社会学方法論-計量的数理的アプローチを中心にして-

抄録

(1) Scaling, sampling and data analysis : a review
After the Second World War the methods of sociological researches in Japan were affected by those of American sociology. Most of Japanese sociologists mainly used the method of case study for sociological researches before the War, but they were not well acquainted with statistical methods. After the War Japanese sociologists had their great wish to develop sociology as an empirical science. They were interested in scaling, random sampling and data analysis in a word.
These sociologists constructed unidimensional scales in various studies such as social consciousness, social stratification and measuring the structure of rural community, etc. To improve good-poor analysis and Likert scale was suggested. Subsampling (especially probability proportionate sampling) was applied to large area surveys like a kind of a sampling survey of social stratification and mobility. Moreover, a panel survey was carried out for the study of election and a replicated sampling design was adopted in a survey of slums.
Sociological data are qualitative as well as quantitative. So, sociologists paid their attention to how to treat qualitative data. According to the development of computer and algorism, they utilized multidimensional scaling and other approaches, for example latent structure analysis, log-linear analysis and Hayashi's quantification methods, etc. Hayashi's quantification methods were systematically set up by C. Hayashi and his cooperators to clarify the interrelations among objects or items by the use of a score allocated to each category of qualitative items. In 1987 fuzzy quantification methods are shown by some experts of fuzzy set theory.
In short, Japanese sociologists laid the foundation of quantitative approaches for sociological researches in three decades after the War.
(2) A review of quantitative and mathematical approaches to the sociological researches of Japan since 1970
These approaches have mainly developed on the study of the social stratification and mobility especially concerning the data of the SSM survey since 1955.
The topics of these approaches are : 1) the problem of the consistency (or inconsistency) of the social stratification 2) the quantitative study of the social mobility and the analysis of the status formation process 3) the study of the use of the log-linear-model 4) the present condition and the prospects of the mathematical sociology in Japan.
The tremendous efforts have been made for the SSM survey and the analyses of these data. The multivariate analysis revealed a consistent trend of the social stratification of Japan, and further investigations were made as to the applications of various statistical models and how to use these models.
Yet in general, statistical or mathematical models have hardly generated and invented, and there have been few original studies (except for Yasuda's index) of quantitative and mathematical sociology in Japan. Accordingly, we hope for the 'originality', JAMS (Japanese Association for Mathematical Sociology), started in 1986, motto.