Event Title
Comparing Relative Efficiency of Nonparametric to Parametric Methods using Monte Carlo Simulation
Date of Publication
2-3-2021
Document Type
Presentation
Abstract
It is well-known that Type I or Type II error control in parametric statistical inference is related to the tenability of various model assumptions. The purpose is to demonstrate situations in which nonparametric tests may be more powerful than parametric tests under varying degrees and types of model assumption violation. The method is simulation-based, using a new, mixture-based simulation method for generating distributions with varying degrees of nonnormality. The results of this study will provide applied researchers with important information that can be used for designing studies and/or getting the most statistical power out of the available data.
Keywords
Nonparametric to Parametric Methods, Monte Carlo Simulation
Persistent Identifier
http://hdl.handle.net/10950/2957
Comparing Relative Efficiency of Nonparametric to Parametric Methods using Monte Carlo Simulation
It is well-known that Type I or Type II error control in parametric statistical inference is related to the tenability of various model assumptions. The purpose is to demonstrate situations in which nonparametric tests may be more powerful than parametric tests under varying degrees and types of model assumption violation. The method is simulation-based, using a new, mixture-based simulation method for generating distributions with varying degrees of nonnormality. The results of this study will provide applied researchers with important information that can be used for designing studies and/or getting the most statistical power out of the available data.