# Likert item statistical evaluation

Lets say I have a group of student and I assign them to both conditions - condition 1: View videos from lecture A and condition 2: View videos from lecture B.

After each students view the videos from both lecture, the student have to complete survey for each lecture which is a 5 point Likert item (a single question)

Q: I learn effectively from the lecture (5 points Likert item: Strongly agree, agree, neutral, disagree, strongly disagree)

The question is if I want to determine if there is significant difference between both lecture (e.g. if lecture A is better than B) from the likert response should I be using a Wilcoxon signed-rank test ? Since this is a within group design (i.e. the same group of student is expose to BOTH lecture) and the response from the Likert item is an ordinal scale?

• Technically, I think the data have to be assumed to be interval in nature, because the first step in the test is to subtract the observations for B from those for A. If the measurements were truly ordinal, you wouldn't be able to perform the subtraction unless you first imposed an interval-level numbering scheme. Jun 12, 2018 at 18:30