I do not have experience with ordinal data, please help me out with my questions?

I have conducted a survey which consists of 25 Likert-items, these items are all statements for certain conditions that affect the learning of the participants (N=173). They were required to give a score for these items, and the score range was 1 to 9 (1=fully disagree, 5= neutral, 9=fully agree). The sampling distribution was not normally distributed so I am focusing on non-parametric tests.

I am interested in the question whether my socio-demographic variables do affect the outcomes for these 25 Likert-items differently, or rather I would like to test for this. I do not know in which way the socio-demographic variable could affect the outcome, only that there might be some correlation according to previous studies. I thought of performing a Kruskal-Wallis test but I am not sure whether it is the right test for my type of data. The outcome variable is not just one answer for each participant, it consists of 25 separate answers (25 columns with a score) for each participant (row). I have 5 categorical socio-demographic variables: age (3 groups), gender (2groups), education (3 groups), years of employment (3 groups), and profession (3 groups).

  1. Is it right to perform the Kruskal-Wallis test for these 25 outcomes and 5 socio-demographic variables or is there a more suitable test for it in SPSS?
  2. And what if more than one socio-demographic variable shows differences between groups for the same Likert-item, should I then check whether the socio-demographic variables are inter-related and which test would be suitable for this?

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.