Normalized power priors always discount historical data

TL;DR

This paper reveals that normalized power priors with beta-distributed power parameters inherently fail to fully incorporate historical data when current data perfectly mirror past data, due to their marginal posterior behavior.

Contribution

The paper provides a novel theoretical analysis showing the limitations of power priors in Bayesian analysis, especially their inability to fully leverage historical data under certain conditions.

Findings

Marginal posterior of alpha does not concentrate at 1 even with large sample sizes.

Normalized power priors cannot fully pool historical and current data.

Complete data pooling is impossible with beta-distributed power priors.

Abstract

Power priors are used for incorporating historical data in Bayesian analyses by taking the likelihood of the historical data raised to the power $\alpha$ as the prior distribution for the model parameters. The power parameter $\alpha$ is typically unknown and assigned a prior distribution, most commonly a beta distribution. Here, we give a novel theoretical result on the resulting marginal posterior distribution of $\alpha$ in case of the the normal and binomial model. Counterintuitively, when the current data perfectly mirror the historical data and the sample sizes from both data sets become arbitrarily large, the marginal posterior of $\alpha$ does not converge to a point mass at $\alpha = 1$ but approaches a distribution that hardly differs from the prior. The result implies that a complete pooling of historical and current data is impossible if a power prior with beta prior for $\alpha$ is used.