Date of Award


Degree Name



Public Health


Leah D. Whigham


Background: School gardens are an educational tool used to increase children's exposure to, and intake of, fruits and vegetables (F/V). Diets rich in F/V reduce the risk of heart disease, stroke, some cancers, and type 2 diabetes, and can aid in weight management as a strategy to reduce obesity prevalence. Objective data of the impact of school gardens on F/V intake are lacking, as most studies have relied on dietary self-reporting to validate F/V intake. Self-reporting of dietary intake is inherently an unreliable method for assessing F/V intake. Carotenoids, a family of phytochemicals, are a reliable biomarker of dietary intake of F/V. Reflectance spectroscopy (VEGGIE METER) allows assessment of carotenoids in humans that is non-invasive, cost-effective, and efficient.

Purpose: The goal of this study was to evaluate the impact of a high school garden and its curriculum on F/V intake among participants from baseline to follow-up using carotenoids as a biomarker.

Methods: In this 9-month intervention, 149 high school students (99% Hispanic, 54% male, age 14-19y) residing in El Paso, TX participated in classrooms using a school garden and related curriculum or control classrooms. Measures included reflectance spectroscopy for skin carotenoids (OD, optical density) and self-efficacy for F/V intake. Means were compared by paired t-test. Multiple linear regression analysis was used to determine if participation in the garden intervention predicted a change in F/V intake or self-efficacy, while controlling for sex, age, and baseline carotenoid levels and self-efficacy.

Results: There was no significant change in skin carotenoid levels in response to the intervention or in the control group (% change, control vs. garden: 4.0% vs. -2.6%, p = NS). There was also no change in self-efficacy for F/V intake (% change, control vs. garden: -1.4% vs. 4.8%, p = NS). Multiple linear regression analysis to predict change in skin carotenoids including three predictors (intervention, sex, and age) was not significant (R2 = .039, adjusted R2 = .001, F(3, 77) = 1.037, p = 0.4). The analysis to predict change in self-efficacy for F/V intake including three predictors (intervention, sex, and age) was not significant (R2 = .039, adjusted R2 = .009, F(3, 95) = 1.297, p = 0.3). A small effect size was calculated for change in skin carotenoids and change in self-efficacy for F/V intake using Cohen's Æ?2 (Æ?2 = 0.041). Further regression analyses showed significance (p < 0.05), but only baseline carotenoids and baseline self-efficacy were significant predictors. Post-hoc qualitative analysis discovered that the curriculum in use did not target nutrition knowledge or behavior change related F/V intake.

Discussion: This study found no impact on F/V intake (as measured by a valid biomarker) or self-efficacy to eat F/V from participation in school garden activities. However, curriculum did not sufficiently target nutrition knowledge or behavior change. It is important to use behavior change theory in curriculum design for school-based approaches to increase F/V intake; exposure to F/V and a school garden alone is unlikely to impact intake. The ability to non-invasively measure carotenoids provides an objective tool for assessment of the impact of school garden programs on F/V intake.




Received from ProQuest

File Size

69 pages

File Format


Rights Holder

Patrick Vaughn Hopkins