I collected data from an instrument based on a likert scale questionaire. Each participant answered the instrument two times and I want to define what kind of test can I use to analyze my data if I want to compare the mean score variations for each question. Can I use a t test or a Mann-Whitney test?
I scale usually means that you have combined several Likert-scored items. If that is the case, then you can use a paired comparisons t-Test.
Thank you David, but what if I want to compare each item separately? Can I keep the t test? What if the variance is not enough to assume normality?
You can use non-parametric test for ordinal data such as Mann-Whitney U test or Wilcoxon sign rank test for paired samples if normality can not be assumed. Likert scale mean you need to sum up the scores. however, you can test individual items.
There have been so many questions here about Likert-scored items that I have collected a set of resources on this topic:
https://www.researchgate.net/post/How_can_one_help_people_create_simple_scales_from_Likert-scored_items
- Badges
- Science topic