Re: graphing Ketan's weight

[John Carrow <jecarrow@houston.rr.com>]

Den:

With all due respect to those who have tried to analyze Ketan's data, I
beg to differ. It shouldn't be done.  It is all ancient history and
there is no accompanying information that might explain any signals of
special cause that may show up.

First, what do we know about his measurement process? Was it stable over
the time period?  Were there any recalibrations of the scale?  Was the
scale ever checked with a check weight at all?  Were the data even taken
using the same scale?

Even if we assume the measurement process was in a state of statistical
control and therefore we actually have measurements, what about the
Ketan variables?  Did he make his measurements at the same time each
day?  We know our bodyweights change during the course of a day.  Did
his dietary habits vary? His pattern of exercise?

Once again, we have an inappropriate application of the Shewart
technique which was meant to be carried out in real time, with
immmediate investigations when signals appear, and annotations of the
findings right on the chart.

I'm afraid this sort of thing sets a bad example.  It may be useful for
demonstrating the mathematecal techniques but I have seen too darn many
instances of  people engaged in after-the-fact charting and thinking
they have "applied SPC" to let this pass.

Just my two cents.

John