0
$\begingroup$

I have had members of staff from two different departments complete surveys on their opinions of waiting times:

Example question: Do you think waiting times have: increased, decreased, stayed the same?

Results: Department A: 30% staff felt waiting times had decreased. Department B: 50% staff felt waiting times had decreased.

I want to compare the opinions of department A to department B. My confidence intervals overlap so I thought I could do a t-test to prove that there is a statistical difference between the data sets. Would this be a suitable test to do? If yes, what type of t-test?

Or would a Chi-squared test be more appropriate? As my data is categorial e.g. agree or disagree and my quantative values are proportions of each category.

Thanks!

$\endgroup$
2
$\begingroup$

The most straightforward method would be to use a chi-square test. You have two variables for each individual: department (A and B) and response (increase, decreased, same). You want to see if they are related, which is to say if the pattern of responses in department A is different from the pattern of responses in department B. A basic test for the relationship between two categorical variables is a chi-square test. A significant test statistic (i.e., a small p-value) is evidence against the lack of relationship, i.e., evidence in favor of a relationship or a difference between the departments. A chi-square test is an omnibus test, though; knowing that there is a difference doesn't tell you in which categories the two departments differ. Follow-up tests can be used to answer this question.

There are more complicated analyses you can perform, like multinomial or ordinal logistic regression, but those are quite advanced and probably overkill for your task, though they will give you more complete answers to your questions and can be used in flexible ways (e.g., to see whether characteristics of the individual affect their perceptions beyond just what department they're in, or if the effect of department is actually explained by the composition of the departments).

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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