Updated on 2022/05/21

写真a

 
SAIGO, Hiroshi
 
Affiliation
Faculty of Political Science and Economics, School of Political Science and Economics
Job title
Professor

Concurrent Post

  • Faculty of Political Science and Economics   Graduate School of Economics

  • Faculty of Political Science and Economics   Graduate School of Political Science

Research Institute

  • 2021
    -
    2022

    データ科学センター   兼任センター員

  • 1989
    -
     

    現代政治経済研究所   兼任研究員

Education

  •  
    -
    1992

    Waseda University   Graduate School of Economics   economic theory and economic history  

  •  
    -
    1992

    Waseda University   Graduate School of Economics   economic theory and economic history  

  •  
    -
    1984

    Waseda University   School of Political Science and Economics   division of economics  

Degree

  • 早稲田大学   経済学(修士)

  • Master in Economics

Research Experience

  • 1999
    -
     

    Professor, School of political science and economics, Waseda University

  • 1999
     
     

    Professor, School of political science and economics, Waseda University

  • 1994
    -
    1999

    Associate Professor, School of political science and economics, Waseda University

  • 1992
    -
    1994

    Assistant Professor, School of political Science & Economics, Waseda University

Professional Memberships

  •  
     
     

    The Canadian Statistical Society

  •  
     
     

    The American Stalistical Association

  •  
     
     

    The Japanese Economic Association

  •  
     
     

    The Japan Statistical Sciety

 

Research Areas

  • Economic statistics

  • Statistical science

Research Interests

  • Statistics

Papers

  • The independence assumption in the mixed randomized response model

    Hiroshi Saigo

    Statistics & Probability Letters   170   108986 - 108986  2021.03  [Refereed]

    DOI

  • 標本調査の基本理論 第6回

    西郷 浩

    経済統計研究   38号 ( III ) 42 - 48  2010.12

  • Comparing Four Bootstrap Methods for Stratified Three-Stage Sampling

    Hiroshi Saigo

    JOURNAL OF OFFICIAL STATISTICS   26 ( 1 ) 193 - 207  2010.03  [Refereed]

     View Summary

    For a stratified three-stage sampling design with simple random sampling without replacement at each stage, only the Bernoulli bootstrap is currently available as a bootstrap for design-based inference under arbitrary sampling fractions. This article extends three other methods (the mirror-match bootstrap, the rescaling bootstrap, and the without-replacement bootstrap) to the design and conducts simulation study that estimates variances and constructs coverage intervals for a population total and selected quantiles. The without-replacement bootstrap proves the least biased of the four methods when estimating the variances of quantiles. Otherwise, the methods are comparable.

  • 標本調査の基本理論 第5回

    西郷 浩

    経済統計研究(経済産業統計協会)   37 ( 1 ) 1 - 12  2009.04

  • 標本調査の基本理論 第4回

    西郷 浩

    経済統計研究(経済産業統計協会)   35 ( 4 ) 2 - 8  2008.03

  • 平成19年全国物価統計調査に期待する

    西郷浩

    統計   58 ( 9 ) 2 - 8  2007.09

  • 標本調査の基本理論 第3回

    西郷浩

    経済統計研究   35 ( 1 ) 15 - 24  2007.06

  • Mean - Adjusted bootstrap for two - Phase sampling

    Hiroshi Saigo

    SURVEY METHODOLOGY   33 ( 1 ) 61 - 68  2007.06  [Refereed]

     View Summary

    Two-phase sampling is a useful design when the auxillary variables are unavailable in advance. Variance estimation under this design, however, is complicated particularly when sampling fractions are high. This article addresses a simple bootstrap method for two-phase simple random sampling without replacement at each phase with high sampling fractions. It works for the estimation of distribution functions and quantiles since no rescaling is performed. The method can be extended to stratified two-phase sampling by independently repeating the proposed procedure in different strata. Variance estimation of some conventional estimators, such as the ratio and regression estimators, is studied for illustration. A simulation study is conducted to compare the proposed method with existing variance estimators for estimating distribution functions and quantiles.

