The significance of a model is affected by outliers.The outliers can affect the effectiveness of structural equation modeling (SEM). Here we describe and investigate the behavior of the nonparametric single and double residual bootstrap (DRB) methods in the presence of outliers when applied to SEM. Our study also intends to shorten the computational time of the standard double bootstrap by using an alternative double bootstrap approach. We demonstrate our proposed method by conducting a Monte Carlo experiment series for clean normal Gaussian distributions and contaminated data. The simulation studies were manipulated with different sample sizes, effect sizes, and 10% of contamination in the Y direction. The performance of the proposed method is evaluated using standard measurements and the construction of confidence intervals. The reasonably close parameter and bootstrap estimates suggest that the nonparametric single and double residual bootstrap is an excellent method. The DRB method showed a robust declining pattern for standard measurement estimates and shorter confidence intervals compared to the single residual bootstrap method in both normal and contaminated data. Also, the double bootstrap method takes twice as long as the single bootstrap method to compute. The DRB method is straightforward but demands slightly more computational time and better prediction approximation. This study offers additional perspectives to fellow researchers considering using the nonparametric single and alternative DRB methods with contaminated data.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.