I have a 39 items questionnaire with 9 demographic questions.Since my items variables are scalar,ordinal(transformed into scalar) and nominal(multiple answear type), can I use in my PhD thesis, for differences between groups,Mann-Whitney test for scalar variables and Chi-Square for multinominal items?Would this be wrong?
Thank you!
Please also mention your sample size. Sure you can use Mann-Whitney test , and Chi-square.
Dear Hom, thank you for your prompt answer. My sample size is 303 respondents. From everything I have documented before, I didn't found anything similar. I know now, that I was wrong when I created the questionnaire, but now is to late to make changes.
With such a large set of questions I have found it is helpful if you reduce the size of the response field by eliminating the questions using chi-squares binary response test. This helps to run regressions with the systems classified and systems integrated data. This may also take care of your design problem. How you pair depends on your problem.
S.K.Mallick
- Badges
- Science topic
- Similar topics
- Philosophy
- Esthetics
- Interpretation