Übung Political Sociology: Applied Causal Inference

Course Details

Course Instructor: Nan Zhang (Office hours by appointment)

Course Website: https://nanzhangresearch.github.io/QMPS_FSS25

Date and Time: Mondays, 13:45-15:15 in C 217 (A 5, 6 Bauteil C)

Course Description

This course will teach students how to analyze questions from the field of political sociology through the application of causal inference methods. We will begin by asking what it means for X to cause Y using the framework of potential outcomes. We will then look at some of the most popular research designs in causal analysis including experiments, regression discontinuity designs, difference-in-differences / two-way fixed effects, and instrumental variables.

Students will learn to apply these methods to real and simulated data in R.

The overarching goal is to provide students with the foundation to perform their own analyses (e.g. for their BA theses) by transferring the acquired skills to their research interests.

Course Structure and Requirements

Class-time will consist of a mix of lecture, “live-coding” exercises, reading discussion, and student presentations of problem sets.

To encourage class attendance, I will only post lecture slides and code examples after we have covered the material in class.

In addition to in-class participation, students must complete the following assignments outside of class:

1. Reading memos. Students are required to read applied research articles that implement the methods we cover in class and submit reading memos.

Your memos should consist of questions about the methods in the article (e.g. data analysis, the assumptions involved, etc.). There is no need to summarize of the readings. I will assume that you have read them.

You should work on the memos in your groups. That is, discuss the articles within your group beforehand, and then come up with a common set of questions that you would like to discuss in class.

There are 5 required reading memos in total. Your memos should be submitted to me via email by 8:00am on the following Fridays:

  • 21 Feb
  • 28 Feb
  • 21 March
  • 2 May
  • 16 May

2. Take-home problem sets will provide students the opportunity to practice the methods they learn in class by analyzing real research data. Students are required to work in groups.

There are 3 problem sets in total. Problem sets must be submitted to me via email by 23:59 on the following Sundays:

  • 16 March
  • 6 April
  • 11 May

I will randomly select students to present their problem set answers during our class sessions. Problem set answers will be posted in the following week.

3. Diligent preparation and active participation are essential for the success of this class! The material is hard, especially if you are encountering it for the first time, so don’t be afraid to make mistakes and ask questions! This is how you will learn.

To encourage participation from everyone, I will randomly select several students who are “on call” to answer questions during each class session. The idea is not to cause anyone stress or anxiety, but rather to “break the ice” and foster a more active environment where everyone can feel comfortable to participate.

If you are worried about this “cold calling” policy, please come see me.

Of course, situations may arise when you cannot attend class or did not have time to prepare. As a courtesy, please let me and group members know beforehand.

Assessment

Technically, the Prüfungsleistung is a Hausarbeit.

This course will graded somewhat differently than what you are used to from other classes. You are still expected to complete assignments, and you will still receive a numeric grade on your transcript at the end of the semester. However, your grade will be determined based on your own self-assessment of your learning and effort in the class, with the possibility of adjustments up or down from me.

During the semester, you will be asked to complete two self-evaluation exercises (once at midterm, and once at the end of the semester). These exercises will include a series of questions about your work in the course. Specifically, you will be asked to reflect upon:

  • The amount of work you have done in the course. How much effort did you commit to:
    • the readings and memos
    • the group problem sets
    • participating in class discussions
  • Your successes and struggles in the course:
    • What did you learn each week?
    • What challenges did you encounter with the concepts, the assignments, or the logistics of the course?
  • Opportunities for improvement:
    • What steps could you (or did you) take to deepen your understanding of course concepts or increase your engagement in the class?
  • Propose the grade you feel you should receive in the course (up to that point), based on your:
    • attendance
    • completion of all required assignments
    • the quality of your contributions to (i) group homeworks, and (ii) class discussion
    • your overall comprehension and mastery of the course material
  • Explain and justify why you would give yourself that grade.

When assigning final grades, I will strive to honor your end-of-semester assessment of your own performance and progress in this course. However, I reserve the right to alter your proposed grade as appropriate, based on my own evaluation of your performance and progress in the course as a whole. If such an alteration seems warranted, I will contact you to set up a meeting to discuss your work in the course.

To help you keep track of your progress throughout the semester, it is helpful to keep a weekly “journal”. In particular, at the end of each class session, take 5 minutes and write down (1) what is one thing that you learned over the past week, and (2) what is something that you still do not yet understand?

Please include this journal as an appendix to your final reflection paper.

The final reflection paper is due TBA.

A Note on the Course Workload

This course is worth 6 ECTS. That’s 180 hours of “work” in total.

We will spend 21 of those hours in class. This means that you should work about 160 hours outside of class during the semester.

The course is set up so that you will be spending the majority of these 160 hours on group assignments.

I know that these assignments are time consuming.

For instance, not only will you have to read an academic article, but you will have to discuss it within your group and write up a joint response paper.

As a result, it will feel like you are doing a lot more work for this course in comparison to your other courses.

This is intentional. The assignments are designed so that you can learn not only from me, but also from each other.

On the other hand, you will not have to spend much time studying or preparing for the final exam. Again, this is intentional.

Rather than focusing on “cramming” at the end of the semester, I believe that you will learn more by working at a more constant pace for the entirety of the course.

