Comments on: Statisticians take themselves too seriously

By: John

John — Fri, 05 Aug 2011 12:55:04 +0000

Stan, I agree with your points, except that I think mass delusion, or at least willingness to be deluded, may come first. The DDT research came out at a time when people were primed to accept its conclusions. On the other hand, people were unwilling to believe that smoking causes cancer, and it took a tremendous amount of evidence to change public opinion. I'm hesitant to call false conclusions the result of "faulty" statistics, since that implies that we can always be correct if only we're more careful. The best statistical analyses will reach wrong conclusions fairly often, and there's no getting around that. See Why most published research findings are false.

By: Stan Young

Stan Young — Fri, 05 Aug 2011 12:31:17 +0000

When one person being wrong becomes a mass delusion, e.g. Type A personality and heart attacks, there can be much downstream mischief. DDT killing birds, so far as I know wrong, led to millions of deaths in Africa. Much of this wrongness comes from (faulty) statistics. The author, best, and the rest of us need to know how to recognize claims that are very likely wrong.

By: AnnMaria

AnnMaria — Thu, 04 Aug 2011 03:21:35 +0000

I have never been accused of taking myself too seriously. On the contrary, in graduate school I wrote an “article” on the Rousey Expectorant Distance Intelligence test with measures like internal consistency reliability among the items. Items included expectorant distance while sitting, expectorant distance while standing, etc. I included test-retest reliability, inter-rater reliability and concurrent validity.

My advisor commented disapprovingly that “It was a light treatment of a very serious subject.”

He told me my article on comparative factor structures of the WISC-R and WISC-R Mexicano was better and I should stick to that.

By: On being wrong — The Endeavour

On being wrong — The Endeavour — Thu, 04 Aug 2011 02:56:33 +0000

[...] Statisticians take themselves too seriously ? [...]

By: John

John — Fri, 29 Jan 2010 15:48:41 +0000

John, it was in a psychology course from The Teaching Company. I'm not certain, but I think it was this one.

By: John MacIntyre

John MacIntyre — Fri, 29 Jan 2010 15:46:29 +0000

“I’m reminded of a talk I heard one time in which the speaker listed a number of embarrassing things that people used to believe”

Sounds interesting. I don’t suppose you saw it online & have a url? (wishful thinking on my part)

By: Ito

Ito — Fri, 29 Jan 2010 10:26:47 +0000

Nice idea, but in my opinion too short for a blog post.
You could either compress this idea to fit in a tweet, or (better) developp it, giving some examples of concrete cases. where statisticians take themselves too seriously.

By: Mike K Smith

Mike K Smith — Thu, 28 Jan 2010 14:46:24 +0000

Sounds a bit like the archetypical ISTJ Myers-Briggs type for statisticians: Think deeply, apply logic, trust evidence, use the methods that worked last time…