News for hackers. Stuff that matters.
Re:Scientists Decry "Flawed" and "Horrifying" Nationality Tests by dkarl on Wednesday September 30, 2009 @12:12PM (#852800)
IIRC there's 3 categories they can tell on average; where you grew up, where you lived and where you've lived recently. Basically if you grew up in London, lived in Manchester and have lived in Scotland for a half-year or so, they'll know.

If linguistic specialists can work this well....

Law enforcement have a bad track record of having great faith in techniques that don't actually work. One recent example in the news was the bogus forensic arson investigation techniques that were used to convict and execute Cameron Todd Willingham. Confessions are another example -- how many cases have been closed prematurely because police believed it was implausible for a person with no obvious mental abnormalities to confess to a heinous crime they didn't commit? That's why it's important to scientifically validate the techniques that are used. The fact that police have used a method and found it completely satisfactory doesn't say much about its validity.

Re:Scientists Decry "Flawed" and "Horrifying" Nationality Tests by tokenadult on Wednesday September 30, 2009 @05:19PM (#853727)
Law enforcement have a bad track record of having great faith in techniques that don't actually work.

Yes, actually most police investigation techniques have never been subjected to proper validation studies.