2003 年 72 巻 p. 151-169
C. Wright Mills once criticized Lazarsfeld's work as abstracted empiricism. I understand the current state of the sociology of education in Japan as resembling this kind of empiricism. We want the sociology of education to be policy science or applied science, and empirical data, especially numerical values, seem to be useful. However, inside of the field of sociology, sociologists are divided according to their methodologies, so there is no common language or communication method. If they insist on one methodology, especially statistics or multivariate analysis, they will lose sight of the linkage between the empirical data and sociological theory, or forget the sociological significance for using that type of method. In addition, they may make statistical errors because the packages (for example, SPSS) enable people who do not understand statistical theory or are poor at mathematics to use statistical methods, and statistical methods have improved rapidly.
Incidentally, sociologists who want sociology to be a policy science take some risk. Sociological data seem to be objective and scientific. If they want sociology to be useful, sociology will be influenced by power because power decides the criterion for usefulness. Ideologies always try to take advantage of objective and scientific data and theory. Furthermore, sociological knowledge, which seems to be objective and scientific, influences social conditions, sometimes changes people's actions and causes self-fulfilling prophecies. We sociologist must always be conscious of that kind of problem.