ANALYSIS OF VARIANCE PROCEDURES

Abstract

Research on analysis of variance procedures is described. A new series of orthogonal maineffect plans is presented, allowing for the testing of (2(s to the nth power -1)/(S-1)) factors each at s levels in 2 s to the nth power observations and two classes of main-effect plans for factors with unequal numbers of levels. Previous work on the analysis of variance is extended by developing general rules for the structure of the analysis of variance and expectations of mean squares for balanced samples from balanced populations which involve nesting crossingAND RANDOM CONFOUNDING. Results of research are given on one class of partially balanced structures, the symmetric Latin cube. Early results are made on an attempt to obtain a general formulation for partially balanced structures, based on orthogonal partitions. The role of covariance structures induced by randomization and the relevance of these to the use of least squares are considered. Some initial work on the power of randomization tests is described. Some preliminary work on constrained randomization is given. (Author)

Document Details

Document Type
Technical Report
Publication Date
Nov 01, 1961
Accession Number
AD0272399

Entities

People

  • G. Zyskind
  • O. Kempthorne

Organizations

  • Iowa State University

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Analysis Of Variance
  • Computing-Related Activities
  • Covariance
  • Data Science
  • Information Science
  • Interdisciplinary Science
  • Mathematical Analysis
  • Mathematics
  • Observation
  • Statistical Analysis

Fields of Study

  • Mathematics

Readers

  • Graph Algorithms and Convex Optimization.
  • Regression Analysis.