Unidimensional Item Response Theory with
Estimation of the Latent Density
Carol M. Woods
Center for Applied Statistics and Psychology
Washington University in St. Louis
Friday, September 25, 2009, 12:30–1:30 pm
GEMS classroom, 3rd Floor in
Shriner's Building
Coffee, tea, and cookies will be provided
Abstract
This talk describes a program of research related to the distribution of the random latent variable in unidimensional item response theory (IRT), when the model is estimated using marginal maximum likelihood estimation. Often, the latent variable is presumed to follow a normal distribution, but this may not be tenable for all problems. Two key methods for fitting these models wherein the latent density is estimated simultaneously with the item parameters instead of presumed normal are described and evaluated.