Hi, I'm currently analysing data for my quantitative dissertation and I'm struggling with how to do it. My data has come from 3 questionnaires with 148 questions in total and it is at an ordinal level.
I have added sections on to my data set that is each participants total score for each questionnaire, rather than the individual scores for all 148 questions. Would I use those total scores to conduct the spearmans correlation?
I hope this makes sense and thank you in advance for any help :)
Hello Leonie,
Yes, if you are convinced that the total score for each questionnaire is a meaningful, unitary variable or representation of a construct and your research question/s indicate that these are the desired variates, you should evaluate the correlations among the three measure totals.
If you're not convinced of this, then you might think about either a logical or rational basis for combining scores within a questionnaire, or consider factoring the set to have some support for your decision to use either subscores or total scores.
However, evaluating correlations among 148 individual items is probably not worth your time or energy (that's nearly 11,000 correlation coefficients to examine!), unless your research question/s specifically indicate that doing so is important to your research aims.
Good luck with your work.
Leonie,
To add to David's excellent answer above, if I were a member of your dissertation committee, I would not take your word for it that the total score from your questionnaire is meaningful. I would want empirical evidence. There are several ways of getting that. The simplest would probably be to run a factor analysis to demonstrate that there is only one factor. This would also produce an estimate of reliability which I would also require. An additional advantage of this approach is that it would tell you the number of factors your questionnaire is measuring. It would certainly be less than 148 and greatly reduce the number of correlations you would need to compute.
Good luck,
J. McLean
so easy using SPSS, but make sure your test variale is ordinal data..pm me..
- Badges
- Science topic