multivariate analysis factor analysis censored regression

Multivariate Analysis

1. Overview of Multivariate
Analysis Methods

2. MULTIVARIATE ANALYSIS
These techniques are important in marketing
research because most business problems are
multidimensional and can only be understood when
multivariate techniques are used.
statistical techniques used when there are multiple
measurements of each element/concept and the
variables are analyzed simultaneously.
17-2

3. Classification of Multivariate Methods
17-3

4. Summary of Multivariate Methods
17-4

5. DEPENDENCE VS INTERDEPENDENCE METHODS
Examples: multiple
regression analysis,
discriminant analysis,
ANOVA and MANOVA
Dependence –multivariate
techniques appropriate when
one or more of the variables
can be identified as
dependent variables and the
remaining as independent
variables.
17-5
Examples: factor analysis,
cluster analysis, and
multidimensional scaling.
Interdependence –multivariate
statistical techniques in which a
set of interdependent
relationships is examined –The
goal is grouping variables in
some way.

6. FACTOR ANALYSIS
Purpose –to simplify the data.
Dependent and independent variables are analyzed
separately, not together.
. . . used to summarize information contained in a large
number of variables into a smaller number of subsets or
factors.
All variables being examined are analyzed together –to
identify underlying factors.
17-6

7. FACTOR ANALYSIS PROCESS
Steps
Examine factor loadings &
percentage of variance
Interpret & name factors
Decide on number of factors
17-7

8. Factor loadings are calculated between all
factors and each of the original variables.
These are the starting point for interpreting
factor analysis.
Factor Loadings are correlations between the
variables and the new composite factor.
They measure the importance of each
variable relative to each composite factor.
Like correlations, factor loadings range
from +1.0 to –1.0
17-8

9. CLUSTER ANALYSIS
classifies or segments objects into groups that
are similar within groups and as different as
possible across groups.
classifies objects into relatively homogeneous
groups based on the set of variables analyzed.
identifies natural groupings or
segments among many variables,
does NOT include a dependent variable.
17-9

10. 17-10
CLUSTER ANALYSIS

11. 17-11
SPSS DIALOG BOX FOR CLUSTER ANALYSIS

12. Coefficients
17-12
CLUSTER ANALYSIS COEFFICIENTS

13. New
cluster
variable
17-13
NEW CLUSTER VARIABLE

14. DISCRIMINANT ANALYSIS
Dependent variable –nonmetric or
categorical (nominal or ordinal).
It’s a dependence technique used for
predicting group membership on the basis of
two or more independent variables.
Independent variables –metric (interval or
ratio), but non-metric (nominal) dummy
variables are possible.
17-14

15. 15
DISCRIMINANT ANALYSIS
Characteristics
Discriminant function –a linear
combination of independent variables
that bests discriminates between the
dependent variable groups.
Develops a linear combination of
independent variables and uses it to
predict group membership.
Predicts categorical dependent variable
based on group differences using a
combination of independent variables.
17-15

16. DISCRIMINANT ANALYSIS
Multipliers of
variables in the
discriminant function
when variables are in
the original units of
measurement.
Estimates of the
discriminatory power
of a particular
independent
variable.
Discriminant
Function
Coefficients
17-16

17. DISCRIMINANT ANALYSIS
.
.
Classification (Prediction) Matrix –
shows whether the estimated discriminant
function is a good predictor.
Shows the number of correctly and
incorrectly classified cases .
The prediction is referred
to as the hit ratio.
17-17

18. 17-18
DISCRIMINANT ANALYSIS
SCATTER PLOT

19. 17-19
SPSS DIALOG BOX FOR
DISCRIMINANT ANALYSIS

20. 17-20
SPSSDISCRIMINANT
ANALYSIS OUTPUT

21. 17-21
SPSSDISCRIMINANT ANALYSIS
OUTPUT CONTINUED

22. Sample Conjoint Survey Profiles
17-22

23. Importance Calculations for
Restaurant Data
17-23

24. Conjoint Part-Worth Estimates for
Restaurant Survey
17-24