Detection and handling of outliers in surveys

2019 Fiscal Year Final Research Report

• Project/Area Number: 17K03665
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Economic statistics
• Research Institution: Waseda University
• Principal Investigator: Saigo Hiroshi 早稲田大学, 政治経済学術院, 教授 (00205626)
• Project Period (FY): 2017-04-01 – 2020-03-31
• Keywords: 統計調査法 / 外れ値 / 世帯調査 / 事業所調査

Outline of Final Research Achievements

We focused on methods for detecting and handling outliers in survey statistics. To obtain practical information on outliers in real survey data, we organized a meeting called Economic Statistics Meeting on a regular basis, four meetings in a year, twelve in three years. In the meeting, we discussed survey statistics in which outliers occur frequently and influence estimation tremendously through questions and answers with those who make the statistics. These discussions are a source of our research. Some papers are presented in Statistics published by the Japan Statistical Association. Currently, two manuscripts are under review in refereed journals.

Free Research Field

経済統計

Academic Significance and Societal Importance of the Research Achievements

統計調査における外れ値は、集計結果に影響を及ぼす。その的確な処理は、調査結果の集計精度向上に役立つ。統計調査における外れ値は、調査対象（世帯調査・事業所調査）に応じて性質が異なる。また、外れ値の発生頻度は、統計基準（産業分類など）によっても左右される。数理的な性質とともに、統計調査の周辺の条件も考慮しながら外れ値の検出と処理方法を研究することが本研究の特徴である。この研究の結果、統計調査の集計結果の安定することが期待できる。