Indian Society of Genetics & Plant Breeding


Published on

Factor analysis as a branch of multivariate analysis useful to explain the
inter-correlations of variables is well known (Holzinger and Harman, 1941 ;
Maxwell, 1961; Thurstone, 1947; Lawley and Maxwell, 1963; Rao, 1964).
It helps to find out the number and nature of causative influences on which more
intensive work can be concentrated. As pointed out by Cattell (1965), its
utility lies not only at the exploratory stages of research but also at later stages
where the simultaneous action of several factors influencing a variable is to be
critically analysed. While the principal component analysis breaks down a
covariance matrix into a set of orthogonal components equal in number to the
number of variates irrespective of the distribution of the variates or even their
randomness, a factor model assumes that the
p correlated variables follow a
multivariate normal distribution and that their inter-correlations can be adequately accounted for by
k factors (k<P) which are linear and additive (Maxwell,
1961; Rao, 1964). Thus, in factor analysis, the matrix of covariances can be
explained by a smaller number of hypothetical variates or factors. Such an
approach is important in studies on biological evolution where the experimenter
is unlikely to have
a priori knowledge of the causal influences.  


Year: 1967
Volume: 27
Issue: 1
Article DOI: NA
Print ISSN: 0019-5200
Online ISSN: 0975-6906



Download PDF