Analysis of Variance: Selection of a Model and Summary Statistics
Abstract
The topics of this paper are models for the analysis of variance (ANOVA) (fixed, random, or mixed models) and the subsequent summary statistics (F ratio, quasi-F ratio, and magnitude of treatment effect) that may be computed following the ANOVA. ANOVA is a useful method for assessing the statistical significance of treatment effects. But the significance of an effect is a function of two decisions. The first decision is the selection of a model and an appropriate sampling plan for elements within each of the treatment factors. The second decision is the choice of summary statistics that indicate the extent of significance achieved. In this paper, comparisons will be made between models and between summary statistics. Specific issues will be clarified concerning the interpretation of results when various models and summary statistics are used on the same set of data.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1978
- Accession Number
- ADA077965
Entities
People
- Frederick H. Steinheiser Jr.
- Kenneth I. Epstein
Organizations
- U.S. Army Research Institute for the Behavioral and Social Sciences