Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


Animal Nutrition and Food Science

First Advisor

Rachel K. Johnson


The majority of U.S. children do not meet recommendations for fruit and vegetable (FV) consumption. FV confer numerous health benefits including intake of critical shortfall nutrients, reduced risk of chronic diseases, and maintenance of healthy weights. Identifying both policy and behavioral approaches to increase children's consumption of a variety of FVs has become a public health priority. When evaluating policy and interventions it is integral to apply validated dietary assessment methods to measure FV selection, consumption, and waste. The three aims of this dissertation addressed the feasibility of dietary assessment methods and their application to evaluate policy change and behavioral interventions.

Aim 1: Apply validated dietary assessment methods to evaluate the effect of national school lunch policy change on elementary school children's FV consumption. As of the 2012 school year, the United States Department of Agriculture (USDA) requires school children to select either a fruit or a vegetable with a reimbursable meal. The objective of the first study was to compare children's FV choices in two school cafeteria environments a year before these new USDA regulations took effect. Elementary school (grades 3-5) children's FV choices were measured during Spring 2011 at two northeast elementary schools (NES-A and NES-B). NES-A had a self-serve salad bar, whereas NES-B served pre-portioned FV and pizza daily. Of 555 trays assessed (n=284 NES-A, n=271 NES-B), 15.3% (n=85) had no FV selected. A higher percentage of trays from NES-A had no FV (23.6%, n=67) in comparison to NES-B (6.6%, n=18) (P<0.0001). On average children selected more processed FV (PFV) (80.8g) than whole FV (WFV) (40.5g, P<0.001). The mean amount of FV selected was lower in NES-A (111.4g) than NES-B (131.5g, P<0.01). When trays without a FV were removed, quantities selected were not significantly different between schools (P=0.46). For PFV, 100% fruit juice was on 41.4% of trays (n=230) and pizza was on 42.1% of NES-B trays (n=114). Trays with pizza or 100% fruit juice were less likely to have a WFV (P<0.001, P<0.0001 respectively). In the second study, elementary school children's FV selection, consumption and waste were compared in two northeast elementary schools before the USDA rule in spring 2012 (10 school visits, tray observations n=498) and following the USDA rule in spring 2013 (11 school visits, tray observations n=944). More children selected FVs in higher amounts when FVs were required compared to when they were optional (0.69 cups vs. 0.89 cups, P<0.001); however, consumption decreased slightly (0.51 cups vs. 0.45 cups, P=0.01) and waste increased (0.25 cups vs. 0.39 cups, P<0.001).

Aim 2: Address the feasibility of non-research volunteers (teachers and parents) collecting digital imaging (DI) dietary assessment data in a sample of elementary schools. Two Northeast Elementary Schools (NES-A and NES-B) were recruited and a parent-based dietary assessment team (PDAT, n=5) and teacher based dietary assessment team (TDAT, n=4) were formed. We compared data collected by the PDAT and TDAT with the university-based dietary assessment team (UDAT). Feasibility was measured based on the total number of DI pairs collected out of the total number of eligible DI pairs across all data collection days for each of the three teams. Using binary logistic regression, at NES-A, the PDAT was less proficient at collecting DI pairs (74.1%, n=218 of 294 LD) than the UDAT (81.9%, n=262 of 320 LD, P<0.05). At NES-B, the TDAT was better able to capture DI pairs (95.9%, n=257 of 268 LD) than the UDAT (91.3%, n=366 of 401 LD, P<0.05).

Aim 3: Apply weighed plate waste (WPW) as a validated dietary assessment method to evaluate a behavioral intervention addressing pre-school aged children's FV consumption during afternoon snack time using older elementary school children as 'FV Mentors'. Two Northeast pre-school classrooms (NEPC-A and NEPC-B) at the Burlington, VT YMCA were recruited for the study. Children in grades 3-5 who participated in the Live Y'ers Afterschool program were recruited to model FV consumption and use FV verbal cues during the intervention period in addition to teacher FV verbal cues. In NEPC-A (n=15) based on a repeated measures ANOVA with mixed design analysis, there was a significant main effect of time on FV consumption as measured by the mean amount of cups of FVs consumed [F(2,10)=7.89, P=0.009] across study periods. Mean consumption was lowest at baseline at 0.16 cups (95% CI: 0.10-0.22) and increased during both the intervention period (M=0.26 cups, 95% CI: 0.17-0.36) and the follow-up period (M=0.33 cups, 95% CI=0.28-0.38). The main effect of time (study period) was qualified by a significant interaction between time and type of FVs consumed [F(8,10)=3.10, P=0.049] indicating that the effect of study period on FV consumption depended on the type of FVs consumed. In NEPC-B (n=16) there was not a significant main effect of time on FV consumption [F(2,10)=1.10, P=0.372].



Number of Pages

133 p.

Available for download on Monday, September 11, 2017

Included in

Nutrition Commons