Is there research which shows that infants intuitively stabilize towards a nutritionally balanced diet, if they are given a wide range of foods to choose from by themselves?

I ask because I recall reading this from some popsci article years ago, but cannot find neither the source nor the paper itself.

The research paper I had in mind went like this:

Method: Get several newborn (or at least very young) babies, put them away in a room away from any other people, this room contains nothing but a wide wide variety of foods, wait until they decide to eat for their selves and see what happens.
Results: Though the infants initially ate very randomly, they eventually stabilized towards a nutritionally balanced diet (albeit a bit unusual, apparently a subject drank fish oil raw)

I didn't read the paper first hand, as it was quoted by Cracked. While I can find the same article, I don't think they would pull fake research out of their arse (they have(had) a reputation of being a pretty legit popsci source I recall).

I'm looking for either the exact paper I described, or a paper showing a very similar result.

  • Found it! The study is really old, published 1939, somewhat lacking proper scientific methods and data, and your Cracked jumped to the wrong conclusion: They had nothing really to kill themselves (it was a fairly foolproof diet), and the results of the study is somewhat limited. By no means it recommends to let infants decide themselves what to eat. The author clarifies that herself.
    – Narusan
    Oct 11, 2017 at 19:01
  • 1
    thank you so much! a follow up question, but how did you find it? it would be much better if I could find a bit of research by myself instead of asking the folks on S.E. Oct 12, 2017 at 3:52
  • Mostly luck, I think :P. I used google as search engine (opposed to DuckDuckGo which I normally use, but Google finds more), and my successful keywords were infant dieting habits study natural. I had to look at the first 20 results or so. I didn’t find the study itself, but a meta analysis, which had the pdf linked at the bottom.
    – Narusan
    Oct 12, 2017 at 4:40
  • And I specially used GoogleUk.
    – Narusan
    Oct 12, 2017 at 4:54

1 Answer 1


The Study in Question: Davis: Self-Selection of Diets (1939)

Doing for themselves specifically meant permitting newly weaned infants to choose how much or how little to eat of 33 available foodstuffs. As she emphasized to her Quebec audience, no adult was allowed even to hint to the children what might be a proper choice or portion amount. “The nurses' orders were to sit quietly by, spoon in hand, and make no motion,” she said.

Accordingly, Davis devised the experiment to let children do for themselves because she suspected that children's bodies instinctively “knew best” what the individual child should eat. Her intellectual model, a view that would later be called “the wisdom of the body,” likened a child's instinctive appetite to the way various autonomic body systems effortlessly adjust themselves to compensate for external challenges — think of sweating on a hot day, and breathing faster when you start to run.

Initially, it seemed that this conceit didn't apply to Davis's test children and their food preferences. None of the eat-what-and-how-much-of-what-you-want infants had the same diet on any given day, week or month. “Every diet differed from every other diet, 15 different patterns of taste being presented, and not one diet was the predominantly cereal-and-milk diet, with smaller supplements of fruit, eggs and meat, that is commonly thought proper for this age,” she told her Montréal audience.

Yet, she and others later saw that the infants' fanatical heterodoxy turned into what appeared to be 15 uniformly well-nourished, healthy children.

How could eating drastically different diets achieve uniform health and nutritional balance? Body wisdom was the only likely explanation Davis concluded. “Such successful juggling and balancing of the more than 30 nutritional essentials that exist in mixed and different proportions in the foods from which they must be derived suggests at once the existence of some innate, automatic mechanism for its accomplishment.

However, this does not mean infants should not be taught what to eat.

The foods she offered the children were varied, but all were generally thought to be healthy. Their intrinsic goodness meant that it would have been difficult for her small charges to veer too far from the nutritional straight-and-narrow.

“Errors the children's appetites must have made — they are inherent in any trial-and-error method — but the errors with such a food list were too trivial and too easily compensated for to be of importance or even to be detected.” The key thing was to provide healthy food and let children eat as much or as little of it as they wanted.

“The results of the experiment, then: Leave the selection of the foods to be made available to young children in the hands of their elders, where everyone has always known it belongs,” she told her peers in Montréal.

While an interesting double-hinged interpretation of her results, it was, Davis recognized, more a comforting argument than a true demonstration of the limitations of baby body wisdom. She did not present her little ones with a foolproof diet, just a not-intrinsically-foolish one.

The study is lacking statistics, graphs and all that.

All that Davis did was to tell what she found. There are absolutely no graphs, no charts, no individual breakdowns of any sorts for any of the children. It's a summary paper of an Everest of data with next-to-no data in it.

All quoted text taken from the analysis:
Stephen Strauss: Clara M. Davis and the wisdom of letting children choose their own diets. doi: 10.1503/cmaj.060990 CMAJ November 7, 2006 vol. 175 no. 10 1199

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.