Page 2 of 2

Re: Betwetting

PostPosted: Thu Apr 09, 2015 11:22 am
by yomikoma
Obligatory xkcd: https://xkcd.com/1138/

Re: Betwetting

PostPosted: Sat Apr 11, 2015 7:04 pm
by vvn
Obligatory xkcd http://xkcd.com/552/

Re: Betwetting

PostPosted: Mon Apr 13, 2015 7:06 pm
by DanielH
Flint_A wrote:But we have a saying in the social sciences: "If you torture your data enough, they will confess."

(Well, I usually hear it as "it will confess", but "data" is plural damn it.)


The word “data” is plural, but no individual datum is confessing. I think the proper fix would be be “If you torture your data set¹ long enough, it will confess.”; the entire collection, treated as a single object, does the confessing.

¹ Do statisticians use this term, or just computer scientists? if not, then it should just be read as “collection or data”

Re: Betwetting

PostPosted: Mon Apr 13, 2015 10:04 pm
by Flint_A
They do use that term, yes. And sure, that sounds good to me.

Re: Betwetting

PostPosted: Fri Apr 24, 2015 8:54 pm
by sparr
Flint_A wrote:How strong is the correlation? It's obviously not 1.00, literally every bedwetter does not become a serial killer.
Although, the reverse correlation could be 1.00 (every serial killer could be a bedwetter), which would still be amazingly important an useful information, regardless of the correlation factor in the other direction.

Re: Betwetting

PostPosted: Sun Apr 26, 2015 8:50 am
by Flint_A
Eh. If someone's been accused of being a serial killer and yet they don't wet the bed, that wouldn't be enough evidence for them to go free. It'd just change our statistics. Induction is bad for legal judgement.

On the other hand, we could use their bedwetting to taunt serial killers we know but haven't yet caught, which I suppose is still useful.