« Wrong Tomorrow | Main | Incentives, Allies Cut Bias »

April 02, 2009


I believe the MR article was written by Alex.

So we know the women Hannah hung out with viewed breast feeding as a sign of social status and/or conformity to their group's norms. That explains their behavior, but why the pro-breastfeeding bias (assuming there is one; having neither breasts nor kids its not something I've ever read about before) in the media at large?

It seems that few people advertising this article (and assuming the author's conclusions are correct) seem to actually look up the data concerning the subject. Even if you accept the author's conclusions, it is still inappropriate that she neglected to mention many of the studies that conflicted with her argument. She tosses the common trope that "studies conflict" (on what topic are there no conflicting "studies?" Do we ever give all studies equal weight?), but doesn't mention the findings of the largest studies with the best design and highest statistical power.

For example, the PROBIT study, which was conducted by Canadian and American researchers in Belarus (where most women use formula), did actually conduct an randomized controlled trial, and found that the two groups had a statistically significant 6 point difference in IQ (which, in my opinion, is a large difference). It should be noted that the groups were not divided into breastfed vs. formula but into a group that learned about breastfeeding and childcare and another group that spent an equivalent amount of time just on child care. It is certainly possible that the manner of instruction or some other variable accounted for the difference in IQ. However, it is at least suggestive of the notion that, all other things being equal, breastfeeding is preferable for the child, and that we should look more into these findings, before we throw up our hands and decide that the journalist suddenly showed us what's what.

In general, I have a request for the writers at "Overcoming Bias," when correcting a bias about a specific topic, to actually read some of the related literature, before posting about it. Making a seemingly enthusiastic endorsement, when referencing a magazine article, could possibly only lead us to be "even more wrong." In other words, increasing neutrality about breastfeeding seems more fair, but it is more incorrect if the best evidence we have seems to indicate that breastfeeing is actually beneficial. Introducing an uncessesary degree of uncertainty ("vaccines may cause autism") is a form of error, so we should not feel confident, even if the only thing we did was cast more doubt. Call it the "Increasing Uncertainty Un-necessarily" bias (one which you've likely covered on this blog).

Of course, even if there is an IQ difference, this doesn't mean that women should be demonized for weighing their priorities, and then deciding to go to work and use formula instead. But that is a whole other issue.

there is no such thing as scientific consensus. appeals to the majority are not science, regardless if the majority are researchers. It does not matter what the researchers say. all that matters is: does the data support the hypothesis?

Robin, I love your posts and have learned a lot from them and find them very insightful - right up until you latch onto some crazy crusade about how some well-supported health care idea doesn't really work, cherry-pick a few studies that you then misinterpret, and on the basis of a few hours of research accuse everyone who disagrees with you (including every doctor, scientist, and expert ever) of "signalling". Please stop it. Please. It would let me appreciate your other posts so much more.

Overweight is indeed one of the few conditions in which the literature has found little to no positive effect for breastfeeding. Blood pressure and atopy are two others. But going from there to "breastfeeding isn't that great" is like citing a study saying exercise doesn't cure AIDS as proof that you don't need to go to the gym.

Every meta-analysis on every topic ever has complained of bias, missing evidence, and inconsistent studies. This is a commentary on today's scientific institutions and journals, not on the particular topic. I think you may have an unrealistic view of exactly how perfect evidence has to be before it's generally accepted. God only knows how you got such a view by being an economist, but somehow you seem to have managed.

My very quick (yes, I admit, I didn't spend so long either) review of the literature tells me breastfeeding increases IQ several points, slashes the rates of dozens of childhood diseases (one example: SIDS), and decreases the rate of some diseases of later life (including celiac disease, diabetes, coronary artery disease, and breast cancer (in the mother). These studies I'm citing include meta-analyses and work by the top researchers in the field.

On a societal level, the studies say that minor increases in breastfeeding rates could prevent 1.3 million childhood deaths each year (worldwide) and save $3.6 billion in health care costs (US alone).

Would also like to point out that every doctor and epidemiologist with whom I've spoken interprets the literature supports breastfeeding. I know you're not usually one to care, but your thesis here is the simple point that the medical literature doesn't support breastfeeding, not the more complicated point that the medical literature supporting breastfeeding is wrong, so you do need to pay attention this time.

Maybe the problem is misconception of what a "small" effect means in medicine. Yes, most of the effects of breastfeeding are on the order of a five percent risk reduction of coronary disease or something like that, and five percent doesn't sound too impressive, but considering the numbers involved even a slight increase in breastfeeding saves thousands of lives each year (or, in the case of that developing world study, millions).

