For me personally, I consider prejudiced action to be 'bad' when it fails to update based on countering evidence, and 'fine' if it does update.
SOME EXAMPLES:
3
3
56
Conversation
Your stereotype of redheads being untrustworthy comes from an anti-redhead movement from 50 years ago. You personally don't know any redheads to support this stereotype, and there's no statistical evidence to support the stereotype.
If you still believe it after this: BAD.
1
2
28
You, a small woman, are afraid of men, because men are the ones who follow you home and make you feel unsafe, not women.
You meet Joe, who behaves consistently respectfully. If you are still afraid of Joe after evidence to the contrary: BAD.
3
1
31
Things that are not bad: General prejudices based on evidence. Men are more likely to commit assault upon others, this is statistically supported. It's ok to be afraid of a man until you are provided evidence to the contrary, like Joe.
3
3
29
Some stereotypes aren't statistically supported because there hasn't been proper research done - but a prejudice from consistent individual experience is still not necessarily bad, as long as you are careful to check your biases and continue to update on counter information!
1
32
Replying to
Is it good or bad if your risk assessment greatly exceeds (or falls short of) the actual risk posed by a person who represents a particular demographic group? (e.g. to be *very* afraid of all men, despite knowing only a small percentage act violently.)
1
Replying to
Totally agree. Stereotypes are pooled experience disseminated as a meme across culture - of course it won't apply to everyone, but they are useful until they are not (i.e. evidence to the contrary) - 'knowledge' should always be contextualized/updated in favor of tabula rasa
1
Replying to
What about concerns of the form "Even when I think they're safe, ___s are still more likely to harm me than non-____s". That meta-level principle may be evidentially justified, but it undermines the case-by-case updates you want.
Replying to
But they're not seperate right? We are social creatures so couldn't acting on statistical analysis, impact your statistics? That is to say, if everyone keeps walking away from Joe, fearful of Joe, not open to Joe, wouldn't that impact Joe? Couldn't it even make Joe TURN scary?