How the quality of school lunch affects students’ academic performance

Students get their lunch from a salad bar at the school cafeteria as some of more than 8,000lbs of locally grown broccoli from a partnership between Farm to School and Healthy School Meals is served at Marston Middle School in San Diego

In 2010, President Barack Obama signed the Healthy, Hunger-Free Kids Act. The main goal of the law was to raise the minimum nutritional standards for public school lunches served as part of the National School Lunch Program. The policy discussion surrounding the new law centered on the underlying health reasons for offering more nutritious school lunches, in particular, concern over the number of children who are overweight. The Centers for Disease Control estimates that one in five children in the United States is obese.

Surprisingly, the debate over the new law involved very little discussion as to whether providing a more nutritious school lunch could improve student learning. A lengthy medical literature examines the link between diet and cognitive development, and diet and cognitive function. The medical literature focuses on the biological and chemical mechanisms regarding how specific nutrients and compounds are thought to affect physical development (e.g., sight), cognition (e.g., concentration, memory), and behavior (e.g., hyperactivity). Nevertheless, what is lacking in the medical literature is direct evidence on how nutrition impacts educational achievement.

We attempt to fill this gap in a new study that measures the effect of offering healthier public school lunches on end of year academic test scores for public school students in California. The study period covers five academic years (2008-2009 to 2012-2013) and includes all public schools in the state that report test scores (about 9,700 schools, mostly elementary and middle schools). Rather than focus on changes in national nutrition standards, we instead focus on school-specific differences in lunch quality over time. Specifically, we take advantage of the fact that schools can choose to contract with private companies of varying nutritional quality to prepare the school lunches. About 12 percent of California public schools contract with a private lunch company during our study period. School employees completely prepare the meals in-house for 88 percent of the schools.

To determine the quality of different private companies, nutritionists at the Nutrition Policy Institute analyzed the school lunch menus offered by each company. The nutritional quality of the menus was scored using the Healthy Eating Index (HEI). The HEI is a continuous score ranging from zero to 100 that uses a well-established food component analysis to determine how well food offerings (or diets) match the Dietary Guidelines for Americans. The HEI is the Department of Agriculture’s preferred measure of diet quality, and the agency uses it to “examine relationships between diet and health-related outcomes, and to assess the quality of food assistance packages, menus, and the US food supply.” The average HEI score for the U.S. population is 63.8, while the median HEI score in our study is 59.9. In other words, the typical private company providing public school lunch in CA is a bit less healthy than the average American diet.

We measure the relationship between having a lunch prepared by a standard (below median HEI) or healthy (above median HEI) company relative to in-house preparation by school staff. Our model estimates the effect of lunch quality on student achievement using year-to-year changes between in-house preparation of school meals and outside vendors of varying menu quality, within a given school. We control for grade, school, and year factors, as well as specific student and school characteristics including race, English learner, low family income, school budget, and student-to-teacher ratios.

We find that in years when a school contracts with a healthy lunch company, students at the school score better on end-of-year academic tests. On average, student test scores are 0.03 to 0.04 standard deviations higher (about 4 percentile points). Not only that, the test score increases are about 40 percent larger for students who qualify for reduced-price or free school lunches. These students are also the ones who are most likely to eat the school lunches.

Moreover, we find no evidence that contracting with a private company to provide healthier meals changes the number of school lunches sold. This is important for two reasons. First, it reinforces our conclusion that the test score improvements we measure are being driven by differences in food quality, and not food quantity. A number of recent studies have shown that providing (potentially) hungry kids with greater access to food through the National School Lunch Program can lead to improved test scores. We are among the very few studies to focus on quality, rather than food quantity (i.e., calories). Second, some critics of the Healthy, Hunger-Free Kids Act worried that by raising the nutritional standards of school lunches that fewer children would eat the food, thereby unintentionally harming the students that the law was designed to help. Our results provide some reassurance that this is not likely to be the case.

Finally, we also examine whether healthier school lunches lead to a reduction in the number of overweight students. We follow previous literature and use whether a student’s body composition (i.e. body fat) is measured to be outside the healthy zone on the Presidential Fitness Test. We find no evidence that having a healthier school lunch reduces the number of overweight students. There are a few possible interpretations of this finding, including that a longer time period may be necessary to observe improvements in health, the measure of overweight is too imprecise, or that students are eating the same amount of calories due to National School Lunch Program calorie meal targets.

Education researchers have emphasized the need and opportunity for cost-effective education policies. While the test score improvements are modest in size, providing healthier school lunches is potentially a very cost-effective way for a school to improve student learning. Using actual meal contract bid information we estimate that it costs approximately an additional $80 per student per year to contract with one of the healthy school lunch providers relative to preparing the meals completely in-house.

While this may seem expensive at first, compare the cost-effectiveness of our estimated test score changes with other policies. A common benchmark is the Tennessee Star experiment, which found a large reduction in the class size of grades K-3 by one-third correlated with a 0.22 standard deviation test score increase. This reduction cost over $2,000 when the study was published in 1999, and would be even more today. It is (rightfully) expensive to hire more teachers, but scaling this benefit-cost ratio to achieve a bump in student learning gains equal to our estimates, we find class-size increases would be at least five times more expensive than healthier lunches.

Thus, increasing the nutritional quality of school meals appears to be a promising, cost-effective way to improve student learning. The value of providing healthier public school lunches is true even without accounting for the potential short- and long-term health benefits, such as a reduction in childhood obesity and the development of healthier lifelong eating habits. Our results cast doubt on the wisdom of the recently announced proposal by Agriculture Secretary Sonny Perdue to roll back some of the school lunch health requirements implemented as part of the Healthy, Hunger-Free Kids Act.