My paper “The choice of effect measure for binary outcomes: Introducing counterfactual outcome state transition parameters” has just been published in the journal Epidemiologic Methods. In this paper, we propose a new class of effect measures for causal effects. While this new effect measure is not generally identified from the data, we argue that it can be used to formalize biological knowledge relevant to effect modification and to the choice between standard, identifiable measures of effect such as the risk ratio and the risk difference.
The published version is available here (behind paywall, a preprint is available on arXiv).
An informal discussion of COST parameters has been posted to Less Wrong.
I would very much appreciate any comments, especially critical ones. As always, I invoke Crocker’s Rules for all discussion of this paper.