Path Analysis, Structural Equations and Causal Inference for Biologists
https://prstats.org/course/path-analysis-structural-equations-and-causal-inference-for-biologists-pscb04/ Delivered by Prof. John William (Bill) Shipley, Professor Emeritus at Université de Sherbrooke and an internationally recognised expert in causal modelling, path analysis and structural equation modelling in biology. Prof. Shipley is the author of Cause and Correlation in Biology and has published extensively on causal inference and statistical methods for biological research. Learn how to use path analysis, structural equation modelling (SEM) and causal inference in R to investigate complex cause-and-effect relationships in marine mammal and ecological data. Marine mammal research frequently involves observational datasets in which environmental conditions, prey availability, habitat characteristics, anthropogenic pressures, behaviour and biological characteristics interact through multiple pathways. This course provides a rigorous framework for representing these relationships as explicit causal hypotheses and evaluating them statistically. What you'll gain * Understand causal inference using experimental and observational data * Translate biological hypotheses into directed acyclic graphs (DAGs) * Understand d-separation and statistical conditioning * Construct and test causal path models * Distinguish direct and indirect causal effects * Conduct piecewise structural equation modelling * Use the piecewiseSEM package in R * Incorporate mixed models into structural equation models * Understand equivalent models and compare competing causal hypotheses * Use AIC to compare non-equivalent models * Conduct covariance-based path analysis * Fit structural equation models using lavaan * Construct and interpret latent variables and measurement models * Work with nested and non-normal biological data * Fit and interpret multigroup models * Understand causal heterogeneity * Evaluate model fit and statistical power * Deal with missing data and small sample sizes * Report SEM results appropriately in scientific publications Course format * Dates: 19–30 October 2026 * Duration: 35 hours * Approximately 20 hours of lectures and 15 hours of practical work * Format: Live online * Practical analyses conducted in R and RStudio * R scripts, datasets and supporting materials provided * Opportunities to work with and receive feedback on your own data * All sessions recorded * 30 days' access to recordings * 30 days of post-course support * Course fee: £450 Who is this course for? The course is particularly suitable for students, researchers and professionals working in: * Marine mammal science * Cetacean and pinniped ecology * Marine ecology * Behavioural ecology * Habitat-use and distribution research * Conservation biology * Population ecology * Human impacts and anthropogenic disturbance * Environmental and climate-related research * Quantitative ecology and ecological modelling It will be particularly useful for marine mammal researchers working with observational datasets containing multiple interacting variables, where the aim is to move beyond identifying statistical associations and formulate explicit hypotheses about the causal processes connecting them. Participants should have experience using R and RStudio for statistical analysis and a basic understanding of statistical inference and regression. Experience with GLMs or mixed models is useful but not required. Why take this course? Marine mammal research often involves complex networks of potential drivers. For example, environmental conditions may influence prey availability, which can subsequently affect marine mammal distribution, behaviour or demographic responses. Anthropogenic pressures may have direct effects while also operating indirectly through changes in habitat, prey or behaviour. Analysing each relationship independently can obscure these interconnected pathways. Path analysis and SEM provide a framework for representing multiple hypothesised relationships simultaneously, allowing researchers to distinguish direct and indirect effects and formally evaluate whether observed data are consistent with a proposed causal structure. The course is especially applicable to marine mammal research because it covers piecewise SEM, which allows mixed-effects models and other model structures commonly needed for ecological observational data to be incorporated into a causal network. Participants will also learn covariance-based SEM and latent-variable approaches, while developing a deeper understanding of when causal interpretations of observational data are—and are not—justified. Learn more & enrol Path Analysis, Structural Equations, and Causal Inference for Biologists (PSCB04) https://prstats.org/course/path-analysis-structural-equations-and-causal-inference-for-biologists-pscb04/ Questions? Email: [email protected]<mailto:[email protected]> Best wishes, Oliver Hooker PR Stats Sent from Outlook for Mac
_______________________________________________ MARMAM mailing list [email protected] https://lists.uvic.ca/mailman/listinfo/marmam
