Hybrid (onsite and online): Introduction to GLLVM
- With spatial or temporal dependency -
Institute for the Sea and the Atmosphere (IPMA) Portuguese. Av. Dr. Alfredo Magalhães Ramalho. Lisbon, Portugal
22 - 26 September 2025
Course flyer
This is an onsite course, but you can also participate online via a Zoom connection (same price).
The central theme of this course is the analysis of multiple correlated response (or dependent) variables using GLMs and GLMMs. Rather than applying multiple univariate GLMs or GLMMs, we will focus on multivariate GLMMs, particularly generalised linear latent variable models (GLLVMs), for the simultaneous analysis of all variables.
During the course, we cover a wide range of exercises with examples such as biomass data from fish species sampled at multiple sites, count data from 250 freshwater benthic species across 200 sites, abundances of multiple parasite species on fish, abundances of multiple spider species in traps, various morphometric variables from honeybees, presence/absence of diet variables from brown bear faecal samples, and multivariate behavioural data from caribou.
In all these examples, we can analyse each variable with a univariate GLM(M). Although these analyses are relatively simple, there are also some problems:
- Extra Work: Individual analyses are computationally less efficient and require separate validation, interpretation, and reporting.
- Lack of Multivariate Relationships: Analysing the variables individually neglects the interconnected relationships and interactions between them.
- No Shared Variation: Univariate models might overlook consistent residual patterns across species, while multivariate models can capture shared variations due to common environmental factors.
- Multiple Testing: Conducting separate analyses increases the risk of Type I errors, especially when the response variables are highly correlated.
- Loss of Community-Level Insights: Analysing species separately misses out on a comprehensive, community-level viewpoint and can lead to inconsistent conclusions.
This is an applied and non-technical course that focuses on the practical implementation in R.
Pre-required knowledge
Participants should be familiar with data exploration, linear regression and basic GLM and GLMM (i.e. Poisson and negative binomial GLM) in R. The course website contain revision/preparation material with on-demand videos covering these topics. This is a non-technical course.
1 hour face-to-face
The course includes a 1-hour face-to-face video chat with the instructors (to be used after the course). You are invited to apply the statistical techniques discussed during the course on your own data and if you encounter any problems, you can ask questions during the 1-hour face-to-face chat.
A discussion board (access for 12 months) allows for interaction on course content between instructors and participants.
Course content
Monday:
- General introduction.
- A short theoretical presentation revising linear mixed-effects models.
- One exercise on linear mixed-effects models.
- Theory presentation on generalised linear latent variable models (GLLVM) for the analysis of multivariate data.
Tuesday:
- Two exercises on GLLVM for the analysis of count data (Poisson/negative binomial).
Wednesday:
- Theory presentation on constrained GLLVM (reduced rank regression and concurrent ordination).
- One exercise on constrained GLLVM.
- Tweedie GLLVM for the analysis of multivariate continuous data.
Thursday
- Catching up
- Gamma, Bernoulli and beta GLLVMs for the analysis of continuous, binary and proportional multivariate data.
- Time allowing: Adding spatial correlation to GLLVMs.
Friday:
- Adding spatial or temporal correlation to a GLLVM.
We reserve the right to change the exercises. Pdf files of all theory material will be provided. All exercises consist of data sets and annotated R scripts. Access to the course website is for 12 months. The Monday-Friday material does not contain on-demand video.
For terms and conditions, see: