Author Topic: analyzing DIF with multidimensional models  (Read 386 times)

gfulmer

  • Guest
analyzing DIF with multidimensional models
« on: October 25, 2012, 07:44:47 AM »
I am trying to analyze DIF for items in a multidimensional model.  In the Manual, it suggests adding an option for the estimate command to be stderr=full.  However, this produces an error message that full standard errors cannot be estimated for the multidimensional model.

Should I examine DIF for each dimension separately (having already tested the multidimensional model overall), or is there some other approach suggested?

many thanks,
gavin

Eveline Gebhardt

  • Administrator
  • Full Member
  • *****
  • Posts: 103
    • View Profile
    • Email
Re: analyzing DIF with multidimensional models
« Reply #1 on: October 30, 2012, 04:38:35 AM »
Dear Gavin,

Yes, I think it is best to estimate DIF for each dimension separately with the full standard error.

Eveline