‘Metrics Monday: Advanced Econometrics–Causal Inference with Observational Data
(Update: As far as I can tell from the link below, the course is now full. It remains possible, however, to put your name on the waiting list if you are interested. I suspect some people will be moved from the waiting list to the course eventually.)
I will be co-teaching a course titled Advanced Econometrics: Causal Inference with Observational Data at the University of Copenhagen from May 14-18, 2018 with my colleague Arne Henningsen. Though the link lists Arne as the instructor, I will be teaching the lecture part of the course, and Arne, who will be teaching the lab part of the course, writes:
[T]he website states that I am the “lecturer” of this course. I asked our administration to change it to you or to both of us. However, they cannot change this, because only staff at our University can be responsible for our courses and, thus, can be mentioned as lecturers of our courses. A complete list of lecturers (including you) is given further down of the course website.
If you are interested in taking the course, enrollment is open to students outside of the University of Copenhagen, and as of writing, registration is about $165, which is a bargain (though if you register, you are obviously responsible for your travel and accommodation costs, but I hear Copenhagen is lovely in May).
Here is the course description:
Social science researchers are usually interested in investigating causal relationships. The analysis of causal relationships is generally easiest based on experimental data. The use of experiments in social sciences, however, has many limitations, and most empirical studies are based on observational (i.e., nonexperimental) data. The participants of this course will learn the theory and practice of state-of-the-art empirical approaches used for investigating causal relationships with observational data. The course participants will also learn how to evaluate and discuss the appropriateness of identification strategies for analyzing causal relationships and to choose the most appropriate identification strategy for analyzing a specific research question. All this will help the participants obtain more credible and reliable results in their empirical work and to publish their work in better journals.
I will start from the assumption that you are familiar with the standard methods for causal inference, viz. instrumental variables, panel data and difference-in-differences, regression discontinuity designs, but also discuss limited dependent and discrete choice variables, count data, duration data.
My goal with this class is to go beyond the analytics, and to teach what people do in applied work, i.e., the tests they run, the robustness checks they present, the figures they show, and so on to make a convincing use of the various techniques just enumerated. I like to think about this as the live version of this ‘Metrics Monday series of posts. If anything, it’ll just be a good opportunity to hang out and talk about applied econometrics for a week!