andrea ruggeri
professor of political science
dipartimento di scienze sociali e politiche
università degli studi di milano
Causal Inference
in Social & Political Science
This course teaches students to move beyond correlation and defend genuine causal claims using the kind of data political scientists actually have since randomized trials are rarely an option when studying war, institutions, or voters.
Each session pairs a technique with a live empirical puzzle: does ethnic diversity really drive civil war onset, or is the regression just laundering omitted confounders? Do UN peacekeepers cause local peace, or are they simply sent where peace was already likely ? Students learn to interrogate published findings with real skepticism before they ever build their own model.
The course then climbs a conceptual ladder — from what regression and matching can extract from observables, through a sensitivity-analysis "hinge" asking how large a hidden confounder would have to be to overturn a result, to design-based strategies (fixed effects, difference-in-differences, event studies, synthetic control, instrumental variables, regression discontinuity) that exploit natural sources of variation instead. Along the way, the questions get sharper: did economic shocks cause civil conflict in Sub-Saharan Africa? Does the transatlantic slave trade causally explain mistrust in Africa today? Did narrowly winning a British parliamentary seat cause a politician's later wealth?
summary topics/readings
Topic Required reading Suggested Slides
The Experimental Ideal Angrist & Pischke, ch. 2 (2009) M1 Slides
Causal inference & social sciences Imbens 2024 M2 Slides R2
Regression back to basics Angrist & Pischke, ch. 3 (2009) Fearon & Laitin 2003 M3 slides R3
Pre- & post-treatment controls Bellemare, Masaki & Pepinsky 2017 Chaudhry 2022 M4 Slides R4 data4
Matching I Sekhon 2009 GilliganSergenti 2008 M5Slides R5 data5
Matching II Rosenbaum 2020 Ruggeri et al 2017 M6 Slides
Sensitivity analysis Cinelli & Hazlett 2020 Hazlett 2020 M7 Slides
Fixed Effects I Angrist & Pischke, ch. 5 (2009) Acemoglu et al 2019 M8 Slides
Fixed Effects II Imai & Kim 2021 M9 Slides
Difference-in-Differences I Angrist & Pischke, ch. 5 (2009) Fouka 2019
Difference-in-Differences II Bertrand et al 2004 Acemoglu et al 2011
Event Study Miller 2023 Gutmann et al 2023
Synthetic Control Abadie et al 2015 Costalli et al 2017
Placebo & falsification tests Eggers, Tuñón & Dafoe 2024
Counterfactual estimators Liu, Wang & Xu 2024 Xu 2017
Instrumental Variables I Angrist & Pischke, ch. 4 (2009) Miguel et al 2004
Instrumental Variables II Morgan & Winship, ch. 9 (2007) Nunn & Wantchekon 2011, AER
Regression Discontinuity I Angrist & Pischke, ch. 6 (2009) Eggers & Hainmueller 2009
Regression Discontinuity II Valentim et al 2021 Esberg 2021
Causal Mediation Imai et al 2011 Acharya et al 2016