4 ms·
Looking at it the other way, even if you have a 99% chance that the positive result is correct, is a 1% rate of sending innocent people to jail really acceptabl
by mikeash 10y ago
Looking at it the other way, even if you have a 99% chance that the positive result is correct, is a 1% rate of sending innocent people to jail really acceptable?
- LoSboccacc 10y agoOr with some context if you put the usa population under test you'd get 3 MILLION innocent convicts. small percentages adds up scary fast over continents
- justicezyx 10y agoOther factors should be considered. For example, it is not a random sampling. Polices target those who are statistically more likely to be positive, according to their empirical or historical statistics.
- mikeash 10y agoI think that's already accounted for in my hpothetical, as those are factors that might bring a 33.2% rate up to 99%.
- bryondowd 10y agoDepends how good of a filter the police are using. Even if they are only targeting people who are 10 times as likely to be users as the general population in the above example, that only brings the accuracy to about 84%. And I'd wager their filtering isn't that good. What exactly are they using to determine who is a likely user? Are those people really more likely at all, and if so, how much? I expect it would be highly optimistic to even assume a person targeted by police is even three times as likely as the general population to be in possession of drugs, which would only be about 60%
- analog31 10y agoIndeed, the person could be less likely. From what I've read, affluent whites have the highest rates of illegal drug use, but the police direct their enforcement efforts towards blacks.
- cleaver 10y agoLast time I checked the data, there were approximately 80,000 aircraft flights in the USA each day. If 1% crashed, there would be 800 airplane crashes every day. While 1% seems like a small error rate, I find this comparison a good way to convey the impact of a small error rate combined with a serious outcome.
- vacri 10y agoMaybe on the first day, but with that many planes taken out of action each day, it'd tail off pretty fast... :)
- jackcarter 10y agohttp://www2.law.ucla.edu/volokh/guilty.htm http://www2.law.ucla.edu/volokh/guilty.htm '"Better that ten guilty persons escape than that one innocent suffer," says English jurist William Blackstone. The ratio 10:1 has become known as the "Blackstone ratio.'
- k7wqcyfgmu 10y agoThis is not the same ratio, the other 100 people are not guilty in this case.
- guard-of-terra 10y agoWe can probably work out a formula then to figure out maximum acceptable false positives given apparent usage stats.
- wahern 10y agoThe maxim is much older than Blackstone, and has been variously recited as 2:1, 100:1, 1000:1, etc. The point is that the justice system should error on the side of protecting the innocent. It's not quantifiable. It's not a statement of social utility, and no utility function (e.g. which related community wealth to innocents jailed) could replace it. The maxim is a bedrock principle of how society defines justice. It's an articulation of the values and preferences that go into defining a utility function. That is, how just you _believe_ the government is a utility in and of itself, regardless of objective efficacy, and the maxim outlines one of the criteria for how you guage justness. The notion is that a government that emphasizes protection of the innocent is believed to be most capable at defending fundamental liberties, in part (in as much as people rationalize it) because it incentivizes rigor in how government prosecutes criminals. Putting criminals away benefits everybody, so there's already substantial incentive to get it done. Imposing constraints on how it's done allows society to shape the process and, ideally, push it toward a global optimum rather than falling into a local optimum. So it's an articulation that society places substantial value on liberty even if to the detriment of some security. (Compare that to cultures like Japan.) Second and simultaneously it articulates a belief that security _depends_ on the protection of liberty; that is, effective governments safeguard liberty, and those which do not safeguard liberty are unlikely to be effective at providing long-term security. Neither of those statements are absolute truths, and they certainly don't depend on empirical efficacy. They're normative value judgments in and of themselves, therefore they help define the utility function, both in the aggregate and of individuals in the society, the vast majority of whom will have internalized those normative judgments.
- nkrisc 10y agoI'd rather the guilty went free than the innocent convicted.
- rasz_pl 10y agoThis is exactly the problem with those kinds of parameters. YES, almost all doctors and judges will tell you 99% success rate is worth it. It takes a lot of work to educate them this paper 1% translates into 66% error rate in the field.