Take Action
Research Details
Choosing Sample Sizes
Abstract
How much data about an unknown parameter would a designer provide to a decision maker (DM) in order to convince the DM that the parameter value is sufficiently high? We study this question for DMs who are unbiased or Bayesian statisticians and for data which are Bernoulli experiments governed by the parameter value. We establish that in many environments the designer's optimal sample size is the largest one satisfying that one or more --- or a simple majority of --- favorable data realizations would convince the DM that the parameter value is sufficiently high.
Type
Working Paper
Author(s)
Yuval Salant, Sanket Patil
Date Published
2020
Citations
Salant, Yuval, and Sanket Patil. 2020. Choosing Sample Sizes.
LINK