Developing computational tools for nonparametric summaries of income mobility

该奖项是美国国家科学基金会社会、行为和经济科学博士后研究奖学金（SPRF）计划的一部分。SPRF计划的目标是为学术界、工业界或私营部门和政府的科学事业准备有前途的早期职业博士级科学家。SPRF奖励包括在知名科学家的赞助下进行为期两年的培训，并鼓励博士后进行独立研究。美国国家科学基金会寻求促进科学界各阶层的科学家，包括那些未被充分代表的群体的科学家，参与其研究项目和活动；博士后阶段被认为是实现这一目标的一个重要的专业发展阶段。每个博士后必须解决各自学科领域的重要科学问题。在加州大学洛杉矶分校珍妮·布兰德博士的赞助下，这个博士后奖学金奖支持一个早期职业科学家研究计算方法来研究收入流动性。广泛的社会科学研究记录了个人一生中以及家庭中几代人的收入是如何变化的。这些研究大多依赖于线性模型。该项目通过使用统计学和机器学习中开发的新计算方法，扩展了我们对收入流动性的理解。这些方法允许在高收入和低收入人群中有不同的模式，并允许对分布进行更全面的总结。因为收入对美国家庭的幸福是必不可少的，该项目的目标与NSF推进国家健康、繁荣和福利的使命一致。有两个具体目标。该项目的第一个目标是将美国家庭收入的年变化可视化，并将其记录在现有的调查数据中。一方面，收入可能表现出稳定性：去年的收入可能是今年收入的一个很好的预测指标。另一方面，收入可能是不稳定的：升职和裁员等事件可能会产生意想不到的变化。过去的工作主要是用均值和方差的线性模型来总结稳定性和波动性。这个目标是建立在研究分位数模式和允许非线性关系的基础上的。该项目的第二个目标是考虑几代人的收入模式。如果儿童在童年时接触到更好的物质条件，这将如何影响他们成年后的收入？为了回答这个因果关系问题，研究调整了其他可能影响儿童收入暴露和成年收入实现的家庭背景指标。在以往工作的基础上，本项目明确考虑了非线性因素，即与高收入家庭相比，每增加一美元的收入可能对低收入家庭产生不同的影响。除了对代际流动这一实质性主题的贡献外，该项目的这一组成部分还为估计和可视化连续治疗变量的因果效应提供了方法。总之，这些目标带来了计算社会科学的新发展，以解决有关美国家庭收入的经典问题。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

This award was provided as part of NSF's Social, Behavioral and Economic Sciences Postdoctoral Research Fellowships (SPRF) program. The goal of the SPRF program is to prepare promising, early career doctoral-level scientists for scientific careers in academia, industry or private sector, and government. SPRF awards involve two years of training under the sponsorship of established scientists and encourage Postdoctoral Fellows to perform independent research. NSF seeks to promote the participation of scientists from all segments of the scientific community, including those from underrepresented groups, in its research programs and activities; the postdoctoral period is considered to be an important level of professional development in attaining this goal. Each Postdoctoral Fellow must address important scientific questions that advance their respective disciplinary fields. Under the sponsorship of Dr. Jennie Brand at the University of California, Los Angeles, this postdoctoral fellowship award supports an early career scientist researching computational methods to study income mobility. Extensive social science research has documented how incomes change over the life course for individuals and across generations within families. Much of that research relies on linear models. This project extends our understanding of income mobility by using new computational methods developed in statistics and machine learning. These methods allow distinct patterns among those with high and low incomes and allow fuller summaries of the distribution. Because income is essential to the well-being of American families, the goals of the project align with NSF’s mission to advance national health, prosperity, and welfare.There are two specific aims. The first aim of the project visualizes year-to-year changes in the incomes of American families, as recorded in existing survey data following those families over time. On one hand, incomes may exhibit stability: income last year may be a good predictor of income this year. On the other hand, incomes may be volatile: events like promotions and layoffs may create unexpected changes. Past work has primarily summarized stability and volatility with linear models for means and variances. This aim builds on that work by examining patterns in quantiles and allowing nonlinear relationships. The second aim of the project considers patterns of incomes across generations. If children were exposed to better material conditions in childhood, how would this affect the incomes that they realize as adults? To answer this causal question, the research adjusts for other measures of family background which may affect childhood income exposure and adult income attainment. Expanding on past work, this project explicitly considers nonlinearities such that each additional dollar of income may have a different effect among low-income as compared with high-income families. In addition to its contribution to the substantive topic of intergenerational mobility, this component of the project contributes to methodology for estimating and visualizing the causal effects of continuous treatment variables. Together, these aims bring new developments in computational social science to bear on classic questions about the incomes of American families.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.