hi i recently aksed about ghq28 - below a summary of replies
regards
ann
1.I don't believe that anyone should use a test, survey or whatever before they
understand what the test is supposed to measure. The scoring method should
be known beforehand.
If the user has porposed certain hypthesis then possibly the test/survey will
provide the necessary information but this should be done before
administering the test/survey.
Just a thought
2.I'm getting into this area too. A useful reference is
"Health Measurement Scales, A Practical Guide to their Development and Use" by David L. Streiner and Geoffery R Norman.
I'd be grateful if you would copy any replies that you receive.
3.You might find some thing on either of these websites www.outcomes-trust.org
or www.atsqol.org. They are unusually good starting places.
Normally with these questionnaires, you sum the individual compoments and
then scale them as a percentage of the maximum total score (adjusting for
missing components).
After that if you have sufficient numbers of responders than an analysis of
variance type approach is sufficiently robust to be used, otherwise a
non-parametric approach is advicable. The dimensions are usual designed to
be uncorrelated too.
4.Whether you will calculate the overall score or the four sub scores will
depend on the objectives of your study. I mean if sub-group analysis is
necessary, then it is better. However, in case of overall score, the
internall consistency (Chron back Alpha) will be ambiguous. I helped one
student who did her MSc in Psychology using GHQ 12.
5.it can be analysed in total - and this is the way to use it as a
screen for psychiatric caesness - or the sacles can be totalled
separately
If you but a manula from nelson it explains - better to do this as
there are also two or more different scoring sytems
the users need to be clear what the research question was!
thanks to all
regards
ann
----------------------
ann e fitzmaurice
medical statistician
dugald baird centre
department of obs and gynae
university of aberdeen
[log in to unmask]
tel 01224 553876 (direct)
fax 01224 404925
|