Which covariates to adjust for: An introduction to directed acyclic graphs

Graduate Programme in Public Health

Contents & aim
Researchers typically seek to answer causal questions using observational studies due to randomized controlled trials are either unethical or impractical in many situations. 

Since confounding, selection bias, and information bias may lead to spurious statistical associations in observational studies. Therefore it is important to adequately address confounding, selection bias, and information bias for making valid causal inferences from observational studies. 

Directed acyclic graphs (DAGs) are increasingly used in modern epidemiology to visually present causal knowledge and assumptions between variables. Once one can manage the rules for translating the causal knowledge and assumptions into a DAG and reading off the expected statistical associations from the causal knowledge and assumptions represented in a DAG, it can facilitate a number of tasks, such as choosing regression covariates, understanding selection bias, and information bias. Using DAGs makes it easier to recognize and avoid mistakes in a number of analytic decisions.

The course aims to provide participants with an introduction to the use of directed acyclic graphs (DAGs) as a tool to control confounding. The course is  also the pre-course for future course – an short introduction to g-method

Course credit: 1,8 ECTS

Teaching arrangement: Lecturing and workshops

Course leader:  Associate professor Chunsen Wu, University of Southern Denmark

Place: SDU, Odense

Max number of participants: 20 PhD students

Course Fee:
The course is free of charge for PhD students enrolled in Universities that have joined the "Open market agreement".
For other participants there is a course fee of DKK 4,754,-

To give you the best possible experience, this site uses cookies Read more about cookies

Accept cookies