As a statistician, married to a university research professor, I have seen
this problem in research students frequently.
1. I strongly recommend NOT Allowing 'multiple' answers to the same
question.
2. If you are already there, then create a unique code for the multiple
response and apply that code universally to all the records.
3. Start at the End and work backwards. (End --> Data
Interpertation -->Statistical Package --> InPut Format (spreadsheet) -->
InputFormat (Database) -->InputFormat (paper Forms)
4. Buld a 'dummy set' of data to work through the process, just to find out
where the snags are (you will be frequently surprised that you need to Count
where you had thought you wanted to average, etc.)
The 'real' question is: What are your really trying to do?
1. Build an electronic database so people can fill out a questionaire at a
computer and will have seveal thousand questionaires. (msAccess problem)
2. Enter data from A paper questionaire then analyize the data (excel
problem)
3. Build an electronic form you can email to people to complete and return
(XML Problem, msInfoPath, AdobeForms, msWord)
4. Build an electronic form posted on a web page for people to fillout
(msAccess with webPages (asp etc.), or Java, or ......)
In any case you have
ResearchSubject -->Questionaire -->Question --> Response
So your final data table for analysis should be three or four fields wide
(four if you have more than one questionaire to analize) maybe a couple more
if you need to capture date or other stuff.
Subject Questionaire Question Response
1005 1 1 0
1005 1 2 5
1005 1 3 2
------------------------------etc.------------------
1006 1 1 1
1006 1 2 3
1006 1 3 3
-----------------------------etc.-------------------
etc.
Lots of luck
Ed Warren