Clustering of Youth Empowerment and Health Behaviors Among Sixth to Eighth Grade Adolescents

Study Design, Settings, Participants, Intervention: The sample consisted of n¼263 students nested in 18 classrooms in North Carolina and Ohio. Nine classrooms used FMI and nine used traditional mathematics and science curriculum in a quasi-experimental design. Outcome Measures and Analysis: Multi-level models were used to assess the relationship between student-level variables including: gender, free and reduced lunch status (as a proxy for student socioeconomic status), and baseline nutrition knowledge; classroom-level variables including: control or intervention status, teacher experience and teacher nutrition training on student nutrition knowledge at the conclusion of the program. A total of five models were built and analyzed using PROC MIXED and estimated using maximum likelihood in SASv9. 4. Conclusions and Implications: Classroom-level variables including curriculum and teacher characteristics accounted for $56% of variability in nutrition knowledge after the program. The FoodMASTER intervention was found to be a significant predictor of nutrition knowledge. Students exposed to FMI score on average 4.32 points higher than control classrooms (p¼. 0001). Teachers with 8+ years of experience were found to improve predicted post-nutrition knowledge scores (p¼. 006). Based on these analyses, integrated nutrition education such as FoodMAS-TER significantly increases student nutrition knowledge. Further, there is a moderating influence of teacher experience indicating that effectiveness of nutrition education programs may be improved when administered by experienced teachers