  • 標本調査の基本理論 第2回

    西郷浩

    経済統計研究   34 ( 3 ) 15 - 23  2007.03

  • 標本調査の基本理論 第1回

    西郷 浩

    経済統計研究   34 ( 3 ) 101 - 108  2006.12

  • Bernoulli Bootstrap for Stratified Multistage Sampling

    Funaoka, F, Saigo, H, Sitter, R. R, Toida, T

    Survey Methodology   32 ( 2 ) 151 - 156  2006.12  [Refereed]

  • Jackknife variance estimator with reimputation for randomly imputed survey data

    H Saigo, RR Sitter

    STATISTICS & PROBABILITY LETTERS   73 ( 3 ) 321 - 331  2005.07  [Refereed]

     View Summary

    Reimputing after each deletion of a unit in the jackknife is known to lead to inconsistent variance estimators in the presence of random imputation in complex surveys. We use rescaling to adjust the reimputation and show that the resulting variance estimator performs somewhat better than the adjusted jackknife variance estimator. (c) 2005 Elsevier B.V. All rights reserved.

    DOI

  • 層別抽出のためのジャックナイフ法

    西郷浩

    早稲田政治経済学雑誌   ( 252-253 ) 217 - 251  2003

  • 平成14年全国物価統計調査の分析の視点

    西郷浩

    統計(日本統計協会)   ( 9月号 ) 32 - 37  2002.09

  • 複数の項目をもつ標本調査のための最適配分

    西郷 浩, 美添 泰人, 竹村 彰通, 坂巻 敏夫

    日本統計学会誌   32 ( 3 ) 303 - 314  2002  [Refereed]

  • 無回答補完の手法と推定量の精度評価

    西郷浩

    早稲田政治経済学雑誌   ( 349 ) 155 - 181  2002

  • 社会現象における交互作用の分析

    西郷浩

    早稲田政治経済学雑誌   ( 350-351 ) 131 - 157  2002

  • A Repeated Half-Sample Bootstrap and Balanced Repeated Replications for Randomly Imputed Data

    Saigo, H, Shao, J, Sitter, R. R

    Survey Methodology   27 ( 2 ) 189 - 196  2001  [Refereed]

  • 層別抽出における分散推定のためのBRR法

    西郷浩

    早稲田政治経済学雑誌   ( 341 ) 526 - 555  2000

  • Thompson, M. E. (1997), Theory of Sample Surveys, Chapman and Hall

    西郷浩

    日本統計学会誌   29 ( 3 ) 368 - 371  1999.12

  • 確率比例抽出における母平均の推定

    西郷 浩

    早稲田政治経済学雑誌   ( 338 ) 119 - 140  1999

  • 年齢別フロー率からみた雇用の動向(労働力調査の特別集計結果の報告)

    西郷浩

    総務庁統計局労働力統計課     1 - 83  1998.03

  • 有限母集団からの非復元無作為抽出のためのブートストラップ法

    西郷浩

    早稲田政治経済学雑誌   ( 334 ) 119 - 154  1998

  • 無回答を補完した調査データにもとづくノンパラメトリックなブートストラップ区間推定

    西郷浩

    早稲田政治経済学雑誌   ( 333 ) 215 - 247  1998

  • 確率比例非復元抽出における有限母集団修正

    西郷 浩

    早稲田政治絵経済学雑誌   ( 331 ) 194 - 221  1997

  • 無回答処理へのリサンプリング法の応用

    西郷浩

    早稲田政治経済学雑誌   ( 332 ) 160 - 195  1997

  • 誤差項の系列相関が漸次的に変化するスイッチング回帰モデル

    西郷 浩

    早稲田政治経済学雑誌   ( 315 ) 260 - 282  1993

  • 有限過去自己相関誤差回帰モデルにおける初期標本のウェイト

    西郷 浩

    早稲田政治経済学雑誌   ( 300 ) 282 - 311  1989

  • 自己相関誤差回帰モデルのベイズ推定

    西郷 浩

    早稲田政治経済学雑誌   ( 297,298 ) 308 - 338  1989

  • コクラン=オーカット法の効率について

    西郷浩

    早稲田政治経済学雑誌   ( 295-296 ) 313 - 341  1989

▼display all

Research Projects

  • Detection and handling of outliers in surveys

    Project Year :

    2017.04
    -
    2020.03
     

  • Research on statistical models for official statistics

    Project Year :

    2014.04
    -
    2017.03
     

     View Summary

    The objective of this research is to develop suitable statistical models for producing official statistics based on survey data. In particular, the present research focuss on (1)doubly protected imputation, (2)stratum jumpers, and (3)estimation under changes in the population. For (1), we made a presentation at the 2014 Japanese Joint Statistica Meeting. For (2), we have develped a model based on small area estimation. For (3), we currently construct a model to accommodate a shifting population

