Exploring New Frontiers in Statistical Analysis using Micro Data Sets

1980年代，欧洲和北美逐渐向公众和学术界提供微观数据集的政府调查统计数据。然而，日本当局严格限制微观数据集的使用。为了跟上这一国际趋势，统计理事会于1995年建议采用更灵活的方式使用保护隐私和商业秘密的微数据集。实现这一政策存在许多问题和障碍。我们的项目是在三个研究议程下，从院士那里寻求实践和理论上的解决方案，以实现这一目标：1)统计和计算的理论方面，2)从政府调查中获得的微观数据集的实验分析，3)使用微观数据集的法律和制度方面。1)根据日本统计法，我们的项目申请了“指定统计”的特别许可，使用由日本统计局控制的重要统计调查，并使用这些数据进行数据分析。我们获得了5个部门的17个统计数据的使用许可。2)秘书处小组处理了这些微数据集，作为订制制表的中心。1999年，我们的项目秘书处组对整个项目进行了总结，并开始了新的数据处理程序，使用了全国家庭收支调查和就业状况调查中20%的重新抽样数据，参与了这个新项目的18名研究人员。我们的项目开发了一个系统，在这个系统中，项目的每个参与者都可以交换信息，而不会泄露数据集中的私人信息。此外，我们从法律和实践两个方面对各发达国家统计调查微观数据集的披露现状进行了全面调查。我们的计划于一九九九年七月二五日至五日举行研讨会，介绍过去三年进行的个别计划的最终结果，向政府统计人员展示我们以研究方式保障私隐的有用性和安全性。作为总结，我们发表了“利用微数据集探索统计分析新领域”项目的总结报告（日语和英语）。为了供公众使用，我们计划出版五个关于微观经济统计的系列讲座，前三卷已于2000年2月出版。少

In the 1980s, government survey statistics in micro data sets were gradually made available for public and academic use in Europe and North America. However, Japanese authorities strictly restrict the use of micro data sets. In order to catch up such an international trend, Statistics Council made recommendation to adopt a more flexible way of using micro data sets protecting privacy and business secrets in 1995. There are many problems and obstacles to realize such a policy. Our project is to set up to pursue practical and theoretical solutions for this goal from the academicians and organized under the three research agenda : 1) theoretical aspects of statistics and computation, 2) experiments to analyze micro data sets obtained from governmental survey, 3) legal and institutional aspects of using micro data sets.1) Under the Japanese Statistics Law, our project applied for special permission to use "designated statistics", important statistical surveys controlled by the Statistics A … More ct, and conducted data analyses using these data. We got permissions of using seventeen statistics from five ministries and agencies.2) The secretariat group had processed these micro data sets as a center for order-made tabulation. In 1999, the secretariat group of our project summed up the whole project and started new data handling procedure by using 20%, re-sampled data from National Survey of Family Income and Expenditure and Employment Status Survey. 18 researchers were involved in this new project.Our project developed the system in which each participant of the project could exchange information without leaking private information from the data sets.In addition, we conducted the comprehensive survey on the current situation of the disclosure of the micro data sets of the statistical surveys in the various advanced countries, both from legal and practical aspects.Our project held the seminar on July 2-5, 1999, presenting the final results of the individual projects conducted in the past three years to demonstrate usefulness and safety of protecting privacy by our way of research against the statisticians of the government.As a conclusion, we published the summary report of our project on "Exploring New Frontiers in Statistical Analysis Using Micro Data Sets" (both in Japanese and English). For public use, we plan to publish the five lecture series on microeconomic statistics, and the first three volumes were published in February 2000. Less