IRT Parameter Estimation using the EM Algorithm

Bradley A. Hanson

Date: October 28, 1998
Revised: September 27, 2000

Summary: These are notes to accompany a lecture I gave in a seminar at the University of Iowa taught by Mike Kolen. A detailed description is given of using the EM algorithm to compute maximum likelihood parameter estimates in IRT models for dichotomous items. An example of applying the procedures described is given using a generalized Guttman scale model. Brief descriptions of how to use the EM algorithm to compute Bayes modal parameter estimates and maximum likelihood parameter estimates for polytomous IRT models are given in the final section.

PDF Download paper in PDF format. Version 3.0 or later of Adobe Acrobat Reader (which is available for free) is needed to view this paper.

Links: A paper I wrote with David Woodruff provides more details in regard to parameter estimation in IRT models for polytomous items. I have written a paper that describes using the EM algorithm to compute maximum likelihood or Bayes modal estimates of a discrete latent variable distribution.


Return to Papers by Brad Hanson

Return to Brad Hanson's Home Page

Return to OpenIRT Home Page

Last updated: November 16, 2014.