
Accession Number : AD0675558
Title : LINEAR MODELS AND ANALYSIS OF VARIANCE RESEARCH PROCEDURES,
Corporate Author : IOWA STATE UNIV OF SCIENCE AND TECHNOLOGY AMES
Personal Author(s) : Zyskind,G. ; Kempthorne,O. ; Martin,F. B. ; Carney,E. J. ; West,E. N.
Report Date : JUL 1968
Pagination or Media Count : 188
Abstract : Research on some related linear model theory and analysis of variance procedures is described. Chapter I is introductory, giving a general outline of topics described in the report. Chapter II presents a formulation of aspects of best and simple least squares linear estimation in linear models with arbitrary, possibly singular, covariance structure. Chapter III develops a generalization of the famed GaussMarkoff theorem, applying to situations including a singular variancecovariance structure of the observations. Chapter IV deals with simple combination of information in linear models originating from uncorrelated distinct sources of information. Chapter V deals with some formulations of sampling from balanced complete experimental structures, and with theoretical and computational aspects arising in the calculation of variances of variance components. Chapter VI presents results of a Monte Carlo investigation of significance levels generated by the BehrensFisher fiducial procedure and by the Welch Aspin procedure. Chapter VII presents a brief discussion of certain nonparametric test procedures based upon ranks and, in particular, points up the problems arising from the inevitable grouping error of measurement. (Author)
Descriptors : (*ANALYSIS OF VARIANCE, MATHEMATICAL MODELS), LINEAR SYSTEMS, LEAST SQUARES METHOD, DECISION THEORY, SAMPLING, MONTE CARLO METHOD, VECTOR SPACES, STATISTICAL DISTRIBUTIONS, STATISTICAL TESTS, CORRELATION TECHNIQUES, COMBINATORIAL ANALYSIS, THEOREMS
Subject Categories : Statistics and Probability
Distribution Statement : APPROVED FOR PUBLIC RELEASE