Does Posting a Calorie Count Change How People Eat?

Some time ago, we wondered if New York City’s new law requiring certain restaurants to post calorie counts might provide good material for academic researchers who care about obesity.

The answer: yes!

Brian Elbel, an assistant professor of medicine and health policy at the NYU School of Medicine and the NYU Wagner School of Public Service, was the lead researcher on a new study called “Calorie Labeling and Food Choices: A First Look at the Effects on Low-Income People in New York City” (abstract here; pdf here).

Elbel sent researchers into the field to measure the caloric habits of low-income people who eat at fast-food restaurants including McDonald’s and KFC. They gathered data before the calorie-posting law went into effect, in order to measure the change, and in addition to gathering data from New York fast-food restaurants, they also gathered data from equivalent restaurants in Newark, N.J., which doesn’t have a caloric-posting law, to serve as a control.

What did they learn?

“We found that 27.7 percent [of people] who saw calorie labeling in New York said the information influenced their choices.”

Great news, right? Um, no:

“However, we did not detect a change in calories purchased after the introduction of calorie labeling.”

Yikes. This is a great example of how well-intentioned incentives, even if they are successfully presented, fail to move the needle. There are obviously a million more questions to be asked in this realm. For instance:

  • As nudges go, is this one simply too subtle? How might a calorie nudge compare to a price nudge (some kind of a fat tax — on fatty foods, that is, not people)?
  • How might the response to caloric postings affect the choices of middle- and upper-income people?
  • If you want to curb obesity, especially among lower-income people, what do you do if it turns out that lower-income people are less responsive to the calorie nudge than other people?
  • Since this particular study relied on the subjects’ participation — the researchers intercepted them, inspected their receipts, and talked to them about their buying choices — what should we think about how representative these participants are? We write a good bit in SuperFreakonomics about how participants in most experiments are a) not likely representative of the general population — i.e., they are more cooperative — and b) likely to be influenced by the observers’ scrutiny and approval.

Personally, I think the calorie-posting law isn’t a bad idea at all. But I suspect that the people who will be most responsive to it, especially in the long run, are those who are already the most vigilant about their health and well-being. Think of it this way: what if the safest drivers on the road were the only ones to wear seat belts?

Leave A Comment

Comments are moderated and generally will be posted if they are on-topic and not abusive.



View All Comments »
  1. Matt says:

    The results were actually even worse than what you mention – people actually ate *more* calories after the calorie counts were posted:

    Thumb up 0 Thumb down 0
  2. Brad says:

    Not really a big shocker here, since the whole point is that if people have the option and they have caloric information they will make the right choice. But if you don’t have the money to make economic choices, then the caloric information is irrelevant.

    Plus really, are people really that worried about calories when you roll in the Micky D’s or KFC? What would be more interesting is go to a non-fast food restaurant like Chili’s and Applebees and see how the selection of salads changed. Since in general people get salads to “be healthy” but if you don’t know the calories it is hard to gauge how you are doing in this, and here is where I would expect there to be a real impact. Also the add ons purchases at Starbucks, I would expect an impact as well.

    Thumb up 1 Thumb down 0
  3. Steve Kalman says:

    Your subhead asks if the new law would provide good material for researchers. The answer to that was YES with a 100% certainty factor. No matter what behavioral changes, if any, occur, the before and after observations are grist for many academic papers.

    For one anecdotal bit of evidence, I saw the calories in the optional flavorings Starbucks and similar coffee bars put into a medium latte. It made me switch to cappuccino. (I had guessed it would be high, but not THAT high.)

    Thumb up 0 Thumb down 0
  4. Ratna says:

    I suspect that those calorie counts will make the healthy healthier and, at best, not impact the unhealthy.

    I was in NYC for several weeks this summer and that calorie information really did help me walk away from many unnecessary treats (who knew, really, really, knew, that a carrot muffin at Starbucks is over 300 calories?! I suspected it, but didn’t really know till I saw it up there…urk). Can you guess whether I consider myself a healthy person? :-))

    Thumb up 0 Thumb down 0
  5. science minded person says:

    I have been counting calories for more than 30 years. For me, it works. When I started, I figured that I could eat anything I wanted up to 1700 calories. For my height, Fifteen hundred or less, I was losing wait. 6 Entemens doughnuts later, I knew that if I ate more than salad, I would gain wait and I started to dislike eating sweets too much. Just enough to keep me satisfied at times. Now, I try to eat healthwise most of the time–except those good times and difficult times when chocolate or chocolate cake or hazelnut icecream and lemon sorbet just feels good.

    What helped me is knowing the limits of what I can and cannot eat to maintain my weight. And now, I’m focused upon limits (diet and exercise wise) as far as maintaining my health.

    But my guess is, every individual movesmore or less to the tune of their own drum. IF it works for you, that’s what counts.

    Thumb up 0 Thumb down 0
  6. Joshua says:

    When I was in college, I’d always go and read the calorie information carefully on the muffins/snacks/drinks in the on-campus convenience store so I could optimize my per-dollar calorie count and get a meal (or a meal and a half!) for less than two bucks (milk and a muffin, usually…about 750 calories! 4 g protein)

    Thumb up 0 Thumb down 0
  7. David says:

    My guess is that people have a general estimate in their head of which items on a menu are the worst, even if they don’t have exact calorie counts, and to some extent that knowledge already guides their decision.

    So when they say that the calorie count influenced their choice, they’re really just replacing their estimate of an item’s fat factor, which was already guiding their decision, with the exact calorie figure.

    Over a large sample size, some estimates will be high, some will be low, and overall they will even out to result in the same purchasing behaviors. What surprises me the most about the lack of change in calories purchased is that it doesn’t reflect any optimism bias in people’s estimates.

    Thumb up 0 Thumb down 0
  8. Erik says:

    I agree with both salient points in the comments above.

    I’d love to see this study done at a Starbuck’s and a Friday’s or Applebee’s. People know what they are in for at a true fast food restaurant; the calorie counts are probably not unexpected. It’s places outside of McDonald’s and KFC that would be more interesting to study.

    Also, I think it would be more interesting to study this across the income spectrum. Diets are so tied to income that isolating a certain population introduces a separate set of inputs.

    Everyone I know has a story about how they skipped a muffin or switched away from frappuchinos or ordered better when they boss took them to Friday’s. Anecdotally to me and the people I interact with, the calorie counts have made a difference.

    But when I crave a Big Mac, I’m still going to get that Big Mac.

    Thumb up 0 Thumb down 0