Beyond accuracy: Epistemic flaws with statistical generalizations

Abstract

What, if anything, is epistemically wrong with beliefs involving accurate statistical generalizations about demographic groups? This paper argues that there is a perfectly general, underappreciated epistemic flaw which affects both ethically charged and uncharged statistical generalizations. Though common to both, this flaw can also explain why demographic statistical generalizations give rise to the concerns they do. To identify this flaw, we need to distinguish between the accuracy and the projectability of statistical beliefs. Statistical beliefs are accompanied by an implicit representation of the statistic's modal profile. Their modal profile determines the circumstances in which they can legitimately be projected to unobserved instances. Errors in that implicit content can be compatible with the accuracy of the "bare" statistic, whilst systematically leading to downstream errors in reasoning, in a manner which reveals an epistemic flaw with an important aspect of the belief state itself.There are some generalizations about demographic groups which we may hesitate to endorse, even as they are borne out by empirical data. These include claims like the following: "Black Americans are almost eight times more likely to have a homicide conviction than white Americans;" 2 "teenage girls perform less well at mathematics than boys;" 3 "gay men have far higher rates of sexually transmitted diseases than straight men;" 4 "conservative political views are associated with lower IQ." 5 These sound like classic racist, sexist or homophobic claims, and yet they are statistically supported. Nonetheless, we tend to feel uneasy with such beliefs, and to suspect that someone who drew on them in their reasoning about the people they met would be rationally remiss. 6 This gives rise to the following puzzle: what explains that unease? Is it due merely to a socially commendable but epistemically 228

Varieties of Moral Encroachment

“…Fritz and Jackson (n.d.) argue that some theorists—specifically, advocates of Radical Moral Encroachment—who don't advocate moral encroachment on credences are nevertheless committed to it. Finally, several authors who deny encroachment on credences (Bolinger, 2018; Fritz, 2019; Gardiner, 2018; Munton, 2019) have argued that in many of the cases used to motivate moral encroachment, the agent's evidence does not even license rational high credence by ordinary epistemic standards.…”

Can Pragmatists Be Moderate?

“…That is why I have introduced the case here, and why I consider it in what follows. However, let me be clear that even if I am right in my eventual contention that there is no good explanation of how to keep one's pragmatism moderate, it doesn't follow that there is nothing wrong with 22 Examples of such explanations of this or similar cases include Gardiner (2018), Munton (2019), and to some degree Fritz (2019: §4). 23 Cf.…”

Moral Encroachment and Positive Profiling

“…6 Others have suggested, along broadly similar lines, that what goes wrong in cases like social club has to do with the particular sort of statistical inference that is being drawn. Munton (2019) suggests that the problem with certain accurate statistical generalizations, like those that we appeal to when racial profiling, is that we use them to project a regularity observed among past instances of a kind to a novel, unobserved instance to which it does not apply. Still others have claimed that the tools of mainstream epistemology can be used to explain the bad feature of social club.…”