Mplus (Statistical Package): Basics

Mplus (www.statmodel.com) is a syntax based commercial statistical software package for estimating models containing latent,or unobserved, variables. It has grown out of the tradition in statistical modelling known as SEM - Structural Equation Models, sometimes called LISREL models (this stands for linear structural relations modelling). However Mplus is much more comprehensive that this and can handle models for multivariate analysis with both continuous and categorical observed and latent variables e.g. probit and logistic regression models and their SEM equivalents. It also has capabilities for estimating latent classification and trajectory models, multilevel models, conducting Monte Carlo simulations, combining multiply imputed data sets, and handling complex data from survey samples. This course is intended to introduce new users to the basic code required to run half a dozen simple examples.

• An interest in and motivation to learn statistical modelling using latent variables e.g. latent class analysis (LCA)
• latent trait analysis (LTA), latent factor analysis, latent growth analysis, and mixture regression modelling

• How to get data into Mplus
• How to run the programme and an overview of its functionality and structure
• How to conduct basic analyses in the demo (free) and network version

Presentation/demonstrations with practical.

• Mplus can estimate a wide range of complex statistical models. Only a small number (about 6) will be covered in this session.
• A huge amount of relevant material and the free version of the software can be downloaded from http://www.statmodel.com (please peruse this material before coming).

1 full day (morning and afternoon).