Experiments offer more reliable evidence on causation than observational studies, which is not to gainsay the contribution to knowledge from observation. Experiments should be analyzed as experiments, not as
observational studies. A simple comparison of rates might be just the right tool, with little value added by “sophisticated” models. This article discusses current models for causation, as applied to experimental and observational data. The intention-totreat principle and the effect of treatment on the treated will also be discussed. Flaws in perprotocol and treatment-received estimates will be demonstrated.
If you read this article, tell us what you thought!
Awful title (tells you virtually nothing about the argument), obscure journal… but this was a great article. Written with a Wittgensteinian sort of precision, and extremely clear and accessible, this is a fascinating introdcution to the endless debates that rage between social and biological scientists over the natures of proof and causality. Particularly useful to epidemiologists and medical researchers, it also de-mystifies some of the impenetrable jargon around the randomised controlled trial (RCT): ever wondered just exactly whose ‘intention’ the ‘intention-to-treat’ refers to? All is revealed here. There’s also a fairly annhilating critique of most of the assumptions made by epidemiologists in creating ‘experimental’ models from population data (e.g. by controlling for variables using simultaneous equations): probably not for those who are overly comfortable in their epistemological worldview…
Excellent paper, Mithu, thanks for the recommendation!