If you flip a coin six times and get four heads, that doesn't mean the coin is biased against tails. But the AI ethics crowd would say yes, and proceed to "debias" the coin by rigging the flips.
-
The analogies I can think of are more like "flipping the coin 6000 time times and getting 4000 heads". Not sure where you're coming from with this one, but would be interested to hear more.
The point is that there are all sorts of reasons the outcome might not be 50/50 that have nothing to do with discrimination.
-
-
Okay ya sure, but I think most will read your analogy as saying "researchers think there's unequal outcome but there actually isn't/isn't statistically significant", which is very different from "researchers are correct it is unequal but I disagree about its cause"
-
https://en.m.wikipedia.org/wiki/Animistic_fallacy … It's not just about statistical significance, but also about attribution to outside or external factors as a de facto cause of all disparities which don't show 50/50 outcomes. My 2c anyway
- Näytä vastaukset
Uusi keskustelu -
-
-
Bias is not the same as discrimination otherwise a Selection Bias would be called Selection Discrimination. Launching a crusade against the English language while being ESL is a bold move!
Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
-
-
For example when men receive 40% higher prison sentences and go to prison 10x more often than women. No bloke ever complains about that.
-
Complained
Keskustelun loppu
Uusi keskustelu -
-
-
Nobody expects coin flips to be exactly 50/50. We DO expect a high p-value for the difference, especially on repeat attempts. If your coin yields stat sig different results depending on the race or gender of the person flipping it, you probably made a biased coin.
Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
Lataaminen näyttää kestävän hetken.
Twitter saattaa olla ruuhkautunut tai ongelma on muuten hetkellinen. Yritä uudelleen tai käy Twitterin tilasivulla saadaksesi lisätietoja.