Eugene Volokh had up an interesting article that touch on the intersection of stats and law. It was on the topic of campus tribunals that hear sexual assault cases, and I thought it showed a fundamental principle of stats fairly nicely: when in doubt, put it in words. He does this with 3 legal standards for evidence: beyond a reasonable doubt (95% confidence), clear and convincing evidence (75% to 80% confidence) and a preponderance of evidence (51% or more confidence). He then says to determine the standard we should convert this in to words:
- Better that 19 students guilty of sexual assault remain at the university, with no discipline imposed, than one innocent student be expelled or otherwise disciplined
- Better that 4 students guilty of sexual assault remain at the university, with no discipline imposed, than one innocent student be expelled
- These outcomes are about equally bad for both students and the university
There’s some other interesting legal discussion in his post, but I thought the conversion of legal standards and probabilities in to clear sentences was a particularly helpful way to frame the discussion.
graph paper diaries 
Link to Volokh's article?
LikeLike
Oops, fixed!
LikeLike