What Survived: The Food-Environment Findings That Held Up
KEY TAKEAWAYS
- A retraction says a paper can’t be trusted, not that its idea is false. The test used here is whether people outside the Cornell lab found the same thing.
- The sturdiest finding is about how much is offered. A 2015 Cochrane review found that larger portions, packages and tableware went with higher consumption, with effects the authors called modest.
- The most repeated tip, smaller plates, has the thinnest support. A 2019 editorial concluded there is no substantive evidence to recommend it, and a 2024 trial found no meaningful difference.
- Nothing here is a plan. These studies describe what surroundings do to people on average, mostly in short lab experiments. They don’t tell anyone what to do.
Somewhere in your kitchen there is a plate that came with advice. Smaller, the advice said. Keep the fruit where you can see it. Most of it circulated as settled science, and much of it traces back to one lab, whose papers now read as a list of retractions and corrections.
The lab was the Food and Brand Lab at Cornell, run by Brian Wansink. By the time BuzzFeed News reported on six more retractions in September 2018, Retraction Watch’s database counted 13 retracted papers, one of them retracted twice, and 15 corrected. The six papers had appeared in JAMA journals between 2005 and 2014. When the journals asked Cornell to verify them, the university wrote that, because it did not have access to the original data, it could not assure the journals that the results of the studies were valid. Wansink left Cornell that same month, after the university found academic misconduct.
That leaves a fair question, and it is the one this piece answers. Which of the kitchen ideas were reproduced by other people, and how big were the effects when they were?
A retraction is not a verdict on the idea
A retraction says a paper cannot be trusted. It does not say the claim is false. That gets settled the slow way, by other groups running their own experiments with their own data.
A 2018 article on Phys.org made this point about Wansink’s work and said that other researchers had reproduced some of his lab’s ideas, larger bowls leading people to eat more among them. I read only a summary of that article, it is a secondary source, and it is a pointer, not a finding. So the working test used here is simple: an idea counts as having held up only if people with no connection to that lab found it too.
This is also why the retraction notices matter less for what they removed than for what they cannot say. Cornell’s statement was about missing data, which means the six papers cannot be checked, not that their claims were checked and failed. The gap between “cannot be verified” and “was disproven” is where every honest sentence in this area has to live, and it is why this piece keeps saying “reproduced” and “modest” rather than “proven” and “debunked.”
Which food-environment findings held up, and which didn’t
Bigger servings, bigger appetites
The sturdiest of the ideas is also the least glamorous. In 2015, Gareth Hollands and colleagues published a Cochrane review of experiments on portion, package and tableware size. It found that when larger portions, packages or tableware were on offer, people consumed more, and the authors described the effects as modest, more pronounced in adults than in children.
One expert quoted in response, Prof. Ratcliffe, suggested a reason: people are reluctant to waste food, and larger portions can look more attractive. That is a hypothesis, not a result.
Mind what this kind of evidence measures. Most of it is what people take and eat in one sitting, in an experiment. It cannot tell you what happens across a month of dinners in a house with a partner, a dog and a deadline.
The small plate: the tip that outran its evidence
Here is the part that surprises people. The single most repeated piece of advice, smaller plates, is where the evidence is thinnest.
A 2016 meta-analysis by Holden and colleagues, in the Journal of Consumer Research, found the answer depended on who was serving and who was looking, which is a polite way of saying it was not a clean effect. In 2019, Dana Lee Olstad and Clare Collins wrote an editorial in the International Journal of Behavioral Nutrition and Physical Activity about a pre-registered trial that found no clear evidence of a difference between smaller and larger plates. Their conclusion was blunt: at present there is no substantive evidence on which to recommend smaller dishware. They added that the wide uncertainty in the data does not rule out small or medium effects.
A 2024 pre-registered crossover trial in Appetite, by Langfield and colleagues, tested 50 adult women aged 18 to 77 who ate breakfast twice, once with smaller and once with larger dishware. It found no meaningful difference between conditions. The trial was done in a lab, and its authors noted it lacked the power to detect small effects.
Notice the pattern. The finding that held up is about how much is offered. The one that faded is about the plate the same amount sits on.
Looks matter, at least in photographs
There is one place where plate design does show something. McClain and colleagues, in a 2013 study in the International Journal of Obesity, showed people photographs of paired plates and asked which held more food. Wider rims made the food look larger than it was, and colored rims had a smaller effect. It is a real perceptual effect, but the participants were looking at pictures. Nobody was eating.
