This paper contributes to the debate on front-of-pack nutritional labeling for food. We implement an incentivized laboratory experiment to assess the relative performance of Guideline Daily Amount (GDA) and Traffic Light (TL) labeling schemes in assisting consumers to build a healthy daily diet.

GDA provides detailed nutritional information per serving, as % of an officially recommended daily consumption basket. TL displays color-coded coarser information about the amounts of key nutrients.

From a behavioral point of view, GDA and TL sit at two ends of a continuum. GDA gives detailed information, but requires high computational and memory skills to be effective in shaping consumer behavior. TL gives coarser information, but requires lower cognitive skills. GDA is designed for careful, skilled, unconstrained decision makers, and would be perfect for homo oeconomicus. Behaviorally, though, it has been observed that too much information can lead to bad choices, especially if the information is multidimensional; TL could fare better for cognitively constrained consumers.

Previous studies have compared the performance of GDA vs. TL (for a review see Grunert 2007). Most of the existing literature, though, is based on simple tasks consisting in ranking two products according to their perceived healthiness. Moreover, a thorough survey of the literature reveals that the question asked determines the relative performance of the labeling schemes. When the subjects are asked to rank the products' relative healthiness, or to classify them into healthy/medium/unhealthy, TL wins. When instead the subjects are asked absolute levels, i.e. to evaluate how much of a nutrient is present in each product, GDA wins. This is not surprising: labels perform best when the question asked is the one they have been designed to answer.

But the real goal of nutritional labels is to help consumers to build healthy diets. In our experiment we ask the subjects (engineering students and a representative sample of the Grenoble population) to do exactly that. Our subjects act as hired nutritionists of a refectory. They must compose a daily diet, choosing from a predetermined set of products, and are paid in cash only if the diet they built satisfies a set of nutritional goals. To guide the subjects in their choices, nutritional labels are provided. We implement three treatments: GDA, TL, and a mixed GDA+TL label.