Open Science.

Meta Analyses

IV Designs in Political Science

We conduct a meta-analysis of instrumental variable (IV) designs published in three top journals in Political Science (APSR, AJPS, JOP) and identify several concerning patterns:

  • Researchers often overestimate the strength of their instruments due to non-i.i.d. error structures such as clustering.
  • 2SLS estimates are often highly uncertainl; commonly used t-test frequently underestimate the uncertainties.
  • 2SLS estimates are much larger in magnitude than OLS estimates in many observational studies and their ratio is inversely related to the strength of the instrument, suggesting violations of unconfoundedness or the exclusion restriction in the former.

Here, we present replication results for each paper we reanalyzed.

Empirical Studies

Legality & Legitimacy

This RMarkdown file, with R code and results embedded, replicates the main analyses in Fu, Xu and Zhang (2024): “Does Legality Produce Political Legitimacy? An Experimental Approach.”

Credit: Jinwen Wu

Revisiting LaLonde (1986)

This Quarto Book tutorial, with R code and results embedded, replicates the main analyses in Imbens & Xu (2024): “LaLonde (1986) after Nearly Four Decades: Lessons Learned.”

Credit: Zihan Xie & Jinwen Wu

Disguised Repression

This RMarkdown file, with R code and results embedded, replicates the main analyses in Pan, Xu and Xu (2024): “Disguised Repression: Targeting Opponents with Non-Political Crimes to Undermine Dissent.”

Credit: Jinwen Wu

Victimization Narrative

This RMarkdown file, with R code and results embedded, replicates the main analyses in Xu and Zhao (2023): “The Power of History: How A Victimization Narrative Shapes National Identity and Public Opinion in China.”

Credit: Jinwen Wu

Propaganda through Framing

This RMarkdown file, with R code and results embedded, replicates the main analyses in Pan, Shao and Xu (2022): “How Government-Controlled Media Shifts Policy Attitudes through Framing.”

Credit: Jinwen Wu

Legality & Legitimacy

This RMarkdown file, with R code and results embedded, replicates the main analyses in Fu, Xu and Zhang (2024): “Does Legality Produce Political Legitimacy? An Experimental Approach.”

Credit: Jinwen Wu

Disguised Repression

This RMarkdown file, with R code and results embedded, replicates the main analyses in Pan, Xu and Xu (2024): “Disguised Repression: Targeting Opponents with Non-Political Crimes to Undermine Dissent.”

Credit: Jinwen Wu

Victimization Narrative

This RMarkdown file, with R code and results embedded, replicates the main analyses in Xu and Zhao (2023): “The Power of History: How A Victimization Narrative Shapes National Identity and Public Opinion in China.”

Credit: Jinwen Wu

Propaganda through Framing

This RMarkdown file, with R code and results embedded, replicates the main analyses in Pan, Shao and Xu (2022): “How Government-Controlled Media Shifts Policy Attitudes through Framing.”

Credit: Jinwen Wu

Revisiting LaLonde (1986)

This Quarto Book tutorial, with R code and results embedded, replicates the main analyses in Imbens & Xu (2024): “LaLonde (1986) after Nearly Four Decades: Lessons Learned.”

Credit: Zihan Xie & Jinwen Wu

Disguised Repression

This RMarkdown file, with R code and results embedded, replicates the main analyses in Pan, Xu and Xu (2024): “Disguised Repression: Targeting Opponents with Non-Political Crimes to Undermine Dissent.”

Credit: Jinwen Wu