Q: Should I do a univariate or multivariate ANOVA if I have one independent variable?

Detailed Question -

Hi, so I'm currently working on a study that has one independent variable with three stages and five DVs. I know this usually calls for a MANOVA but the reason I'm confused is that my survey only has one question for each of the DVs, i.e. if I'm measuring comfort then there's only one question in the survey regarding that; for example: "Did you feel comfortable during the interactions?" Is that enough or do I need more questions? And what omnibus and post-hoc tests can I use for such a situation?

Write an answer...
1 Answer to this question
Answer:

I think univariate ANOVA will be applicable in this case since there is only one independent variable. MANOVA is applicable only if there are multiple independent variables.

 

Regarding the number of questions for each dependent variable (DV), you should be best able to understand if one question is sufficient to generate the kind of data that you need to support your research hypothesis. It might be a good idea to go over the questionnaire once again and think whether some of the DVs require more than one question.

 

It’s difficult to recommend the post hoc tests but one option that you can consider applying is Tukey’s HSD which is generally used for oneway ANOVA.  You can also consult your supervisor or a professional statistician about this if you are not sure.

 

Do go through these articles before you conduct your statistical analysis. You might find them useful: