Center Director: Yoshiaki Kobayashi (Professor, Faculty of Law)
Main Campus: Mita
Center Overview
The research objective of this center is to contribute broadly to researchers in Japan and abroad by building a social science data archive, which is an urgent task for empirical analysis in the social sciences. Specifically, while supplementing the vast amount of data accumulated by the principal investigator and co-investigators—such as census data by municipality mesh, election results by municipality, election pledges, laws and judicial precedents, minutes of the National Diet and all prefectural assemblies, and long-term opinion surveys—we will build a comprehensive data management system. This system will enable multilingual searches in seven languages (Japanese, English, Chinese, Korean, Russian, Indonesian, and Malay), making it accessible to international students and researchers from abroad.
Through this research, we aim for the database to be widely used as the largest database on Japanese politics by researchers not only in political science but also in law, economics, sociology, and other fields, both in Japan and abroad, thereby serving to establish an infrastructure for Japanese studies. Furthermore, by accumulating diverse data into a single data archive, we aim to acquire new knowledge through the integration of data. As a contemporary demand for the construction of a social science data archive, although empirical analysis of social phenomena has become widespread, further development requires the creation of a data archive where data necessary for empirical analysis is accessible to everyone.
The reasons for this include the following:
When submitting to top-tier international journals, the disclosure of data that allows for the replication of the paper's findings is often a requirement.
Data is often lost upon the completion of research periods or retirement, leading to inefficient use of research funds.
Valuable data, such as municipal administrative materials from municipal mergers and judicial precedents after a certain period, is being lost.
Amid the rapid internationalization of academia and the push for data archiving, the difficulty in accessing data related to Japan has become a contributing factor to the decline of Japanese studies overseas. In the natural sciences, projects such as the sharing of related data by academic societies and the bioresource backup project by the National Institute for Basic Biology have been successful. In the humanities, the project to build an international collaborative research network for historical Japanese texts by the National Institute of Japanese Literature is also progressing. In contrast, because the social sciences do not have an Inter-University Research Institute Corporation, a comprehensive data archive including aggregate data has not been built, with efforts limited to partially collecting opinion survey data. To solve these problems, the research objective of this center is to collect data on census results, election outcomes, minutes of the National Diet and local assemblies, opinion surveys, laws, and judicial precedents from Japan and other countries, manage it in an XML database, and build a data archive with a multilingual (Japanese, English, Chinese, Korean, Indonesian, Malay, and Russian) search function to make it accessible to everyone.
In collaboration with the Subcommittee on Political Process, Committee on Political Science, Science Council of Japan, Cabinet Office, and the MEXT Grant-in-Aid for Scientific Research (A) "Construction and Expansion of a Political Data Archive," we held a symposium on the construction and expansion of a social science data archive and presented our research findings both in Japan and internationally.
Keywords and Main Research Themes
Data archive, Democracy, Multilingual search
FY2019 Business Plan
■ Continuing Activities from the Previous Fiscal Year: Background, Rationale, and Goals
To explore survey methods that can replace face-to-face interviews, which lack responsiveness, we will compare three survey methods—face-to-face, mail, and internet—to determine whether internet or mail surveys have less bias compared to face-to-face interviews, and whether this bias is statistically significant. If internet surveys have less bias than mail surveys compared to face-to-face interviews, and this bias is not statistically significant, the superiority of internet surveys as an alternative to face-to-face interviews will be established. We will then experiment to see how much the bias between internet and face-to-face surveys can be minimized by correcting for biases in the attributes of the collected sample.
Specifically, based on a comparison between face-to-face surveys conducted in previous fiscal years and internet surveys with similar questions:
01. If the difference from face-to-face surveys is due to sample bias, we will control the collected sample using a triple quota based on social attributes (gender, age, urban scale of residential area) to match the distribution of social attributes in the face-to-face survey's collected sample as closely as possible.
02. If the difference from face-to-face surveys is due to self-selection bias, we will use a multi-stage sampling method. Then, for each segment combining urban scale, gender, and age, we will estimate the bias in political attitudes (e.g., voting participation, party support, political satisfaction) between face-to-face and internet surveys, and between face-to-face and mail surveys, using the data collected from the aforementioned surveys.
If the resulting bias between the methods is within a statistically acceptable range, internet surveys will be established as an alternative survey method to face-to-face interviews. An important objective of this research is to determine the optimal sampling method for this purpose.
In case 01, the advantages of internet surveys will be fully retained. In case 02, advantages such as implementation costs (excluding sampling costs) and duration will still be retained.
■ New Activity Goals, Content, and Background for FY2019
Install the election results by municipality and candidate for the 25th regular election of the House of Councillors in 2019.
Conduct the JES VI opinion survey targeting voters nationwide for the 25th regular election of the House of Councillors in 2019.
If the 49th general election of the House of Representatives is held concurrently with the 25th election of the House of Councillors in FY2019, install the election results by municipality and candidate for that election.
In collaboration with the Subcommittee on Political Process, Committee on Political Science, Science Council of Japan, Cabinet Office, and the MEXT Grant-in-Aid for Scientific Research (A) "Construction and Expansion of a Political Data Archive," hold a symposium on the construction and expansion of a social science data archive and present research findings both in Japan and internationally.
FY2019 Business Report
■ Implementation Details, Research Outcomes, and Degree of Achievement for the Fiscal Year's Business Plan
In FY2019, the following new activities were conducted.
We installed the election results by municipality and candidate for the 25th regular election of the House of Councillors, held on July 21, 2019.
In conjunction with the 25th regular election of the House of Councillors held on July 21, 2019, we conducted the JES VI opinion survey targeting voters nationwide. Specifically, as a pre-election survey for the House of Councillors election, we conducted the JES VI Wave 2 survey from July 13 to 15. The survey subjects were men and women aged 18 and over nationwide. After stratification by residential area (block) and urban scale based on the instructions of the Study Group on Voting Behavior, quotas were set for gender and age, resulting in 3,000 valid responses. Furthermore, as a post-election survey, we conducted the JES VI Wave 3 survey from August 2 to August 4. This was a panel survey targeting the respondents of the JES VI Wave 2 survey, with approximately 1,800 valid responses. The 49th general election of the House of Representatives was not held in conjunction with the 25th election of the House of Councillors in FY2019.
Through a multi-method comparison of opinion surveys, we developed a survey method that can replace conventional face-to-face interviews and published the research results in the 2019 issue of "Hogaku Kenkyu" (Journal of Law, Politics and Sociology).
With the change of the era name from "Heisei" to "Reiwa" on May 1, 2019, many of the 47 prefectural assembly minutes publication systems were modified. We responded by changing and updating our data archive's automatic collection system.
On October 31, 2019, we presented a report on "Building a Data Archive in the Social Sciences" at the "Committee on the Examination of the Deepening and Promotion of Open Science," a committee by issue of the Science Council of Japan, Cabinet Office (held at the Science Council of Japan). This was intended to disseminate the project's results to all cross-disciplinary fields, going beyond the initially planned Subcommittee on Political Process, Committee on Political Science. President Yamagiwa of the Science Council of Japan was also present and engaged in enthusiastic discussion. Based on the report from that day, the committee decided to issue a "Proposal: Proposal for the Deepening and Promotion of Open Science" from the Science Council of Japan, and we were invited to participate in its drafting.
On June 22, 2020, an academic exchange symposium was held by faculty from the Keio University Faculty of Law and the Yonsei University College of Social Sciences. This project was introduced and received great interest from Yonsei University affiliates.
On February 21, 2020, a symposium was held based on the results of joint research with the Graduate School of Public Administration (GSPA) at Seoul National University in South Korea to discuss strategies and challenges for building a data archive in the social sciences and promoting open science. In addition to Center Director Kobayashi, the symposium was attended by co-investigators Takeshi Iida (Professor, Faculty of Law, Doshisha University) and Yuta Kamahara (Associate Professor, Yokohama National University), as well as Dean Im and many other professors from the GSPA at Seoul National University, and two professors from Hungary.
Based on the above, we were able to exceed our business plan and achieve our goals for FY2019.
Achievements in Social Contribution, including Published Papers, Conference Presentations, and Events
Number of published papers: 18
Main publication venues: The International Journal of Community Well-Being, Japanese Journal of American Studies
Number of conference presentations: 14 (6 domestic, 8 international)
Events, etc.: Joint Conference with GSPA, Seoul National University, held at the Graduate School of Public Administration, Seoul National University on February 21, 2020
Special Achievements through Center Activities
Traditionally, face-to-face interviews and mail surveys have been commonly used to investigate human social behaviors such as consumer behavior, social consciousness, and political attitudes. However, since the enactment of the Act on the Protection of Personal Information, the attitudes of potential survey subjects have changed, with an increasing number of people refusing to be surveyed when visited by an interviewer. Furthermore, the recent increase in apartment buildings with auto-lock security has led to more cases where building managers deny entry to interviewers, preventing them from even meeting potential subjects. Additionally, as more companies switch from less responsive face-to-face interviews to telephone surveys for consumer behavior research, it has become difficult for interviewers to make a living as they once did, making it hard to secure interviewers without paying higher compensation. These combined factors have caused the cost of face-to-face interviews to soar, making it practically difficult to conduct them without top-tier funding such as MEXT's Specially Promoted Research, New Academic Fields, or Grant-in-Aid for Scientific Research (S). The decline in response rates for face-to-face interviews due to these circumstances also introduces response bias. Generally, if the response rate falls below 60%, the difference between the population and valid respondents becomes large, potentially casting doubt on the survey's reliability. To secure a certain response rate, the survey period for face-to-face interviews has become prolonged (e.g., 16 days including three weekends), which can lead to changes in circumstances between the beginning and end of the survey. This raises the problem of analyzing data collected over a long period as a single dataset. Therefore, in this project, we developed our own multiple quota method to ensure that there is no statistically significant difference between the results of internet surveys and face-to-face surveys using the same questions. As a result, the internet survey with multiple quotas conducted in this study can not only serve as an alternative to face-to-face interviews but also leverage the advantages of internet surveys to respond quickly to events and be conducted at a lower cost. This enables young researchers and graduate students without large research grants to compete on an equal footing with senior researchers. The results of this research were presented at the 2019 annual meeting of the Japanese Political Science Association on October 5, 2019, and received a significant response.
Project Members
Principal Investigator
Yoshiaki Kobayashi
ProfessorFaculty of LawKohei M. Itoh
ProfessorFaculty of Science and Technology