Multilevel random mediation analysis: A comparison of analytical alternatives

Abstract

The present article focuses on the multilevel random mediation effects model (1-1-1) and examines its various analytical procedures. The performances of these procedures under a variety of conditions were compared using Monte Carlo simulations. We compared the multilevel random mediation model with two compact models: the multilevel fixed mediation model and the single-level traditional mediation model. The results showed better performance for the multilevel random mediation model. The results also indicated that we can obtain unbiased estimation of the mediation effect, the correct standard error, and proper hypothesis testing results from the multilevel random mediation model. Moreover, the differences between the multilevel fixed mediation model and the single-level traditional mediation model are minimal. Several implications and recommendations for this application are discussed.