Related Experiment Video
Updated: Jan 14, 2026

Errors as a Means of Reducing Impulsive Food Choice
Published on: June 5, 2016
Food label granularity and working memory: effects on food choice in a randomized controlled trial
1Department of Social Statistics, University of Manchester, Manchester, UK. constanza.avalos@postgrad.manchester.ac.uk.
Objective:
While prior research on front-of-package labels has focused on their design, the impact of their informational granularity-or level of detail-on consumer choice remains understudied. This study aimed to investigate how the informational granularity of food calorie labels, in conjunction with consumers' working-memory capacity, influences choice behavior.
Methods:
A randomized controlled trial was conducted (UK adults, N = 498) in which participants completed a cereal-shopping task after being assigned to one of three conditions: coarse label, detailed label, or no label. Working-memory capacity was evaluated via a three-level n-back test. The primary outcomes measured were the average number of calories per chosen product and the probability of selecting low-calorie options. Multilevel models were employed to test both main and interaction effects, with the Benjamini-Hochberg adjustment applied to control the false discovery rate.
Results:
Compared with the control condition, food labels reduced mean calorie selection by 4.56%. Each one-unit increase in 3-back performance resulted in an additional 14 kcal reduction under coarse labels and 18 kcal under detailed labels. Individuals with high working memory capacity exhibited a preference for moderate-calorie products over the lowest-calorie products when detailed labels were displayed. The effects of capacity under coarse labels were less pronounced and primarily limited to avoiding the lowest-calorie option. These findings provide conceptual insights into information optimization and offer practical guidance for policymakers on designing effective FOP schemes by specifying when detail helps versus hinders consumer choice.