Distance and options
In 2019, Hollands and colleagues reviewed 24 studies on the availability and proximity of food, six on availability and 18 on proximity. Broadly, when fewer options were on offer or food sat farther away, people selected and consumed less. The authors rated the certainty as low. Most studies were done in laboratories over short periods, there were relatively few of them, and many were poor quality. Real-world settings such as supermarkets, cafeterias and restaurants were barely represented.
This describes what surroundings do to behavior. It is not a suggestion to take options off the table. Nothing in these reviews is a recommendation for how anyone should eat.
What “held up” is really saying
Line the findings up and a pattern appears. What survived is modest, comes mostly from short experiments, and describes the average of many people, not the person standing in your kitchen on a Tuesday night. A small effect on one meal in a lab does not add up to a plan, and none of these papers claims otherwise. The strongest of them says the effects are modest, and the review of proximity says the certainty is low.
There is a second reason to keep the claims small. Advice built on food-environment findings can slide into rules about what belongs on a plate, and rules like that cause trouble for a lot of people. This site does not go there. The question here is how a room, a serving habit and a week shape what people do, not what anyone should do about it.
That is the honest account, and it is more useful than either extreme. The confident version, in which the right plate fixes everything, went with the retractions. The dismissive version, in which none of it means anything, is not what the evidence says either. What is left is smaller: rooms and serving habits nudge people, a little, and the size of the nudge is still being measured.
The claim to be suspicious of next
The next time someone sells the rearranged kitchen as the answer, ask three questions. Who tested it besides the person selling it? Did anyone eat, or did they look at photographs? And how long did the experiment last?
The studies stop at the edge of the kitchen, which is a strange place for them to stop, because that is not where the week begins. It begins on the way there.
References
- Hollands GJ, et al. Portion, package or tableware size for changing selection and consumption of food, alcohol and tobacco. Cochrane Database of Systematic Reviews, 2015. cochranelibrary.com/cdsr/doi/10.1002/14651858.CD011045.pub2/full
- Hollands GJ, et al. Altering the availability or proximity of food, alcohol, and tobacco products to change their selection and consumption. Cochrane Database of Systematic Reviews, 2019. pubmed.ncbi.nlm.nih.gov/31452193
- Holden SS, et al. Whether smaller plates reduce consumption depends on who’s serving and who’s looking: a meta-analysis. Journal of Consumer Research, 2016. journals.uchicago.edu/doi/abs/10.1086/684441
- Olstad DL, Collins C. Smaller dishware to reduce energy intake: fact or fiction? International Journal of Behavioral Nutrition and Physical Activity, 2019. link.springer.com/article/10.1186/s12966-019-0831-4
- Langfield T, Clarke K, Sadab N, Jones A, Robinson E. Dishware size and energy intake in adult females: a pre-registered randomized crossover trial. Appetite, vol. 197, 2024. sciencedirect.com/science/article/pii/S0195666324000977
- McClain AD, van den Bos W, Matheson D, Desai M, McClure SM, Robinson TN. Visual illusions and plate design: the effects of plate rim widths and rim coloring on perceived food portion size. International Journal of Obesity, 2013. nature.com/articles/ijo2013169
- BuzzFeed News. Cornell University food scientist Brian Wansink just had six more studies retracted. buzzfeednews.com/article/stephaniemlee/brian-wansink-jama-six-retractions-cornell
- Science Media Centre. Expert reaction to Cochrane review looking at portion size and consumption of food and drink, 2015. sciencemediacentre.org/expert-reaction-to-cochrane-review-looking-at-portion-size-and-consumption-of-food-and-drink/
- Phys.org. Retraction of a journal article doesn’t make its findings false, October 2018. phys.org/news/2018-10-retraction-journal-article-doesnt-false.html
Diana Woods researches and writes about sleep, environment and everyday habit. She has no clinical training, and that shapes what this site does: instead of explaining what might be wrong with you, she reads the behavioural and environmental research and writes about the parts you can act on — the light in the hallway, the fifteen feet to the bathroom, what the morning does to the night that follows.
Every article here cites its sources, and she tries to include the parts researchers themselves flag as uncertain. Studies disagree, sample sizes are small, and correlation gets reported as cause more often than it should. Saying so is part of the job.
