Monday, October 14, 2013

Child behavior, bedtimes, and study design

The NY Times reports on a Pediatrics article on regular bedtimes and child behavior. As you might expect, parents' self-reports of regular bedtimes is associated with better children's behavioral scores, but who knows whether it's having a regular bedtime that causes better scores, or better behavior causing the scores and the regular bedtime. Not revealed in the Times article is that the study looks at differences-in-differences or changes in these variables over time in the panel, which will remove any unchanging sources of variation. But without a clearer source of identifying variation, it's hard to accept such results reveal a causal lever.