Tuesday, December 10, 2013

Well, I'm not surprised

From a recent article at New Scientist [ht: realclearscience.com]:
There are major flaws in the statistical tools for assessing whether people with serious psychological problems will reoffend, a psychiatrist says
WOULD you be surprised to learn that many imprisoned people have serious personality disorders, such as psychopathy? Probably not. But I'd confidently predict you would be surprised to find that popular methods to predict the risk of reoffending to aid decisions on sentence length and release date for such people are no better than a coin toss.
This is the worrying verdict on long-established tools to assess the risk of violent reoffending, the latest blow to their utility.
This is an interesting, though demanding, article.  And if you like it, you might like these (here, here) indirectly related OLS posts as well.