I would need some help in analyzing a survey that I have been asked to run in my workplace. The survey is run in successive waves on the same group of people. I know how big is the universe and I know the number of respondents, but I don't have their identity (no paired test possible). The number of respondents can be different from wave to wave even if the universe is the same.
The survey is made of four questions that are answered via a Likert scale (from 1=strongly disagree to 4=strongly agree).
I would like to be able to test if the results of, i.e. wave 1 and wave 2, are significantly different and eventually which are the questions that are driving the change. Thanks.