CRoNoS MDA 2018: Start Registration
View Submission - CRONOSMDA2018
A0151
Title: Latent variable models and pairwise likelihood framework: old and new developments Authors:  Irini Moustaki - London School of Economics (United Kingdom) [presenting]
Abstract: Latent variable models and factor models are frequently employed in social sciences where the main interest lies in measuring and relating unobserved constructs, such as emotions, attitudes, beliefs and behavior. The models become complex with the increase of the number of observed variables and the number of factors. The aim is to discuss some developments of applying a pairwise likelihood framework for the purpose of estimating the parameters of latent variable models, but also for model testing and treatment of missing values. Pairwise likelihood is a special case of composite likelihood methods that uses lower order conditional or marginal log-likelihoods instead of the full log-likelihood. Simulated and real examples will be used to illustrate the methods presented.