My economics research is mostly about undernutrition in Africa and Asia, but I am also keen to learn about nutrition in the US, and maybe help improve policies closer to home.  Recently I had the opportunity to collaborate with Sue Roberts and others on a question that’s puzzled me for some time:  why do we so often leave restaurants feeling regretful that we ate too much?

Restaurants provide a steadily rising share of food consumption in the US and around the world, so making restaurant food healthier is increasingly important for overall diet quality.  The study from Sue Roberts’ group showed that, whatever one thinks of the ingredients and nutritional composition, all kinds of restaurants usually bring much too much food to the table. The headline was that 92 percent of measured servings exceeded recommended calorie requirements for a single meal.

People generally eat what’s served, and people don’t fully compensate by eating less at later meals.  Large portion sizes therefore play a causal role in over-eating.  Our paper documented how big portion sizes actually are, and made the case for asking restaurants as well as diners to take responsibility for the problem by offering smaller portions.

Our paper appeared recently in JAND, and pushed by a well-written press release it added one more study to the daily blizzard of nutrition news, like this article + video.  Friedman’s Marissa Donovan did a particularly nice piece for the wonderful Friedman Sprout website, here:

In reporting her article, Marissa asked me a few questions about the study — here is my full response to Marissa’s enquiry:


Hi Marissa – sorry for very slow response, I was traveling in Ethiopia and am writing this on the flight back.  If you’re still working on the story, here are some answers:

  1. What was a surprising finding of the study?

What was most surprising to me about this study is that no one had done it before.  I think pretty much anyone who ever eats out has seen how large portion sizes are, even in independent restaurants.  But nutrition researchers took this to be inevitable, so not worth measuring — like everyone else, dieticians just knew that restaurants were dangerous for your waistline.  With menu labeling comes the possibility of actually controlling portion sizes, so it’s finally worthwhile to actually measure and publish the data.  Measuring something is a key first step towards improving it.

The one small result that’s surprising but not really a finding is that see few differences among types of restaurant.  We do find that one virtue of Mediterranean (in this case, Greek) restaurants is smaller serving sizes, but the study was not powered to detect differences among neighborhoods and price points.  If we had funds to collect and test many more samples, I expect we’d find that meal sizes are larger in restaurants that serve low- and middle-income people.  That’s certainly my experience from eating in all kinds of restaurants around the US, but it would take a lot of sample meals to detect a statistically significant difference since the variation among dishes is so large.

  1. What do these findings mean for restaurant goers?

I often see diners advised to commit themselves, before they see or smell the meal, to taking home half of what they will be served.  Making the decision early gives power to your far-sighted self.  The key is to make these decisions before you’re hungry, and especially before your appetite is revved up by an oversized dish.  But it’s very difficult to actually follow this advice, mainly because packing up and taking food home is such an awkward step.  In practice, I think it’s much smarter just to choose menu items that will come in small enough sizes for you to be comfortable eating the whole thing.  Use your far-sighted self to identify restaurants that offer delicious foods in portions suitable for your body size and activity level, then praise them for it on Yelp and Tripadvisor.

  1. How do the findings of this study change advice you would give to consumers (if at all)?

The standard recommendation is to stay away from restaurants and cook at home instead.  This helps you control portion size, and also the mix of ingredients.  But you can exercise some of that control in the restaurant by only ordering dishes whose composition and size are both OK.  I am confident that restaurants will eventually find ways to offer all kinds of food in appropriate portion sizes, and with appropriate ingredient ratios.  Until then, we just need choose restaurants that serve at least one good main dish in a reasonably-sized portion that fits our needs.

  1. What changes should be made on a policy level based on these findings?

I think many small steps will be involved.  Like so many policy problems, there’s no one magic bullet.  Making restaurant meals healthier will involve a lot of local steps, like municipal ordinances and state laws. Massachusetts regulations pioneered how to make restaurants healthier for people with food allergies and we can now do the same for nutrition and portion size. There is also room for many voluntary steps by individuals, including food writers and restaurant reviewers as well as restaurants, groups and associations.

A key first step is to understand that serving excessively large meals causes overeating and diet-related disease.  This may sound obvious but it’s not, since many people believe that overeating when served a big meal is just the diner’s fault.  A next step is transparency, with menu labeling so customers can know ahead of time how big each dish will actually be.  Then there’s right-sizing, through various steps to help restaurants serve more dishes in sizes that fit everyone not just their largest and hungriest customers.  Ultimately, I think segregating menus to have some “healthy” or “diet” foods that are served in small portions will be a thing of the past.  Almost all menu items can be served in appropriate sizes.

One specific idea to accelerate the transition to transparency and right-sizing many dishes is for local ordinances and state laws to give restaurant diners the right to order a partial portion at partial price.  As explained in the paper, we know that restaurants would not like doing that.  If such an ordinance were actually passed, most would reduce the default size of their largest dishes for which many customers ask to exercise this right, to avoid having to actually serve too many partial portions at partial prices.  They could then adjust other aspects of the menu so that those who decide that they want to eat more can order additional side dishes.  The problem of excessively large portion sizes can be solved.  A first step is just to realize that it is a problem, to measure what’s served and think carefully about what customers really want.

I hope that’s helpful!

All the best,



Comments are closed.