This book discusses important applications of the Behrens-Fisher statistic and the False Discovery Rate (FDR). Covered applications include ANOVA and MANOVA under potentially non-normal errors and heteroscedasticity; and an intuitive method of analyzing s x r contingency tables when the column variable is ordinal. This book also explores the novel possibility that these applications may be deemed nonparametric.
Author(s): Tejas A. Desai
Series: SpringerBriefs in Statistics
Publisher: Springer
Year: 2022
Language: English
Pages: 87
City: Cham
Contents
1 The Behrens-Fisher Statistic
References
2 One-Way ANOVA with Possible Heteroscedasticity of Variances and Non-Normality of Errors
2.1 The One-Way ANOVA Problem and the Behrens-Fisher Statistic
2.2 Our Methodology for One-Way ANOVA When k≥3
2.3 The Classical ANOVA F Test
2.4 The Kruskal-Wallis Test
2.5 Comparative Analyses for One-Way ANOVA
References
3 One-Way MANOVA with Possible Heteroscedasticity and Non-Normality of Errors
3.1 The One-Way MANOVA Problem and Our Methodology For It
3.2 The Classical Wilks' Lambda
3.3 A Robust Statistic for One-Way MANOVA
3.4 Comparative Analyses for One-Way MANOVA
3.5 Data Analysis Example: The Oslo Transect Data
References
4 Statistical Analysis of s r Tables When the Column Variable Is Ordinal
4.1 Methodology for a Single 2 r Table
4.2 Simulations for a Single 2 r Table
4.3 Example of a Behrens-Fisher Analysis of a Single 2 r Table
4.4 Sets of 2 r Tables
4.4.1 Methodology
4.4.2 Simulations of Sets of 2 r Tables
4.4.3 Example of an Analysis of a Set of Two 2 r Tables
4.5 Behrens-Fisher Analysis of s r Tables, s ≥ 3 and Column Variable Ordinal
4.5.1 A Single s r Table When s ≥ 3 and Column Variable Ordinal
4.5.2 Simulations of a Single Set of s r Table, s > 2 and Column Variable Ordinal
4.5.3 Simulations of More Than One s r Table, s > 2, and Column Variable Ordinal
4.5.4 Example of an Analysis of a 3 3 Table When the Column Variable Is Ordinal
References
5 Concluding Remarks