Monday, June 9, 2008

Randomised baby data

Following on from last week's post about data collected by parents, Ian Ayres suggests using sites like Trixie Tracker to learn more about which parenting techniques work. His first suggestion is on controlled crying (something that I have written about here and here). The idea is that you need to 'negotiate' with your baby some sensible sleeping arrangements. I am an advocate but only based on economic theory and my own experience. But I am not sure it is for everyone and for every child.

What is interesting about the possibility of randomised experiments here is that there are no real ethical issues. Put simply, lots of people are giving parents advice to try various things. All this would do is ask parents to record what they actually did and a research could map that systematically with formal advice they are receiving. You could also control for diet and illness (you know when your baby is a little sick and uses that an excuse to halt the controlled crying standoff). This would put to rest (literally) random advice from parents like me with at most a few data points at their disposal and get us some real knowledge.