This page focuses on the course 14.387 Applied Econometrics: Mostly Harmless Big Data as it was taught by Prof. Joshua Angrist and Prof. Victor Chernozhukov in Fall 2014.
This course covers empirical strategies for applied micro research questions. Topics include:
Mostly my audience is headed for careers in research - either in academia, government, or NGOs. There’s also a growing private-sector contingent, people headed for jobs at places like Google, Microsoft Research, and Amazon.
This course is usually taught once aew year. Up until 2014, it was taught in the spring. However, it is now taught in the fall.
The course features three graded problem sets, which had to be submitted on time to be graded for credit. The course is graded pass/fail.
Most students were from Ph.D. or professional schools (like the Harvard Kennedy School). We typically get some precocious research assistants. Most of JPAL's (Abdul Latif Jameel Poverty Action Lab) local full-timers seem now to take the course. SEII (School of Effectiveness & Inequality Initiative) full-timers also take the course.
Many students have read at least some of the Angrist and Pischke book(s). These provide a good screen. If you like that stuff, and find it accessible and relevant, you'll like the course.
During an average week, students were expected to spend 12 hours on the course, roughly divided as follows:
Met 2 times per week for 1.5 hours per session.
In addition, several recitation sessions were held throughout the year; these were facilitated by a teaching assistant.
The course was co-taught by two instructors, Prof. Josh Angrist taught topics I-VII and Prof. Victor Chernozhukov started teaching from topic VIII.
Students spent time outside of class working on three graded problem sets as well as the readings.
In the column, “Mastering Metrics: Teaching Ecnomometrics,” Prof. Angrist and Jorn-Steffen Pischke wrote about the need to "overhaul" econometrics pedagogy. Below, Prof. Angrist shares aspects of his pedagogical practice in 14.387 Applied Econometric: Mostly Harmless Big Data.
We try to keep the content real by using authentic empirical questions that can be answered with real data. We think logically, not abstractly.
14.387 Applied Econometrics: Mostly Harmless Big Data is a graduate-level course. Although many students come wanting to just "sit in," all students are given name cards and required to participate in discussions. We also give long challenging problems sets - good performers earn prizes and recognition by completing these problems.