The Data Analysis team has significant
experience in handling large, complex databases and performing
a range of statistical analysis as follows:
| · |
Methods of Regression
(explanation and prediction) |
| · |
Factor Analysis (dimensions
reduction) |
| · |
Cluster
Analysis (grouping by multiple characteristics) |
| · |
Conjoint
Analysis (preferences' modeling) |
| · |
Correspondence
Analysis & Multidimensional Scaling |
| · |
Discriminant
Analysis |
| · |
Perceptual
Maps |
The data processing and the statistical analysis are carried
out by using specialised software such as SPSS,
MS Office (Access).
Sampling Methods
The sampling methods used are as follows: stratification
by region and by location size, random selection of
cities and starting points within a city or a village
according to a strict rule. Households to be surveyed
are chosen according to the random route method (using
KISH method or left-right method).
Data Cleaning Process
The data cleaning process is a key step in our quality
assurance procedures. Data is inputted based on computerized
forms with electronic validations which eliminate logical
errors. All data inputted is randomly checked in a percentage
of 10%. Each data input operator is checked on every
project to identify any possible errors in the input.
|
![Mercury Research [logo]](../images/common/logo.gif){kind=logo}