  • 公的統計の集計における統計モデルの利用に関する理論的・実証的研究

    科学研究費助成事業(早稲田大学)  科学研究費助成事業(基盤研究(C))

    Project Year :

    2014
    -
    2016
     

  • Research on statistical models for official statistics

    Project Year :

    2014
    -
    2016
     

     View Summary

    The objective of this research is to develop suitable statistical models for producing official statistics based on survey data. In particular, the present research focuss on (1)doubly protected imputation, (2)stratum jumpers, and (3)estimation under changes in the population. For (1), we made a presentation at the 2014 Japanese Joint Statistica Meeting. For (2), we have develped a model based on small area estimation. For (3), we currently construct a model to accommodate a shifting population

  • Cohort Analysis for Micro-data in Periodical Statistical Survey: Empirical Analysis of Cultural Demand

    Project Year :

    2011.04
    -
    2015.03
     

     View Summary

    This study empirically analyzes the factors affecting cultural demand such as arts participation and cultural expenditures. The data used in the analysis are based on the Survey on Time Use and Leisure Activities, National Survey of Family Income and Expenditure, Family Income and Expenditure Survey - which was conducted by the Statistics Bureau of Japan - and Survey on Japanese University Students’ Attitudes toward Arts and Appreciation of Arts conducted by some of co-investigators in this project. The results of the empirical analysis are as follows: 1) Regarding arts participation, comparing the effects of age, period, and cohort, the effects of cohort are relatively greater than that of age and period for some of arts participation. 2) The correlation between items of household expenditures, including cultural expenditures, is not stable over time and structural breaks in correlation exist. 3) Cultural capital significantly affects arts participation of college students

  • Methodology of inference on the population for economic survey

    Project Year :

    2011.04
    -
    2014.03
     

     View Summary

    The Economic Statistics Meetings and The Reading and Discussion on surveys were held on a regular basis. Three themes were studied through these meetings: (a) Missing in censuses were reviewed using Economic Census and Census of Commerce; (b) A model for stratum jumpers based on small-area estimation was developed and presented at Japanese Joint Statistical Meeting 2012. Quantile estimation under doubly protected hot deck imputation was studied through simulations and presented at Joint Statistical Meeting 2013 in Motreal as an invited paper

  • Dissemination of Official Statistics, Theory and Practical Analysis

    Project Year :

    2009.04
    -
    2013.03
     

     View Summary

    We conducted extensive micro data analysis using several sets of official statistics. Our research covered household classification, database construction for dissemination, analysis of "Survey on Time Use and Leisure Activities", etc. We presented some of our findings at several academic meetings, and published some papers in journals. We also proposed a desirable facility for micro data analysis open for the publi

  • Methodology of inference on the population for economic survey

    Project Year :

    2011
    -
    2013
     

     View Summary

    The Economic Statistics Meetings and The Reading and Discussion on surveys were held on a regular basis. Three themes were studied through these meetings: (a) Missing in censuses were reviewed using Economic Census and Census of Commerce; (b) A model for stratum jumpers based on small-area estimation was developed and presented at Japanese Joint Statistical Meeting 2012. Quantile estimation under doubly protected hot deck imputation was studied through simulations and presented at Joint Statistical Meeting 2013 in Motreal as an invited paper.

  • Methodology of inference on the population for economic survey

    Project Year :

    2011
    -
    2013
     

     View Summary

    The Economic Statistics Meetings and The Reading and Discussion on surveys were held on a regular basis. Three themes were studied through these meetings: (a) Missing in censuses were reviewed using Economic Census and Census of Commerce; (b) A model for stratum jumpers based on small-area estimation was developed and presented at Japanese Joint Statistical Meeting 2012. Quantile estimation under doubly protected hot deck imputation was studied through simulations and presented at Joint Statistical Meeting 2013 in Motreal as an invited paper.

