Estimating Confidence Intervals of Mediating Effects by Using the Distribution of the Product, Bootstrap and Markov chain Monte Carlo Methods

FANG Jie,ZHANG Min-Qiang,LI Xiao-Peng

DOI: https://doi.org/10.3724/SP.J.1042.2011.00765

2011-01-01

Advances in Psychological Science

Abstract:Because the estimators of mediating effects are generally not normally distributed, it would be better use asymmetric confidence intervals to analyze mediating effects. There are three approaches to obtain the asymmetric confidence intervals of mediating effects: 1) Based on the distribution of the product (including M method and Empirical-M method); 2) Bootstrap methods (nonparametric percentile Bootstrap method, Bias-corrected nonparametric percentile Bootstrap method, parametric percentile residual Bootstrap method and bias-corrected parametric percentile residual Bootstrap method); 3) Markov chain Monte Carlo (MCMC) methods. After introducing each of the methods in details, we compared them and found the following results: 1) the behaviors of the three approaches were approximate. 2) Compared with Distribute of the product methods, bias-corrected percentile Bootstrap method was better. 3) The mean square error (MSE) of the MCMC with prior information was smaller than Distribute of the product methods. Directions for further research on asymmetric confidence intervals of mediating effects were discussed.