Current course offerings

Please Note: Course profiles marked as not available may still be in development.

Course description

Module 1: Introduction to causal concepts (2 weeks) Potential outcomes and counterfactuals Definition of causal effects Assumptions required for measures of association to have a causal interpretation General concepts of estimation of causal effects in randomised controlled trials and observational studies Module 2: Causal diagrams and directed acyclic graphs (2 weeks) Drawing a causal diagram Association versus causation in causal diagrams Graph terminology, properties, and interpretation Recognising biases in diagrams and using diagrams to guide design and analysis Eliciting and refining the structure of causal diagrams Module 3: Time-invariant exposures and propensity scores (3 weeks) Introduction to propensity scores ¿ the `why and what¿ General operation of propensity scores Propensity score adjustment methods: matching, stratification, regression, and inverse probability weighting - the `how¿ Guidelines for the estimation of causal effects using propensity scores Module 4: Marginal structural models (MSM) for time-varying treatments (2 weeks) Key characteristics of the source data The problem with standard methods Specifying marginal structural models and their estimation Required assumptions Issues specific to estimation of MSMs for survival outcomes Module 5: Causal mediation analysis (2 weeks) Critique of the standard ¿Baron and Kenny¿ approach used in the social sciences Modern approaches to mediation Definition of controlled direct, natural direct, and natural indirect effects Estimation methods Interventional effects Multiple mediators Module 6: Further topics (1 weeks) Propensity scores with continuous exposures Complier average causal effects and survivor average causal effects Introduction to G-methods (Generalised-treatment methods)

Archived offerings