I asked a random group of people from an organisation a question. I then have a proportion of favorable responses (e.g. 80% of 100 participants gave favorable responses).

6 months later, I asked another random group of people from the same organisation, the same question. I then have a proportion of favorable responses (e.g. 85% of 120 participants gave favorable responses).

The question (e.g. the organisation was a good place to work in) was in likert scale: strongly disagree, disagree, neutral, agree, strongly agree. Favorable responses are number of responses to agree and strongly agree.

How do I know if the difference between the first time point and second time point on the same question is significantly different? What test can I use for it? I tied calculating z-score for 2 Population Proportions but am not too sure if it is the right stats to use.

  • 1
    $\begingroup$ @IWS the link was posted as an answer. I converted to a comment as the OP posted the same answer twice. $\endgroup$
    – Glen_b
    Commented Nov 4, 2017 at 10:19

1 Answer 1


Rather than simplifying your data by splitting it into favourable and not, keep it as scores on your 5 point likert scale.

Then you can do a repeated measures t test to determine whether there is a difference between the two question times.

Then when you interpret the results you can then answer the question how much more/less favourable were the scores? Rather than simply saying the proportion changed.

Edit: it's not clear from your question, but if the two sets are from different employees rather than asking the same employees twice, then you'd need to use an independent samples t test instead of repeated measures t test


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.