Dr Daniel Phipps
Associate Professor Sama Low-Choy
A series of workshops and seminars on Structural Equasion Modelling.
SEM1 - What is Structural Equation Modelling
Tuesday 13th September | 10:30 am - 12:00 pm
This seminar is designed to help participants understand what a structural equation model is, and what it can be used for. The seminar will incorporate an active learning approach, with several group-based activities involving discussion and short presentations. The seminar will end with an activity, so that those wishing to stay can put their new learnings into practice. For absolute beginners in SEM. It is recommended attending this seminar if you wish to attend the following series of SEM workshops.
SEM2 -The Meaning of Structural Equation Modelling
Tuesday 20 September | 10:30 am - 12:00 pm
This session starts with a non-mathematical introduction to the underlying logic: revisiting variables and dependence, association vs causality, assumptions vs evidence of causality, and issues such as falsification or accumulation of evidence, counterfactuals and experiments. Then we distil the basic building blocks of Structural Equation Modelling (SEM), and their representation both in pictures and in the simple math of a regression model, avoiding the math of computation. To consolidate these ideas, we interpret existing and news SEMs in terms of basic logics, graphics and mathematics.
There will be 3 parts to this session:
Logical meaning of SEM;
Representing SEM using pictures;
Interpreting the basic maths in the model (not the computing) of SEM.
The seminar will end with an activity, so that those wishing to stay can put their new learnings into practice.
For absolute beginners. Attendance of the What is SEM seminar is recommended.
SEM3 - Latent Constructs in Structural Equation Modelling (SEM)
Wednesday 28 September | 9:00 am - 3:00 pm
In this workshop we discuss how latent constructs in SEM are used to create an indirect measurement of a concept or theme via multiple items. We show how validity of constructs can be evaluated by confirmatory factor analysis (CFA). We will also discuss how CFA differs from alternative methods of forming a construct such as a simple average of items or an exploratory factor analysis (EFA). To illustrate these ideas we also run and interpret a CFA in R.
This workshop will consist of 3 sessions:
A basic refresher in setting up an R workspace
The Theory Behind Latent Variables for representing constructs
Running a CFA in R, and Interpreting
SEM4 - Structural Equation Modelling for Basic Pathway Models
Monday 5 October | 9:00 am - 3:00 pm
This workshop follows on from "Latent Constructs in SEM". In this workshop we use a pathway analysis, implemented in R, to relate several variables as well as constructs created via CFA. This includes discussion of what these pathways mean, how they are interpreted, and how to estimate mediation and total effects.
At the conclusion of the workshop we will discuss how one might apply SEM to their own research.
This workshop will consist of 4 parts:
Creating a simple pathways analysis in R
Calculating Total and Indirect Effects in R
An Introduction to Interpreting Pathway analysis in SEM;
Discussing SEM and how it applies to your research
SEM5 - Mediation and Model-Checking in SEM
Wednesday 12 October | 9:00 am - 3:00 pm
This workshop follows on from "SEM for Basic Pathway Models". After running a simple pathway analysis using SEM, we provide practical advice on how to detect inadequate model performance via diagnostics, and then suggest some remedies. This includes examining the possibility of extending an SEM to include moderation.
This workshop has 4 parts:
How to Address Poor Model Fit (Theory and Computing);
Running an SEM with Moderating Variables in R;
Interpreting SEMs (Part 2);
Discussing SEM and how it applies to your research.