The credence of statistical generalisations should not categorically license outright rational belief

Abstract

In this paper, I am expressly concerned with issues emanating from the taking of p to be the case in evidence-based medicine. I interrogate the tenability of statistical generalisations with regard to reference classes and, in particular, demographic groups. Upon the grounds of my estimations, I contrarily make a moral argument that the credence of statistical generalisations should not categorically license outright rational belief in evidence-based medicine as supported by the opacity of bare statistical generalisations, the problem of reference classes and inappropriate credence projection, even when such generalisations may appear epistemologically sound.