The Health Effects of Hazardous Waste Sites: Evidence from Superfund Clean-Ups

描述（由申请人提供）：项目摘要/摘要本项目有三个目标。首先，我们将提供新的统计证据，说明接触危险废物场地污染与婴儿健康之间的关系。其次，我们将努力确定对婴儿健康最有害的特定化学物质和途径。第三，我们将量化超级基金计划对婴儿健康的影响。本研究将实施几项研究设计，以减轻遗漏变量的影响。首先，分析将利用特异质差异的时间超级基金清理比较前和清理后的健康结果，而非参数调整的时间趋势。清理时间的变化允许进行事件研究式分析，比较清理前后的年份。第二，该项目将利用选择规则的知识，该规则确定了最初计划用于超级基金清理的690个危险废物场地中的前400个。选择规则是一个非线性函数的一个连续的和嘈杂的措施的风险，这使得准实验回归不连续设计的实施。第三，可以利用我们将获得的保密的生命统计数据，查明同一妇女在清理前后的生育情况。识别兄弟姐妹的能力使我们能够控制固定的母亲特征（例如，遗传混杂因素），如果清理导致移徙，这一点可能特别重要。还将利用母源识别器直接评估清理工作是否导致个人迁移到以前受污染场地周围的地区。这项分析将使用有史以来收集的关于超级基金网站和婴儿健康的最全面的数据进行。超级基金数据文件包含每个超级基金站点的详细信息，包括经度、纬度和清理完成日期。重要的是，该项目还将编制一个电子数据库，其中包括在所有超级基金地点发现的化学品及其浓度，以及化学品可能影响婴儿健康的途径（空气、土壤、水）。超级基金的数据将与数百万个关于婴儿健康的个人层面的国家生命统计记录相匹配（例如，妊娠、出生体重、先天性异常和APGAR评分）和死亡率。这些数据将来自六个州（加利福尼亚州，佛罗里达州，密歇根州，新泽西州，宾夕法尼亚州，德克萨斯州），占美国出生人数的近40%，覆盖不同的地理区域和人口。我们将使用机密版本的生命统计数据中的地址数据来确定每个母亲的住所与最近的超级基金站点之间的距离，以1/4英里为增量。总的来说，该项目旨在评估接触国家危险废物场地对婴儿健康的影响。宾夕法尼亚州的初步结果表明，我们将有能力发现婴儿健康结果的微小变化。这些结果将与正在进行的关于在多大程度上应该以婴儿健康危险为动机来清理危险废物的辩论极其相关。本项目将提供新的统计证据，说明接触危险废物场地的污染对婴儿健康的影响。该分析是基于对危险废物场地周围地区在超级基金方案下清理场地之前和之后的婴儿健康状况的比较。它将使用有史以来收集的关于超级基金网站和婴儿健康的最全面的数据进行。

DESCRIPTION (provided by applicant): Project Summary/Abstract This project has three objectives. First, we will provide new statistical evidence on the relationship between exposure to pollution from hazardous waste sites and infant health. Second, we will endeavor to identify the specific chemicals and pathways that are most deleterious to the health of infants. Third, we will quantify the benefits of the Superfund program in terms of its effects on infant health. This research will implement several research designs to mitigate the influence of omitted variables. First, the analysis will exploit idiosyncratic differences in the timing of Superfund clean-ups to compare pre and post cleanup health outcomes, while nonparametrically adjusting for time-trends. The variation in timing of the clean-ups allows for an event-study style analysis that compares the years immediately preceding and following clean-ups. Second, the project will exploit knowledge of the selection rule that determined the first 400 hazardous waste sites out of 690 that were initially slated for Superfund clean ups. The selection rule was a nonlinear function of a continuous and noisy measure of risk, which allows for the implementation of a quasi- experimental regression discontinuity design. Third, it is possible to use the confidential vital statistics data we will obtain to identify births to the same women before and after clean-ups. The ability to identify siblings allows us to control for fixed maternal characteristics (e.g., genetic confounders), which may be especially important if clean-ups induce migration. The mother identifiers will also be used to assess directly whether the clean-ups lead individuals to migrate to the areas surrounding the formerly polluted sites. The analysis will be conducted with the most comprehensive data ever assembled on Superfund sites and infant health. The Superfund data file contains detailed information on every Superfund site, including their longitude, latitude, and the date the clean-up was completed. Importantly, the project will also compile an electronic database of the chemicals and their concentrations found at all Superfund sites, as well as the pathways through which the chemicals might affect infant health (air, soil, water). The Superfund data will be matched to millions of individual-level state vital statistics records on infant health (e.g., gestation, birth weight, congenital abnormalities, and APGAR scores) and mortality. The data will come from six states (CA, FL, MI, NJ, PA, TX) representing almost 40 percent of U.S. births, and covering diverse geographic regions and populations. We will use the address data in the confidential versions of the vital statistics data to determine the distance between each mother's residence and the nearest Superfund site in 1/4 mile increments. Overall, this project aims to assess the consequences of exposure to the nation's hazardous waste sites for infant health. Preliminary results for Pennsylvania indicate that we will have the power to detect small changes in infant health outcomes. The results will be extremely relevant for the on-going debate about the extent to which cleanup of hazardous waste should be motivated by dangers to infant health. OTHER PROJECT INFORMATION - Project Narrative This project will provide new statistical evidence on the impacts of exposure to pollution from hazardous waste sites on infant health. The analysis is based on comparisons of infant health in the areas surrounding hazardous waste sites before and after the sites were cleaned up under the Superfund program. It will be conducted with the most comprehensive data ever assembled on Superfund sites and infant health.