The Short-Term Impact of Warning Labels on Sugary Drinks

Research suggests that graphic warning labels can decrease sugar consumption.

Posted Aug 28, 2018

 slackrhackr CC0/Wikimedia Commons
Source: slackrhackr CC0/Wikimedia Commons

Public health campaigns spend a lot of time trying to influence people’s behavior. The campaign to reduce smoking rates is a great example. From the 1960s to the present, smoking rates have been cut from over 50 percent to under 20 percent in the United States.

More recently, there has been a focus on sugar consumption. Many people select drinks with a lot of added sugar such as sodas, fruit drinks, and energy drinks. High sugar consumption has been linked to a number of health problems including diabetes and obesity.

One tool that public-health advocates have at their disposal is warning labels that can be placed on products that might influence use of that product. For example, most people are familiar with the warning labels that are printed on cigarette packages. Although these warning labels are well-intentioned, they do not always have the desired impact. That is why it is always important to collect data on the impact of an intervention before going ahead with it.

This strategy was behind a paper in the August 2018 issue of Psychological Science by Grant Donnelly, Laura Zatz, Dan Svirsky, and Leslie John that explored the impact of different kinds of warning labels on sugary drinks.

They were focused on three types of warning labels. A calorie label had a message about the number of calories in the drink relative to the typical daily calorie intake. A text warning label had a message that drinking sugary beverages is associated with health problems such as obesity, diabetes, and tooth decay. A graphic warning label had the same message as the text warning label, but also had images of an obese stomach, an individual injecting insulin, and bad teeth.

The researchers did a field study to explore whether these labels affected real behavior in a hospital cafeteria. They gathered data about the proportion of sugary and non-sugary drinks sold at that cafeteria for the previous two years. Then, they got a baseline measure of purchases for two weeks before the study. Next, they labeled the sugary drinks in the cafeteria with one of the labels for a two-week period during which they measured purchases followed by a two-week period with no labels to see whether there was any long-term impact of each type of label.

Because they had access to only one cafeteria, the calorie label was used first then (after a two-week delay), the text warning label was used and then (after another two-week delay) the graphic warning label was used. To ensure that people were not just finding other places to buy drinks at that hospital, the vending machines were also given the labels used in the corresponding experimental condition, but purchases at those vending machines was not tracked.

The calorie and text warning labels had no significant influence on the number of sugary drinks purchased. In the baseline, and also in these two conditions, roughly 21.5 percent of the drinks sold at the cafeteria were drinks with added sugar. In the two weeks with the graphic warning label, though, that decreased to about 18 percent. While this is not an enormous percentage decrease, it does reflect about 150 fewer purchases in the two-week period. Of course, as soon as the warning labels were removed, the purchases returned to their normal level.

In another study, people imagined either that they were purchasing a drink at lunch or that they were purchasing a drink and that they were confronted with the graphic warning label from the previous study. Those who imagined seeing the label felt more negatively about their drink and thought more about their health, and that led to a greater intention to purchase water in the future.

These studies suggest that showing people the potential hazards from sugary drinks can have a small impact on sales—at least in the short-term. There was a reliable decrease in purchases over the two weeks that the study was conducted. On the basis of this study, I would not run out and plaster warning labels on sugary drinks, though.

First, these labels did not have a long-term impact on sales. As soon as the labels were removed, sales returned to their baseline levels. Presumably, then, at least some people changed their behavior only while the labels were present.

Second, over time people find ways to ignore information in the environment they do not want to see. If warning labels on sugary drinks become the norm, it is likely that their overall impact will likely decline (and may not have that big an effect to begin with). Instead, a targeted use of graphic warning labels might be an effective part of a more comprehensive plan to influence behavior.


Donnelly, G.E., Zatz, L,.Y., Svirsky, D., & John, L.K. (2018). The effect of graphic warnings on sugary-drink purchase. Psychological Science, 29(8), 1321-1333.