Causal Inference and Bayesian Analysis

Kursöversikt

  • KurskodFDAT002
  • ECTS-poäng7,5
  • InstitutionDATA- OCH INFORMATIONSTEKNIK
  • ForskarskolaData- och informationsteknik
  • Startar2025-09-12
  • Slutar2025-12-19
  • SpråkEngelska
  • AnmälanKontakta kursansvarig

Kursansvarig

Om kursen

The expectation is that this will be a useful course for you, and hopefully it will inspire you to further interest in this topic and, thus, also start using these approaches in your own research. The aim is that you will conduct a more advanced analysis of data you have collected in your own research.


Content
This course is for PhD students and researchers who are curious about quantitative analysis. We will focus on causal inference and Bayesian analysis. We do not expect you to have any statistics background (likely a good thing).
The course contains:
  • Descriptive and inferential statistical techniques.
  • Causal analysis.
  • Usage of statistical tools.
It is a very hands-on course, which will allow you to up your skills in data analysis.


We will meet on Fridays 13.00-15.00, Sept 12 to Dec 19, i.e., 15 occasions. 
Before we meet, we expect participants to have read the relevant chapter(s) in the course book and watched accompanying videos. It is also good if you keep an eye on the lecture notes and exercises for each week.

Preliminary schedule.

Date

Title

Material*

Teacher

Sept 12

Introduction

E1, LN1, CH1-3

Richard

Sept 19

Design of models I

E2, LN2, CH4-6

Richard

Sept 26

Design of models II

CH4-6 (focus on DAGs)

Julian

Oct 3

Information theory and maximum entropy

E3, LN3.1, LN3, CH7

Richard

Oct 10

Interactions and sampling

E4, LN4, CH8-9

Richard

Oct 17

GLMs

E5, LN5, CH10-11

Richard

Oct 24

Funky distributions

E6, LN6, CH12

Richard

Oct 31

Multilevel models

E7, LN7, CH13

Richard

Nov 7

Covariance and Gaussian Processes

E8, LN8, CH14

Richard

Nov 14

Measurement error and missingness

CH15

Richard

Nov 21

Generalized Linear Madness

CH16

Richard

Nov 28

Causality in research

-

Fredrik

Dec 5

Causality in research

-

Clemens

Dec 12

Multiverse analysis

-

Robert

Dec 19

Causal discovery

-

Moritz

* E = Optional exercises, LN = Lecture Notes, CH = Chapter in course book.


Objectives

Knowledge and understanding

  • Describe and explain the concepts of probability space (incl. conditional probability), random variable, expected value and random processes, and know a number of concrete examples of the concepts.
  • Describe Markov chain Monte Carlo methods such as Metropolis.
  • Describe and explain Hamiltonian Monte Carlo.
  • Explain and describe multicollinearity, post-treatment bias, collider bias, and confounding.
  • Describe and explain ways to avoid overfitting.

Skills and abilities
  • Assess suitability of and apply methods of analysis on data .
  • Analyse descriptive statistics and decide on appropriate analysis methods.
  • Design statistical models mathematically and implement said models in a programming language (R is used in the course, but several other options exist).
  • Make use of random processes, e.g., Bernoulli, Binomial, Gaussian, and Poisson distributions, with over-dispersed outcomes.
  • Make use of ordered categorical outcomes (ordered-logit) and predictors.
  • Assess suitability of, from a ontological (natural process) and epsitemological (maximum entropy) perspective, various statistical distributions.
  • Make use of and assess directed acyclic graphs to argue causality.

Judgement and approaches
  • State and discuss the tools used for data analysis and, in particular, judge their output.
  • Assess diagnostics from Hamiltonian Monte Carlo and quadratic approximation using information theoretical concepts, e.g., information entropy, WAIC, and PSIS-LOO.
  • Judge posterior probability distributions for out of sample predictions and conduct posterior predictive checks.


Examination
Examination will be done through a case that you submit in written form. The case is an analysis that you either want to redo (from a previous study) or that you want to do for the first time (in a coming study). Our hope is that the written material can be part of a replication package, or the actual paper, when you later publish it.


Course literature
McElreath, R. (2020), 2nd edition, Statistical Rethinking: A Bayesian Course with Examples in R and Stan, CRC, Boca Raton, Florida. ISBN: 9780367139919
(We believe the book can be downloaded from Chalmers library.)
Accompanying videos for the book on YouTube https://www.youtube.com/playlist?list=PLDcUM9US4XdPz-KxHM4XHt7uUVGWWVSus

Support for students with additional needs
In case you need any special support do not hesitate to contact the course responsible when the course starts or before the course.

Contacts
Course responsible and examiner: Richard Torkar torkarr@chalmers.se
Teachers:
Richard Torkar
Julian Frattini
Fredrik Johansson (guest lecture, CSE)
Clemens Wittenbecher (guest lecture, LIFE)
Robert Felt (guest lecture, CSE)
Moritz Schauer (guest lecture, MV)

Kurslitteratur

McElreath, R. (2020), 2nd edition, Statistical Rethinking: A Bayesian Course with Examples in R and Stan, CRC, Boca Raton, Florida. ISBN: 9780367139919
(I believe the book can be downloaded from the library.)

Föreläsare

Richard Torkar, Julian Frattini, Fredrik Johansson, Clemens Wittenbecher, Robert Feldt
Causal Inference and Bayesian Analysis | Chalmers