Working out Consumption: Designing a Robust Information Intervention for Healthful Eating

Providing consumers with calorie information has received a great deal of attention, as municipalities such as New York City have mandated greater disclosure and retailers (including Panera and McDonalds) now provide calorie information on their menu boards nationwide. However, tests of providing decision makers with information on the calorie content of food has yielded very mixed results (Bollinger, Leslie, & Sorensen, 2010; Elbel, Gyamfi, & Kersh, 2011; Loewenstein, 2011). Using the idea of channel factors (Leventhal, Singer, & Jones, 1965), we develop a low cost and robust mapping intervention (Thaler & Sunstein, 2008) which makes the welfare consequences of choices more easily evident to the decision maker. In a field study conducted at the end of December and beginning of January, museum visitors (n=1127) were offered a choice of candy bars before completing an unrelated survey as a compensation for their time. The survey was about gift card preferences, and a few demographic questions were asked at the end. Visitors chose between four different candy bars whose calorie content ranged from 70 to 280 calories in increments of 70 calories (Snickers, 280; Hersheys, 210; Chips Ahoy, 140; M&Ms 70). Notably, the candy bars all disclose calories on their packaging. The museum patrons chose a candy bar from one of four boxes which either showed in large font (a) just the names of the candy bars, (b) the names and calorie information, (c) names with calories and “exercise equivalents” (e.g., number of minutes of running or biking required to burn the calories). Ordinal regression revealed that, compared to the no information disclosure condition, visitors consumed significantly fewer calories, on average, in the conditions where exercise equivalents were additionally made available (mean = 177 calories vs. 199 calories; Wald=9.4, p.01). However, the reduction in calorie consumption was only marginally significant when calorie information alone were disclosed. The effect of exercise equivalents was robust across age, income, and gender.