Hacker News Reader

Comment by krallja | Hacker News Reader

It doesn’t matter, either way. You can’t use statistical measurements on individuals.

Of course you can. To take it to the extreme, should doctors (or public health campaigns) assume that men and women have an equal chance of breast cancer?

Again, it doesn’t matter whether there’s a different chance. As the man’s physician, should I rule out breast cancer because the patient is a man?

> As the man’s physician, should I rule out breast cancer because the patient is a man?

Not exactly. But even given the very same inconclusive evidence, you probably should assign a lower probability to a man having breast cancer than you would for woman. Simply because the priors are different.

But you wouldn’t ignore the evidence completely, which is what happened to the NFL in this case. They decided “on average Black players have less dementia, as a race they have lower baseline cognitive capacity” which is at once racist, stupid, and incorrect. The hypothetical doctor needs to consult radiology, despite the patient being male. And the NFL needs to find a more individualized test for dementia, taking into account each individual’s former mental capacity, not lazily using a single evidence-free racial marker that also happens to save them a bunch of money.

While another reply shows that you're principally wrong, it may also be practically wrong.

The author implies that the test was designed with malice, when infact it may be quite appropriate.

Reply on news.ycombinator.com