Applied Epistemology and Argumentation in Epidemiology

On other occasions I have argued that ‘informal logic’ should not really be seen as a kind of ‘weak’ form of logic, but rather as ‘applied epistemology.’ This categorization is intended to create an analogy with applied ethics. Applied ethics has created a robust research project and stimulated ethical thinking both in and outside philosophy. As with applied ethics, I believe that as philosophers explore the actual application of their principles and theories (ethical or epistemological) they will discover new insights into the powers and limitations of their theories. Application is not just about philosophy being ‘useful,’ it is also an intellectual and theoretical challenge and to a discipline that often suffers from undo abstraction. In this paper, I will argue that those who are interested in philosophy of science and applied epistemology should look not at physics but at epidemiology for a model of how a ‘hard’ science actually establishes causal claims. Epidemiology is a very epistemologically self-conscious and highly successful science. It is not characterized by the over arching laws à la Newton, nor does the Popperian principle of falsifiability work at all well within the discipline. Falsification is as elusive as proof, not only because epidemiology is fundamentally a stochastic science, but also because no experiment is sufficiently conclusive to falsify a claim. No one, though, would deny the enormous success of epidemiology in contributing to both an understanding of and enhancement of human health. While research in epidemiology is characterized by the use of elaborate statistical methods, claims are not established simply by the ‘statistically significant’ results of particular studies or experiments. This claim may seem surprising to anyone who has looked at medical research. Most such research uses the mathematical tests developed in statistics to assess the likelihood that a result of the study is ‘real’ -not merely a result of chance. But because of inevitable confounding factors and because few studies actually meet the random sampling criteria for the application of these statistical methods, researchers must still argue for the plausibility and significance of their results. When we observe the epistemological practices of epidemiology we see that a primary tool of this successful science is argumentation and judgement. Claims are established not by critical experiments or the confirmation of precise predictions, rather they are established (as they are in many sciences) by an evaluation of all relevant experimental and study results. Establishment of a causal claim typically involve arguments:  about quality and significance of results,