  • Dissemination of Official Statistics, Theory and Practical Analysis

    Project Year :

    2009
    -
    2012
     

     View Summary

    We conducted extensive micro data analysis using several sets of official statistics. Our research covered household classification, database construction for dissemination, analysis of "Survey on Time Use and Leisure Activities", etc. We presented some of our findings at several academic meetings, and published some papers in journals. We also proposed a desirable facility for micro data analysis open for the public

  • Dissemination of Official Statistics, Theory and Practical Analysis

    Project Year :

    2009
    -
    2012
     

     View Summary

    We conducted extensive micro data analysis using several sets of official statistics. Our research covered household classification, database construction for dissemination, analysis of "Survey on Time Use and Leisure Activities", etc. We presented some of our findings at several academic meetings, and published some papers in journals. We also proposed a desirable facility for micro data analysis open for the public

  • Research on use of auxiliary information in making official statistics in Japan

    Project Year :

    2008
    -
    2010
     

     View Summary

    We studied use of auxiliary variables to provide accurate official statistics given statistical resources. In particular, we proposed doubly protected regression random imputation, compared four bootstrap methods for stratified three-stage random sampling through simulation, and examined the structure of questions in the U.S. Economic Census. The results have been presented at the annual joint statistical meeting in Japan and published in refereed journals.

  • Research on use of auxiliary information in making official statistics in Japan

    Project Year :

    2008
    -
    2010
     

     View Summary

    We studied use of auxiliary variables to provide accurate official statistics given statistical resources. In particular, we proposed doubly protected regression random imputation, compared four bootstrap methods for stratified three-stage random sampling through simulation, and examined the structure of questions in the U.S. Economic Census. The results have been presented at the annual joint statistical meeting in Japan and published in refereed journals.

  • The Time Series Analysis on the Effect Caused by Art-related Service Industry as a Result of the Growth of Japan's Service Industry

    Project Year :

    2007
    -
    2010
     

     View Summary

    The 5th Japan's Artists Survey was conducted to construct the total database including the four rounds surveys conducted during the past twenty years. Artists' annual income was analysed by their household patterns. The results were published at several academic meetings both in Japan and overseas. The use of ten percent sampled micro data of Japan's Population Census from 1985 to 2005 was granted just before the final project year. We constructed its database, and finished the initial data processing by the household patterns

  • The Time Series Analysis on the Effect Caused by Art-related Service Industry as a Result of the Growth of Japan's Service Industry

    Project Year :

    2007
    -
    2010
     

     View Summary

    The 5th Japan's Artists Survey was conducted to construct the total database including the four rounds surveys conducted during the past twenty years. Artists' annual income was analysed by their household patterns. The results were published at several academic meetings both in Japan and overseas. The use of ten percent sampled micro data of Japan's Population Census from 1985 to 2005 was granted just before the final project year. We constructed its database, and finished the initial data processing by the household patterns.

  • Studies on Incomplete Data in Statistical Surveys

     View Summary

    Throughout our research period in the past two years, we regularly held seminars once or twice a month. Besides the project memebers and a few graduate students, official statisticians from divisions of the government joined our seminar. The main results we obtained are as follows :1. Most members of the project participataed in collecting and summarizing previous resarch in the field of sampling from the finite population, mostly in the field of economic surveys. Our viewpoint was to develop a useful guideline to design new sample surveys, especially those conducted by the government when we take into account the effect of nonresponse explicitly.2. As for statistical analysis of incomplete data, we surveyed actual methods of correcting incompleteness of the data, employed by official statisticians. At the same time, we discussed theoretical aspects of estimation and evaluation of the sampling errors when nonresponse is present.3. The papers presented at the seminar are compiled, so that they are easily available for future studies

  • Statistical Methods of Micro data Analysis

     View Summary

    A practical problem of economic statistics is divided into (1) collection / processing stage of data and (2) an analysis stage. In particular you must examine reliability about observation when you analyze the micro data. The typical problem is to cope with incomplete data supplementing with external information. In this study, investigate technique to analyze the incomplete data which is supplemented by statistical matching.1. Some theory models and an experiment of statiscal matching based on the government examine statistics to do so, me compare complete matching and statiscal matching and investigate a possibility of disclosure of micro data to the public.2. As examples of incomplete data are missing values, termination of data collection, the curancy, and grouping of continuous data. And aecrecy of the individual information is a kind of incomplete data. In particular, we investigate the processing method of missing values.3. An application of bootstrap method to the case of incomplete data. Suggest a simple non parametric method of bootstrap to supplement incomplete data

