A Sneaky Experiment About Calorie Counts

By
Green Traffic Light
Photo: Con Tanasiuk/Corbis

What’s the best way to get people to eat healthier foods? Calorie counts? Traffic-light-style indicators of nutritiousness? Both? At the moment, a fair amount of research suggests traffic lights are better, but a new study in the Journal of Public Policy and Marketing complicates things a bit.

For the study, a team led by Dr. Eric M. Van Epps of the University of Pennsylvania worked with an “on-site corporate restaurant,” where they “developed an Internet-based system through which employees could place lunch orders, which they then picked up at a central location.” This allowed them to “present participation in the study as a pilot testing period for the online system, without drawing attention to calories or nutrition,” meaning that it served as a somewhat natural experiment. Participants were wooed with discounts on lunches, as well as small fees for filling out entrance and exit surveys.

Over the course of a month and a half, participants were randomly assigned to either a control group with no nutritional information, or one of three other groups: “numbers [calories] only, traffic lights only, and numbers plus traffic lights.” The researchers found that, compared to members of the control group, the members of the other groups did order significantly fewer calories, but they didn’t find much support for their idea that combining numbers and traffic-light imagery would be the best approach (it was merely as good as the others).

This study’s interesting partly because it runs counter to prior research suggesting that calorie counts don’t really work. But those studies have been conducted in restaurant settings, so it’s possible that the environment matters — the researchers suggest there may be something different about online ordering that makes calorie counts more effective. Until we know more, all we can say is that this latest study should slow down some of the pessimism about calorie-count approaches.