Comments on: Falsehood: Correlation Implies/Does Not Imply Causality

By: Jj

Jj — Sat, 18 Aug 2018 02:32:02 +0000

In reply to dean. Mansplained again!

By: dean

dean — Fri, 27 Oct 2017 20:33:25 +0000

In reply to dean.

One more for you Greg:

Why do all frequentists become Bayesians after five drinks?

Because that’s when they start pulling conclusions from their posteriors.

By: Greg Laden

Greg Laden — Wed, 18 Oct 2017 16:53:34 +0000

In reply to dean. :)

By: dean

dean — Wed, 18 Oct 2017 14:07:42 +0000

The head of the statistics department at a large university was walking across the grounds to a lunch meeting when he saw several students standing around a light pole. Seeing that they were having an animated conversation he walked closer to hear, and determined that they were trying to figure out a way to measure the height of the pole as part of a class assignment. He walked over to them and said

“Look, the pole is bolted to this metal plate on the bottom. Get some wrenches, unbolt it, lay it down, then measure it.” He then walked off.

The students looked at each other, laughed and one said “Isn’t that just like a statistician. He knows you want the height of something and he tells you how to figure out how long it is.”

By: Li D

Li D — Wed, 18 Oct 2017 09:30:29 +0000

A scientist, an engineer, and a politician go hunting for rabbits… and after a bit they come across some tracks.
Scientist. ” These are are rabbit tracks. We will soon get them. ”
Engineer. ” Nah, idjit! Kangaroos made these .”
Pollie ” You people are nuts. These were made by a wombat. ”
They were still arguing an hour later, till a train ran over em…..

By: Gregory Laden

Gregory Laden — Tue, 17 Oct 2017 21:44:32 +0000

In reply to dean. A scientist, an engineer, and a politician go hunting for rabbits...

By: dean

dean — Tue, 17 Oct 2017 15:42:40 +0000

It is cool.

I don’t consider my resume highly impressive – it was meant to support my comment.

At the other extreme: last summer an in-law told me that he no longer believed anything scientists or statisticians claimed was “supported by research” — because

If they were really right they wouldn’t have to have any nonsense about “plus or minus” stuff: they’d give the right answer and that would be it.

I’ve started using comments like that in class to stress the importance of grasping why variation matters.

By: Greg Laden

Greg Laden — Tue, 17 Oct 2017 15:14:19 +0000

In reply to dean.

Not as you would apply it (I taught advanced multivariable stats to grad students at Harvard, as long as we are trading highly impressive resume items, though I don’t consider my self a statistician).

As noted, you can equate the F statistic and R, which is generally known as the coefficient of correlation.

But mainly, I liked the layout of the formula and the fact that the background on the image is transparent so it can be used elsewhere.

By: dean

dean — Tue, 17 Oct 2017 13:46:54 +0000

I realize that Greg – I am a statistician. My first two papers were on robust measures of correlation and multiple determination. My point (originally intended to be tongue in cheek) was that two-way ANOVA itself doesn’t automatically relate to correlation.

By: Gregory Laden

Gregory Laden — Tue, 17 Oct 2017 12:28:48 +0000

In reply to dean.

Dean, brilliant! I was. Wondering who would bring that up.

At some level all statistics are the same as some other statistics. Ie. you can convert an R-squared value to an F statistic without redoing the stats on the same data.

Regression analysis and ANOVA are two faces of the same underlying model.