On the so-called “frequentist” interpretation of probabilistic knowledge, it is knowledge tied inextricably to the consideration of a series of instances of whatever it is about which knowledge is being claimed. For example, to speack of having established empirically that the probability of attendance by “highly extraverted” persons at the next upcoming party is 0.7 is to say that, within some sufficiently large sample of persons, 70% of those who were found to be “highly extraverted” attended the next upcoming party and 30% of them did not. Such a finding, it should be noted, could serve as the basis for a testable–which is to say empirically falsifiable–proposition in a replication study.
However, to assert on the basis of findings of this sort that the probability is 0.7 that some particular “highly extraverted” individual will attend the next upcoming party is to express a knowledge claim that is not empirically testable. As a claim to knowledge, such a statement would be disconfirmed neither by the finding that the individual in question in fact attended the next upcoming party nor by the finding that s/he did not attend, and those are the only two empirical possibilities. Note, moreover, that a knowledge claim of this sort is equally incorrigible for each and every individual case considered as such. The very essence of the frequentist view of probabilistic knowledge generated through population-level studies is that such knowledge quite literally applies to populations and not to individual cases within those populations (cf. Bakan, 1955). In short, we must understand that such knowledge is, quite literally, knowledge of no one. (p. 9)
Lamiell, J. T. (2013). On psychology’s struggle for existence: Some reflections on Wundt’s 1913 essay a century on. Journal of Theoretical and Philosophical Psychology, 33, 205-215.