  • THEORY AND APPLICATION FOR IMPROVEMENTS IN EFFICIENCY OF STATISTICAL SURVEY

     View Summary

    WE HAD RESEARCH MEETIGS ONCE A MONTH TO HEAR THE RECENT TRENDS IN STATISTICAL SURVEY FROM GOVERNMENT STATISTICIANS.TO IMPROVE EFFICIENCY IN STATISTICAL SURVEY, INTERNET AND OTHER MAGNETIC DEVICES CAN BE USED TO COLLECT QUESTIONNAIRES FASTER THAN NOW.TO USE COMPUTERS EFFICIENTLY IN STATISITICAL SURVEYS, FOLLOWIG PROBLEMS SHOULD BE EXAMINED FURTHER : (1)TO ESTABLISH COMMON CODINGS SYSTEM BETWEEN NATIONAL AND LOCAL GOVERNMENTS, (2) TO INTRODUCE AUTOMATICAL REPORTING SYSTEM TO FIRMS AND ESTABLISHMENTS.AS FOR MAK1NG STATISTITICAL TABLES EFFICIENTLY, FOLLOWIG PROBLEMS SHOULD BE EXAMINED FURTHER : (1) TO COMPILE AND USE AN OVERALL BUSINESS FRAME, (2) TO USE MICRO STATISTICAL LINKAGE FOR DIFFERENT SURVEYS.INTENSIVE COMPUTER EXPERIMENTS ARE NEEDED TO MAKE AN INQUIRY INTO ABOVE PROBLEMS BY USING DESIGNATED STATISTICS COMPILED BY DIFFERENT MINISTRIES.AS THE USE OF DESIGNATED STATISTICS IS RESTRICTED BY STATISTICS LAW, FURTHER RESEARCH IS NEEDED FROM THE VIEWPOINT OF LAW

  • Developing and Organising Statistical Indexes for Arts and Culture Policies

     View Summary

    (1) We have conducted comprehensive research in order to find out present situations on arts and culture in Japan. There are two major streams in our research ; one is the reuse and re-construction of existing government statistics on arts and culture, and the other is to conduct our own questionnaire surveys on both demand and supply of arts and culture.(2) We also went on a research trip to Australia to collect various kinds of materials on statistical surveys on arts and culture in Australia.(3) We made an analysis on daily time use and activities from the result of SURVEY ON TIME USE AND LEISURE ACTIVITIES.(4) We made an analysis on expenditure for arts and culture from the result of FAMILY INCOME AND EXPENDITURE SURVEY.(5) We made an analysis on arts education from the result of BASIC SURVEYON SERVICE INDUSTRIES.(6) For the last fifteen years we have conducted four periodical questionnaire surveys on (a) professional artists, (b) arts organizations and halls and (c) students demand on art activities. These database are re-organised for easier access, and then made a time-series analysis respectively.(7) We made a basic frame for 'Survey on Art NPO's in Japan' in order to grasp the actual situation of art NPO's

  • Application of recent statistical theory and use of IT in Japanese official statistics

     View Summary

    We held 7 meetings on the Japanese official statistics from April, 2002, to January, 2003. There we discussed (a) new regional price indexes in the National Survey on Prices ; (b) theoretical aspects of regional comparisons ; (c) prefecture-based unemployment estimator from the Labor Force Survey ; and (d) a new estimation method in GDP statistics. In particular, we proposed a new resampling method for the National Survey of Prices, which was presented at the annual meeting of the Japan Statistical Society. At the meeting, we also made two presentations on (a) use of micro data from the Family Income and Expenditure Survey and (b) new economic indicators. In 2002, we had no talks on use of IT in Japanese economic statistics at the regular monthly meeting.From April, 2003, to January, 2004, we had 7 meetings on the Japanese official statistics. This time, we discussed use of IT in Japanese economic statistics : (a) use of portable devices on the Retail Price Survey ; (b) IT statistics from the Survey on Time Use and Leisure Activities ; and (c) New Generation Statistics System by METI. At the same time, we continued our research on the resampling method for the National Survey on Prices. Our method was applied to NSP micro data and the results will be submitted to a refereed journal.In summary, we have obtained several satisfactory accomplishments on application of recent advanced statistical theory on surveys. Use of IT, on the other hand, will be studied continuously. We had obtained useful comments and suggestions from our monthly meetings

