breakpoint regression, policy evaluation, machine learning, result variables
Abstract
Breakpoint regression is closer to quasi-natural experiments in design, and the estimated results are more accurate, which has attracted widespread attention from the academic community in recent years. This study systematically sorts out the theoretical and applied research of breakpoint regression in the field of policy evaluation. The research results show that the theoretical research on breakpoint regression mainly focuses on model applicability, bandwidth selection, data accumulation problems, etc., and breakpoint regression is in higher education, ecological environment, fiscal taxation, scientific and technological innovation, etc. The field has been widely used. This article aims to provide new ideas for the application scenarios of breakpoint regression, and research combined with machine learning can be added in the future.