|
ANALYSIS TECHNIQUES
Analysis techniques are continuously updated
through research with leading academic institutions. Development of new
tools and techniques is regularly monitored and evaluated. Selection of data
analysis procedures is guided by the detailed research model and hypotheses
that are developed prior to questionnaire design. Some of the techniques
include:
 Principal components and common factor analysis
Analyses
interrelationships among a large number of variables and condenses these into a
small number of factors
Multiple
regression
Predicts the changes in a single
dependent variable in response to
changes in the independent variables
Multiple
discriminate analysis
Used where the single
independent variable is dichotomous,
e.g. male-female
Multivariate
analysis of variance and covariance
Explores the relationship
between several categorical
independent variables and two or
more metric dependent variables
Conjoint
analysis
Assesses the importance of
attributes as well as levels of each
attribute while consumers evaluate
only a few product profiles, e.g.
evaluation of new products, services
or ideas
Canonical
correlation
Simultaneously correlates
several metric dependent variables
and several metric independent
variables; extension of multiple
regression analysis
Cluster
analysis
Classifies individuals or
objects into a small number of
mutually exclusive groups which have
been identified on the basis of
similarities
Multidimensional
scaling
Transforms consumer judgements
of similarity or preference into
distances represented in
multidimensional space
Correspondence
analysis
Provides a multivariate
representation of interdependence
for non-metric data that is not
possible with other methods, e.g.
brands that are perceived as similar
are plotted close to one another
Linear
probability models (logit analysis)
Combination of multiple
regression and multiple discriminant
analysis in which one or more
variables are used to predict a
single dependent variable
|
Structural
equation modeling
Estimates and models the
interrelationships between a set of
dependent and independent variables
using a structural model and a
measurement model
Data
mining
Quantifies relationships among
large amounts of information with
minimal pre-specification of the
nature of the relationships
Neural
networks
Flexibly performs both
relationship identification or data
reduction and structure analysis
Resampling
and bootstrapping
Eliminates the need for the
statistical assumptions of sampling
distributions (such as normality) by
resampling the original sample and
generating estimate of the sampling
distribution
Text
mining
Considers text as numerical data
and identifies frequencies and
interrelationships between words and
phrases and can generate concept
mapping of qualitative information
|
