Module 9: Causal Inference

Dates - Venue - Description - Target audience - Exam - IMPORTANT: Incorporation in DTP and reimbursement by DS
Course prerequisites - Teachers - Course material - Fees - Enrol


Three full days in May 2019: May 20, 21 and 22, 2019, from 9 am till 4 pm.
Please note: the deadline for UGent PhD students who want a refund to open a dossier on the DS website (Application for Recognition) is April 19, 2019.


Faculty of Science, Site Sterre, Krijgslaan 281, building S9, Ghent, lecture room 3.1, 3rd floor.


Within fields spanning drug testing, epidemiology and social sciences, researchers are often faced with the challenge of assessing the effect of an exposure (e.g. asbestos exposure) on an outcome (e.g. lung cancer). Standard statistical methods are commonly used for this purpose, but often not targeted towards the causal question of interest, or even misleading. In this course we will introduce modern causal inference methods to infer causal effects from data.

The course will in particular introduce popular tools such as causal diagrams, standardisation, propensity score methods, marginal structural models for evaluating the effect of time-varying exposures, mediation analysis for inferring causal pathways and finally instrumental variable methods. All theoretical concepts will be set into the context of real life research problems, taken from medicine, epidemiology, economy and the psychological or social sciences. Exercise sessions in R throughout the course will ensure that participants actively use the just taught concepts.

Day 1

  • Introduction to causal inference and causal diagrams
  • Lab session on causal diagrams
  • Evaluating the effect of time-fixed treatments: estimands / regression / standardisation / propensity scores / stratification and matching
  • Lab session on standardisation / stratification / matching

Day 2

  • Evaluating the effect of time-fixed treatments: regression adjustment for the propensity score / inverse weighting / confounder selection
  • Lab session on regression adjustment for the propensity score / inverse weighting
  • Evaluating the effect of time-varying treatments: problems of standard regression / marginal structural models
  • Lab session on marginal structural models

Day 3

  • Introduction to modern mediation analysis: traditional mediation analysis / natural effects / mediation formula
  • Lab session on the mediation package
  • Introduction to instrumental variables methods
  • Lab session on instrumental variables

Target audience

This course targets anyone who is interested in the evaluation of the effects of treatment, interventions or exposures from randomized experiments or observational studies.


Participants can, if they wish, take part in an exam. Upon succeeding in this test a certificate from Ghent University will be issued.

Please note: For UGent PhD students it is no longer necessary to succeed in this exam to be able to incorporate the course in the DTP.

Incorporation in DTP and reimbursement from DS for UGent PhD students

As a UGent PhD student, to be able to incorporate this course in your Doctoral Training Program (DTP) and get a reimbursement of the registration fee from your Doctoral School (DS) you need to follow strict rules: please take the necessary action in time. The deadline to open a dossier on the DS website (Application for Recognition) for this course is April 19, 2019. Please note that opening a dossier does not mean that you are enrolled. You still need to enrol via the registration form on this site.

Course prerequisites

Participants should have a good statistics background and basic R experience. In particular, participants should be very familiar with linear regression and have some familiarity with logistic regression. For participants who are not familiar with logistic regression, reading materials will be made available prior to the start of the course.


Tom Loeys is Professor of Statistics at the Department of Data Analysis, Faculty of Psychology and Educational Sciences, Ghent University. His PhD focused on causal inference for survival data. After his PhD, he worked as a statistical consultant in the pharmaceutical industry. His current research focuses on mediation analysis and dyadic data analysis. He has authored over 90 peer-reviewed publications in international journals varying from statistics to psychology and medicine.

Beatrijs Moerkerke is Professor of Statistics at the Department of Data Analysis, Faculty of Psychology and Educational Sciences, Ghent University. During her PhD, she worked on multiple testing problems in statistical genetics. Her current research interests include causal mediation analysis and the analysis of brain imaging data.​ She teaches courses in statistics and methodology to students in psychology and educational sciences.​

Stijn Vansteelandt is Professor of Statistics at Ghent University, and Professor of Statistical Methodology at the London School of Hygiene and Tropical Medicine. He did postdoctoral research at the Harvard School of Public Health and Ghent University. He is expert in causal inference, where he focuses on the development of statistical methods for inferring the causal effect of an exposure on an outcome from experimental and observational data under minimal and well-understood assumptions, primarily in collaboration with epidemiologists. He has authored over 150 peer- reviewed publications in international journals on a variety of topics in biostatistics, epidemiology and medicine, such as the analysis of longitudinal and clustered data, missing data, mediation and moderation/interaction, instrumental variables, family-based genetic association studies, analysis of outcome-dependent samples and phylogenetic inference. He is Co-Editor of Biometrics, the flagship journal of the International Biometric Society, and has previously served on the editorial boards of Biometrics (2006-2012), Biostatistics (2010-2015), Epidemiological Methods (2011- 2015), Journal of Causal Inference (2011-2015), and Epidemiology (2013-2015).

Oliver Dukes received an MSc in Medical Statistics in 2014 from the London School of Hygiene and Tropical Medicine, and from 2014-15 worked at the Farr Institute of Health Informatics Research, University College London. In 2016 he began a PhD in Statistics with Prof. Stijn Vansteelandt at Ghent University. He works on developing statistical methods for inferring causal inference from large, routinely collected medical datasets.

Johan Steen is a postdoctoral researcher / statistician at the Intensive Care department of the Ghent University Hospital. He did doctoral research on causal mediation analysis at Ghent University (2012-2016) and is author of the medflex R package for flexible mediation analysis. Johan has been involved in teaching in the advanced master programme in statistical data analysis (Ghent University) and several hands-on workshops on causal mediation analysis abroad. His current research focuses on drawing and improving causal inferences from routinely collected hospital data to better inform clinical decision making. The time-dependent nature of exposures, medical interventions and confounding factors often requires trading traditional for more advanced statistical methods. In his work, he hopes to further develop and to foster widespread usage of these state-of-the-art statistical methods for causal inference in the biomedical sciences and hospital epidemiology.

Course material

Participants receive a printed copy of the notes and slides used in the presentations and of the example computer programs.


A different price applies, depending on your main type of employment.

EmploymentModule 9Exam
Private sector/Industry1 900 30
Non-profit, government, university outside AUGent2 405 30
(Doctoral)student outside AUGent2 315 30

1 If three or more employees from the same company enrol simultaneously for this course a reduction of 10% on the module price is taken into account.

2 AUGent-personnel and AUGent doctoral students who pay through use of an SAP internal order/invoice can participate at these special rates.

Enrol for this course