The Globe and Mail reports on a (very) long-term followup of a study from the 1970s that looked at the effects of protein on the brain development of children in Guatemala.
Missing from the newspaper report:
1. Information about the design of the study. Was it a randomized trial? One imagines so, but the article never actually says so.
2. Information about differential non-response by treatment status. The article, to its credit, does note the overall follow-up response rate of 0.60 but says nothing about whether that differed by treatment status, which is key.
3. A smell test. The reported effect on wages (what about non-workers?) is huge at 46 percent. Is this really plausible given the other evidence in the literature and/or what is known about the underlying biology?
4. A link to the article in the Lancet.
Replicating this and trying out various econometric methods for dealing with the non-response - e.g. the "Lee bounds" - would be a good second or third-year paper for some aspiring gradual student.
Hat tips: Dan and Susan Black
Who was my favorite student this term?
4 months ago