Abstract
Design-based causal inference has reshaped the social sciences, and observational data give public administration scholars a complementary way to identify causal effects in real-world settings. Yet we know little about how often the field uses these designs, how well studies execute them, and whether they advance theory. To fill this gap, we systematically review 5,246 articles published in 10 leading public administration journals from 2016 to 2025 and analyze 224 studies that apply design-based methods to observational data. These studies account for 6% of quantitative articles. Of them, only 33% pass a validity test of their identifying assumption. Nearly half of the studies address causal mechanisms only through theoretical framing. We conclude by offering propositions for advancing public administration theory with observational causal inference methods.
Figure 1: The landscape
