"Paddle Effect"

[Kromkowski@aol.com]

Jonathan Siegel <jmsiegel@yahoo.com> writes

>But as Deming explained in TNE, he kept careful
>records over the years and the average number of red
>beads obtained from each of 4 paddles he used varied
>from 9.2 to 11.3, a much larger variation than
>statistical theory would predict. (Try doing a
>simulation yourself and verifying this!) 

It's interesting that you ascribe it to the paddle. 
In my head, I've always ascribed it to attempts by
willing workers to "aim" their paddles toward an area
which "looks" like it has less red beads. The smaller
the paddle, the easier to aim.

Is my "explanation" (or the "paddle effect" explanation)
an attribution of a special cause explanation where
one is not warranted or is an attribution of a common cause
explanation -- after all, the workers were exhorted 
to "get less red" beads.

When you say a variation much larger than statistical
theory would predict, could you elucidate? How does
statistical theory "predict" variation of a manmade
"process" beyond what would be predicted by 
control charts? (In this case, you want a control chart
on the performance of paddles.)

JDK

Message posting through the Clemson CQI Web Server.