Logo image
A representation of orthogonal components in analysis of variance
Journal article   Open access   Peer reviewed

A representation of orthogonal components in analysis of variance

B.R. Clarke
International Mathematical Journal, Vol.1(2), pp.133-147
2002
pdf
orthogonal_components_in_analysis_of_variance.pdfDownloadView
Published (Version of Record) Open Access

Abstract

The history of orthogonal components in the analysis of variance and of the Helmert transformation is allied to more recent mathematical representations of Kronecker products. The latter are well known tools useful in the teaching of factorial designs. Presented here are quick derivations of the usual orthogonal projection matrices associated with independent sums of squares for some familiar balanced designs. Succinct representations used to form orthogonal contrasts clearly illustrate the proofs of independence of sums of squares and give obvious interpretation to degrees of freedom. Resultant central and noncentral chi-squared distributions follow easily. A relationship between recursive residuals and the Helmert transformation is also noted.

Details

Metrics

373 File views/ downloads
120 Record Views
Logo image