Friday, November 12, 2010

Bedbug -sniffing Dogs

From the Atlantic:
http://www.theatlantic.com/national/archive/2010/11/bed-bug-sniffing-dogs-are-causing-false-alarms/66448/

No grand impacts from this particular story but... I am documenting *all* false positive stories I get my mitts on, not just the really salient ones. The issue here is money and reputation. It will hurt the dogs' reputation if they don't have a good rate of having been right. So this phenomenon has some salience because it suggests that calm, Occam's Razor and a shrug about power structures is sufficient - if they do a bad job, their reputation goes in the toilet. "What are you worried about - the proof will be in the pudding. If a police chief turns out to have been wrong one too many times, the guy is sacked."


My complaint is that it makes those *first few* incidents into canaries in a coalmine. Or to repeat another phrase I have repeated in a few different posts, "some kid always has to get killed at the crosswalk to get the new traffic light put in." In the case of this news story, let's say for argument's sake that authorities could compel a hotel to do expensive fumigations because the dog sniffed bedbugs. It turns out the dog was wrong, and by that time, the hotel has already spent money. So the remedy is that the reputation of the dogs suffers. And this is better than nothing, but it sucks for the first few examples who were instrumental in bringing that problem to light. And if the domain you're talking about is something irretrievable - money can be refunded, but let's say we're talking about the death penalty based on DNA evidence which is later found to be wrong - it is a problem because you aren't going to be able to bring that person back. Even if the reputation of DNA evidence duly suffers, the actually-not-guilty person has already been executed.

To add one final thing to this messy post, I want to mention the idea of "it could be you." Someone who tries to point out the fallibility in a given system can use the argument, "it could be you, you might feel differently if it were you or someone you know." The vast middle class comes back with Occam's Razor.

"It's just not going to be."

"Why not?"

"It's just not going to."

So we are more comfortable and less concerned with "canary in a coalmine" situations where something can be remedied gradually, when it is at someone else's expense. Possibly race and economics comes into it, because if the person saying "you're being an alarmist," is well off and/or not part of an ethnic group that tends to be poorly treated, they are subtly taking advantage of that, even though they might decry it out loud.

So to be more specific, a poor black man is a suspect for murder. Through DNA evidence, he is found guilty, gets the death penalty and is executed. He's later exonerated. People start to have second thoughts about DNA evidence. But if someone was trying to drum up more attention for the flaws in the DNA-evidence process before the wrongful execution happened, they wouldn't have gotten very far. And one of the reasons they wouldn't have gotten very far is that "it could be anyone, even you!" arguments wouldn't have caught fire because the center can fire back, "it won't be me so it's not my problem." Ultimately it seems to me like a subtle exercise of might makes right.

No comments:

Post a Comment