Intrinsic tests for the equality of two correlated proportions

TL;DR

This paper introduces intrinsic prior-based tests for comparing two correlated proportions in longitudinal studies, addressing limitations of Bayesian methods by ensuring fair hypothesis comparison and providing practical strategies with real data examples.

Contribution

It develops two objective prior strategies for testing equality of correlated proportions, improving prior centering and interpretability over existing Bayesian approaches.

Findings

New intrinsic prior methods for correlated proportions

Sensitivity analysis demonstrates robustness of the proposed tests

Comparison with existing methods shows improved fairness and interpretability

Abstract

Correlated proportions arise in longitudinal (panel) studies. A typical example is the ``opinion swing'' problem: ``Has the proportion of people favoring a politician changed after his recent speech to the nation on TV?''. Since the same group of individuals is interviewed before and after the speech, the two proportions are correlated. A natural null hypothesis to be tested is whether the corresponding population proportions are equal. A standard Bayesian approach to this problem has already been considered in the literature, based on a Dirichlet prior for the cell-probabilities of the underlying two-by-two table under the alternative hypothesis, together with an induced prior under the null. In lack of specific prior information, a diffuse (e.g. uniform) distribution may be used. We claim that this approach is not satisfactory, since in a testing problem one should make sure that the prior under the alternative be adequately centered around the region specified by the null, in order to obtain a fair comparison between the two hypotheses. Following an intrinsic prior methodology, we develop two strategies for the construction of a collection of objective priors increasingly peaked around the null. We provide a simple interpretation of their structure in terms of weighted imaginary sample scenarios. We illustrate our method by means of three examples, carrying out sensitivity analysis and providing comparison with existing results.