See my added to the post.

Grant, thanks, fixed.

Robin, your ability to publicly change your point of view is quite admirable.


I took the main point of the article to be the one you conceded - that she shouldn't be considered a bad mother for choosing not to breastfeed. Even if IQ means anything at all, and if 6 points is really significant, it's still an open question whether it's best for the mother to endure the hardship she describes.

Thom, I agree with you about that. And I wish that some of my colleagues in the medical profession would realize that "increasing marginal health" is not always someone's first priority. Doctors ought to present information as accurately as possible, without bullying patients to take on certain behaviors, which may run counter to other values and plans.

> why the pro-breastfeeding bias


It's a display of status by conspicuous consumption (of time and effort).

A poor working woman taking two buses each way to get to her cleaning job has to pass her new infant off to a mother, sister or friend.

An Ivy grad woman with a doctor/lawyer/whatever husband can stay home / take her infant to work at her web 2.0 job / take frequent breaks to breast pump during the work day.

This is my theory for why "bo-bos" prefer organic food, despite the lack of evidence showing that it's better: they conspicuous consumption of the food cements their place in the primate pecking order WRT other mothers - their husband is a good provider.

Yvaine writes:

> breastfeeding increases IQ several points

From the article you link to:

Long-term, exclusive breastfeeding appears to improve children's cognitive development ...

some of the research has struggled to identify whether the findings were related to the fact that mothers from more affluent backgrounds were more likely to breastfeed

Intelligence leads to affluence.

Affluence leads to breastfeeding.

Intelligence is heritable.

and now ... wait for it ... we learn that children of intelligent mothers are intelligent.

I don't see this as a killer "gotcha" argument in favor of breast feeding.

There is a general problem when laymen look at technical literature, which is that it almost never as certain as popularizations, but they assume that it's only the particular corner they notice that has this discrepancy. Rosin does provide a little evidence that her corner may be special by looking at long enough history to see fashions.

No, Rosin does not mention the studies you mention, but she doesn't reach very different conclusions. She (half-heartedly) concedes 5 IQ points, compared to the 7-10 that people harangue her with. Mainly she complains about all other effects that are much weaker.

TJIC, the 6 IQ points is from a randomized study.

Even if IQ means anything at all, and if 6 points is really significant, it's still an open question whether it's best for the mother to endure the hardship she describes.

For 6 IQ POINTS? If that's not worth breastfeeding your child, then you shouldn't be having children. I'm sorry if that sounds harsh but I'm afraid to even think about how far I would go to avoid losing 6 IQ points.

Eliezer: Can you quantify the difference 6 points makes? Is it like having one drink (for an average person)? Two? Would my spelling get, on average, worse?

I think I might (maybe) have a good handle on what 50 IQ points would cost me, but you sound like you'd kill a truckload full of kittens for 6 additional points of IQ.

I was breast-fed for 20 months, no allergies, no cholesterol, no diabetes, no overweight,.............etc. It was good for me.

The link labeled "MR" just points back here. I'd like to read that comment.

"Robin, your ability to publicly change your point of view is quite admirable."

Seconded. My belief in Robin's weird medical opinions now goes up substantially, since I know he's be willing to change his opinion when the evidence suggests it. Still not that high, I'll admit, but substantially higher.

Here is synopsis of the study:


The key here is that the study was initially focused on determining the effects on GI infections.

Here is a relevant an important limitation:
"Because protection against infections in developed country settings does not have the life-and-death implications for infant and child health that it does in less-developed settings, cognitive benefits may be among the most important advantages for breastfed infants in industrialized societies," the study authors write. "The consistency of our findings based on a randomized trial with those reported in previous observational studies should prove helpful in encouraging further public health efforts to promote, protect, and support breastfeeding."

However, I wonder what that means really. It sounds like they are trying to infer too much. If, in fact, the effects of lower infection rates are a "life and death" matter for less developed countries that suggests o me that the limitation on GI infections alone could have an impact on cognitive development independent of the breast feeding. That is, if the infections are making children weaker and compromising their physical (and associated mental development) then really how can you say that "breast feeding" is the real cause?

I expect rage to follow.

I am intrigued by the emotions that this issue produces. It appears to be a good cypher for uncovering all kinds of underlying prejudices and preconceived notions. As if parents who bottle feed are "bad" parents who fail to give their children every possible advantage in the world. I guess these stay-at-home-breast-feeding mom's can be proud of the fact that they gave their children 6 extra IQ points even though it meant sacrificing their careers! Of course, maybe they should look at studies that determine whether the loss in income resulting from their staying at home has a negative effect on IQ? What if studies showed that for every $100,000 loss or gain in household income a child loses or gains 10 IQ points. My goodness they are actually harming their children by staying home!

