Introduction to Structural Equation Modelling (SEM) in R

PrerequisitesNew

Structural equation modelling in the statistical software R, with Dr Jan Štochl, Dr Kate Xu, and Dr Adam P. Wagner (all three are members of the Department of Psychiatry at the University of Cambridge).

R is freely available software, with versions for all major platforms (Microsoft Windows, Linux, Mac, etc.), and is growing in popularity in academia and beyond for carrying out statistical analysis and modelling. This course aims to give an introduction on how to use R for structural equation modelling (SEM), a range of techniques that are common in psychology and psychiatry, and that are being used ever more widely. We will focus on the package lavvan, which is becoming the most sophisticated package for carrying out this type of analysis within R.

This course will consist of four sections. Firstly, participants will be briefly introduced to the basic concepts of factor analysis and structural equation modelling (from 09:30 till approximately 11.30). Secondly, the audience will be shown using R through Rstudio (http://www.rstudio.com/), a free and open source integrated development environment for R. Thirdly, we will demonstrate how SEM can be carried out in lavaan. Participants will actively work through examples of using lavaan to get first hand user experience of doing factor analysis and other, more complicated, SEM models (for both continuous and categorical variables). Finally, we demonstrate the application of lavaan to measurement invariance analysis – a technique often applied in SEM analysis when multiple groups or longitudinal data are involved.

Those that are familiar with the concepts of factor analysis and structural equation modelling can join the course from 11.30; however, the theory covered in the first two hours will be assumed knowledge for the later sections.

Please ensure to read the Notes section about the use of laptops on the course!

All postgraduates, Research Assistants and Research Associates, who want to learn how to carry out SEM in R using the lavaan package. Participants will be expected to have, at minimum, some experience with factor analysis (FA) or structural equation modelling (SEM) for the afternoon sessions. Those with no such experience should ensure to attend the first morning session (09:30-11.30). Though not essential, participants will benefit from having experience with using R.

• Participants will be expected to have, at minimum, some experience with factor analysis (FA) or structural equation modelling (SEM); where participants do not, they should ensure to attend the first morning session where the basic concepts of factor analysis and structural equation modelling will be covered.
• Though not essential, participants will benefit from having experience with using R.

Number of sessions: 1

• Basic concepts and theory of factor analysis (FA) and structural equation modelling (SEM)
• Using R with RStudio
• Structural equation modelling in R
• Measurement invariance analysis in R

By the end of the course, participants should have a basic understanding of using the lavaan package, within the statistical software R, to carry out factor analysis (FA) and structural equation modelling (SEM).

To introduce people to carrying out structural equation modelling (SEM) within the statistical software R, using the lavaan package.

Presentations and interactive sessions where participants will get first hand experience of using the software.

Please email Adam Wagner (apw40@medschl.cam.ac.uk) indicating if you wish to use your own laptop or one of the 12 laptops available for use at the workshop that have been pre-loaded with R and RStudio. After these laptops have been allocated, participants will either need to share, or bring their own. These laptops will be allocated to the first individuals to express an interest in using them. Those participants bringing their own laptops should:

• Ensure to have installed the latest version of R from http://cran.r-project.org/
• Download and install the lavaan package within R
• Install RStudio from http://www.rstudio.com/ide/download/

The course convenors have little experience with Macs and so will be less able to support their use.

One day session of 7.5 hours (there will be a break for lunch of around 30 minutes and some shorter comfort breaks).

Once per term

• An Introduction to R: Software For Statistical Analysis
• Bioinformatics: R object-oriented programming and package development
• Bioinformatics: An Introduction to Solving Biological Problems with R