  • 世帯データの統計的分析基盤の整備と応用

     View Summary

    世帯データを中心として,官庁統計におけるミクロデータの利用可能牲が広がってきている。その中で,今後の広範な研究利用を実現するためには,いくつかの現実的な問題が解決されなければならない。特に,研究上の多様な要望に対応しやすい一次統計データベースの構造を検討すること,さらにミクロデータを提供するためのデータ処理に関わる作業をできるだけ効率的に設計するとが必要である。すでに初年度の研究において,世帯のミクロ統計データに対象を絞って,研究基盤を整備する効果的な条件を検討した。昨年度と今年度の研究においては、その成果に基づいて,家計調査および家計消費状況調査を対象として標本設計に関する詳細な情報を収集し,調査の特性を明らかにする作業を実施した。家計消費状況調査に関しては統計法上の目的外利用の手続きによって,ミクロデータを利用した分析を実施した。現在,その内容の取りまとめ作業に当たっている段階である。関連して,研究者の視点から要求される問題として,ミクロデータを分析する場合に必要なデータに関する情報の整備方法についても検討した。具体的な例としては,個票における欠測値ないし外れ値の発生などの不完全性と,それに対処するための補完ないし外部情報の利用に関する情報がある。世帯データの提供に密接に関連する問題である「統計データアーカイブ」については,諸外国の事例を参照しつつ,わが国の官庁統計に対する課題を検討している

  • Theoretical and Empirical Research for Organizing the System of Economic Statistics in Japan

     View Summary

    (1)Research Meeting on Economic StatisticsFrom May 2004 to January 2006, we have held 15 regular Research Meetings on Economic Statistics, including three conducted during May 2004 to September 2004, when they were not funded yet. Among the topics were : the first occasion of survey practice by a private survey company in Japan's official statistics ; a new estimation method for the monthly Labour Statistics taking its time series structure into account ; the new survey on small- and medium sized firms in Japan ; use of resampling methods and density estimation in the National Survey of Prices ; Economic Census in Japan ; and Business Register in US and UK. Particularly, we organized a registered session entitled "Reforming the Official Statistics System in Japan" at the Joint Statistical Meeting 2005 in Japan. Two key note speakers (Mr. Tatsuo Inoue and Professor Yoshiro Matsuda) and two discussants (Professor Hideo Fukui and Professor Kei Takeuchi) were invited to the session and many researchers and official statisticians participated to discuss.(2)OutputSaigo and Sitter (2005) proposed a residual-rescaling jackknife variance estimator with re-imputation for randomly imputed survey data. Yoshizoe (2005) discussed economic statistics in Japan in a historical perspective. Hiromatsu, et al. (2004) and Katsuura (2005) studied combination of multiple statistics for empirical analyses.In addition to articles appearing journals and books, we made several presentations at the Joint Statistical Meetings 2005 and 2006. Among them were : Research on Statistical Matching using Japan's Household Survey (by Yoshizoe) ; A Bootstrap Method for Two-Phase Sampling (by Saigo) ; Use of kriging for interpolating Tokyo's land prices (by Saigo and Masunari) ; and Effects of changes in questionnaire on survey outcome-An empirical study on Survey on Time Use and Leisure Activities (by Katsuura). Some of them are submitted for publication in research journals

  • Research of theoretical problems for collecting, disclosing, and utilizing official statistics

     View Summary

    This research issues related to the official statistics, this research was focused on issues directly related to improvement of the current survey method and also wide-range utilization of the survey results, and was intended to present theoretical solutions.The actual research activities were carried out jointly with other research groups Research meetings were held about once every fiscal year, and exchanges of opinions were carried out with researchers in peripheral regions and staff members from various government agencies involved with official statistics. The results of the research were announced by academic societies and other bodies in fields related to official statistics, such as the Japan Statistical Society.Along with the death of Satoru Kanoh (professor at the Institute of Economic Research, Hitotsubashi University)who planned this research, the research representative was replaced with Inaba in the final fiscal year. Kanoh configured this research organization after arguing that theoretical study concerning official statistics was an urgent research issue and that it was necessary to carry out total research on mutually related research topics. In May 2007, the Statistics Law was amended, and stipulations covering the provision of statistics by tabulation carried out under contract and the provision of survey sheet information concerning the measures to guarantee anonymity were newly enacted. The contents of these stipulations are directly related to the issues of this research, enabling the foresight and importance of this research planned by Kanoh to be confirmed. We believe that in order to use the results of this research for the disclosure and utilization of official statistics, it is necessary to carry out exchanges of opinions with various government staff members to a greater extent than before, and it is also desirable to carry out further research based on the results of these exchanges

