Structural Equation Modelling for Longitudinal Data
Ethan McCormick (University of Leiden)
27&28 March 2023
This event has already happened
Longitudinal models are a broad class of approaches for understanding change and stability over time. There are many different approaches for understanding repeated measures data, but individual researchers are often exposed to only one framework (at best) and may be unaware of the full range of options for longitudinal modeling. In this workshop, we will navigate these issues, focusing on the multi-faceted decision process needed for model-selection among various popular modeling frameworks, specific modeling decisions that are required for appropriate analysis and inference, and steps for best connecting theories of change with model implementation. Where they arise, we will also compare the relative strengths of various modeling frameworks, including mixed effect (multilevel and generalized additive) and structural equation (latent curve and latent change score) models for fitting the various model options. The workshop will have two parts, one focused on developing a theoretical understanding of the models and their assumptions, and the second focused on gaining practical experience with fitting the models in R (with mentions of additional options available in other programs).
We will reference the preprint McCormick et al., PsyArXiv 2021 and associated online codebook if attendees wish to get an early sense of the topics covered and have R and the relevant packages installed. For a “cheat sheet” of model terms, notations, and a guide on how to read model equations more generally, please see the following Primer & Glossary.
The workshop is free but requires registration (see below). It will take place in the Lecture Theatre of the MRC Cognition and Brain Sciences Unit.
Schedule
Monday, March 27
9:00 – 10:30: Introduction to Longitudinal Modeling (Goals, Approaches, Data)
10:30 – 10:45: Break
10:45 – 12:00: Approaches to Incorporating Time
12:00 – 13:30: Lunch (not provided)
13:30 – 14:45: Determining the Optimal Shape of Change Over Time
14:45 – 15:00: Break
15:00 – 16:30: Hack-a-thon Part 1
16:30 – 17:00: Q&A and Wrap-up
Tuesday, March 28
9:00 – 10:30: Covariates & Distal Outcomes
10:30 – 10:45: Break
10:45 – 12:00: Nested Data, Advanced Applications
12:00 – 13:30: Lunch (not provided)
13:30 – 14:00: Q&A and Wrap-up
14:00 – 16:30: Hack-a-thon Part 2
In order to register please send an email to olaf.hauk@mrc-cbu.cam.ac.uk with subject "Structural Equation Modelling workshop", including your full name, affiliation, and a couple of sentences to justify your attendance (Max 100 words). You will be notified whether you have a place in due course. Spaces for the workshop are limited, so please make sure that you have space in your diary at these dates before registering interest, and let us know as soon as possible if you need to cancel.
If you are attending this workshop then please read the preparation instructions.