Not blinded in Belarus

When this study on breastfeeding and IQ was hyped in the press (MSNBC; see WebMD for a better take) why did no one mention that the pediatricians who rated the IQ of the children in question (more vs. less breastfeeding) were not blinded to which group the children were in?

I would understand if the findings of the study were a slam dunk, but they aren't: verbal IQ was the only category with a statistically significant difference (performance and total were not statistically significant), and the IQ difference is probably not behaviorally significant in any case (what's five or seven points?).

Given that the differences are small and possibly not statistically significant, it behooves those reporting the news to give a sense of potential sources of bias: factors which could skew the results in one way or another. And one way in which results can be skewed is for raters to "know" or "guess" while rating the children in one group that their IQ should be higher. We are all inherently biased creatures - blinding is the way, in modern epidemiology, to reduce bias. When small results are claimed (as they are here) a source of bias like this is quite worrisome.

The cognitive benefits of breastfeeding remain unproven.

(Of course my son is being breastfed - with formula too. But the reason he'll grow up smart, God willing, is because his mother is smart.)

No comments:

Post a Comment