One of the criticisms of Oster’s work is that she cherry-picked data. But as an economist, Oster is especially well-suited to identifying how a study that initially seems well-constructed might be flawed. She looks, for example, at a paper published in Pediatrics in 2001, in which researchers concluded that light drinking during pregnancy impacts children’s future behavior.
Oster summarizes the authors’ conclusion: “When the authors compared women who didn’t drink during pregnancy to those who had one drink or less per day, they found more evidence of aggressive behavior (although not of other behavior problems) among the children of women who drank.”
That sounds like damning evidence for the danger of alcohol consumption during pregnancy. One drink a day and your child will be the school bully.
What’s missing from the researchers’ conclusions, Oster notes, is that nearly half of the study’s drinking mothers were also using cocaine, while only 18 percent of the non-drinking mothers were. Oster posits that perhaps it’s the difference in cocaine use that made the impact on childhood behavior.
You can read more here.