Statistical Modeling, Causal Inference, and Social Science

2 Comments

1. revo11 says:

   November 28, 2011 at 2:17 pm

   thanks for sharing, it’s frustrating how easily “can’t reject a null hypothesis for a specific measure of efficacy” gets turned into “proof that the drug is ineffective for all populations” (of course similar fallacies happen too frequently when the null is rejected as well)
2. Dave Backus @ NYU says:

   November 28, 2011 at 5:13 pm

   Great idea, thanks.