Tuesday, November 19, 2019

Statistics does not require randomness



Do you buy this?
"Statistics does not require randomness.

The three essential elements of statistics are measurement, comparison, and variation.


Randomness is one way to supply variation, and it’s one way to model variation, but it’s not necessary.

Nor is it necessary to have “true” randomness (of the dice-throwing or urn-sampling variety) in order to have a useful probability model."
Andrew Gelman
 

How about this one?
" ... the #1 neglected topic in statistics is measurement "
All this, and the image, are what Andrew Gelman asserts in his posting here.



Statistics does not require randomness (just variation) --- an arresting statement to be sure. It's what caught my eye.

Gelman is on to a different agenda, however. His point is that in his experience -- which may be more academic than empirical -- measurement is given a back seat, almost neglected in some respects. (Measurement occurs here; lets press on)

But of course, if you are going to stand in front of your team or sponsor and spout statistics, it would be good if you had real observations to back you up. And, shocking as it may seem, you have to take measurements of real events and phenomenon in order to have back-up data

In turn, that leads toward accuracy and precision, cause and effect, and correlation. Perhaps all that is a bridge too far for one posting. I am leaving the building now...


Buy them at any online book retailer!

No comments:

Post a Comment