DEN Discussion List Archive
[Date Prev][Date Next][Date Index]
[Thread Index]
[Author Index]
Relation between 3 STD and approximated from XmR
- Subject: Relation between 3 STD and approximated from XmR
- From: "John Constantine" <thesfg1@cox.net>
- Date: Tue, 31 Dec 2002 10:24:24 -0700
"This is suppose to be simple so that small businesses as well as large
businesses can apply it.
Minimal training to each employee."
The answer, if there is one, lies in the use of XmR in periodic, sequential
time series (monthly, etc.) applications and activities. The X chart portion
will indicate the limits of the actual data using the appropriate scaling to
define the upper and lower limits, while the mR chart portion will contain
the data in sequence.
The Xbar will still be the centerline of the xchart, and the mRbar will
still be for the moving range chart. The data will determine what is
happening, what the limitations are, and how far apart the spread is. This
will allow the manager, or rather the employee closest to the "action", to
see what is occurring in the process under scrutiny as soon as one month
after a significant event has occurred.
How much better can you get? As Wheeler points out, this can be used even if
the data are themselves suspect. But simply plotting the data will never be
enough. The crux of the matter will be in the interpretation. (I use it for
keeping track of my various utilities, such as gas, electricity, water, and
I've had real time examples of how useful it can be.)
If we are not using time series as our base data, the data will still
determine what the limitations are using other types of charts. This is the
beauty of Shewhart's charting. Prediction works best when the process itself
is random. If the data show that the process is statistically stable, it
(the process) will tell you what its limits are. If not, the
managers/employee's job will be to get it to that point of stability, from
whence prediction can be made.
In the event that probability produces material which coincides with the
actual data, the results will be evident. But, if the probability flies in
the face of the actual data, the data will win. You can't ask for more.
________________
John Constantine
thesfg1@cox.net
Phoenix, AZ
DEN Home |
Main Index |
Thread Index |
Author Index