StiffMittens wrote:TheProwler wrote:There is about a 99% chance that the guy was telling the truth about the dog. I'm pretty sure it is quite rare that those dogs make mistakes
Not as rare as you might think:
http://www.cbsnews.com/stories/2004/01/05/60minutes/main591477.shtml
http://www.k9fleck.org/reltrn.htmThe dog must be trained, certified and reliable. The dog does not have to be 100% accurate or perfect. The courts have recognized the fact that “false responses”, “false positives” or “false alerts” occur and dogs can be as low as 54% accurate.
Not to get too far off track, because I'm pretty sure in this case the dog did not even alert them to drugs or "dead people", but think about it.
Falco's usage reports showed that in controlled tests he could be highly accurate -- correctly alerting more than 90 percent of the time. But in real life situations the results were starkly different.
“The court found that Falco was 35.5 percent accurate,” says Gaines. And a judge ruled that such a low success rate didn't give the police probable cause to make the search.
Okay, so couldn't the slip from 90% to 35.5% be explained by the situation we are seeing in this thread? The cops says he gave a positive, when he didn't. The cop does this to justify a search. Hence, the dog's accuracy rating slips.



