Self-refilling bowls: An idea whose time should never come

[Originally posted in April 2007]

ResearchBlogging.orgOne "trick" dieters often use is to put their food on a smaller plate. The idea is to fool yourself into thinking you're eating more food than you really are. But doesn't our stomach tell us how full we are?

Actually, it doesn't. Brian Wansink has devoted his career to studying how perception of food intake relates to actual eating behavior. Together with James Painter and Jill North, he's come up with a dramatic demonstration of how wrong our stomachs can be.

Volunteers were recruited to participate in a soup-only lunch in a room adjoining the school cafeteria. They filled out a form asking about color preferences, then were seated a table with four different-colored bowls. The colors were just a distraction: the real purpose of the study was to see how much people would eat when their soup bowls refilled automatically.

Two of the participants ate from self-refilling bowls; the other two had their bowls refilled by a server. Everyone was encouraged to eat as much as they wanted. The self-refilling bowls involved a fair bit of cooking technology -- plastic tubes connected a soup pot next to the table to the underside of each bowl. The refill rate of the bowls was adjusted so that the bowls could be filled completely in 20 minutes -- the duration of the study. Technically the bowls could be nearly empty by the end of the session, but each bowl held 18 ounces of soup, so this would have required consuming over a quart of soup!

Despite the fact that everyone's bowls were refilled, the people eating from self-refilling bowls ate 73 percent more soup. Even more surprising is that they didn't feel any different from people who ate from manually-refilled bowls:


None of these other measures were significantly different -- even though the people eating self-refilled soup indicated that it seemed they couldn't possibly eat all their soup, they didn't estimate they'd eaten significantly more than those who had the visual cue of a server refilling their bowl every time it was less than 25 percent full.

When asked to rate hunger on a 1-9 scale, again, there was no significant difference between the two groups. In all, a dozen ratings were collected, asking questions about whether they monitored their food intake during the study, whether they generally try to clean their plate, and how the presence of others affects their eating. In every case, there was no difference between the two groups -- the only difference was how much they ate.

The team also controlled for gender, body mass index (BMI), and other factors, and still found the same results (though since they didn't study an extremely wide range of BMIs, the results might be different for dramatically over/underweight individuals).

Wansink et al. argue that this demonstrates that the primary way people decide how much to eat is visual: when there is a visual indicator of how much food is consumed, then people are accurate at determining how much to eat. The problem comes when social norms of "reasonable" portions change: as portion sizes in restaurants and stores increase, people expect to eat more at each meal--leading to unhealthy eating. The team argues that restaurants and retailers should present food in smaller portions to reinforce the idea of eating less. Parents could repackage snacks for their kids in individual bags to reinforce the idea that just a small portion is reasonable.

And in no case should you install a self-refilling soup bowl in your kitchen!

[For more on the idea that we rely on our perception and memory of what we eat to decide when we're full, check out How do we know when we're hungry?]

Wansink, B., Painter, J.E., & North, J. (2005). Bottomless Bowls: Why Visual Cues of Portion Size May Influence Intake Obesity Research, 13 (1), 93-100

More like this

One "trick" dieters often use is to put their food on a smaller plate. The idea is to fool yourself into thinking you're eating more food than you really are. But doesn't our stomach tell us how full we are? Actually, it doesn't. Brian Wansink has devoted his career to studying how perception of…
It's that time of year again when the most exciting set of awards for us here at Omni Brain are announced. If you don't know what the Ig Nobels are you're missing out on a very important part of science! According to some article at MSNBC "The annual no-rules awards ceremony, where flying paper…
More tales of hilarious heuristics that lead us astray and make us fat: An appalling example of our mindless approach to eating involved an experiment with tubs of five-day-old popcorn. Moviegoers in a Chicago suburb were given free stale popcorn, some in medium-size buckets, some in large buckets…
Getting more than one helping of food is obviously a draw for patrons of all-you-can-eat restaurants.   But can one predict how many trips to the buffet a particular diner will make?  Does scoping out the buffet before grabbing a plate lead to more trips?  How does sitting in a chair that faces the…

What happens when you put the pump in reverse and have a bowl of soup to drains over time instead of fills?

By Jason Woertink (not verified) on 17 Jun 2009 #permalink

I just had to tell you - when I saw the title of this post on my blogspot sidepanel, I would have sworn it said "Self-Refilling Bowels."

And I was like "Wait.....aren't they all?"


Isn't this why women in France tend to be slimmer than women here in America? They eat wonderful rich high-fat food, but just less of it.

Just a nitpick - how many servers were there and how quickly did they react when a bowl needed to be filled? If there was time spent waiting around while the server was serving another participant that could affect the results.

I think there is some risk that people perceive liquids differently but I did see a similar study where they were providing hot wings to people watching a football game. They would continually clear away the dishes with bones from some people and not others. The people who had visual feedback from the bones nearby ate much less than the people who's plates were kept clear.

So if I understand this correctly

The refill rate of the bowls was adjusted so that the bowls could be filled completely in 20 minutes

a subject with a self-refilling bowl might not be aware that the bowl was slowly refilling itself over the course of the trial? At that rate, I don't know if I'd notice that the bowl was refilling - I'd notice that it wasn't getting emptier and I'd just think that I was eating slowly.

There is a typo in the study--the bowls completely refill in 20 seconds, not minutes. At that rate, the soup is constantly burbling and jumping up, not unlike a "paint pot" at Yosemite. The increased fear and fascination of the diners, then, is presumably responsible for their enhanced appetites.

Larry Niven had a similar device in a short story (one of his ones from the Ringworld universe). It involved teleportation, and beer instead of soup, but he also warned of overconsumption.

By Dan Miller (not verified) on 17 Jun 2009 #permalink

I once got a friend drunk by swapping glasses with her every time hers contained less than mine did. Then when she noticed mine had less in, she'd drink to catch up... and then i'd wait til she wasn't looking, and swap again. She didn't stop drinking because she 'wasn't thirsty' or because she was getting drunk, she was entirely basing her drinking on the visual cue from the amount left in the glasses..

I wonder how much the study was affected by the soup, though. If it was piped in by tube it meant the soup was a fairly homogeneous liquid, like a thin potage. That removes a lot of other potential cues for satiety, such as the amount you chew or the mechanical resistance to the stomach working.

I mean, beer, milk and drinking yoghurt is also nutritious and yet people have no problem "overeating" those since it drains through you much faster than solid foods.

teiana, I do that (drinking, not swapping). Mainly because if I don't base my drinking of alcohol on visual clues, I'll drink like it's water, and end up sitting there with an empty glass feeling silly.

I saw the results of this study in the newspaper sometime, but read it again to reinforce the point with myself.

By Katherine (not verified) on 17 Jun 2009 #permalink

@Teresa: If I remember this study correctly only one participant tweaked to their bowl being *magic* - and this was only because they felt the pipe against their leg and looked under the table.

I think your recommendation for parents to individually pack food is a good one. A similar argument could be made for purchasing smaller lunchboxes. Also has ramifications for buying food in bulk (I only ate half a packet of chips...)
People clearly think in proportions, not in actual amounts.