행정학 및 정책학 연구에서의 회귀단절 설계의 활용 가능성 탐구

초록

This paper aims to introduce a regression discontinuity design (RDD) in order to prompt causal inference studies in the fields of public administration and public policy. RDD recently gains popularity in social science as a natural experiment that is useful for causal inference, in that it estimates an non-experimental but as-if-random effect that would be close to any experimental effect derived from a randomized controlled trial (RCT). The paper sections explain the basic principles, necessary assumptions, identification strategies, and the modeling methods in RDD, in addition to the comparison of sharp and fuzzy RDD types. At the end, the authors show the current underutilization of RDD in the public administration and public policy fields in Korea, and they provide an issue checklist that would be useful for those who want to apply RDD to their studies.

키워드