-
-
5/ So if there's a base rate of X accidents per thousand (constant across all demographics) and Y rate of violence to children in middle class households, and Z rate in poor households, it might very well be the case that 10% of middle class kids coming to hospitals are abused
Show this thread -
6/ and 80% of poor kids coming to the hospital are abused. (or 1% and 99%, or 49.1% and 49.2% ; I don't have hard data here) But either way, based on data we have, we SHOULD expect that a higher percentage of poor kids arriving at hospitals w broken bones are there bc abuse
Show this thread -
7/ So now we've got a classic Type I / Type II or, in my preferred terminology, false negative / false positive problem One algorithm you could use is "define heuristic H, and apply it to all kids". E.g. "if the situation is 5+ points on a scale of dubious evidence, call cops"
Show this thread -
8/ But heuristic H is likely going to UNDER protect poor kids and OVER protect middle class kids. So a better algorithm is to use heuristic H1 for poor families and H2 for middle class families. I mean "better" here if we want to minimize child abuse and false allegations.
Show this thread -
9/ So...THAT SUCKS. We've just talked our way into explaining why, if we want to maximize justice, we HAVE to discriminate by family income....which in turn means, effectively, by race. whew lads! anyway, tweets like this are why I'll never be elected to office.
Show this thread -
New conversation -
-
-
poor people have less to lose; a middle class man will walk away from a fight because going to jail for ten days will cost him his job. Some classes of poor people gain status in their culture by being incarcerated.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.