 data. Age, sex, father presence, and maternal employment were entered into models as covariates. than born non-Hispanic lower-SES Hispanic youth report roughly equal levels of use across the full SES spectrum. This finding is in line with the MDRs framework and may reflect factors such as structural racism, social stratification, and the marginalization of ethnic minority families in the US.

differences, the MDRs framework makes frequent use of moderated-mediation models and introduces non-additive effects of race/ethnicity and SES, allowing SES effects to vary across groups. These MDRs models are thus more realistic than those that assume universal effects of SES or that "one size fits all". Finally, the MDRs framework explains why racial and ethnic health gaps may widen rather than narrow over time and why they may increase rather than decrease as SES increases (Assari, 2017;.

Aims
While there are several reports documenting MDRs in the associations between parent educational attainment and different health behaviors, it is not currently known whether this phenomenon applies to youth cannabis use. To address this gap in the literature, we conducted a secondary analysis of the 2019 Monitoring the Future (MTF 2019) survey data to determine the multiplicative effects of race, ethnicity, and parental educational attainment on youth marijuana use frequency. First, we hypothesized an inverse association between parental educational attainment and marijuana use frequency. Second, consistent with the MDRs framework, we hypothesized that this inverse association would be weaker for Hispanic and non-Hispanic Black than non-Hispanic White youth. As a result, we expected that non-Hispanic Black and Hispanic youth with highly educated parents would still report high levels of use. Conversely, we expected non-Hispanic White youth with highly educated parents to report significantly lower marijuana use than peers with less highly educated parents.

Materials and Methods
Design and Setting. MTF 2019 is a cross-sectional survey that provides nationally representative estimates for substance use in US youth population. Response rates for 2019 were 89% and 86% for 8 th and 10 th -grade students, respectively. We included 8 th (n = 14,223) and 10 th graders (n = 14,595) who were either Hispanic, non-Hispanic Black, or non-Hispanic White for this analysis.
Sample and Sampling. About 130 US high schools participate in the MTF study each year. These schools include both public, private, and Catholic schools. Schools are selected using a multistage sampling design. When a school refuses to participate in the MTF study, that particular school is then replaced with another school with similar geographic and demographic characteristics. From participating schools, 8 th , 10 th , and 12 th grade youth are selected via a three-stage sampling procedure: classrooms and students within schools within geographic areas, which were the primary sampling units.
In the MTF study, school participation rates commonly range between 66% and 80% across years.
Main Independent Variable. The main independent variable was maternal education level. This variable was measured using the following item: What is the highest level of schooling your mother completed?
Items included 1="Completed grade school or less", 2="Some high school", 3="Completed high school", 4="Some college", 5="Completed college", and 6="Graduate or professional school after college". This variable was treated as a continuous measure, with a higher score indicating higher maternal educational attainment. Outcome Variable. Our outcome was measured using the following item: "The next major section of this questionnaire deals with various other drugs. There is a lot of talk these days about this subject, but not enough accurate information. Therefore, we still have a lot to learn about the actual experiences and attitudes of people your age. We hope that you can answer all questions, but if you find one which you feel you cannot answer honestly, we would prefer that you leave it blank. Remember that your answers will be kept strictly confidential; they are never connected with your name or your class. On how many occasions (if any) have you used marijuana (weed, pot) or hashish (hash, hash oil) . . . in your lifetime?" Responses were 1="0 occasions"; 2="1-2 occasions"; 3="3-5 occasions"; 4="6-9 occasions"; 5="10-19 occasions"; 6="20-39 occasions"; and 7="40 or More". This variable was treated as a continuous measure, with a higher score indicating higher marijuana use frequency.
Covariates. For the demographic variables, we included age (1 = 16 years or older, 0 = less than 16 years old), sex (male = 1, female = 0), father presence (1 = father present, 0 = father absent), maternal employment, metro area (MSA: geographical regions with a relatively high population density at its core), and region of the country.
Race and ethnicity. We examined racial and ethnic group membership as a potential moderator, operationalized as two dichotomous variables: non-Hispanic Black and Hispanic. Non-Hispanic White was the reference group for both variables.

Analytic Strategy
We used the Survey Documentation and Analysis (SDA) at the Inter-University Consortium for Political and Social Research (ICPSR) to conduct our data analysis. We used the weighted logistic regression analysis for multivariable models to produce Odds Ratios (OR) and the corresponding 95% Confidence Intervals (CI). First, we ran models without interactions. Then we ran models with interactions between race/ethnicity and parental education. All analyses were weighted using the MTF 2019 individual-level sampling weights to make the estimates representative at the national level for the US 8 th and 10 th -grade population. We considered P-values <0.05 as statistically significant, and all tests were two-sided.

