Chris Blattman

Intuitive yet important results

2 Responses

  1. This insurance trial is much better designed than the last big one in Malawi. This trial compared a control group that got a payout equal to the expected value of the insurance to a treatment group that received free insurance. The Malawi trial compared control regions that got nothing to a control group than got free insurance – so it really wasn’t apples to apples.

    The one thing that worries me is that they conveniently moved the goal posts on their threshold for statistical significance when their main finding came in with a p-value of 0.055. Chris why don’t controlled trial-folks just go Bayesian so they can avoid these shenanigans?

Comments are closed.

Why We Fight - Book Cover
Subscribe to Blog