NBER Working Papers · 2026 · Kelly M. Jones、Mayra Pineda-Torres
中文摘要
许多公共政策旨在扶持人口中的特定定向子群体 (或改变其行为)。为衡量此类政策的因果影响, 社会科学家采用准实验的政策评估方法, 但常常受制于数据可得性问题, 尤其是涉及定向人群时——这些人群往往处于弱势地位和/或较为脆弱。我们以堕胎禁令对经济福利的影响为例, 利用功效计算来量化基于人口层面数据检测政策对规模较小的少数人群的影响所面临的挑战。我们深入剖析了这一挑战, 并提供了一份路线图, 说明研究者如何能够事前 (ex-ante) 计算其研究是否具备足够的统计功效来检测出较为有意义的效应量。
Abstract
Many public policies aim to support (or change the behavior of) a specific, targeted sub-group of the population. To measure the causal impacts of such policies, social scientists use quasi-experimental methods of policy evaluation, but are often hampered by data availability issues, especially regarding targeted populations, who are often disadvantaged and/or vulnerable. We use power calculations to quantify the challenge of detecting a policy’s impact on a small minority population based on population level data, using the impact of abortion bans on economic welfare as an example. We provide insight into this challenge and a roadmap for how researchers can calculate, ex-ante, whether their study is sufficiently powered to detect reasonably meaningful effect sizes.