Realism and probability

Malcolm Williams*, Wendy Dyer

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

11 Citations (Scopus)

Abstract

Although most discussions of methodological positions in the social sciences will cite realism alongside positivism and interpretivism, in empirical practice the first position is uncommon. The best-known realist programme in social science, Roy Bhaskar's critical realism, has many supporters yet, despite its commitment to scientific naturalism, it has had little or no impact on quantitative research in social science. The only important exception has been Ray Pawson's (1989) A Measure for Measures, which is a reconstruction of survey method along realist lines. Now there may be good sociological reasons for this lack of realist penetration into quantitative research, but there are also good ontological grounds for a rejection (by quantitative researchers) of critical realism as it currently stands. In this paper we will defend the overall project of scientific realism in social research, but we will be critical of its current principal manifestation (of critical realism) in one crucial respect, that of necessity and what this then implies for probability. We will claim that the particular view of necessity critical realists adopt leads them to criticise empiricist versions of probability, but to offer nothing back in return. We take it as axiomatic that quantitative research needs some theory of probability in order to be able to explain and predict.
Original languageEnglish
Title of host publicationMaking realism work
Subtitle of host publicationrealist social theory and empirical research
EditorsBob Carter, Caroline New
Place of PublicationLondon
PublisherRoutledge
Chapter3
Pages67-86
Number of pages20
Edition1st
ISBN (Electronic)9780203624289
ISBN (Print)9780415347716, 9780415300612
DOIs
Publication statusPublished - 19 Aug 2004
Externally publishedYes

Fingerprint

Dive into the research topics of 'Realism and probability'. Together they form a unique fingerprint.

Cite this