I'm fitting Rasch models with the mirt package. Upon measuring item-wise $Q_1$ using mirt::itemfit(model, fit_stats="X2"), I noticed that items have different degrees of freedom. I thought the degrees of freedom would, in all cases, equal the number of trait bins minus the number of fit item parameters ($1$ for Rasch models). Where do the differing degrees of freedom between items come from?

  • $\begingroup$ There's more than one way to make a bin.... did you read the group.bins argument in the package documentation? $\endgroup$ – philchalmers Nov 8 '19 at 2:04
  • $\begingroup$ Even if I set group.bins=10, the items will have varying degrees of freedom even though they all have the same number of parameters, that's what I don't understand (Item 1 has 3 df, item 2 has 0, item 3 has 2, ...). I should also mention I have a lot of missing data. $\endgroup$ – Bananin Nov 8 '19 at 16:43

Your Answer

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

Browse other questions tagged or ask your own question.