5 ms·

I am not denying that there are major problems with the tools used by police and that juries fail to understand their limits, however, I take issue with 30k bog

I am not denying that there are major problems with the tools used by police and that juries fail to understand their limits, however, I take issue with 30k bogus arrests as indicative of a bad test without further context. For example, how many tests were given? What is the false positive rate, false negative rate, etc. i.e. the ROC characteristics. The tests may be quite accurate, we do not know from just that number. The question of whether the tests should favor reducing false positives of false negatives is a policy decision that must take into account public safety risk (false negative) vs harm to the individual (false positive and arrest) vs lasting harm from the false arrest (imprisonment). A moderate false positive rate might be good for public safety if used as an initial screener, but once arrested, further tests should be used that emphasize reduction of false positives. Then the harm would be limited to the disruption of an evening, not a life.

yieldcrv 3y ago
The article said 4%, and contends that 4% is too high and goes further to notice that convictions are mostly plea deals, and that prosecutors in many places don’t have to verify if the plea deal is based on any lab based evidence and further finds a bunch of departments that arrest people no matter what the test result is, making your standard about “bad tests” completely moot its a system based on total apathy and the article is saying thats the problem that should be fixed
godelski 3y ago
> The tests may be quite accurate, we do not know from just that number. Be careful as this might go against exactly what your prior concern was. It's the Bayesian Trap. FWIW, the article says there's 773k arrests each year and gives around a 4% bogus rate. I'd say that's too high for something we all can admit comes at a great cost. > The question of whether the tests should favor reducing false positives [or] false negatives is a policy decision that must take into account public safety risk We have a written code that says innocent until proven guilty. Clearly false positive rates is of grave concern. There are many references to Blackstone's Ratio[0] by the founders and the premise is that it is better to err on the side that a guilty person go free than a free man lose his freedom. But Bayes is critical here too. [0]https://en.wikipedia.org/wiki/Blackstone%27s_ratio https://en.wikipedia.org/wiki/Blackstone%27s_ratio
- Closi 3y ago
  Depends - I would say 4% is probably accurate enough evidence to support an arrest (i.e. they are detained someone in order to investigate a crime they are suspected of committing), but not enough to support a conviction. Arrest != conviction. However 4% probably isn't accurate enough to support jailing someone (in my view, depending on other evidence), so there is nuance here. Besides, while you are innocent until proven guilty, this doesn't mean that you can't be arrested (i.e. detained for investigation) if you are a suspect.
  - godelski 3y ago
    > Besides, while you are innocent until proven guilty, this doesn't mean that you can't be arrested (i.e. detained for investigation) if you are a suspect. I was with you till here. While I disagree with a bit prior, like you say, nuance. But this sentence is egregious. An arrest record is still particularly damming to one's life. Even if one is never convicted it can still greatly impact one's ability to get a job, credit, or other things. You may say that that is a different problem, and in part I will agree, but this aspect is likely easier to address and I do not think it is wrong to suggest we should have a high bar for detaining someone and placing them on record.
    - Sohcahtoa82 3y ago
      > An arrest record is still particularly damming to one's life. The solution then, IMO, is that if you're arrested for being suspected of a crime, and later found innocent or the charges are dropped, then the arrest should be removed from your record.
      - godelski 3y ago
        I'm perfectly fine with this solution and will even advocate for it. But many others have been trying to unsuccessfully. So I think let them continue their efforts, do bring attention to is, but also not be reliant on them fixing the problem. I should also add that arrests are public. So even if you are not convicted the arrest record may be found by a third party who is going to be less incentivized to expunge that data. So there's extra complexity to this too and why we should also not be reliant upon that solution.
sneak 3y ago
https://en.m.wikipedia.org/wiki/Blackstone%27s_ratio https://en.m.wikipedia.org/wiki/Blackstone%27s_ratio
standardUser 3y ago
> A moderate false positive rate might be good for public safety Do you really think the ~0.5% of drug users who actual get arrested is improving public safety in any meaningful way? If we actually care about public safety, let's stop forcing people to buy their drugs from criminals. It's the prohibition of drugs in the first place that brings the violence.
forgotmypw17 3y ago
> Then the harm would be limited to the disruption of an evening, not a life. It's not just an evening, because even if your charges are immediately thrown out, your arrest remains public record forever.

