Mathematical Sciences: Robust Estimation and Testing of Econometric Models for Panel Data

该奖项将支持计量经济学和统计学领域的合作努力。特别是，调查人员将为面板数据开发强大的技术。面板数据由对几个人中每一个人的观察的时间序列组成。这样的数据在经济学和某些生物统计学的背景下经常被使用。面板数据的标准分析方法对数据中的粗略测量误差不是很稳健，基于这些方法的分析结果可能对粗略误差高度敏感。使用《当前人口调查》等来源的小组数据的调查人员对此非常了解。在评估模型和检验假设之前，这些研究人员通常会使用各种非正式的方法来“清理”数据。提出的研究将证明这一点的正式方法。该奖项将支持计量经济学和统计学领域的合作努力。特别是，调查人员将为面板数据开发强大的技术。面板数据由对几个人中每一个人的观察的时间序列组成。这样的数据在经济学和某些生物统计学的背景下经常被使用。

This award will support a collaborative effort in the areas of econometrics and statistics. In particular, the investigators will develop robust techniques for panel data. Panel data consist of a time series of observations on each of several individuals. Such data are used frequently in economics and certain biostatistical contexts. Standard methods for analysis of panel data are not robust to gross measurement errors in the data, and the results of analyses based on these methods can be highly sensitive to gross errors. This is well known to investigators who use panel data from sources such as the Current Population Survey. These investigators routinely use a variety of informal methods to "clean" the data before estimating models and testing hypotheses. The research proposed will proved formal methods for doing this. This award will support a collaborative effort in the areas of econometrics and statistics. In particular, the investigators will develop robust techniques for panel data. Panel data consist of a time series of observations on each of several individuals. Such data are used frequently in economics and certain biostatistical contexts.