Friday, March 14, 2014

Bayesian estimation and precision as the goal for data collection (expanded)

Precision as the goal for data collection, talk at U.C. Irvine, March 14, 2014.

Part 1: Rejecting null is not enough, also need estimate and precision. Bayesian estimation supersedes confidence intervals and "the new statistics".


Part 2: Two Bayesian ways to assess a null value. Highest density interval with region of practical equivalence. Bayesian model comparison and Bayes factor.


Part 3: Biased estimation in sequential testing and optional stopping.


Part 4: Monte Carlo study of biased estimation in sequential testing and optional stopping.


Parts 3 and 4 are elaborations of this previous blog post.

1 comment:

  1. Thanks for posting - I hope the conference was informative!

    Robin

    ReplyDelete