Political scientists increasingly view multiple imputation as a superior way to handle missing data compared with the widely used practice of listwise deletion. Yet few systematic studies have asked how switching to multiple imputation changes the body of published empirical findings.
🧪 Reanalyzing CIPE Studies From Top Journals
- Scope: almost every quantitative study in the comparative and international political economy (CIPE) subfield published during a recent five-year period in International Organization and World Politics, two leading CIPE journals.
- Method: substituted multiple imputation for the original use of listwise deletion and reestimated the reported analyses.
- Claim: the first large-scale examination of how replacing listwise deletion with multiple imputation affects empirical results in political science.
📊 Key Findings
- The outcome is striking: in almost half of the studies, key results “disappear” (by conventional statistical standards) when reanalyzed using multiple imputation.
- This pattern indicates that the choice of missing-data strategy can materially change whether published effects appear statistically robust.
🌍 Why This Matters
- The common use of listwise deletion may have shaped a sizable portion of CIPE empirical knowledge; substituting a widely recommended alternative produces markedly different conclusions in many cases.
- These results raise broader questions about the robustness of published findings and the importance of transparent handling of missing data in future research.
