Abstract
The Problem
Researchers have described challenges and problems in reporting research that uses only p values and power to make decisions to reject the null hypothesis. Confusion about how to interpret null hypothesis statistical tests has resulted from mixed information presented in statistics articles and textbooks.
The Solution
Combining evidence from data with initial beliefs, Bayesian inference techniques help to provide uncontroversial support of a null hypothesis or alternative hypothesis. An overview of the limitations associated with only using p values and power to make decisions to reject or retain the null hypothesis are presented. Analyses across multiple studies with common parameters can be pooled using Bayesian techniques as a means for conducting meta-analysis. Examples using Bayesian techniques are given.
The Stakeholders
When designing a research study, researchers often use external elements and likelihood to make powerful inferences using Bayesian techniques. Those performing research in populations requiring sampling.
Get full access to this article
View all access options for this article.
