First, specify whether the data produced by questionnaire is continous or categorical. Chi-square is better option for categorical data. If continuous, then you need to specify whether your data is parametric or non-parametric. T-test assumes normality while Mann-Whitney U test substitutes it in the case of non-parametric data. If you have five domains, you can go for ANOVA for inter-domain comparsions for parametric data and Kruskal-Wallis test for non-paramteric data.

Regards

Harman

t test can be used for testing the significance of difference between each pair.

Analysis of variance with LSD test can also be applicable.

Another critical issue to consider is multipe testing. To avoid alpha error accumulation, you could go, e.g., with a MANOVA for comparisons between gender classes (five domains as five DVs and gender als IV). Post-hoc, you have to perform five ANOVAs or t-tests (they will show the same results).

Another idea is to perform a 5 (domains) x 2 (gender) ANOVA with repeated measurements on factor 1. There you could use an appropriate post-hoc test to test for simple main effects. I recommend Dunn's multiple comparison test or Student-Newman-Keuls, since they are more liberal, when many pairwise comparisons are possible.

I agree with Brian Schwartz .

It is possible to suitably apply MANOVA