Exploratory Restricted Latent Class Models with Monotonicity Requirements under Pòlya–gamma Data Augmentation

psychometrics
journal
Authors

Balamuta, J.J.

Culpepper, S.A.

Published

January 1, 2022

Doi

Summary

This paper develops Bayesian methods for exploring latent structure in educational and psychological assessment data, relaxing overly restrictive monotonicity conditions used in prior work and adding a computationally efficient logit-link formulation via Pòlya-gamma data augmentation. Four new model formulations with different link functions and sparsity-inducing priors are presented, validated through simulation, and applied to the Standard Progressive Matrices.

My contributions

(Your role on this paper, to be filled in by operator.)