Matthew S. Fritz
Date of this Version
Chen, D. (2018). A comparison of alternative bias-corrections in the bias-corrected bootstrap test of mediation. Public Access Theses and Dissertations from the College of Education and Human Sciences.
Although the bias-corrected (BC) bootstrap is an oft recommended method for obtaining more powerful confidence intervals in mediation analysis, it has also been found to have elevated Type I error rates in conditions with small sample sizes. Given that the BC bootstrap is used most often in studies with low power due to small sample size, the focus of this study is to consider alternative measures of bias that will reduce the elevated Type I error rate without reducing power. The alternatives examined fall under two categories: bias correction and transformation. Although the bias correction methods did not significantly decrease Type I error rate, the associated confidence intervals were similar to the original BC bootstrap. The transformations, however, did not produce confidence intervals with more accurate Type I error rate.
Advisor: Matthew S. Fritz