3 ms·

I'm curious how they determined that the 1% of cases were gender and race-based and not something else. With such a small number, it seems like it would be very

by troupe 9y ago

I'm curious how they determined that the 1% of cases were gender and race-based and not something else. With such a small number, it seems like it would be very easy for the variations to be caused by other factors. It is probably good PR for them, but the math behind the decision would be pretty interesting.

lallysingh 9y ago
1% of a large number is still a large number.
- troupe 9y ago
  Yes, but with a small enough number you might be able to find that people with blue eyes are paid more than people with brown eyes. You'd expect some level of random variation and such a small number makes it seem that it may really not have anything to do with their gender. Think of it this way. If you flipped a coin a large number of times and 51% of the times it came up heads would you assume the coin was biased or would you suspect that it could just be random variation?