SAEON Graduate Student Network
SAEON Graduate Student Network
Resources for the SAEON GSN Causal Inference for the Biological Sciences with James Hagan, December 2023
What is causal inference? Most simply, causal inference is a sub-discipline of statistics and computer science that aims to infer causes from data. I will go through some basic principles of causal inference, illustrate why many standard analyses conducted in the biological sciences (e.g. multiple regression, model selection) frequently violate causal inference principles, show you why experiments are not safe from thinking causally and, finally, provide some recommendations about how to improve causal inference in the biological sciences.
Directed Acyclic Graphs (DAGs) are the simplest possible causal model. They specify which variables influence each other and in what direction.
This tutorial is meant to introduce the basics of using causal inference methodologies for use in your own research. First, we will learn how to play around with the DAGitty web tool that will allow us to create our own DAGs. Second, we will learn how to transfer our DAGs into R. Third, we will learn how to analyse our DAG in R. By analysing our DAG in R, we will be able to derive testable predictions to check whether our DAG is consistent with the data. Moreover, through this analysis, we will obtain information about the correct statistical models to fit to obtain a given causal estimate. And, we will gain information about how to interpret the coefficients obtained from any given statistical model given the DAG.
All the workshop materials (including R code) is available here.
Using the resources above try and follow along!
