April 20, 2009

This month's Atlantic Monthly has a controversial article about breastfeeding, specifically making arguments against breastfeeding.  The arguments are of two types.  One is social and normative, and the other is on the evidence.  My thoughts, naturally, are on the evidence.  The journalist makes a classic error in understanding and interpreting "treatment effects."  Here is a clip:

"Both the Kramer study and the sibling study did turn up one interesting finding: a bump in “cognitive ability” among breast-fed children. But intelligence is tricky to measure, because it’s subjective and affected by so many factors."

The Kramer study was a randomized control trial (a clustered encouragement design).  The sibling study is an econometric study which tries to control for parental fixed effects by comparing siblings (but can't control for a series of stories one could dream up about ordering effects).  Both find positive effects on IQ.  The RCT finds a 5 point increase.

But Hanna Rosin dismisses the evidence that breastfeeding causes a "bump in cognitive ability" by five points because it is "affected by so many factors."  This is akin to saying that even though we know that eating better as a child leads to higher height, it should be dismissed because, well, aging has so much more to do with height than eating a bit more protein!  The fact is, five points is significant, and the fact that IQ tests bounce around day to day is simply a matter of having sufficiently large sample sizes in the study, something Kramer has done and thoroughly addresses.  

Rosin also makes reference to Kramer's valid point that he cannot separate the milk from the cuddling, and that is fine... we may lack clear knowledge on the mechanism, but we still know the end result: breastfeeding leads to a five point increase in IQ, on average.

The irony of course that Rosin did not address is that the real "cost" of breastfeeding is actually imposed by society... the embarrassment factor.  If more women openly breastfed while sitting in the park, at the bank, in the office, or on the bus, we probably would be having a completely different conversation.  In developing countries, breastfeeding in many places is the norm.  I'd guess in a typical village bank microfinance meetings, one out of 10 women breastfeed while making their loan payments, or answering all our questions we ask when conducting a survey.  

But it is the case that in many parts of the world, women still lack important information about breastfeeding.  Freedom from Hunger, one of our favorite NGO's because of their dedication to the poorest, and their dedication to evidence-based development work, works diligently to integrate breastfeeding education modules into microfinance programs around the world.  We are working with them in Benin and Peru in fact to evaluate the impact of their health education modules.

So back to Rosin, she poses another question which I find much more interesting: why is there a growing social norm that if you do not breastfeed, you should be shunned.  There are countless things we do not do as parents that would be good for our kids, but we just deem them too costly (the safest possible car, e.g., or even the simple choice of how much time to spend with them).  Why do some "costs-benefit" analysis become subject to social pressure and norms, and others do not?  An easy answer would be externalities, but that doesn't seem to be the case.  So what is it about breastfeeding that generates this type of rhetoric?