User participants experienced method A and then method B of a computer interface. Afterwards, we asked them to indicate which method they preferred regarding certain aspects. For example, for ease of use, which do you prefer, A or B, from 1 to 7 (Likert scale), where 1 = definitely method A, 4 = Neutral, and 7 = definitely method B.

Here are my questions:

  1. Since we are directly asking them to compare methods on this Likert scale and thus receive only 1 response, does it make sense to do a statistical analysis? I guess we cannot compare the participants' answers based on different methods because there is only 1 response.
  2. If the answer to #1 is no, then do we simply aggregate the data and see how many users prefer A vs. B?
  3. If the answer to #1 is yes, then do we statistically analyze if the responses vary from some "neutral" distribution? What's the best method/way to analyze this?

1 Answer 1


I believe this is an instance of having 0 independent variables and thus we should use the one-sample median test, a.k.a. the Wilcoxon signed rank test (see this table). This is what I ended up using.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.