Basic and Applied Statistical Inference and Measurement

基础和应用统计推断和测量

• 批准号: 9507926
• 负责人: Joseph Gastwirth
• 金额: $ 15.68万
• 依托单位: George Washington University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1995
• 资助国家: 美国
• 起止时间: 1995-08-15 至 2003-02-28
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9507926&HistoricalAwards=false
• 关键词: Basic; Applied; Statistical; Inference; Measurement

Statistical information plays an important role in establishing the facts upon which public policies are based. For example, the levels of employment and unemployment and other important economic and health statistics are obtained from sample surveys. Statistical evidence is also used in legal cases, especially those concerning equal pay or employment discrimination and trademark infringement. Since there are some imperfections in most real world data sets (e.g., a question in a survey may not be clear to all respondents or it is difficult to obtain information about some potentially relevant variables), it is important to develop methods for assessing the accuracy of the data used in statistical analyses as well as techniques which are less sensitive to modest errors in the data. One goal of the project is to further develop a new approach to determining the misclassification rates in the labor force survey. The effect of these errors on the reported rates of unemployment and employment will be studied. This should enable policy makers to adjust for these misclassification errors when they interpret the monthly economic statistics. Often reasonably sound studies are questioned because some variables that might influence the outcome under investigation (e.g., the efficacy of a drug or educational program) were not included. Methods for assessing the potential impact of an omitted variable are being developed and applied to data arising in legal cases, economics, and public policy. Quite often it can be shown that the main conclusion of the study would not be changed so that action, such as providing a public warning about a potentially harmful drug or introducing a better educational method, can be implemented with further delay and study.

统计信息在确定公共政策所依据的事实方面发挥着重要作用。 例如，就业和失业水平以及其他重要的经济和卫生统计数据都是从抽样调查中获得的。 统计证据也用于法律的案件，特别是涉及同工同酬或就业歧视和商标侵权的案件。 由于在大多数真实的世界数据集中存在一些不完美（例如，调查中的一个问题可能并非所有答复者都清楚，或难以获得关于某些潜在相关变量的信息），因此，必须制定评估统计分析所用数据准确性的方法，以及对数据中的微小误差不太敏感的技术。 该项目的一个目标是进一步开发一种新的方法来确定劳动力调查中的错误分类率。 将研究这些误差对所报告的失业率和就业率的影响。 这将使政策制定者在解读月度经济统计数据时能够对这些错误分类进行调整。 通常，合理的研究受到质疑，因为一些变量可能会影响研究结果（例如，药物或教育计划的功效）不包括在内。 正在开发评估遗漏变量的潜在影响的方法，并将其应用于法律的案件、经济学和公共政策中产生的数据。 很多时候，可以表明，研究的主要结论不会改变，以便采取行动，如提供一个潜在的有害药物的公共警告或介绍一个更好的教育方法，可以进一步延迟和研究。