著者

青井 和夫

出版者

日本社会学会

雑誌

社会学評論 (ISSN:00215414)

巻号頁・発行日

vol.6, no.1, pp.29-50, 1955-07-30

---

1. Although the 'Experimental Method' has played a central role in natural scientific investigation, in the field of the social sciences it has remained in a primitive stage. There are many approaches to the problem of experimental methods in sociological research, and marked differences in this respect between German, French and American methodology. <BR>(1) German methodology (as, especially, in M. Weber's Verstehende Soziologie or H. Freyer's Soziologie als Wirklichkeitswissenschaft) has emphasized the specific character of sociology and asserted that the essential purpose of sociology can be attained only if the experimental method is excluded from the methods of sociological research. This stand-point may be called the 'essential (ist) ' approach to experimental methods. <BR>(2) By contrast, French methodology (as, for example, in E. Durkheim's Les Regles de la Methode Sociologique) has directed attention to the process of logical inference and consequently, following J. S. Mill's system of logic, has claimed that sociological experiment is impossible. (Essentially, Durkheim's 'concomitant method' or 'comparative method' is not experimental.) This standpoint may be called the 'formal logical' approach. <BR>(3) Lastly, in American methodology (see, for example, G. Lundberg's Social Research) 'the methods of social research' means 'the techniques of gathering data' and methodology begins and ends with the consideration of methods of research and has thus been largely concerned with the technical difficulties of sociological experiment. This may be called the 'technical' approach. <BR>2. The origin of these different standpoints may be traced back to the traditional character of sociology in these countries and to the peculiar nature of their social structures, but, at all events, these have been the main barriers to experimental sociology. <BR>What, then, is experiment ? An experiment consists of two elements : the 'verification of hypotheses' and the 'creation of experimental conditions'. The former is the theoretical aspect and the latter the practical aspect. If we neglect either one we shall fail not only to grasp the essential nature of experiment (as, for example, did E. Greenwood, who neglected the practical aspect and included what he calls 'ex-post-facto experiment' in the category of experiment) but also to understand the nature of inference from individual instances to generalisations, the advance from correlation to causation, or the practical significance of experiment. From these points of view, we must, I think, include only Goode and Hatt's 'quasi-experiment', French's 'field experiment', Greenwood's 'projective experiment' and 'stochastical experiment', in the category of experiment proper, and must exclude from it 'natural experiments' ex-post-facto experiment', 'trial and error experiment' and 'controlled observation study. (These are preliminary stages to experiment itself.) <BR>3. The aim of experimental procedure is the determination of geno-typical phenomena or relationships ; the discovery of 'laws'. If such 'laws' can be discovered they will not only influence sociological concepts and theories, they will bring sociology itself out of the stage of verstehen or 'interpretation' to that of 'application'. For the 'futility' of sociological studies depends not only on their purpose, but also on the character of traditional sociological theories and the type of 'laws' they have been concerned to establish.