Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/verenakunz/Causal-Inference-for-Beginning-Undergraduates

An introductory course on causal inference. Designed for undergraduate students with only a working knowledge of R and multiple regression. Light on maths, heavy on intuition and practical examples.
https://github.com/verenakunz/Causal-Inference-for-Beginning-Undergraduates

Last synced: 2 months ago
JSON representation

An introductory course on causal inference. Designed for undergraduate students with only a working knowledge of R and multiple regression. Light on maths, heavy on intuition and practical examples.

Awesome Lists containing this project

README 

        
# Causal-Inference-for-Beginning-Undergraduates

An introductory course on causal inference in the social sciences, that I teach annually at UCL. Designed for undergraduate students with only a working knowledge of R, statistical theory and multiple regression (i.e., 1-2 semesters of a standard sequence). It would also be suitable (but probably too light on technical details) for masters or beginner PhD students. The course is as light as possible on maths, but heavy on intuition and practical examples; matrix notation is not used. The aim is to explain both the statistical theory but also the practicalities and ideas behind causal research designs. Students are taught to design and implement their own studies, and to critique existing papers. The examples are drawn primarily from economics and political science, with a little epidemiology thrown in too. 



This repository contains detailed typed lecture notes, as well as computer exercises in R with solutions and datasets, for a ten-week course covering experiments, matching, instrumental variables, regression discontinuity, difference-in-differences and synthetic control. The lecture notes sometimes reference readings, details of which can be found in the syllabus. I find it helpful to teach lecture 1 to remind students about t tests, omitted variable bias, etc., but it could easily be skipped.



The weekly exercises almost all involve students replicating results from well-known papers that use the techniques. I've provided worksheets (problem sets) as well as solutions and code for R. At some point I will get round to setting these up properly in R Markdown. The R code is mostly at a very introductory level and therefore suitable for those with minimal background in R. For instance, I do not use tidyverse for this course. My assignments do not tell students how to implement the code, but does provide hints to help them get started with new or tricky pieces of code.



I am considering trying to publish this as a textbook, and therefore feedback on any errors and potential additions/improvements would be extremely welcome. You are also very welcome to use these materials for your courses.



Note that lectures 2-4, and some of 6, are designed to be read alongside Gerber and Green's "Field Experiments" textbook and occasionally reproduce equations from the book. Lecture 8 also contains some diagrams that were originally made by Teppei Yamamoto for a similar module at MIT, and some of the material from lecture 9 (including some diagrams) is based on one of his lectures. Having both taken and taught various iterations of causal inference modules at MIT and UCL, it is very possible that some other borrowings have made their way into these lecture notes and exercises: please let me know if anyone else deserves crediting!



Tom O'Grady (Assistant Professor in Political Science, UCL)



14th April 2020