Coverage and interval length of Welch's and Yuen's procedures for shift in location and change in scale for (un)equal sample sizes
The purpose of this dissertation study was to assess the statistical precision for Welch's and Yuen's procedures. Several results have practical implications in applied practice. First, Welch's procedure displayed probabilities of coverage less than the confidence level when size was inversely proportional to heteroscedasticity and skewness was---1.33 or 1.64. Second, lower Type I and Type II error rates have been the basis for recommending Yuen's procedure instead of Welch's procedure (Wilcox, 1994; Luh & Guo, 2000); these results showed that unless skewness and heteroscedasticity were great in magnitude (i.e., | skewness | > 1.25 & σ2 /σ 2 ≥ 4), Welch's procedure displayed better statistical precision than Yuen's procedure. Third, if kurtosis of the population defined by the trimmed mean was less than -1.25, probabilities of coverage were attenuated at the 0.01 alpha level.
|Year of publication:||
|Authors:||Mends-Cole, Saydee Jonathan|
Wayne State University
|Type of publication:||Other|
ETD Collection for Wayne State University