I have a list of responses for a Likert Scale and I want to test if the participants' characteristics somehow influence their responses , if there is any tendency, for example:
Professors tend to Completely agree with Sentence A, while Teachers tend to partially disagree.
Which would be the appropriate statistical test for it?
Many thanks
You say "relation", which means "association" or "correlation", but based on your example, I think you want to know if there is a "difference". If that is the case and there are only two levels of "participant characterization", then you should use a "Mann-Whitney" test because a Likert scale is ordinal data. If you have more than two levels of "participant characterization", then you should use a Kruskal-Wallis test.
An multinomial logit (Mlogit) model would be most appropriate. It essentials generates a partially differentiated model by response level, accounting for the necessity of just one response and recognizing that the distance between scale points may not be linear.
I beg to difer with Dr. Miller. I think a most appropriate approach would be an ordinal logit or probit model, as Likert scale variables are ordered.
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running