A Bayesian Approach to Mediation

Bayesian statistics provide researchers a powerful tool when analyzing data. The purpose of this project is to provide an example how psychological researchers can implement Bayesian theory and methodology to improve research analyses. This project conducted two simultaneous mediation analyses to test the hypothesis that relational interdependent self-construal’s effect on an individual’s self-perceived romantic relationship quality is explainable by perceived partner’s romantic relationship quality. Analysis 1 was conducted using traditional maximum likelihood estimation (ML). Analysis 2 was conducted using a Bayesian Markov Chain Monte Carlo (MCMC) method. The results of this project offer a stark example of the benefits which MCMC provides over ML when conducting mediation analysis. It is suggested that psychological researchers begin to embrace Bayesian theories regarding how data can be analyzed to better the statistical quality of studies