Achievement
RDDs and electoral data
Project
Training Program in Politics, Economics, and Psychology
University
University of California at Berkeley
(Berkeley, CA)
Research Achievements
RDDs and electoral data
In joint work, Jasjeet Sekhon (IGERT faculty affiliate) and Devin Caughey (IGERT trainee) have demonstrated that the "smoothness" assumption required for regression-discontinuity designs (RDDs) is not valid for U.S. elections due to the non-random fashion in which very close elections are decided. This finding suggests that when applied to electoral data, RDDs should not be interpreted as a natural experiment but rather as something more akin to an observational study.
SEE MORE:
- “Research Achievements”
- Achievements for this Project
