MULTIVARIATE ANALYSIS
What is multivariate analysis?
Comprises a set of techniques dedicated to the analysis
of data sets with more than one variable
Usefulness of multivariate analysis
Business research usually involve situations that require
the study of more than two variables. Thus study of univariate or bivariate statistical analysis is not enough. What is required to handle such real life complex situation is mutivariate techniques. this leads us to the discussion of Multivariate analysis
Types of multivariate techniques
Dependency techniques- techniques that deal with
problems involving one or more dependent variables while the remaining variables are independent Interdependency techniques- techniques that deal with problems involving more than two variables where the variables are not segregated as dependent variables & independent variables. These aim at analyzing interrelationships among variables
Dependency Techniques
Explains or predicts one or more dependent variables
based on two or more independent variables Defines relationship between one dependent variable and many independent variables Example: to what extent is the demand for a particular product explained by factors like, price of the product, income of the prospective consumer, advertisement of the product etc.
Multivariate techniques
Multivariate techniques
Are some variables dependent on others
Yes Dependency Methods How many dependent vars one Is data metric? Is data metric? Factor analysis Cluster analysis
No Interdependency Methods Is data metric?
many
Yes
No Non Metric Multi dimensional scaling
Yes Multiple Regression
No
Yes
No MANOVA
Metric Multidimensional scaling
Discriminant Analysis
Canonical Correlation Analysis
Different Dependency Techniques
Multiple Regression
Discriminant analysis
MANOVA Cannonical Correlation Analysis
Criteria for selecting a techniques
Number of dependent variables
Measurement scale used
Criteria for selecting a techniquecontd..
Multiple Regression- if one dependent variable &
measurement scale used is interval or ratio scale Discriminant analysis - if one dependent variable & measurement scale used is non-metric i.e. ordinal or nominal MANOVA if more than one and scale is metric Cannonical Correlation Analysis- if more than one and scale is non-metric
10
Interdependency techniques
Applied in situations where no distinction is made
between independent & dependent variables Interdependent relationships between variables are examined
11
Different Interdependency techniques
Factor analysis
Cluster analysis
Metric multidimensional analysis Non Metric multidimensional analysis
12
Criteria for selecting a techniques
Measurement scale used
13
Criteria for selecting a techniquescontd..
Factor analysis if measurement scale used is metric
Cluster analysis - if measurement scale used is metric
Metric multidimensional analysis- if measurement scale
used is metric Non Metric multidimensional analysis - if measurement scale used is non-metric
14