All I am suggesting is that there are a number of factors that probably influence IQ. Some more than others. I respect people who try to rationally make conclusions about these various issues. However some times I think these seemingly "close up" issues can ignore the bigger picture. (I think there is a name for that bias. I am sure I learned something about it here but cannot remember) Much like some people who religiously save up their cans and bottles for recycling only to drive their Hummer down to the recycling center.

Eliezer pulling out the italicized 'randomized study' to boost legitimacy. I little piece of me just died. A small piece, but still significant. (I'm being 75% sarcastic)

One mans rational, bias dodging journey down the breast is best road. Read his most recent conclusion.


A great book on the topic of development for those interested.


For those without acces to the "rather damning critique," it argues that the PROBIT study had multiple measures of IQ gains, with the largest being the least credible (but most likely to gain headlines). The other three estimates of IQ gain were not statistically significantly greater than zero. The critique authors argue that this is a reason to doubt the larger effect, that giving verbal IQ tests to preschoolers in a (potentially) second language is not meaningful, that the pediatricians who told parents to breastfeed or not and were the same ones who tested the children, and that other geographic/linguistic factors may have been in play. They argue that the most reliable result from the PROBIT study showed a 1-1.5 point increase, with a confidence interval that included zero.

Tony, randomization is a perfectly legitimate technique even from the formal perspective of causal analysis a la Judea Pearl - it has nothing to do with "inherent randomness", you just have to be really sure that you don't believe in any prior link between the "random" variable and the other variables of interest.

BillB, I would eat the kittens.

Zubon, if "the pediatricians who told parents to breastfeed or not and were the same ones who tested the children", then that invalidates the randomization and I retract my use of the study as evidence.

It occurred to me that using breast pumps might be a good way to avoid some of the problems associated with breastfeeding while continuing to take advantage of breastmilk's nutritional value. Any problems?

Yvain & adina,

I appologize if this is a stupid question (not having read the literature), but do the studies try to correct for genetic, class and other differences between breastfeeders and formula-users? It would seem to me that breastfeeders probably interact more with their babies overall (which I believe helps IQ), and are likely more financially successful (suggesting their genes are worth more IQ points than others').

Overall though, I agree with you, Robin does seem to jump to some strange views regarding biases on complex subjects.

Robin's original complaint is about signaling and status, okay. I get that.

But in the complex world we live in, lack of evidence and lack of time are significant issues. Nobody's linking to any studies suggesting that formula is better than breastfeeding, maybe because we were "designed" for breastfeeding. As a parent, you've got to make lots and lots of decisions for your children. On breastfeeding, and many other issues, there are huge, powerful, advertising-funding interests seeking to encourage you to go their way because they make a profit on it. That's a bias, though not one it seems is much liked here on OB (does it smack too much of the radical left's modernism-Luddism?)

I'm saying this argument doesn't consider the fact that we should be holding formula to a higher standard than we do. We don't demand of new things that they prove that they are, genuinely, as good as the old, and we have a hundred years of carcinogenic history to point out that new things often interact in very complex and undesired ways with our carefully-tuned-for-the-savannah bodies.

I think I might (maybe) have a good handle on what 50 IQ points would cost me, but you sound like you'd kill a truckload full of kittens for 6 additional points of IQ.

I don't know about Eliezer, but I sure would, in a heartbeat. And I would singlehandedly wage genocide against felines for 50 additional points of IQ. I'm surprised others wouldn't.

So.... where can I find those kittens?


genocide: by which I mean, err... xenocide or felicide.

Naively applying Robin's usual standards for medical results more generally, I would be forced to conclude that the entire field of economics is total bunk barely better than noise, along with the majority of every other academic discipline but physics, chemistry, and mathematics. I'd probably be right in many cases to do so, but that's a lot of babies out with the bathwater.

It occurred to me that using breast pumps might be a good way to avoid some of the problems associated with breastfeeding while continuing to take advantage of breastmilk's nutritional value. Any problems?

My mother did this with some of my younger siblings. I remember feeding bottles of breast milk to my baby sister, and as an aside, human milk that's gone sour from sitting out too long smells remarkably bad.

Anecdotally, I recall that the primary downsides to this approach is that the pumps are often harsher on sensitive tissue than an infant's mouth, and that a regular schedule of expressing the milk must be maintained regardless because going too long without doing so will cause the mammary glands to produce less milk in response. Obviously benefits to mother-infant bonding are also lost, and if done too frequently the infant may become too accustomed to bottle-feeding and no longer suckle directly.