Figure 2. Distribution of Marijuana Use Frequency by Race and Ethnicity
As shown by Table 2, males and youth above age 16 also reported higher marijuana use frequency.
School in the standard metro area was a risk factor, and father presence was a protective factor against marijuana use frequency in youth. Race and ethnicity did not show any main effects on marijuana use frequency in youth, meaning that non-Hispanic Black, Hispanic, and non-Hispanic White youth did not show any difference in their marijuana use frequency. While maternal education was protective against youth marijuana use frequency, this effect was weaker for Hispanic youth than non-Hispanic White youth. Figure 3 shows the interaction between parental education and ethnicity.

Discussion
In this large, cross-sectional survey completed by a nationally-representative sample of U.S. youth, high parental educational attainment was associated with a lower frequency of youth marijuana use.
However, race and ethnicity moderated this association. We observed a weaker inverse association for Hispanic than non-Hispanic White youth. As a result, Hispanic youth with highly educated parents remained at higher-than-expected risk of frequent marijuana use.
The first finding is in line with fundamental cause theory, the social determinants of health framework, and other SES effects (Clouston & Link, 2021;D. Hudson, Banks, Holland, & Sewell, 2019;Link & Phelan, 1995;Phelan, Link, Diez-Roux, Kawachi, & Levin, 2004). The second finding aligns with recent observations that the effects of SES indicators such as parental education on obesity, depression, anxiety, suicide, internalization, externalization, and self-rated health are all weaker for Hispanic than non-Hispanic White youth and adults. These MDRs may explain why we observe a higher-than-expected risk of chronic diseases ( also shown in high SES non-Hispanic Black and Hispanic people. However, this literature is mainly on tobacco products Assari & Lankarani, 2016). The unique contribution of this work is the expansion of this literature to marijuana use frequency of youth.
A wide range of structural, social, and behavioral mechanisms may explain these MDRs, interfering with the returns of parental educational attainment and how it translates to employment, income, wealth, and residential area. Most of these processes are racialized in the US, so they generate less advantageous outcomes for Hispanic and non-Hispanic Black families (Assari, 2017;. For example, highly-educated parents tend to work in jobs with lower pay and lower occupational prestige if they are Hispanic or non-Hispanic Black. Similarly, highly educated non-Hispanic Black and Hispanic people tend to work in jobs with higher stress and exposure to toxins relative to their White counterparts (S. . The racial composition of the workplace may also be associated with discrimination for highly educated non-Hispanic Black and Hispanic employees (Assari & Moghani Lankarani, 2018b). As a result, highly educated racial/ethnic minority families (Assari, 2017; remain at risk of economic insecurity , stress (Shervin Assari & Mohsen Bazargan, 2019b), living in poor residential areas , and low wealth (Assari, 2020). Thus, interwoven complex social processes may explain why highly educated Hispanic and non-Hispanic Black families remain at behavioral, economic, and health risk.

Limitations
We acknowledge several limitations. First, we used cross-sectional data, which limited our ability to infer temporal ordering or cause. Nevertheless, it is reasonable to assume that parental educational attainment typically precedes youth substance use. Second, the sample size was much more limited for Hispanic and non-Hispanic Black than non-Hispanic White youth, perhaps limiting our ability to compare race and ethnic specific models. However, these differences are present in almost all studies of nationally-representative samples.

Implications
Results suggest that addressing substance use in ethnic minority youth may require more than interventions that seek to simply improve education of ethnic minority families or reduce poverty.
Indeed, we suspect that achieving this aim will require first identifying the processes responsible for the MDRs observed here and in other studies (Assari, 2020) and then seeking to address them with a series of multilevel societal policies designed to equalize the living conditions of individuals and families of different ethnicities. This distinction is important because solutions to the inequalities and disparities than drive the lower returns of SES indicators for Hispanic families (i.e., MDRs) will likely differ from solutions to the problem of lower educational attainment in these families. Unless we go beyond poverty elimination to address the drivers of MDRs, SES indicators such as parental education may continue to operate as a source of ethnic health disparities rather than a solution.

Conclusion
As shown here, SES indicators such as parental educational attainment show different patterns of association with youth marijuana use frequency across ethnic groups. Youth from highly educated Hispanic families remain at risk of frequent marijuana use, a pattern different from their non-Hispanic White counterparts. Thus, ethnic disparities in youth marijuana use frequency remain across the full SES spectrum. These results are consistent with the MDRs framework and suggest that ethnic health disparities should not be reduced to the problem of poverty or low human capital. These MDRs may reflect social stratification, structural racism, and marginalization that negatively impact the behavioral health of ethnic minority families across SES levels.
Assari research is also supported by the following National Institutes of Health (NIH) awards: U54CA229974, U54MD008149, U54MD008149, R25MD007610, U54MD007598, U54TR001627, and CA201415-02.We used the publicly available Monitoring the Future 2020 data for this analysis (Miech et al., 2021). Monitoring the Future is supported by the following NIH grants from National Institute on Drug Abuse (grants R01-DA001411, K24-DA048160, and K01-DA042950).