The World Health Organization considers obesity “one of the most serious health challenges of the 21st century”. In the U.S., nearly one-fifth of children and adolescents are obese, putting them at risk for a range of health problems. Many interventions focus on improving diet quality, which is associated with a lower risk of obesity. Nutrition education has been widely suggested as a key strategy to improve children’s food choices, but research has shown that, at best it leads to modest changes in nutrition knowledge, attitudes, behaviors, or body mass index. Corroborating these results, in an earlier study, my co-author John List and I found that providing short educational messages to students immediately prior to asking them to make a choice about a snack did not improve the healthfulness of the snack selected.
In a recent paper, John List, Chien-Yu Lai and I address this problem by using a random controlled trial (RCT) to explore the impact of using informational prompts in the school lunch-line on milk choice. We conducted the study in several schools with 2500 students in grades 1-6. Children in these schools have an option between white milk or sugar-sweetened chocolate milk. We decided to focus on the choice of milk because research suggests that reduction of sugar-sweetened beverages is one of the primary ways to reduce childhood obesity (see for example here, here and here). Moreover, at baseline, 80% of students in our schools are choosing chocolate milk over white milk and school administrators wished to decrease this percentage, without decreasing the overall consumption of milk (which they worried might happen if chocolate milk were simply taken off the menu). We randomly assigned some children to treatments in which they received a verbal prompt from the cafeteria worker while walking through the lunchline, received an incentive, or did not receive either, and compared their milk choice and consumption.
We were motivated to study prompts in the school lunch-line because unlike nutrition education alone, prompts convey a request to engage in a desired behavior. We hypothesized that directed verbal prompts encouraging a child to “Try the white milk” may be more effective than nutrition education alone. In addition, we in order to explore the channels through which prompts may operate, we also investigated the verbal prompt to “Try the white milk, it is good for you” and “Try the white milk, it tastes good.” We also compared the impact of our prompts with small incentives – glow-in-the-dark wristbands that were attached to the healthier white milk – since incentives have shown large impacts on child food and beverage choice in prior work, yet are not as scalable as a simple verbal prompt (see here and here).
What we found was surprising: to encourage children to forego sugar-sweetened chocolate milk in favor of white milk, you just have to ask. All prompts statistically significantly increased the selection of white milk over chocolate milk from about 20 percent to 30 percent. This effect was nearly as large as the effect of the incentive. Interestingly, all types of prompts – regardless of the information presented – resulted in similar positive effects. We surveyed students and found that most prompts (even when they didn’t mention health) worked by increasing the perceived healthfulness of the milk.
Once the children choose the white milk, how can we be sure they actually drank it? At the end of the lunch period, we asked children to leave their trays at their tables and we weighed each milk carton to measure consumption. We find similar consumption rates for all milk types in all treatments, meaning that the prompt not only improved selection of white milk, it also improved consumption of white milk.
In light of our results, schools should be aware that simple prompts by their staff really do make a difference in the health of students. Schools should make teachers and staff aware of the impact their words can have, and encourage them to prompt children about healthy choices when they can.