Dominik Hangartner: Catalogue data in Spring Semester 2019
|Name||Prof. Dr. Dominik Hangartner|
Professur für Politikanalyse
ETH Zürich, LEH D 4
|Telephone||+41 44 632 02 67|
|Department||Humanities, Social and Political Sciences|
|857-0002-00L||Methods III: Statistical Learning |
Only for MA Comparative and International Studies.
|8 credits||1U + 2S||D. Hangartner, M. Marbach|
|Abstract||Introduction to methods for supervised and unsupervised learning for the social sciences.|
|Objective||The goal of this course is provide students with an introduction to statistical learning methods. Upon completion of the course, students will have an understanding of modern computiational methods for modelling and prediction, the assumptions on which they are based, and be able to use them to address specific research questions in the social sciences.|
|Content||Topics include linear regression with interaction and fixed effects, binary logistic regression, classification, resampling methods, shrinkage approaches, tree-based methods, support vector machines, clustering, principal component analysis, factor analysis, and item response theory.|
|Literature||James, Gareth, Daniela Witten, Trevor Hastie, and Robert Tibshirani. An introduction to statistical learning. Springer, 2013. (7th edition). |
The PDF of the textbook is made freely and legally available by the authors and Springer press and part of the course package.
|Prerequisites / Notice||Methods II|
|857-0102-00L||Methods IV: Causal Inference |
Number of participants limited to 15.
MA Comparative and International Studies are given priority.
|8 credits||2U + 2S||D. Hangartner, D. Ward|
|Abstract||This course provides an introduction to statistical methods used for causal inference in the social sciences, covering both experimental and observational studies.|
|Objective||Familiarity with the key research designs and statistical methods used for causal inference from randomised and observational data.|
|Content||This course provides an introduction to statistical methods used for causal inference in the social sciences. Using the potential outcomes framework of causality, we discuss designs and methods for data from randomized experiments and observational studies. In particular, designs and methods covered include randomization, matching, instrumental variables, difference-in-difference, synthetic control, regression discontinuity, and quantile regression. Examples are drawn from the social sciences.|
|Literature||Angrist, Joshua D., and Jörn-Steffen Pischke. Mostly harmless econometrics: An empiricist's companion. Princeton university press, 2008.|
Rosenbaum, Paul R. Design of Observational Studies. Springer. 2010.
|Prerequisites / Notice||Methods III or equivalent|