PoliSci 451: Overview
Design and Analysis of Experiments
Politial Science 451
Spring Quarter, 2012-13
polisci451.stanford.edu
Instructors
Shanto Iyengar (siyengar@stanford.edu)
Simon Jackman (jackman@stanford.edu)
Location/Time
Wed 9-12
Class Discussion Email List
experiments-seminar@lists.stanford.edu
Summary
Experimental methods have become the standard in many fields of political science. We survey the logic and practice of experimentation as applied to laboratory, field and survey settings. We discuss the strengths and limitations of experiments in relation to observational methods. Design considerations include randomization, the construction of treatments, the use of deception, ethical implications of deception, and new developments in subject recruitment. Turning to the analysis of experimental data, we describe the methods for estimating treatment effects, interactions, and more complex indirect effects stemming from either mediator or moderator variables. We also cover appropriate data analytic strategies for estimating causal effects from natural and "quasi-experimental" designs including interrupted time series, differences-in-differences, the regression-discontinuity design, matching and instrumental variables.
Texts
Elliot Aronson et al. 1989, Methods of Research in Social Psychology. (2nd ed.) McGraw Hill.
Josh Angrist and Jörn-Steffen Pischke. 2009. Mostly Harmless Econometrics: An Empiricist's Companion. Princeton University Press: Princeton, NJ.
Materials
A course website (polisci451.stanford.edu), will be the primary mechanism for distributing course materials and important class announcements. All enrolled students should confirm immediately that they are able to access and view the website. Students are expected to check the site regularly for any updates.
Requirements
Each student will design an experiment (or series of experiments) that advances the state of knowledge in some given substantive field. Your proposal outlining and defending the design will be presented in class during the last two sessions. The proposal will account for 60 percent of the grade and the remaining forty percent is allocated based on a take-home final exam.
Schedule
Please visit the Readings section of this site for an updated class schedule.