|
Home
Key Research Areas
Research Methods
Sophisticated Analysis
Articles
Industries
QSA Principals
Contact Us
|
|
Sophisticated Analysis
QSA professionals know how to use all of the major statistical methods. More
important – we know when to use them and when not to use them. But what
really sets QSA apart is our ability to combine statistical results with our
professional judgment and industry knowledge to produce actionable recommendations.
Multivariate Analysis
•
Cluster Analysis
– Cluster analysis allows us to classify respondents into relatively
homogeneous groups and, thus, is used in segmentation research. NeuroClusterSM, QSA’s version of cluster
analysis, is based on neural network technology and provides higher
resolution and more stable results than conventional cluster analysis.
• Multiple Regression – Multiple
regression assesses the independent impact of each of a number of predictor
variables on a criterion variable. Thus, it is often used to identify the
drivers of corporate image, customer satisfaction, voting behavior and
purchase behavior. QSA is one of the few research firms that uses multi-stage
multiple regression. This powerful technique, which we call Communication ArchitectureSM,
shows both direct and indirect drivers, thus, providing a complete model of
the influence process.
Conjoint Analysis
Conjoint Analysis is an experimental design that simulates the
decision-making process, forcing respondents to trade off valued features
against each other. Choices are analyzed statistically to derive utility
scores for each feature. QSA has developed ways to provide reliable estimates
of the dollar value of features and to calculate a “bang for the buck” ratio
for each feature.
|
|
|
