Files in this item

FilesDescriptionFormat

application/pdf

application/pdf8108688.pdf (4MB)Restricted to U of Illinois
(no description provided)PDF

Description

Title:A Monte Carlo Study: Robustness of Analysis of Covariance to Violation of Selected Assumptions
Author(s):Thomson, David Samuel
Department / Program:Education
Discipline:Education
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:Ph.D.
Genre:Dissertation
Subject(s):Statistics
Abstract:It was the purpose of this study, through the use of a Monte Carlo simulation, to investigate the robustness of the analysis of covariance (ANCOVA) model relative to violations of some of the model's assumptions, and to various experimental conditions. In particular, the effect of violating the following assumptions was examined: (1) homgeneity of within-cell population regression coefficients (i.e., slopes), and (2) homogeneity of within-cell population error variances. In addition, the effect of having a contaminated sampling distribution from which the within-cell population error variances were generated (i.e., 10 or 20% of the sampling distribution has a proportionally larger variance) was examined as an experimental condition. The effects of the violations of these assumptions and of the above mentioned experimental condition were considered for equal, as well as unequal, sample sizes. The actual Analysis of Covariance F distribution generated by using (i) different sample sizes, (ii) heterogeneous within-cell regression coefficients, (iii) heterogeneous within-cell error variances, and/or (iv) a contaminated sampling distribution for the within-cell error variances, was compared to the theoretical F distribution under the null hypothesis, when all assumptions of the ANCOVA model have been met. In addition, analysis of variance was used to analyze the results of the Monte Carlo simulation with heterogeneous within-cell regression coefficients, and heterogeneous within-cell error variances for the two and three group cases. The power of the statistical tests used to detect heterogeneous within-cell error variances and heterogeneous within-cell regression coefficients (slopes) were also examined.
Issue Date:1980
Type:Text
Language:English
Description:148 p.
Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 1980.
URI:http://hdl.handle.net/2142/66019
Other Identifier(s):(UMI)AAI8108688
Date Available in IDEALS:2014-12-12
Date Deposited:1980


This item appears in the following Collection(s)

Item Statistics