THE USE OF STATISTICS IN SOCIOLOGY This essay discusses some widespread critiques of the use of quantitative methods in sociology, with Dag Osterberg's version of such a criticism as starting point. The claim that statistics necessarily imply a certain atomistic ontology is rejected on the basis that also quantitative sociology has to be interpretative, and that the interpretation of identical results may vary a great deal according to theoretical convictions. It is also argued that Osterberg's demands are impossible to accommodate for any empirically based sociology, and that his alternative to quantitative methods is not empirical research based on other methods, but social philosophy. The claim that quantitative methods necessarily imply a deterministic vision of man, and a faith in the possibility to predict human action, is also rejected: Firstly, because quantitative methods interrogate probabilistic relationships (and not deterministic). Secondly, because people are likely to repeat ones own and other people's actions not always, and not in anything near a lawlike fashion, but often enough to result in social and statistical (probabilistic) patterns.
