Week 10

Finding Front Doors

Readings

Further Reading

If you’re interested in learning more about these methods and don’t want to wait until I teach a standalone class on causal inference, check out chapters 19 and 20 in Huntington-Klein (2021).

Overview

This week, the instrumental variable approach to causal inference. We’ll talk experiments, two-stage least squares, and regression discontinuity designs.

Problem Set

In a knitted R script (or Rmd), complete the following problems:

  1. Using the replication data from the Cohn et al. (2019) wallet experiment, estimate the average treatment effect of money for (a) men only, (b) women only, (c) public institutions only, (d) people who seemed to understand what the experimenter was saying. Do there seem to be heterogeneous treatment effects? Can we interpret these estimates as causal? Why or why not?
  2. Estimate the incumbency advantage/disadvantage in other (non-Brazil) Latin American countries from the Klašnja and Titiunik (2017) dataset. Plot the data, faceting by country.
Cohn, Alain, Michel André Maréchal, David Tannenbaum, and Christian Lukas Zünd. 2019. “Civic Honesty Around the Globe.” Science 365 (6448): 70–73. https://doi.org/10.1126/science.aau8712.
Huntington-Klein, Nick. 2021. The Effect: An Introduction to Research Design and Causality. S.l.: CHAPMAN & HALL CRC. https://theeffectbook.net/index.html.
Klašnja, Marko, and Rocio Titiunik. 2017. “The Incumbency Curse: Weak Parties, Term Limits, and Unfulfilled Accountability.” The American Political Science Review 111 (1): 129–48. https://doi.org/10.1017/S0003055416000575.

References