Top Course Tags
A Few Big Assignments
Great Intro to the Subject
This class was tough.
The goal of this course is to provide students with a strong foundation in several commonly used quantitative research designs use for causal inference. A secondary goal is to provide students with exposure to state of the art research methods to enable more critical readers of published research and to transfer tools to carry out such research. Research design refers to the broad strategic framework used to tackle research questions.
The course starts with a thorough review of the experimental research design. This design, which involves random assignment to different research groups, is widely considered to be the most reliable design available. In the course the professor discusses the properties of the experimental estimator, and spends a lot of time on practical issues that emerge when implementing these designs and when analyzing data generated by them. After a thorough grounding in experimental designs, the class explores three quasiexperimental alternatives which are known to have strong statistical properties and which have each evolved considerably in recent years: regression discontinuity, interrupted time series and propensity score analysis. Eventually, the discussion switches gears to discuss the various dimensions of writing a solid research design document and the numerous data analysis details that one needs to pre-specify. The course concludes with a brief discussion of a regression based design which has seen substantial improvement in recent years: instrumental variables analysis. Along the way, the class lectures cover practical data analysis issues that arise in any quantitative research study. As students we “learn by doing” so come out able to analyze data coming out of these research designs in the future.
Hours per week:
Advice for students:
Some advice is to pay attention to the professor, be willing to learn new content, and show interest in the importance of quantitative research design to prove casual inference.