Comparative Social Research: Issues in Comparability

Abstract

Methodological debates in the comparative studies field have become polarized. On the one hand, philosophers such as Winch, Quine and Louch have argued for different reasons that scientific method is inappropriate for social science research. Winch (1958) rules out the possibility of cross-cultural comparison, claiming that outside observers are unable to understand the meaning of rule-governed behaviour, which is dependent on the motives, reasons and decisions of participants in a particular society or culture. Quine's 'translational indeterminacy' thesis implies that equivalence of meaning across cultures is impossible. Louch (1969) argues that observer values will inevitably preclude objective scientific investigation, since social science concepts are moral, not scientific, and that generalization is precluded in any case because activity occurs in such varied contexts. On the other hand, it has been maintained that scientific method in comparative research is conditionally possible, given improved accuracy of measurement (Deutsch 1966), testing of the equivalence of meaning (Levison 1974), or moving to an appropriate and lower level of generalization than scientific laws (MacIntyre 1971). Macintyre argues that the research questions should ask what functions institutions serve in the culture, rather than, as Winch suggests, what institutions mean for participants. With such a move from a subjective towards an objective frame of reference, theory-building becomes possible. The views which support scientific method in social science research are persuasive on both philosophical and practical grounds. Those philosophers who argue that comparative research is impossible, have not only produced incomplete evidence and flawed argument, they have failed to notice that the gap between natural and social science is in practice far narrower than in theory: natural science is subject to uncertainty, based on the difficulty of disentangling causal chains and on the variability of observer perceptions, which is quantitatively, but not qualitatively, different from the uncertainty of social science research. Polarization seems endemic in social science methodology: fact vs meaning; structuralism vs phenomenology; quantitative vs qualitative method. Such debates are likely to continue. Therefore,