I don't doubt there are definitions of privilege which are falsifiable, as those pertaining to discrimination. But those are seldom used, and when challenged, the claimants fall back to some abstract unfalsifiable definition, i.e., a motte and bailey. Given that, perhaps it would be clearer if we stuck to a term that is clearly falsifiable, such as 'discrimination'.

And as for discrimination, it's difficult to study, and a lot of these bedrock studies/theories never even found anything substantial to begin with (the implicit association test, the blind orchestra auditions study, etc) and of course all of the others that have failed to replicate. Never mind that we are fond of using data from the 30s-80s as indicative of attitudes in 2019. Maybe we can finally begin to question whether social science has a groupthink problem?