4 ms·
This article is about errors in a set of population priors used in software. That could affect the outcome of tests, but it also might not. However, given that
by jimrandomh 11y ago
This article is about errors in a set of population priors used in software. That could affect the outcome of tests, but it also might not. However, given that the FBI was recently caught having faked an entire field of forensic science (hair analysis), there is little reason to extend them the benefit of doubt.
The important question is: are these crime labs routinely processing control samples, and if so, are they misclassifying any of the control samples?
- kiba 11y agoThere is a difference between purposefully faking an entire field of science and believing in pseudoscience and not being skeptical. The effect is the same, either way.
- rasz_pl 11y ago>and believing in pseudoscience lie detectors are a big thing for FBI
- meowface 11y agoPolygraphs? They know full well polygraphs are not "lie detectors". And a lot of the general population knows that too. They're still useful when recruiting candidates. Obviously their results are not admissible in criminal cases.