Overall, pumping and storing breast milk is probably the best balance for women who can't breastfeed directly for some reason.


Thanks. Your response to Zubon is what I was driving at. I just didn't have the time or knowledge to articulate it properly. I'm missing more than 6 IQ points I suspect, and I was breast fed!

Eliezer: Really? What do you think 6 points gets you? Would you eat your mother?

the PROBIT study, which was conducted by Canadian and American researchers in Belarus (where most women use formula), did actually conduct an randomized controlled trial, and found that the two groups had a statistically significant 6 point difference in IQ

Where? You seem to mean the randomized PROBIT study or studies in Belarus by Kramer et al. I took a look at the abstracts of 2 papers and 2 followups (Kramer, 2001, 2002, 2004, 2007). None mention any IQ difference.

Eliezer, that is my summary of their argument. I cannot myself address the quality of the original study or its critiques. I was curious, so I had a librarian send me the two-page critique. The surrounding letters are also interesting. One notes that there are three options: breast milk from a breast, breast milk from a bottle, and formula from a bottle (Science! has yet to perfect the fourth permutation); studies often fail to separate the middle category, and may group it with "breast milk" or "bottle fed" inappropriately or unclearly. It goes on to note feeding positions and inner ear fluid, a topic that has been near and dear to OB recently.

How large a truckload are we talking here? I only have so much freezer space.

Billb, like I said... don't ask.

Tom: See http://www.allattamentoalseno.it/lavori/Allattamento_esclusivo_al_seno_e_sviluppo_cognitivo.pdf for a non-gated version of the article (originally published in "Archives of General Psychiatry"), which mentions the IQ differences.

Robin points to an article that contests the findings of these researchers (I can't find an ungated version, so I plan to read it tomorrow, when I'll have access to it at school). The article is located at http://archpsyc.ama-assn.org/cgi/content/full/65/12/1456-a

6 points gets you another correct answer on an iq test, obviously. That's worth a lot of kittens.

@Johnathan Graehl

"6 points"

"Trasande et al. used a value for lifetime earnings (in 2000 dollars) of $1,032,002 for males and $763,468 for females based on the work of Max et al. (2004). . . .Both the U.S. EPA and Trasande use the work of Salkever (1995) to estimate the effect that one IQ point decrement has on earnings. Because the Trasande earnings are sex specific, the authors use Salkever’s sex-specific results. For each IQ point decrement, males experience a 1.93% decrease in lifetime earnings and females experience a 3.23% decrease."

-- US EPA study on effects of IQ detriment due to mercury poisoning in children.

6 IQ points = about US$120K? Check my math, please.

But seriously, you all need to read the whole article. The chick is not actually be-otching about breastfeeding - that's an excuse. Read it with skirt-eyes and pay attention to the inter-text.

It's filled with rage against her husband, what she perceives as inequality in her marriage. The kids & breast-feeding is really an excuse to beat her husband with a stick in public. The nut graf:

"About seven years ago, I met a woman from Montreal, the sister-in-law of a friend, who was young and healthy and normal in every way, except that she refused to breast-feed her children. She wasn’t working at the time. She just felt that breast-feeding would set up an unequal dynamic in her marriage — one in which the mother, who was responsible for the very sustenance of the infant, would naturally become responsible for everything else as well. At the time, I had only one young child, so I thought she was a kooky Canadian — and selfish and irresponsible. But of course now I know she was right. I recalled her with sisterly love a few months ago, at three in the morning, when I was propped up in bed for the second time that night with my new baby (note the my). My husband acknowledged the ripple in the nighttime peace with a grunt, and that’s about it. And why should he do more? There’s no use in both of us being a wreck in the morning. Nonetheless, it’s hard not to seethe."

Yikes! Um, Ms. Rosin, if you actually stopped whining like a wimpy victim about the choices you yourself have made, accepting that you were a full moral agent, and further if you had any respect for your husband as a human being - you would just hire a nanny, get thee to couples counseling, and thus not have to castrate him in public to express your scary anger.

Articles like this make me really ashamed to be a heterosexual feminist sometimes.

Its funny how breast feeding is "medicine" and formula isn't -- when one is obviously a recent technological advancement and the other has been around for ages (Economic benefits of breastfeeding: a review and analysis. US Dept of Agriculture). Supposedly breast feeding reduces costs, and contra what some people are saying, should be easier for poorer women.

