Preventing the pack size effect: Exploring the effectiveness of pictorial and non-pictorial serving size recommendations

People eat more from large than from small packs, which is known as the pack size effect. We hypothesized that providing a serving size recommendation would reduce the influence of the pack size on consumption and would thus diminish the pack size effect. Moreover, we hypothesized that a pictorial serving size recommendation, displaying food amounts visually, would be more effective than a non-pictorial recommendation that communicates the recommended amount in grams only. We tested these hypotheses in two online experiments (N = 317 and N = 324) and in one lab experiment (N = 89). In the online experiments, participants were shown a small or a large pack of unhealthy snacks, with or without a serving size recommendation. The main outcome measure was expected consumption. Replicating the pack size effect in an online setting, we found that participants expected to consume more food from large than from small packs. Furthermore, the pack size effect was considerably stronger for men than for women. Importantly, when including portion size preferences as a covariate, the pictorial serving size recommendation significantly reduced expected consumption, especially when placed on a large pack, as hypothesized. The non-pictorial serving size recommendation had no effect. In the lab experiment, students received a large bag of M&M's which did or did not contain the pictorial serving size recommendation. We again included general portion size preferences as a covariate. The serving size recommendation significantly lowered the amount of M&M's that participants served themselves, but only when participants reported to have noticed the serving size recommendation. We conclude that providing a pictorial serving size recommendation can be an effective intervention strategy to reduce the pack size effect, if it attracts sufficient attention.