Critical realism and statistical methods—a response to Nash

Abstract

This article offers a defence of critical realism in the face of objections Nash (2005) makes to it in a recent edition of this journal. It is argued that critical and scientific realisms are closely related and that both are opposed to statistical positivism. However, the suggestion is made that scientific realism retains (from statistical positivism) a number of elements that result in misleading accounts of social processes and events: indicators are used which do not reflect the close relationship between structure and agency; indicators refer to reified and not real properties of both structures and agents; and indicators do not refer to causal properties of objects and entities. In order to develop a narrative of causal processes, as Nash argues researchers should, then some adjustments need to be made to the principles that underpin scientific realism.

Acknowledgements

I am grateful to Peter Ribbins and two anonymous referees for their constructive criticisms of this article which helped me to sharpen its focus.

Notes

1. The term ‘positivism’ has acquired a variety of meanings, and is even used by some theorists as a term of abuse. Outhwaite (1987) has suggested that there are three principal varieties of positivism, the last of which is relevant in this context. The first, popularised by Auguste Comte, is where causal laws can be derived from observations and these observations are value‐free. However, this does not imply that a common method for the natural and social sciences can be developed. The second variant, known as logical positivism, espoused a form of nominalism, and at the same time suggested that the methods of the natural sciences could be applied to the social sciences. Finally, the third variant, variable analysis, led to the development of statistical explanations for social phenomena in the form of universal laws or generalisations, constructed from the constant conjunctions of events. This third variant has been critiqued extensively by, amongst others, critical realists who have developed a social theory based on a depth and stratified ontology.

2. Claims have been made that certain types of physical phenomena change their form as a result of being observed or investigated. However, no claims have been made that these physical phenomena consciously do so, through a process of reflection and transformation. Social phenomena have this capacity and this can be expressed as learnt behaviour during the process of investigation, as well as reflection on and transformation of those categorical distinctions that structure the social world.