Royal Statistical Society
29 and 30 September 2014
Presented by Irini Moustaki, Jouni Kuha and Sally Stares
Latent variable models are a broad family of models that can be used to capture abstract concepts by means of multiple (continuous and/or categorical) indicators. They are often best known in the form of factor analysis and structural equation models. Day 1 introduces participants to latent trait models (continuous latent variables) and latent class models (discrete latent variables) and Day 2 to multiple group latent trait and latent class analyses. The course provides training in the use of Mplus to carry out the analyses.
By the end of day 1, participants will be able to use Mplus and R routines to fit latent trait and latent class models to sing le groups; read the input from Mplus and produce measures of fit, compare models through model selection criteria; produce plots and interpret results.
By the end of day 2, participants will be able to use Mplus and the R routines developed to fit latent trait and latent class models to multiple groups; compare models; test for measurement equivalence; set model constraints; compare the distributions of latent variables between different groups (e.g. countries); make valid comparisons and interpret results.
Aimed at: Researchers interested in the use of cutting edge statistical methodology for analysing survey data from cross-sectional and for cross-national surveys.