Project:
Breakfast in the classroom, body mass index, and academic outcomes

Year: 2019

Investigator: Thomsen, Michael R., Rodolfo M. Nayga, Jr., and Anthony Goudie

Institution: University of Arkansas

Project Contact:
Michael R. Thomsen
University of Arkansas
Division of Agriculture
217 Agriculture Building
Fayetteville, AR 72701
Phone: 479-575-2256
Email: mthomsen@uark.edu

Low participation in the National School Breakfast Program (NSBP) is problematic because there is evidence that eating breakfast conveys a myriad of benefits to children. Breakfast consumption is associated with improved diet, which has been shown to enhance cognition among school-aged children and may be protective against excess weight gain. A traditional school breakfast schedule requires children to arrive at school before the beginning of classes and is one factor that contributes to lower NSBP participation. In recent years, attention has focused on alternative breakfast delivery schedules that provide “breakfast after the bell” (BAB) as one way to increase participation in the NSBP. BAB has been shown to meaningfully increase participation in school breakfast and reduce the number of schoolchildren who skip breakfast entirely.

There is a need to better understand the impact of BAB on academic and health outcomes. On one hand, BAB could lead to higher achievement, especially among children at risk for poor diets or food insecurity. By increasing the number of meals served, BAB results in additional federal reimbursement dollars, which may translate into higher quality school nutrition programs. On the other hand, some educators have expressed concern that BAB is costly in terms of class time and cleanup, which may adversely impact academic performance. Earlier evidence suggests that some children eat twice in the morning after BAB implementation, which may exacerbate excess weight gain.

This project consisted of two separate studies focused on the impact of BAB on Arkansas elementary schoolchildren. Arkansas is an important case to study because the State ranks among the lowest in student academic outcomes and among the highest in childhood obesity. Moreover, there is a need to understand the impact of BAB adoption in contexts aside from large urban school districts that have been the primary focus of earlier studies.

The first study used administrative data to examine the effect of BAB on standardized test scores in mathematics and English among third-graders. These data show that BAB-adopting schools were systematically different from non-adopting schools in that children in BAB schools tended to be less affluent and had lower baseline test scores. Thus, in order to make valid comparisons, each BAB-adopting school was compared to an optimal “synthetic” control school. The synthetic controls were derived from non-adopting schools by minimizing the pre-adoption differences in observable characteristics. The study focused specifically on 23 elementary schools adopting BAB in 2015. The second study used data from a legislatively mandated BMI screening program to assess the impact of BAB on weight outcomes by the second grade. In the second study, it was possible to use data from all schools that adopted and continuously offered BAB in 2013 or later. In addition to a generalization of the synthetic control method already described, BAB program effects were estimated from a sample of children followed from kindergarten through second grade using propensity-score adjusted regressions.

There was limited evidence that BAB delivery impacts academic achievement in this sample of Arkansas third graders. Estimates that were statistically different from zero were small in overall magnitude and were of limited practical significance. To better understand this “null” finding, three mechanisms by which BAB adoption could impact academic achievement were further explored. First, there was a meaningful association between BAB and NSBP participation, which suggests that a lack of an effect on participation is an unlikely explanation for the limited impact of BAB on test scores. Second, there was no evidence that BAB led to meaningful improvements in attendance as measured by the percentage of days that students attend school before and after BAB adoption. Third, there was strong evidence that infractions for disruptive behaviors like fighting, disorderly conduct, and insubordination declined in schools that adopted BAB. The declines in disciplinary infractions were statistically meaningful and persisted over the entire post-adoption period. In sum, while there was not evidence of meaningful changes in test scores as a result of BAB, there were improvements in both breakfast participation and in student behavior after BAB was adopted.

In the second study, findings from the generalized synthetic control approach showed no meaningful effect of BAB on average BMI, the proportion of children who were obese, or the proportion of children who were overweight or obese. Propensity adjusted regressions similarly showed no significant effects of BAB on second-grade BMI among a sample comparing children in BAB schools to children in non-BAB schools. There were also no significant effects in subsamples by race, ethnicity or school meal status. In sum, these data suggest there is no evidence that BAB is detrimental in terms of excess weight gain.

While it is discouraging that the project did not find stronger evidence of a beneficial link between BAB and academic achievement or BMI, the findings reported here are important because they suggest that BAB can fit into the rhythm of the school day without adversely impacting academic achievement. School is a prominent setting in the developmental milieu for elementary-aged children. By increasing participation in school meal programs, BAB may lead to improved diets and increased food security among at-risk children thereby contributing to better long-term health and quality of life. It is also important to emphasize that the null findings reported here do not contradict earlier evidence showing positive effects of NSBP participation on academic achievement. BAB affects NSBP participation by increasing participation within the school (the intensive margin) while earlier findings on the link between school breakfast and academic performance have been documented by analyzing policies that affect the extensive margin, the number of schools participating in the NSBP. BAB may have meaningful impacts on those students for whom it enables NSBP participation, but this study is unable to identify these specific students in the administrative datasets used here. The average program effects reported thus represent “intention to treat” estimates.