I have 2 IVs( country and gender) and 4 DVs. Should I use both IVs at same time as fixed factors in MANOVA? or Should I conduct separate MANOVA for different IVs?
Dear Bhagat
If your IV's are correlated you must include them both in the Manova model, so check the scatter plot for them and the correlation coefficient then if there's a relationship between them you include them both simultaneously.
Regards
Dear Dr Rana,
Thank you very much for your kind reply. I have taken both IVs simultaneously but I am facing one bog problem. I am getting main effect significant difference between the countries but not for gender. So I am puzzled how to report this result? Will the reviewers be satisfied? i have attached two snapshots of my result.
Looking forward to your kind suggestions.
That's simply indicate that gender didn't affect variances in DV's, while country significantly affects the variances in DV's.
I aggree with Rana, this is a simple explanation. Additionally, you should have a look at the effect sizes of the MANOVA result (since the effect size η^2 for multivariate cases yields larger effect sizes than the univariate cases, for better comparison ηp2 of MANOVAs should be calculated as follows: ηp2 = 1 – Wilk’s Λ1/s with s = min(number of DVs, df effect), but see Tabachnick & Fidell, 2014), as well as for the ANOVA for the single DVs.
I noticed that the univariate effects for your DVs IC and SP showed only a p-value slightly under .05, indicating probably a very small effect size, since you have quite large samples. You should discuss if this effect is of any practical relevance for your research. The p-value alone does not tell you anything, since you will get everything "significant" with just a sufficient large sample size.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics