Simplified Nutritional Labels Reduce Calorie of Purchases in a Cafeteria: An Abstract

To counter the alarming obesity epidemic, public health authorities have adopted front-of-pack (FOP) nutritional labels to promote healthier food choices (FDA 2017). One important FOP strategy is the use of color-coded labeling systems, such as the traffic light, which summarize food nutritional quality using colors (Thorndike et al. 2012). More recently, Europe public policy makers proposed a wider labeling coloring system, the five-color nutritional label (5CNL) (ANSP 2017). Considering the still underexplored research on the impact of the five-color nutritional label system, this research explores how the 5CNL influences the calorie content of purchases made in a university cafeteria during 3 weeks. Importantly, this research assesses the impact of the 5CNL at two moments: at the first moment of exposition to the label and after some days of exposition to the same. In the first week, we measured sales at baseline (control) without any label. Next, we calculated the nutritional grade for the 45 products available (e.g., sandwiches, salads) and labeled them with 5CNL. We then measured sales in the first (5CNL first contact) and second week of exposition (continued 5CNL). Signage describing the 5CNL was also posted in visible areas of the cafeteria. Products’ calorie content (Kcal) was used as our key dependent variable. The data collection generated 2357 purchase observations (781 = control, 899 = 5CNL first contact, and 677 = 5CNL continued contact). A one-way ANOVA with 5CNL condition (control vs. first contact vs. continued contact) entered as a between-subjects factor predicting calories per product purchased revealed a main effect of condition (F(2,2351) = 11.633, p < .001). Specifically, we found no difference between control and 5CNL first contact (p = .231), but the calories per product decreased in the 5CNL continued contact in comparison with the control (M_control = 457.34 vs. M_continued = 415.85, p = .001) and with the 5CNL first contact (M_first=470.74 vs. M_continued = 415.85, p < .001). We also assessed whether the 5CNL impacted the percentage of healthy products (A/B) and unhealthy products (D/E) sold. While we did not find differences in the distribution of sales among all grades (p > .05), we found an interesting phenomenon on the sales within unhealthy products. In the 5CNL first contact, product D sales increased (7.9% to 21.7%, p < .01) and product E decreased (16.3% to 7.8%, p < .05). Meanwhile, in the 5CNL continued contact, product D sales decreased (9.2%, p < .01) and product E slightly increased (11.7%). We propose that this initial increase in product D sales might be associated to the fact that consumers expected these items (e.g., muffins) to be worst in nutritional quality than their actual grade (D), what increased purchases. Overall, this research shows that the 5CNL decreased the average of calories per product purchased. Importantly, the effect of the 5CNL was not immediate but effective after a continued exposition, what reinforces that the impact of FOP nutritional labels is conditioned to consumers’ awareness and capacity to comprehend them. Future studies will explore how consumers’ initial expectations impact 5CNL effectiveness.

References Available Upon Request