Latent Class Analysis (LCA) is a branch of the more General Latent Variable Modelling approach. It is typically used to classify subjects (such as individuals or countries) in groups that represent underlying patterns from the data. In addition to this application LCA provides a flexible framework that can be used in a wide range of contexts: in longitudinal studies (e.g., mixture latent growth models, hidden Markov chains), in evaluation of data quality (e.g., extreme response style, cross-cultural equivalence), non-parametric multilevel models, joint modelling for dealing with missing data.
In this course you will receive an introduction to the essential topics of LCA such as: what is LCA, how to run models, how to choose between alternative models, how to classify observations, how to evaluate and predict classifications. You will also apply this knowledge to a number of more advanced models that look at the relationship between latent class variables and at longitudinal data.
The course covers:
Refresher of basic concepts in categorical analysis: (marginal) probability, odds ratios, logistic regression;
Basic concepts and assumptions of latent class analysis;
Introduction to Latent GOLD software;
Model fit evaluation: global, local and substantive evaluation;
Classification of cases;
Apply these concepts to a number of models looking at: predicting class membership, relationships between latent classes, hidden Markov chains.
By the end of the course participants will:
Know what is Latent Class Analysis;
Be able to estimate and interpret results from Latent Class Analysis;
Be able to choose between alternative Latent Class Models;
Understand latent class classification and how to predict it;
Be able to investigate the relationship between latent class variables.
Pre-requisites
Knowledge of basic categorical analysis: (marginal) probabilities, odds ratios, logistic regression.
Day 1 – introduction to LCA
Refresher of basic concepts in categorical analysis: (marginal) probability, odds ratios, logistic regression;
Basic concepts and assumptions of latent class analysis;
Introduction to Latent GOLD software;
Model fit evaluation: global, local and substantive evaluation;
Classification of observations.
Day 2 – applications of LCA
Predicting class membership,
Modelling multiple latent classes;
Looking at relationships between latent class variables;
Hidden Markov chains.
Each day will run from 10am – 4:30 (approx.)