yieldcrv 3y ago

The article said 4%, and contends that 4% is too high and goes further to notice that convictions are mostly plea deals, and that prosecutors in many places don’t have to verify if the plea deal is based on any lab based evidence and further finds a bunch of departments that arrest people no matter what the test result is, making your standard about “bad tests” completely moot its a system based on total apathy and the article is saying thats the problem that should be fixed

godelski 3y ago

> The tests may be quite accurate, we do not know from just that number. Be careful as this might go against exactly what your prior concern was. It's the Bayesian Trap. FWIW, the article says there's 773k arrests each year and gives around a 4% bogus rate. I'd say that's too high for something we all can admit comes at a great cost. > The question of whether the tests should favor reducing false positives [or] false negatives is a policy decision that must take into account public safety risk We have a written code that says innocent until proven guilty. Clearly false positive rates is of grave concern. There are many references to Blackstone's Ratio[0] by the founders and the premise is that it is better to err on the side that a guilty person go free than a free man lose his freedom. But Bayes is critical here too. [0]https://en.wikipedia.org/wiki/Blackstone%27s_ratio https://en.wikipedia.org/wiki/Blackstone%27s_ratio

Closi 3y ago
Depends - I would say 4% is probably accurate enough evidence to support an arrest (i.e. they are detained someone in order to investigate a crime they are suspected of committing), but not enough to support a conviction. Arrest != conviction. However 4% probably isn't accurate enough to support jailing someone (in my view, depending on other evidence), so there is nuance here. Besides, while you are innocent until proven guilty, this doesn't mean that you can't be arrested (i.e. detained for investigation) if you are a suspect.
- godelski 3y ago
  > Besides, while you are innocent until proven guilty, this doesn't mean that you can't be arrested (i.e. detained for investigation) if you are a suspect. I was with you till here. While I disagree with a bit prior, like you say, nuance. But this sentence is egregious. An arrest record is still particularly damming to one's life. Even if one is never convicted it can still greatly impact one's ability to get a job, credit, or other things. You may say that that is a different problem, and in part I will agree, but this aspect is likely easier to address and I do not think it is wrong to suggest we should have a high bar for detaining someone and placing them on record.
  - Sohcahtoa82 3y ago
    > An arrest record is still particularly damming to one's life. The solution then, IMO, is that if you're arrested for being suspected of a crime, and later found innocent or the charges are dropped, then the arrest should be removed from your record.
    - godelski 3y ago
      I'm perfectly fine with this solution and will even advocate for it. But many others have been trying to unsuccessfully. So I think let them continue their efforts, do bring attention to is, but also not be reliant on them fixing the problem. I should also add that arrests are public. So even if you are not convicted the arrest record may be found by a third party who is going to be less incentivized to expunge that data. So there's extra complexity to this too and why we should also not be reliant upon that solution.

sneak 3y ago

https://en.m.wikipedia.org/wiki/Blackstone%27s_ratio https://en.m.wikipedia.org/wiki/Blackstone%27s_ratio

standardUser 3y ago

> A moderate false positive rate might be good for public safety Do you really think the ~0.5% of drug users who actual get arrested is improving public safety in any meaningful way? If we actually care about public safety, let's stop forcing people to buy their drugs from criminals. It's the prohibition of drugs in the first place that brings the violence.

forgotmypw17 3y ago

> Then the harm would be limited to the disruption of an evening, not a life. It's not just an evening, because even if your charges are immediately thrown out, your arrest remains public record forever.