Likert data, non normal: Chi Square or T-test

[fskywalker]

Hi:

I am working on my dissertation in performing the analysis of the data that I capture through a questionaire, which used a likert scale (1-5). The data captured is not normal (performed a Shapiro test), so know that the tests to be performed needs to be non parametric. In order to fulfill the requirements of some of my hypothesis, I need to check if there is a statistical difference between the data captured for 2 different sub-groups. Should I use Chi-square, Kruskal Wallis or a t-test?

Thanks,

Francisco

[bugman]

Hi, Likert scale analysis has recently been disscussed in the forum. As a starting opint you might wnat to have a look at som eold threads and you might find your answer.

Phil