Agriculture Reference
In-Depth Information
12
Multivariate Analysis
Analysis of information, resulting from different
research programs, particularly the statistical
procedures, may broadly be classified into
Multivariate analyses are more complicated,
as they take care of the system as a whole, having
a number of variables at a time, and consider
their interdependence, relationships, importance,
etc. But with the advent of computer technology
and statistical software, the use of multivariate
technique has become user-friendly and is
gaining momentum day by day. The area and
coverage of multivariate analysis is a huge one,
and it is not possible to include all these here. In
this topic, an attempt will be made to provide an
outline of some of the multivariate techniques
which can be used in different agriculture and
allied fields. The details of analytical/statistical
steps for calculation are also avoided; emphasis
will be provided to place different multivariate
statistical tools using different statistical soft-
ware to solve different problems. Useful
references are provided for interested readers.
univar-
iate analysis and multivariate analysis
.Inunivari-
ate analyses, we consider one variable at a time
contrary to the varied number of variables in mul-
tivariate analyses. The simplest case of multivari-
ate analysis is the bivariate analysis, in which two
variables are considered together. The variables
that we consider in agriculture, economics, anthro-
pology, sociology, psychology, management, etc.,
tend to move together, and as such multivariate
analysis is more useful. Univariate analysis throws
light on one character only, but to explain the
relationship, interdependence, and relative impor-
tance of different variables, multivariate analyses
would be more appropriate.
Let us take the example of analysis of
innovation index which is associated with a num-
ber of parameters like age, gender, education,
economic background, and society. These
components are not independent of each other;
rather these are correlated, interdependent to each
other; these have varied importance towards ulti-
mate innovation index. Univariate analysis can
throw light separately on each of the character,
but to analyze the system as a whole taking due
consideration of their interdependence, relation-
ships, importance, etc., multivariate analysis is
the more acceptable option. Several examples in
other disciplines like agriculture, business, eco-
nomics, management, and medical science can
also be put forward where multivariate analysis
can effectively be used.
12.1
Classification of Multivariate
Analysis
According to different authors, multivariate anal-
ysis may be classified into two broad groups: (a)
dependence method
and (b)
interdependence
method
. In the dependence method of analysis,
relationships of some dependent variables are
worked out with the independent variables, but
in the second method, interrelations among
themselves are considered. The examples of
first group of analysis are regression analysis,
multiple discriminant
analysis, multivariate
