Assessing the effect of health labels on online food choices

• Published in: Food quality and preference Vol. 131; p. 105565
• Main Authors: Wang, Yuanyuan Anna, Grisolía, José M., Ma, Ada H.Y., Ortúzar, Juan de Dios
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.10.2025
• Subjects: China; color; diet; food consumption; food quality; Healthy foods; information systems; logit analysis; Mixed logit; nutrition information; Online-to-offline takeaway; prices; Public health; Stated choice experiment, difference-in-differences; Traffic-light labelling system; China; Stated choice experiment, difference-in-differences; Traffic-light labelling system; Healthy foods; Online-to-offline takeaway; Public health; Mixed logit
• ISSN: 0950-3293
• DOI: 10.1016/j.foodqual.2025.105565

Online takeaway platforms provide a convenient access to food and their importance has increased dramatically in the last years. Our research explores if and how food choices via takeaway apps change after individuals are informed about excessive calories, fats, carbohydrates, and salts. Our results have practical implications for public health and food choices. We designed an efficient stated choice experiment based on actual meals offered online in China and applied it in a Randomized Between-Subjects Design to a sample of 964 respondents across 10 large Chinese cities. We split the sample into two groups: exposed and not exposed to a colour-code, traffic light information system (TLS). Our analysis, using a Difference-in-Differences model and an Error Components Mixed Logit model, revealed that respondents exposed to nutrition information chose takeaway menus with less fat, salt and calories. However, the information did not affect the choice of tasty meals heavy in carbohydrates, as these are far too important in the typical Chinese diet. We also found that price, positive reviews, and delivery time were drivers of the respondents' food choices, but significantly less important than food preferences and tastiness. Regarding TLS, we confirmed that red (i.e. danger) had the most significant impact in dissuading customers from choosing unhealthy food (salt and fat). These findings are helpful in the design of public policies geared toward healthier food consumption habits in the population. •We tested a Traffic-Light Labelling System (TLS) in a food ordering app using a Stated Choice Experiment.•TLS reduced the selection of meals high in fat, salt, and calories but had little effect on carbohydrates.•Red labels were the most effective in discouraging unhealthy choices.•While service attributes matter, food choices are primarily driven by tastiness and preferences.•This is the first study on nutritional labels in food apps, using real meals instead of artificial scenarios.