  • Researches on Comprehensive Applications of Official Statistics

  • Census and Data Archive in the Japanese Official Statistics

     View Summary

    During the term of this project(ie., April, 2006-March, 2008), we had had 13 meetings on economic statistics(Economic Statistics Research Meetings, ESRM henceforth). Among them are : Use of the Japanese Population Census as a micro data(June, 2006) ; Use of private statistics for the Business Frame in Japan(October, 2006) ; Use of information from the National Tax Agency in the Japanese official statistics(May, 2007) ; Preparation of 2010 Population Census(October 2007) ; Recoding of and Access to US Population Census(October, 2007) ; Reforming the Japanese Statistical Law and data archive in the future(November, 2007) ; and Data archive in UK(November, 2007). In addition, we discussed nonsampling errors and panel surveys both of which are particularly useful for reuse of micro data through data archive. Moreover, we had two notable conferences jointly with the Japan Statistical Society : entitled The Japanese Official Statistics in the Future (September 2006, Tohoku University) and The Current Situation and Future Challenges of the Japanese Official Statistical System (September, 2007., Kobe University). Both of the conferences gathered more than 100 participants and provided useful keynote speeches and discussions about the rapidly changing Japanese official statistical system.Other than these 13 ESRMs, we had published 29 papers and made 10 presentations during April, 2006 to March, 2008

  • An empirical analysis of retail pricing and spatial price dispersion

▼display all

Presentations

  • Hanabusa Bunko

    the 10th Annual Meeting of the Japanese Federation of Statistical Science Associations 

    Presentation date: 2011.09

  • Doubly Protected Hot Deck Imputation

    the 8th Annual Meeting of the Japanese Federation of Statistical Science Associations 

    Presentation date: 2009.09

  • Random Regression Imputation Using Response Model

    the 7th Annual Meeting of the Japanese Federation of Statistical Science Associations 

    Presentation date: 2008.09

  • Comparing Four Bootstrap Methods for Stratified Three-Stage Sampling

    the 6th Annual Meeting of the Japanese Federation of Statistical Science Associations 

    Presentation date: 2007.09

  • Jackknifing for Sampling on Two Occasions

    the 5th Annual Meeting of the Japanese Federation of Statistical Science Associations 

    Presentation date: 2006.09

  • Boostrap for Two-Phase Sampling

    the 4th Annual Meeting of the Japanese Federation of Statistical Science Associations 

    Presentation date: 2005.09

  • Application of a Resampling Method for the 1997 National Survey of Prices

    the 3rd annual Meeting of the Japanese Federation of Statistical Science Associations 

    Presentation date: 2004.09

  • Analysis of Variation in Prices in the National Survey of Prices

    the 38th Statistical Seminar (the Japan Statistical 

    Presentation date: 2004.07

  • A Resampling Method for Multistage Stratified Sampling

    the 1st annual Meeting of the Japanese Federation of Statistical Science Associations 

