paired t test for categorical variable

#1
Hi,

I'm working with some survey data. There are baseline and followup data. I'm trying to see if there is a significant difference between the baseline and follow up data. I'm able to use paried t test to compare continuous variables. However, I have some categorical variables so I'm not sure which test to use. I was reading about McNemar test and I think that I can use it. Is this the correc test?

Also, not every individual has base line data or follow up data. Can I still use McNemar test?

Thanks.
 

bukharin

RoboStataRaptor
#2
Yes, you can use McNemar's test as long as the outcome is binary. It only uses cases where the outcome changed between the two time points, so individuals with missing data would be excluded (as are individuals where the outcome didn't change).
 
#3
Thank you for your help! What about questions where the outcome is not dichtomous. For example, some of my questions have answers like "sometimes, often, always, never"?