Northcott, Robert (2012) Partial explanations in social science. In: Kincaid, H. (ed.) The Oxford Handbook of Philosophy of Social Science. Oxford Handbooks in Philosophy. Oxford, UK: Oxford University Press, pp. 130-153. ISBN 9780195392753.
Abstract
How much was the increased murder rate explained by higher unemployment? What was the main cause of the American Civil War? Was it the penetrating offense or the stout defense that was most responsible for the football team’s victory? It is ubiquitous in social science and indeed everyday life that the causes we have identified explain some but not all of an outcome. In such cases, the question of critical interest is to quantify each cause’s contribution to the outcome. The focus is not on how general or deep or transportable a particular explanation or mechanism is, important though those concerns may also be, but rather is narrowly on how much a cause explains an effect in a particular one-off case. This is relevant historically to determine which factors explained an outcome most. It is also relevant as a guide to future intervention—which factors would influence an outcome most? Comparing different causes’ importance, and apportioning responsibility between them, requires making good sense of the 1notion of partial explanation, that is, i.e. of degree of explanation. This turns out to be a delicate task.1 The vast literature on defining causation itself is of no direct help because in the cases of interest here typically all parties already agree on what causes are present. The issue at hand is, rather, degree of causation, which is clearly distinct from mere causation simpliciter. It turns out to be very useful to make our concepts in this area explicit. What do partial explanations amount to, and, thus, what constitutes good evidence for them? How much is degree of causation subjective, how much objective? If the causes in question are probabilistic, how much is the outcome due to them and how much to simple chance? What is the role of contrasts? This chapter is split into four sections. Throughout, the emphasis will be primarily conceptual rather than epistemological. I begin by formulating the notion of degree of causation, or “effect size.” One particular understanding of this is standard across many sciences, and I relate it to influential recent work in the literature on causation. I use this understanding as the basis for my understanding in turn of “partial explanation.” In the second section, I examine to what extent mainstream social science methods—both quantitative and qualitative—succeed in establishing effect sizes so understood. The answer turns out to be, roughly: only to some extent. Next, the standard understanding of effect size, even though widespread, still has several underappreciated consequences. In the third section, I detail some of those. Finally, in the fourth section, I discuss the separate issue of explanandum-dependence, which is essential to assessing any cause’s explanatory importance and yet which has been comparatively neglected.
Metadata
| Item Type: | Book Section |
|---|---|
| School: | Birkbeck Faculties and Schools > Faculty of Humanities and Social Sciences > School of Historical Studies |
| Depositing User: | Robert Northcott |
| Date Deposited: | 06 Aug 2012 10:18 |
| Last Modified: | 02 Aug 2023 16:58 |
| URI: | https://eprints.bbk.ac.uk/id/eprint/4996 |
Statistics
Additional statistics are available via IRStats2.
Archive Staff Only (login required)
