During a sustained attention task, participants were occasionally asked what they were thinking and to press a number indicating their current state of focus. Each instance asked the same question ("characterise your current focus") and there were 5 possible responses (1 = on task, 2 = distracted, 3 = nonalert, etc). I want to explore whether there's a difference between two conditions on the numbers of response for each category. I thought it'd be OK to use mixed models with the following basic model, with n being the counts of responses:
(n ~ condition * probe.type + ( 1 | subject)...)
Are mixed models OK for this, or should I be using something like a Poisson model?
mlogitpackage. $\endgroup$