New research reveals that observational studies may incorrectly attribute oil resources to autocratic survival.
The analysis highlights survivorship bias, a specific type of collider bias, as a potential issue for cross-national political science research. This form of selection bias can lead to false positive findings about the 'Resource Curse' in certain contexts.
Using historical data from the Arabian Peninsula and causal modeling techniques, this paper demonstrates how existing units (countries) are subject to differential survival rates influenced by treatments like sovereignty changes.
When accounting for survivorship bias through proper causal model specification, standard estimates showing oil resources boost autocratic survival disappear entirely. This finding implies that previous conclusions about resource effects on governance may be flawed.
The study urges political scientists toward more nuanced approaches when analyzing observational data and emphasizes the need to account for treatment heterogeneity in comparative politics research.
| Survivorship Bias in Comparative Politics: Endogenous Sovereignty and the Resource Curse was authored by Benjamin Smith and David Waldner. It was published by Cambridge in POP in 2021. |
