Does Neighborhood SES or School SES Matter More for Healthy BMI Among Latinx Pre-K Students in South Texas?

Background and Purpose: The prevalence of obesity amongst Kindergartners in south Texas is above the national average. South Texas has a large Latinx population, and obesity disproportionately affects racial/ethnic minoritized groups. Latinx children in south Texas tend to come from families with lower socioeconomic status (SES). Research on the impact of family SES on obesity is largely consistent, showing that the relationship between SES and obesity in childhood has a negative association in the U.S. (higher SES = healthier status). At the school and neighborhood levels, less research has been conducted. Some have examined one or the other separately with multilevel methods, only a handful have examined the effect of varying levels of SES and how they interact, using cross-classified multilevel models.

Methods: This study is part of a larger ongoing longitudinal evaluation of a parenting intervention in partnership with pre-K classes in a city in south Texas. Data for this study came from survey data collected from parents of pre-K students in the fall of 2019, i.e., baseline (N=269). This cross-sectional, cross-classified multilevel model study consisted of students representing thirty-two census tracts and twelve schools. Our analytic sample only included participants who identified as non-white Hispanic/Latino (n = 237, 89%). The dependent variable was child body-mass-index z-score (BMIz). Individual-level covariates were child’s sex, primary caregiver’s marital status, education level, relationship to child, and family SES expressed as income-to-needs ratio. There were two level-two variables. The first was neighborhood poverty, expressed as percent of residents below federal poverty line in census tract of residence. The second was school-level SES, derived from aggregate individual-level (family) SES data. A bottom-up approach for cross-classified multilevel model building was utilized to examine the unique associations of individual-, neighborhood- and school-level variables with individual student BMIz, and how they interact.

Results: The intra-class correlation coefficient (ICC) was 0.28 and 0.06 for school and neighborhood, respectively. The mean BMI z-score was 0.49, within healthy range, and 24% and 5% of the sample was classified as overweight and obese, respectively. The cross classified multilevel linear regression models resulted in a significant association between school SES and BMIz (B = –0.13; SE=0.06; P<.05), such that individual students’ BMIz decreased by approximately 0.13 points for each one-thousand dollar increase in the school SES, net of controls. Additionally, a significant association between neighborhood poverty and BMIz (B = –1.41; SE=0.49; P<.01), such that individual students’ BMIz decreased by approximately 1.41 points for each 1-point increase in the neighborhood poverty, net of controls.

Conclusions and Implications: The current study provides evidence that SES-related obesity disparities exist as early as Pre-K, indicating it is appropriate to implement a longitudinal intervention in hope of reducing or preventing obesity into adolescence. The study did not provide results to help understand which level of SES may be more important to protecting against obesity in early childhood, but it is clear that neighborhood SES and school SES do play a role. The differing findings provide a foundation for future studies from the larger study.