Advertisement

Behavior Research Methods

, Volume 46, Issue 4, pp 1184–1198 | Cite as

Monte Carlo based statistical power analysis for mediation models: methods and software

  • Zhiyong ZhangEmail author
Article

Abstract

The existing literature on statistical power analysis for mediation models often assumes data normality and is based on a less powerful Sobel test instead of the more powerful bootstrap test. This study proposes to estimate statistical power to detect mediation effects on the basis of the bootstrap method through Monte Carlo simulation. Nonnormal data with excessive skewness and kurtosis are allowed in the proposed method. A free R package called bmem is developed to conduct the power analysis discussed in this study. Four examples, including a simple mediation model, a multiple-mediator model with a latent mediator, a multiple-group mediation model, and a longitudinal mediation model, are provided to illustrate the proposed method.

Keywords

Power analysis Mediation models Nonnormal data Bootstrapping R package bmem 

Notes

Author Note

We thank Scott Maxwell and Ke-Hai Yuan for helpful discussions and David Kenny and one anonymous reviewer for constructive suggestions that have significantly improved this research. Path diagrams used in the article were generated using WebSEM (https://websem.psychstat.org).

References

  1. Beasley, T. M. (2012). Power of product tests of mediation as a function of mediator collinearity. Multiple Linear Regression Viewpoints, 38, 17–23.Google Scholar
  2. Cheung, M. W. L. (2007). Comparison of approaches to constructing confidence intervals for mediating effects using structural equation models. Structural Equation Modeling, 14(3), 227–246.CrossRefGoogle Scholar
  3. Daly, L. E. (1991). Confidence intervals and sample sizes: Don’t throw out all your old sample size tables. BMJ: British Medical Journal, 302, 333–336.PubMedCentralPubMedCrossRefGoogle Scholar
  4. Fritz, M. S., & MacKinnon, D. P. (2007). Required sample size to detect the mediated effect. Psychological Science, 18, 233–239.PubMedCentralPubMedCrossRefGoogle Scholar
  5. Fritz, M. S., Taylor, A. B., & MacKinnon, D. P. (2012). Explanation of two anomalous results in statistical mediation analysis. Multivariate Behavioral Research, 47, 61–87.PubMedCentralPubMedCrossRefGoogle Scholar
  6. Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: Guilford Press.Google Scholar
  7. Hayes, A. F., & Scharkow, M. (2013). The relative trustworthiness of inferential tests of the indirect effect in statistical mediation analysis: Does method really matter? Psychological Science, in press.Google Scholar
  8. Hoenig, J. M., & Heisey, D. M. (2001). The abuse of power. The American Statistician, 55, 19–24.CrossRefGoogle Scholar
  9. Kenny, D. A., & Judd, C. M. (2013). Power anomalies in testing mediation. Psychological Science, in press. (R code at http://www.davidakenny.net/progs/PowMedR.txt)
  10. Knaus, J. (2013). snowfall: Easier cluster computing (based on snow). [Computer software manual]. Retrieved from http://CRAN.R-project.org/package=snowfall (R package version 1.84-4)
  11. MacKinnon, D. P. (2008). Introduction to statistical mediation analysis. New York, NY: Taylor & Francis.Google Scholar
  12. MacKinnon, D. P., Lockwood, C. M., & Williams, J. (2004). Confidence limits for the indirect effect: Distribution of the product and resampling methods. Multivariate Behavioral Research, 39(3), 99–128.PubMedCentralPubMedCrossRefGoogle Scholar
  13. Maxwell, S. E., & Cole, D. A. (2007). Bias in cross-sectional analyses of longitudinal mediation. Psychological Methods, 12, 23–44.PubMedCrossRefGoogle Scholar
