https://github.com/vtraag/wtmc-causality
WTMC method track on causal models
https://github.com/vtraag/wtmc-causality
Last synced: 7 months ago
JSON representation
WTMC method track on causal models
- Host: GitHub
- URL: https://github.com/vtraag/wtmc-causality
- Owner: vtraag
- Created: 2021-02-10T21:22:51.000Z (over 4 years ago)
- Default Branch: master
- Last Pushed: 2021-02-10T21:59:37.000Z (over 4 years ago)
- Last Synced: 2025-01-29T16:43:51.957Z (9 months ago)
- Language: TeX
- Size: 633 KB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Introductory reading
You are asked to read a book chapter to acquaint yourself with causal diagrams. The article that will serve as an introduction to these causal diagrams is:
- Elwert, F. (2013). Graphical Causal Models. In S. Morgan (Ed.), Handbook of Causal Analysis for Social Research (pp. 245–273). Springer. https://doi.org/10.1007/978-94-007-6094-3_13 (Available from http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.364.7505&rep=rep1&type=pdf)
The book chapter is aimed at social scientists who are interested in causal analysis. Unfortunately, the book chapter is somewhat more focused on quantitative social scientists. For example, the article contains some equations, which is not necessary I believe. Please try to ignore most equations, but do attempt to understand the core of the idea. Finally, you may ignore the section "The Sequential Backdoor Criterion for Time-Varying Treatments".
# Lecture and online tool
The concepts introduced in the book chapter may still be relatively difficult to understand without further explanation. In order to help you understand these concepts, there are two things that may help. First of all, I will explain the essence of these causal diagrams in a short lecture. We can get together after two weeks, so that you first have some time to try to understand the introductory paper yourself. If you have any questions, you can raise them during the lecture. After the lecture, you could then try to re-read the article to improve your understanding of the concepts.
**Please fill out this Doodle for picking a time for a lecture: xxx**
Secondly, you can draw causal diagrams, and read off its implications, in an online tool call Dagitty: http://www.dagitty.net/dags.html. A good exercise to make sure you properly understand the causal diagrams, is to draw the diagrams from the book chapter in Daggity. Of course, the results you get from Dagitty should be the same as in the book chapter. But perhaps you think some of the provided example causal diagrams are unrealistic: how do things change if you change the causal diagram?
# Exercises
You are requested to perform two exercises:
1. Please choose a research article that you are well acquainted with. The research article should be aimed at trying to answer some empirical question. Try to draw a causal diagram that you think is reasonable for this research article. Using this causal diagram, try to answer the following questions:
- What are the variables of central interest to the study?
- What variables should you control for to estimate a causal effect?
- Is there any endogenous selection effect (as formulated by Felix Elwert) for this study?
- In light of this causal diagram, what do you think about the conclusions of the study?
2. Please try to draw a causal diagram for your own research, and try to answer the same questions. In addition:
- How could you integrate these insights into your data collection and interpretation of observations?
- Causal diagrams are often more oriented towards quantitative scientists. What opportunity do you see for using causal diagrams in a more qualitative setting?
Feel free to use Daggity for helping you to answer these questions. However, please also do try to make sure you can explain the results yourself.