Before deciding which type of statistical analysis you will use,
there are two other important considerations:
Planning![top of page](img/top.gif)
The analysis of survey data should be considered before the survey
is done, not after. By building into the design of a survey the
type of data analysis that will be done, much time and effort may
be saved in the collection as well as analysis of data.
Databases![top of page](img/top.gif)
Before any type of analysis is performed, data must be stored
in a digital format. Many different types of software are available
to do this, with personal preference being a big factor. The way
in which data is entered and coded has a major influence on the
types of statistical analysis that can be done. For instance, some
statistical programs may not accept letters for identifying different
classes, therefore numerical codes must be used instead.
Think about the structure of the database. Do you need all your
information in the same file? Can different files be linked somehow?
Relation databases (e.g. Paradox, Access) offer many advantages
in storing survey data, as complex queries and searches for different
configurations of data can be made.
|