Generalised Linear Mixed-Effects Models (GLMMs) in R
This is a self-study course. All theory presentations and exercises are provided as on-demand videos, allowing participants to learn at their own pace and access the material whenever it is convenient.
Course overview
Generalised linear mixed-effects models (GLMMs) are among the most widely used statistical tools in ecology, environmental science, fisheries, biology, agriculture and the health sciences. They allow researchers to analyse hierarchical, clustered and repeated-measures data while accounting for dependence among observations. Examples include multiple observations from the same animal, repeated measurements at the same site, patients within hospitals, samples from the same vessel, nests within colonies, plots within fields, or transects within regions.
This course provides a practical introduction to linear mixed-effects models (LMMs) and GLMMs using R. The emphasis is on understanding model structure, selecting appropriate random effects, model validation, and interpretation of fitted values.
The course contains more than twenty worked examples covering a wide range of ecological and environmental applications. Throughout the course we use real data sets and focus on practical implementation using R.
Software
The course primarily uses the R packages glmmTMB, DHARMa and ggplot2, together with a range of supporting packages for data manipulation, visualisation and model interpretation. All analyses are conducted in R.
Required background knowledge
- Participants should be familiar with:
- Basic R programming
- Data exploration and visualisation
- Multiple linear regression & GLM
If you are not comfortable with these topics, we recommend first completing our "Data exploration, regression, GLM and GAM" course. To help participants refresh their knowledge, the course includes preparation exercises covering multiple linear regression, Poisson GLMs, negative binomial GLMs and an introduction to DHARMa.
Course structure
Preparation
The preparation section contains revision material for participants who would like to refresh their knowledge before starting the mixed-model material.
- Revision of multiple linear regression using the Sunfish data
- Introduction to DHARMa
- Poisson GLM applied to Queen conch data
- Negative binomial GLM applied to Queen conch data
Module 1: Introduction to Linear Mixed-Effects Models
- Introduction to matrix notation
- Theory of linear mixed-effects models
- Lilies and beavers: random intercept models
- Bears and ants: linear mixed-effects models
- Painted turtles: repeated measurements
- Orangutan sleep data: mixed-effects models in behavioural ecology
Module 2: Nested, Crossed and Random Slope Models
- Avian malaria: two-way nested random effects
- Baboon grooming behaviour: nested and crossed random effects
- Theory of random intercept and slope models
- Random intercepts and slopes using lilies and beavers
- Heat tolerance in solitary bees
Module 3: Count Data GLMMs
- Hollow oak beetles: Poisson GLMM
- Freshwater plant species richness: Poisson and Generalised Poisson GLMMs
- Brazilian flathead fisheries data: Poisson and negative binomial GLMMs
- Squirrel responses to predator cues: behavioural ecology applications
Module 4: Advanced Count and Binary GLMMs
- UK pollinator data: Poisson GLMM with nested random effects
- Swiss seeds: nested and crossed random effects
- Trout survival: Bernoulli GLMM
- Salmon pathogen eDNA: Bernoulli GLMM with random intercepts and slope
Module 5: Proportional, Continuous and Tweedie GLMMs
- Painted turtles: Binomial GLMM with repeated measurements
- Coral growth and herbivory: Beta GLMM with repeated measurements
- Caribou feeding behaviour: Ordered Beta GLMM
- Warbler movement data: Gamma GLMM with multiple random effects
- Sea trout and salmon lice: Tweedie GLMM
- Pollinator gardens: Tweedie GLMM with hierarchical random effects
What will you learn?
By the end of the course you will be able to:
- Understand the difference between fixed and random effects
- Specify nested and crossed random-effects structures
- Fit LMMs and GLMMs in R
- Analyse count, binary, proportional and continuous responses
- Interpret random intercepts and random slopes
- Validate models using DHARMa
- Construct and interpret fitted values
- Compare alternative random-effects structures
- Deal with repeated-measures and hierarchical data
- Select appropriate distributions for ecological data
- Construct and interpret conditional and marginal fitted values
The course focuses on practical modelling skills that can be applied immediately to your own data.