The Influence of the School Breakfast Program and National School Lunch Program on Child Weight and Academic Achievement

Year: 2012

Research Center: Southern Rural Development Center, Mississippi State University

Investigator: Capogrossi, Kristen

Institution: Virginia Tech University

Project Contact:
Kristen Capogrossi
Christensen Associates, Inc.
800 University Bay Drive, Suite 400
Madison, WI 53705-2299
Phone: 608-216-7156


Two types of childhood misnourishment plague the United States and have far-reaching consequences on children: overweight and underweight. The term ‘misnourishment’ is used because it is usually a child’s dietary and nutritional quality that causes problems of overweight and underweight in the United States as opposed to a lack of food. In 2008, approximately 13 million children and adolescents were obese and 2.4 million were underweight. Some research has linked childhood misnourishment to poor academic performance in school. For instance, conditions associated with overweight and underweight may cause children to miss more school or have lower levels of concentration resulting in poorer performance. Furthermore, other unobserved factors such as certain home and school environment characteristics may be influencing weight and performance causing spurious correlations rather than an actual causal relationship between weight and performance. The school meal programs are now being targeted as potential policy instruments to combat child weight. In 2010, more than 31.7 million students participated in the National School Lunch Program (NSLP) each day at a cost to the Federal Government of $10.8 billion, while the School Breakfast Program (SBP) served more than 11.6 million children daily at a cost of $1.9 billion. While both programs are intended to provide children with nutritionally adequate meals, research has found that much of the food served through the programs has been of low nutritional quality. Because of the links found between child weight and academic performance as well as between school meal program participation and child weight, further research on the relationship among all three is crucial not only for alleviating childhood misnourishment but also for improving U.S. children’s academic performance.

This study was drawn from the Early Childhood Longitudinal Study–Kindergarten Class (ECLS–K) which is a longitudinal study of a nationally representative cohort of 21,260 kindergartners beginning in the 1998-1999 school year and who are followed through 8th grade (2006-07 school year). The study collected data on children in over 1,000 different schools, as well as on their families, teachers and school facilities to examine early childhood experiences, early childhood development and early school performance.

The study used a unique theoretical framework and employed a variety of estimation methods innovative in this area of research to aid in identifying non-spurious relationships. It examined the impacts of SBP and NSLP on child misnourishment status (overweight and underweight) accounting for self-selection into programs and allowing for multiple simultaneous treatments as students have the option to participate in none, one or both programs. A multiple overlapping treatment investigation of school meal programs is necessary because students have the option of participating in SBP and NSLP simultaneously: 25 percent of the students in the sample participated in both programs simultaneously at some point in elementary and/or middle school.

The study specifically utilized average treatment effect on the treated and difference-in-difference methodologies. Furthermore, the study also identified direct, indirect, and total effects of SBP and NSLP on academic performance. Total effects provide the results of a simultaneous change in all inputs from a change in a single exogenous variable. The ability to identify these effects is significant because it allows an answer to different research questions. This is important as policymakers often have several targets of interest such as program participation’s influence on child weight and its consequential influence on academic performance. The study used structural equation modeling to estimate selection equations, child choice equations as well as weight and cognitive production functions simultaneously.

Overall, the study found long-term program impacts to be more significant for obese, overweight, and healthy-weight children. Long-term participation looks at those students who consistently participated in meal programs from 1st to 8th grades. The study showed that while participating only in NSLP decreased the probability of children being overweight, participating in both SBP and NSLP simultaneously increased the likelihood of being overweight and decreased the likelihood of being underweight, particularly for free or reduced-price (FRP) recipients. In examining only FRP eligible students, participation in NSLP decreased body mass index (BMI) z-scores (standardized results of multiple data sets) by approximately 14 percent. The study also found that participating in SBP tended to increase child weight, although the magnitude was not large, and the impact was particularly seen on students paying full-price for breakfast. Further, no direct or indirect impacts of SBP participation on child BMI were revealed when separating by whether or not the child was eligible for FRP. Past BMI, urbanity, television viewing and physical activity levels were the most significant predictors of a child’s BMI in these models.

When examining impacts including per pupil food expenditures as an indicator of food quality as well as based on the percentage of a school’s students eligible for free lunches (the higher the proportion of free-lunch eligible children indicates a greater likelihood of a lower-income school that may devote fewer resources to food services or have more demands for free meals which causes less attention paid to food quality), results do not differ from the original specification. This indicates that the programs have similar impacts on weight whether or not the school is considered “poor.” When examining impacts by urbanity, the study showed that participation in only NSLP increases BMI z-scores in rural areas, and participation decreases the probability of overweight and increases the probability of underweight in the suburbs. In examining impacts by region (Northeast, Midwest, South and West), the study found that participating in both SBP and NSLP increased child weight for those students in the Midwest, and participation in only NSLP increased child weight for 8th grade students in the South and West, indicating that cuisine may play a factor in the nutritional quality of food served. Furthermore, participation rates of rural households were higher in the South with more than 25 percent and 23 percent of all rural households with children in the South participating in NSLP and SBP, respectively.

Findings reveal that effects of program participation on achievement do exist, particularly when examining children by FRP eligibility status. These results further encourage participation in the meal programs. Results indicate that there were both direct and indirect effects from program participation on 8th grade student achievement. When examining all school meal program participants, the study found that there was a large positive direct impact of NSLP participation on math scores while there were small indirect impacts of SBP and NSLP participation on math and English scores. Additionally, attending a Title 1 school (schools with a high percentage of low-income students) had a large negative impact on both math and English scores, which may contribute to the increasing disparities between high- and low-income schools. This could also be evidence of a lack of resources available to students in Title 1 schools.