IDEM 185

Causal Inference With Graphical Models

Start: 25 June 2018
End: 29 June 2018
Location: Max Planck Institute for Demographic Research (MPIDR), Rostock, Germany

Ilya Shpitser

Course description

The advent of computers, the internet, and ubiquitous and inexpensive sensors
has vastly increased the amount of types of data available for analysis. Systematically missing records, unobserved confounders, selection effects, and measurement error present in many datasets make it harder than ever to answer scienti?cally meaningful questions. This course will teach mathematical tools for reasoning about causes, effects, and bias sources in data with con?dence. We will use graphical causal models, and potential outcomes to formalize what causal effects mean, describe how to express these effects as functions of observed data, and use regression model techniques to estimate them. The course will cover ignorable models, mediation analysis, causal inference in longitudinal settings, plug in and inverse probability weighted estimators, and causal decision theory. Time permitting, we will also discuss general identi?cation, structure learning, and inference under missing data, and dependent data.


The course is spread over 5 days, with 4 hours of lecture, and 2 hours of lab (for
practical assignments in R, and conceptual problems/discussions) each day.

Course prerequisites

Participants should be familiar with the R programming language.  Familiarity with statistics (in particular regression models), basic statistical programming in R (in particular manipulating data frames, generalized linear models, and Monte Carlo sampling), and graphical models is also required.


Students will be evaluated on the basis of completion of assignments, including both practical and conceptual parts, and participation in discussions.

Selected Readings From

  • Causal Inference in Statistics: A Primer by Judea Pearl, Madelyn Glymour and Nicholas P. Jewel.
  • Causality: Models, Reasoning, and Inference by Judea Pearl. ([2]).
  • Causal Inference (draft form) by Miguel A. Hernan, James M. Robins. Available at: https://www.hsph.harvard.edu/miguel-hernan/causal-inference-book/
  • Statistical Methods for Dynamic Treatment Regimes (reinforcement learning, causal inference, and personalized medicine) by Bibhas Chakraborty and E.M. Moodie ([1]).


[1] Bibhas Chakraborty and Erica E. M. Moodie. Statistical Methods for Dynamic Treatment Regimes (Reinforcement Learning, Causal Inference, and Personalized Medicine). Springer, New York, 2013.
[2] Judea Pearl. Causality: Models, Reasoning, and Inference. Cambridge University Press, 2nd edition, 2009.

Preparatory material

A good tutorial on causal inference with graphical models is at https://www.microsoft.com/en-us/research/video/tutorial-non-parametric-causal-models/


There is no tuition fee for this course.  Students are expected to pay their own transportation and living costs.

Rostock Retreat on Causality

Please note that this course can be followed in conjunction with the Rostock Retreat on Causality (held July 2 to 4, 2018). If you wish to apply for both, please indicate this in your application (see below). For more information on the Retreat, please visit http://www.rostock-retreat.org/.

Recruitment of students

Applicants should either be enrolled in a PhD program or have received their PhD.
A maximum of 20 students will be admitted.
The selection will be made by the MPIDR based on the applicants’ scientific qualifications.

How to apply

  • Applications should be sent by email to the MPIDR (address below).  Please begin your email message with a statement saying that you apply for course IDEM 185 - Causal Inference With Graphical Models. You also need to attach the following items integrated in *a single pdf file*: (1) A two-page curriculum vitae, including a list of your scholarly publications. (2) A one-page letter from your supervisor at your home institution supporting your application. (3) A two-page statement of your research and how it relates to the course.  Please include a short description of your knowledge of programming in R, regression models, and graphical models. At the very end of your research statement, in a separate paragraph, please confirm that, if admitted, you will be able to come without financial aid from our side.  In the same last paragraph, please indicate whether or not you are also applying for the Rostock Retreat on Causality (http://www.rostock-retreat.org/).
  • Send your email to Heiner Maier (idem@demogr.mpg.de).
  • Application deadline is 28 February 2018.
  • Applicants will be informed of their acceptance by 31 March 2018.
  • Applications submitted after the deadline will be considered only if space is available.