  14. Meehl, P. E. (1997). The problem is epistemology, not statistics: Replace significance tests by confidence intervals and quantify accuracy of risky numerical predictions. In L. L. Harlow, S. A. Mulaik, & J. H. Steiger (Eds.), What if there were no significance tests? (pp. 393–425). Mahwah, NJ: Lawrence Erlbaum Associates.Google Scholar
  15. Micceri, T. (1989). The unicorn, the normal curve, and other improbable creatures. Psychological Bulletin, 105(3), 156–166.CrossRefGoogle Scholar
  16. Muthén, L. K., & Muthén, B. O. (1998-2011). Mplus user’s guide (Sixth Ed.). Los Angeles, CA: Muthén and Muthén. (http://www.statmodel.com)
  17. Muthén, L. K., & Muthén, B. O. (2002). How to use a Monte Carlo study to decide on sample size and determine power. Structural Equation Modeling, 9(3), 599–620.CrossRefGoogle Scholar
  18. Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior Research Methods, Instruments, & Computers, 36, 717–731.CrossRefGoogle Scholar
  19. Qiu, W. (2013). powermediation: Power/sample size calculation for mediation analysis [Computer software manual]. Retrieved from http://CRAN.R-project.org/package=powerMediation (R package version 0.1.3)
  20. Rosseel, Y. (2012). lavaan: An R package for structural equation modeling. Journal of Statistical Software, 48, 1–36.Google Scholar
  21. Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies: New procedures and recommendations. Psychological Methods, 7, 422–445.PubMedCrossRefGoogle Scholar
  22. Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structural equation models. In S. Leinhardt (Ed.), Sociological methodology (pp. 290–312). San Francisco: Jossey-Bass.Google Scholar
  23. Thoemmes, F., MacKinnon, D. P., & Reiser, M. R. (2010). Power analysis for complex mediational designs using Monte Carlo methods. Structural Equation Modeling, 17, 510–534.PubMedCentralPubMedCrossRefGoogle Scholar
  24. Vale, C. D., & Maurelli, V. A. (1983). Simulating multivariate nonnormal disttributions. Psychometrika, 48, 465–471.CrossRefGoogle Scholar
  25. Vittinghoff, E., Sen, S., & McCulloch, C. (2009). Sample size calculations for evaluating mediation. Statistics in Medicine, 28, 541–557.PubMedCrossRefGoogle Scholar
  26. von Oertzen, T., & Brick, T. (2013). Efficient hessian computation using sparse matrix derivatives in ram notation. Behavior Research Methods, in press.Google Scholar
  27. Wang, C., & Xue, X. (2012). Power and sample size calculations for evaluating mediation effects in longitudinal studies. Statistical Methods in Medical Research, in press.Google Scholar
  28. Wang, L., & Zhang, Z. (2011). Estimating and testing mediation effects with censored data. Structural Equation Modeling, 18(3), 18–34.CrossRefGoogle Scholar
  29. Yuan, K. H., Bentler, P. M., & Zhang, W. (2005). The effect of skewness and kurtosis on mean and covariance structure analysis. Sociological Methods & Research, 34(3), 240–258. Retrieved from doi: 10.1177/0049124105280200 CrossRefGoogle Scholar
  30. Zhang, Z., & Wang, L. (2009). Statistical power analysis for growth curve models using SAS. Behavior Research Methods, 41, 1083–1094.PubMedCrossRefGoogle Scholar
  31. Zhang, Z., & Wang, L. (2013a). bmem: Mediation analysis with missing data using bootstrap [Computer software manual]. Retrieved from http://CRAN.R-project.org/package=bmem (R package version 1.5)
  32. Zhang, Z., & Wang, L. (2013b). Methods for mediation analysis with missing data. Psychometrika, 78, 154–184.PubMedCrossRefGoogle Scholar
  33. Zu, J., & Yuan, K. H. (2010). Local influence and robust procedures for mediation analysis. Multivariate Behavioral Research, 45, 1–44.CrossRefGoogle Scholar

Copyright information

© Psychonomic Society, Inc. 2013

Authors and Affiliations

  1. 1.University of Notre DameNotre DameUSA

Personalised recommendations