Current Research

MCMC estimation of parameters of mixture models using Guttman scale data

I am working under the supervision of Professor Peter Bentler.

In Guttman scaling, participants are asked (or tested on) a series of questions of increasing "difficulty" in order to measure some underlying or latent trait. Similar to ordinal scales, data from Guttman scales give a general idea of where a data point may lie, but not its exact location.

Using Guttman data, my research focuses on trying to recover the parameters of the distribution of the underlying trait. The underlying trait could be defined by a mixture model and my research will try to recover the parameters of the mixture. For the moment I am using MCMC methods with synthetic data.