Is it possible to do IRT distractor analysis using R? I've been using the mirt package to analyse a few response matrixes I have through IRT. However, every time I need to perform distractor analysis, I end up having to go back to Classical Test Theory package CTT. Is it possible to use mirt (or any other package, for that matter) to analyse distractors via IRT?
 A: Of course, and I have found this quite useful in the past. Simply fit something flexible that captures information from each category, like the nominal response model, to the items of interest, and inspect the response curves graphically or by looking at the scoring/ordering coefficients and intercepts. That will give you an indication of how the distractors are behaving, and actually whether you correctly scored items in the first place (sometimes, they aren't scored correctly, and this is a good method to detect that because the 'correct' category will not be consistently increasing as $\theta$ increases). 
Note that this has all the benefits of IRT as well (i.e., effective handling of missing data, computation of factor scores, etc) and works for multidimensional testing situations too by inspecting the respective scoring coefficients (the plots can be too high dimensional to create for some tests). See the mirt wiki, specifically the exercise files from the workshop in Zurich, for examples. 