    Presentation date: 2002.09

  • Oprimal Allocation for Stratified Sampling: a Multivariate Case

    the 69th Annual Meeting of the Japan Statistical Society 

    Presentation date: 2001.09

  • Optimal Allocation for Statistical Survey of Census of Commerce

    The 7th Japan-China Symposium on Statistics 

    Presentation date: 2000.10

  • Jackknifing With Re-imputation for Imputed Survey Data

    the 68th Annual Meeting of the Japan Statistical Society 

    Presentation date: 2000.07

  • Variance Estimation for Imputed Data

    the 65th Annual Meeting of the Japan Statistical Society 

    Presentation date: 1997.07

  • Dealing With Nonreponse in Surveys: the present and future

    the 64th Annual Meeting of the Japan Statistical Society 

    Presentation date: 1996.09

  • The Chow Test Under Serial Correlation

    the 62nd Annual Meeting of the Japan Statistical Society 

    Presentation date: 1994.07

  • Bayesian Inference of an AR(1) Error Regression Model

    The Third Japan—China Symposium on Statistics 

    Presentation date: 1989.11

  • Bayesian Inference of an AR(1) Error Regression Model

    the 57th Annual Meeting of the Japan Statistical Society 

    Presentation date: 1989.07

  • Efficiency of the Chochrane-Orcutt Method

    the 56th Annual Meeting of the Japan Statistical Society 

    Presentation date: 1988.07

  • Estimation of the Constant Term in Regression Under Correlation

    the 55th Annual Meeting of the Japan Statistical Society 

    Presentation date: 1987.07

▼display all

Specific Research

  • 統計調査における無回答処理の統計理論的研究

    1997  

     View Summary

    標記課題について、リサンプリング法の応用をサーベイし、いくつかの実験をおこなった。 無回答を補完した調査データにおける推定量の精度を評価するための方法には、これまでのところ3つの手法が考案されている:(1)修正 Jackknife 法;(2)Bootstrap 法;(3)Multiple imputation 法。実験の結果、適用範囲の柔軟さや精度評価の正確さから、Bootstrap 法が有用であるとの結論をえた。Bootstrap 法の利点のひとつは、母数のノンパラメトリックな区間推定がおこなえることである。このことは、無回答を補完した調査データの場合にも志向されるべきである。そこで、この点に関して、Bootstrap-t 法とBCa 法とについて具体的な手法を提案し、実験をおこなった。その結果、提案した手法は、母数の捕捉率が名目値を下回る傾向がみられるものの、母集団の歪みに対応しうることがわかった。 今後は、捕捉率低下の原因を理論的に解明して、それを是正した手法を研究していく。研究成果の発表:1997年10月「無回答処理へのリサンプリング法の応用」『早稲田政治経済学雑誌』第332号 160-195.1998年1月「無回答を補完した調査データにもとづくノンパラメトリックなブートストラップ区間推定」『早稲田政治経済学雑誌』第333号 215-247.

  • 確率比例系統抽出に関する基礎的研究

    1996  

     View Summary

     標記テーマについて、その理論的考察を中心に研究を進めた。まず、確率比例抽出における有限母集団修正について考察した。単純無作為抽出の場合と異なり、確率比例抽出における有限母集団修正は明示的な形ではとらえにくい。この点を、近似的な確率比例抽出による推定量(Rao-Hartley-Cochran推定量)の分散からの類推と、確率比例系統抽出における推定量(Horvitz-Thompson推定量)の分散の近似式(Hartley-Raoの近似)とから吟味した。その結果、前者の類推は基本的に単純無作為抽出の利用に由来し、どのような条件のもとでその類推が妥当であるのか検討が必要であることがわかった。 上で述べた最後の問題を、回帰型の超母集団を導入して考察した。その結果、この超母集団モデルのもとで、いくつかの推定量の分散が導出できた。しかし、この点については先行研究がおこなわれており、今回導いた結果と類似点・非類似点をさらに研究する必要がある。これは、今後の課題である。 このほか、確率比例抽出における最適標本配分についても部分的な考察を始めている。実用的な問題であり、今後検討の価値があるので、さらに考察を進めていきたい。

  • 経済構造の変化の統計的検出に関する研究(モデルの想定過誤との関連について)

    1995  

     View Summary

    経済構造の変化の有無を検定することは,経済モデルを構築する上で重要な意味を持つ。なぜなら,構造的な変化があったにもかかわらずこれを無視して予測をおこなえば,予測の精度が落ちることは避けられない。このため,計量経済学においては,経済構造の変化の検出方法を構築するのが重要なトピックのひとつとなっている。今日の非定常時系列分析においても,単位根の検定に構造変化が大きな影響をおよぼすことから,単位根検定と構造変化の検定とをいかにして組み合わせるかが最近の課題となっている。 従来の構造変化の検出の方法は,「構造変化が生じているという点を除けば,モデルの特定化に誤りがない」ことが前提条件である。ところが,経済分析においては,特定化された分析モデル自体が経験的に組み立てられることが大半であるから,特定化が正しいという前提も検定の対象とされるべきである。にもかかわらず,これまでの構造変化の検定では,モデルの特定化に配慮していたとはいいがたい。 この研究では,構造変化とモデルの特定化と同時に検定することによって,構造変化だけを検出する有効な方法を検討した。まず,構造変化の検出のために考案されている手法と関数形の特定化の検定に考案されている方法を理論的に整理した。双方とも関連文献は膨大であり,理論的な比較対比は容易ではない。今年はこの理論的な整理にとどまり,同時検定の手法については試論の域を出ていない。元来Non-nestedな関数形の検定をNestedな形式に変換して構造変化の検定に盛り込むことが基本的なアイディアである。具体的には,ダミー変数とRESET検定とを組み合わせて同時に使い,前者によって構造変化を,後者によって関数形の検定をおこなおうとするものである。有意水準や検出力を検討するためのシミュレーション・プログラムを試作した。理論的な検討は今年度の課題としたい。

 

Syllabus