Stats-n-such

Date view Thread view Subject view Author view

From: Jeff Bone (jbone@jump.net)
Date: Tue Nov 28 2000 - 09:47:10 PST


> I'd say that anyone with a passing familiarity with measurement would
> say that there is no fair and accurate process to measure a table to the
> 9th decimal place with a footrule.
>

Granted. The point being, though, that selective sampling is what's sometimes referred to as
"favoring your data." And mix-n-matching measurement techniques in a single dataset is, well, it's
unheard of in the lab. Bad science. Too bad this is politics...

Still, if the difference is within the margin of error as it is, the only sensible thing to do is to
go back and review the dataset thoroughly, not selectively. The result may still be statistically
inconclusive, but at least at that point you would've made all the measurements using the same
instrument. Or, hey, here's a thought --- we already did that, twice, with the machine recounts. We
should just use that! Oh, wow, the Republicans are more "scientifically minded" than the Dems!
Who'd've thunk it?

;-)

jb


Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Tue Nov 28 2000 - 09:53:17 PST