Combining Statistical Evidence from Multiple Studies Using Bayesian Evaluation of Inequality Contrained Hypotheses: Can Learning or Control Better Predict Trust?

Abstract

Often researchers have multiple studies available with respect to answering a research question, such as one variable being a better predictor than another in explaining a phenomenon. Usually these studies differ in setup, though are still measuring the same underlying concept. This paper will show how inequality constrained hypotheses can be formulated to answer the research question, as well as how to construct and use the Bayes factor to quantify support for competing hypotheses. The prediction of trust will be used as an example to illustrate this new procedure for combining evidence of multiple studies.