The two clearest benefits I remeber about breast feeding are reduced infections (Breast milk provides immunity to baby, by passing moms immunoglobulins on) and pair-bonding for mom and baby. Benefits to Mom and Baby are clearly summarized, along with the controversies surrounding cognitive improvement in UpToDate review articles.

As many other people have pointed out -- Robin can't be bothered to read the medical literature. He can't be bothered to find a review. He just doesn't want to hear contradictory evidence -- god help anyone who listens to Robin's medical advice.

For the billionth time -- get a damn subscription to uptodate and read the review article before you even try to contradict anything. The review articles cite the primary literature AND the problems with the studies as well as the argument for current medical thinking.

To demonstrate how LAZY and reckless you are, I just want to highlight what happened when I went to UpToDate. For those not familiar -- uptodate is the equivalent of google for medicine. In any other setting You caa't make an argument against a popularly held belief without bothering to google counter arguments. So, I put breast feeding in the search box, and 3 of the top 10 results are
1.) Maternal and economic benefits of breastfeeding
2.) Infant benefits of breastfeeding
3.) The impact of breastfeeding on the development of allergic disease

Each of these reviews has over 50 references. Here is an excert on the part on cognitive development:

A number of studies have shown small neurodevelopmental advantages in children who were breast-fed compared to those who received formula, as illustrated below:
Cognitive development — Although there have been several reports that breastfeeding improves cognitive development later in childhood and adolescence to breastfeeding [76-79], this association remains uncertain [80].

invalidates the randomization

It invalidates the blinding, not the randomization! It seems like a pretty small problem (every study has problems), though the cherry-picking multiple measures of IQ is a big problem.

frelkins, I think that you misread her. You seem to think that she is being sarcastic when she writes, "My husband acknowledged the ripple in the nighttime peace with a grunt, and that’s about it. And why should he do more? There’s no use in both of us being a wreck in the morning." I think that she was being sincere. That interpretation seems to me to be most consistent with the following line: "Nonetheless, it’s hard not to seethe." The "nonetheless" to me implies that she grants the truth of the preceding sentence.

She is saying that she had a very human reaction of seething when she was working in the middle of the night while the person next to her slumbered, but that she saw good reasons for why it should be that way. As a male who doesn't like to be roused in the middle of the night, I can easily understand her irritation at watching her husband sleep while she has to get up. The irritation may not be rational, but I certainly don't think that her feeling it implies that she doesn't have "any respect for [her] husband as a human being."


"She is saying that she had a very human reaction of seething"

Oh no, I don't think she was being sarcastic. No, I did not misread her. Rather you can't seem to see the not-very-subtext that is whacking you upside the head.

You see, this is why so many OB readers don't seem to have successful relationships. Many appear to lack understanding of what constitutes dysfunctional behavior.

"Seething" is not a functional relationship strategy. It may be human, but it will destroy your love and respect for one another in an acid sea of resentment.

Don't "seethe" in your relationships. If you can't talk about X right there on the spot, respectfully and with tenderness, you're screwed. Or rather, you won't be. Ask anyone who's been married, oh 10 or 20 years.

Resenting your husband and writing about it in national magazines isn't my suggested way to go. It's the airing the ugly resentment she has carefully stored for such a monetary occasion - look! how humorous, how self-deprecating! - that indicates lack of respect for her husband as a human being. It's a tantrum, and we all know about tantrums. Don't try this at home, ok?

Resenting your husband for the choices you have made and your inability or refusal to communicate your issues around them shows a lack of maturity, a lack of accepting responsibility for your own behavior. This is repulsive man-hating victim feminism in a not-so-subtle form.

After reading the paper that Hanson cites, my "faith in breasts" is somewhat diminished. However, I do put some stock in arguments from authority, as a "shortcut" regarding topics that I don't know much about. Thus, given the fact that so many medical organizations that claim to have reviewed the evidence strontly favor breastfeeding, I tentatively lean toward breastfeeding being superior to formula, albeit with a bit less enthusiasm. While I recognize that both might be equal, I believe that there is little chance that formula is superior, and that there's a possible chance that breast milk could be significantly superior- a possibilty that would likely scare me into breastfeeding, even given its costs. I'm not sure what I'd recommend to my patients.

Can you quantify the difference 6 points makes?

Note that the link between breastfeeding and IQ is important not only because of the significance of 6 IQ points, but also because we know of very few other interventions that can similarly boost intellectual ability.

>>>BillB, I would eat the kittens.
>>Would you eat your mother?
>Billb, like I said... don't ask.

I think the real question is "Would you eat babies?"...

The comments to this entry are closed.

Less Wrong (sister site)

May 2009

Sun Mon Tue Wed Thu Fri Sat
          1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30