Weekly Schedule

Note: the current version of the syllabus is a work in progress. In fact, it’s probably too much to cover in one semester. I will make adjustments, depending on the pace of the class, so check back often!

All changes to the syllabus will be announced in class.

Session 1 (10 Feb): Course Introduction.

Nan will introduce the course, go over the syllabus, and answer any logistical questions you may have.

We will form groups for the remainder of the semester.

Throughout the course, we will use statistical simulations in R to illustrate statistical concepts.

To motivate this approach, we will end the session by simulating an answer to the famous Monty Hall problem.

Assignments for next week:

  • You should install both R and RStudio before our meeting next week. Make sure the software works on your laptops!

  • If you are new to R, please have a look at this excellent online material from Lion Behrens.

Session 2 (17 Feb): Probability and Statistics Review.

We will review some basic concepts from probability and statistics that will come up again and again in the course.

To illustrate these concepts, we will use simulation to draw marbles from a large bowl.

This example will help us to understand the ideas of expected value, unbiasedness and uncertainty.

Assignment for next week:

  • Read chapter 1 of Angrist and Pischke’s book Mastering ’Metrics (you can skim the Appendix if you like).

  • Email me your first reading memo by 8:00am on Friday, 21 Feb. Don’t forget to put your name on your memo!

Session 3 (24 Feb): Potential Outcomes and Average Treatment Effects

We will start today by simulating a randomized experiment under the potential outcomes framework.

We will use this simulation to illustrate how randomization produces unbiased estimates of the average treatment effect (ATE).

We will also relate the variation in our simulated experimental results to the standard errors produced by regression.

Assignment for next week:

Session 4 (3 March): Analyzing Experiments

We will discuss your questions about Enos’ trains experiment.

Using simulations, we will work through how to conduct randomization inference. This will (hopefully) provide a more intuitive explanation of p-values and the logic behind hypothesis testing.

No assignment this week.

Session 5 (10 March): Analyzing Experiments (continued).

We will wrap up our discussion of Enos and the randomization inference / hypothesis testing from last time.

We will also talk about how to construct confidence intervals.

Assignment for next week:

  • Using Enos’ data, please complete Problem Set 1. Email me your group’s problem set answers by 23:59pm on Sunday, 16 March. Be prepared to present your answers in class next week.

Session 6 (17 March): Student Presentations

Students present their problem set solutions in class.

Assignment for next week:

Session 7 (24 March): Regression Discontinuity

We will discuss your reading questions, and Nan will give a short presentation to introduce the logic of the regression discontinuity design (RDD).

We will spend the bulk of our time on data visualization, which is important for assessing the plausibility of the RDD.

Finally, if there is time, Nan will introduce some parametric estimation techniques.

Assignment for next week:

  • Please fill out the midterm evaluation. You should write 2-3 pages max! Use your weekly “journals” to help jog your memory.

  • Please submit your evaluation by 8:00am on Friday, 28 March.

Session 8 (31 March): Regression Discontinuity (continued)

We will wrap up whatever we did not cover from last week, including both parametric and non-parametric estimation methods.

Assignment for next week:

  • Using Myerson’s data, please complete Problem Set 2. Email me your group’s problem set answers by 23:59pm on Sunday, 6 April. Be prepared to present your answers in class next week.

Session 9 (7 April): Student Presentations

We will discuss your problem set answers, and wrap up any outstanding questions about RDD.

Easter Break (14 April - 25 April): No Class.

Session 10 (28 April): Introducing Instrumental Variables (IV)

Nan will give an introduction to instrumental variables (IV).

Assignment for next week:

Session 11 (5 May): Instrumental Variables (continued)

We will discuss your reading questions.

We will also look at some R code for how to estimate 2sls models.

Assignment for next week:

  • Using White’s data, please complete Problem Set 3. Email me your group’s problem set answers by 23:59pm on Sunday, 11 May. Be prepared to present your answers in class next week.

Session 12 (12 May): Student Presentations

We will discuss your answers to the problem set nd wrap up any outstanding questions from IV.

Assignment for next week:

Session 13 (19 May): Difference-in-Differences (DiD) and Two-Way Fixed Effects (TWFE)

We will discuss your reading questions, and the logic behind the difference-in-differences (DiD) design.

Nan will also give a short presentation showing the relationship between DiD and Two-Way Fixed Effects (TWFE).

No ssignment this week.

Session 14 (26 May): TWFE (continued)

In our final session, we will wrap TWFE and Nan will present some analysis code.

We will also do a short debrief about the course as a whole, and talk about your final Hausarbeit.

Lecture Slides

You can find lecture slides (with code examples) here. Note that these will be posted only after we have covered the material in class.

  • TBA

Problem Set Solutions

You can find my answers to the problem sets here. These are not necessarily “official” – they are just how I would have done the problem set.

  • TBA

References

If you want extra practice with R, check out the free ebook R for Data Science by Hadley Wickam.

If you are looking for more materials on causal inference, you can check out this Andrew Heiss’ Program Evaluation class, or this excellent ebook by Nick Huntington-Klein: theeffectbook.net.

Both resources are free, provide code examples, and are accompanied by youtube videos!