Course Number: 94-835
Applied Econometrics II
The second mini of this sequence teaches how to evaluate causal relationships. Linear regressions and other econometric methods recover statistical relationships among different variables, but this says nothing about whether some variable "x" truly has an effect on another variable "y". The primary challenge for econometric analysis is to infer causation.
This course begins by demonstrating problems that often arise with regressions, such as omitted variables and endogeneity bias. Then we introduce a framework to understand causation based on the analogy of a random experiment. This framework helps to clarify the assumptions required for regression estimates to show true causal relationships. Finally, we will cover methods that can identify causal assumptions under certain assumptions: instrumental variables, and matching.
The course emphasizes learning through application. There will be extensive work with real-world datasets, both in homework and with examples in class. Through an individual project, students will apply methods from the course to analyze data on a subject they choose. As a prerequisite, students are assumed to have competence with statistical software such as Stata, which is taught in the preceding course.
94-834 Applied Econometrics I 6 Credits