2 ms·
Like I said, "I'll take one example". The study is utterly flawed even conceptually (1), but that flaw might just be a result of some law prof not understanding
by yummyfajitas 18y ago
Like I said, "I'll take one example". The study is utterly flawed even conceptually (1), but that flaw might just be a result of some law prof not understanding basic statistics.
The example I provided is merely the simplest and most egregious misrepresentation she makes.
(1) To estimate the number of bankruptcies caused by "medical reasons", you need to compare Pm = P(bankrupt | medical cause) to Pnm = P(bankrupt | no medical cause). The number of medically caused bankruptcies is then (Pm-Pnm) x (number of people with medical cause). She only measures Pm x (number of bankruptcies with medical cause).