Course

A practical, intuitive introduction to multilevel models and visualisation in R

Organizer(s):

Multilevel or hierarchical models have become essential tools across the ecological, behavioural, and social sciences because many research questions involve data that are naturally structured in groups: individuals within sites, repeated measures over time, species within habitats, or households within communities. Traditional statistical approaches often ignore this structure, leading to misleading inferences and overstated confidence in results. Multilevel models provide a flexible and conceptually intuitive framework that allows researchers to account for variability at multiple levels, examine patterns across scales, and better understand complex ecological and socio-ecological systems. This session introduces participants to the logic, applications, and practical implementation of multilevel models using R. Without relying on mathematical detail, it explains when and why these models are useful, how to recognise hierarchical structure in data, and how to interpret results in a clear and meaningful way. The emphasis throughout is on real examples drawn from field studies, long-term monitoring projects, and community-level datasets, making the material accessible to participants with varied disciplinary backgrounds and levels of statistical experience. A key focus of the session is on effective visualisation, an often overlooked but critical component of modelling. Good visualisations help researchers explore data before modelling, communicate model structure, diagnose fit, and present final results transparently and intuitively. Using the ggplot2 package in R, participants will learn how to visualise raw patterns, random effects, uncertainty, and predicted values from multilevel models. The demonstrations will show how thoughtful plotting not only improves clarity but also enhances reproducibility and scientific credibility. By the end of the session, participants will understand the value of multilevel approaches, feel confident identifying hierarchical structure in their own datasets, and gain practical skills in fitting and visualising these models using R. The overarching goal is to empower researchers to use modern, robust analytical tools and communicate their results clearly, ultimately improving the quality of research across disciplines.