RE: Periodic Control Charts

["Dr. W. J. Latzko" <latzko@att.net>]

At 04:28 PM 3/19/2002 +0530, you wrote:
>In x-bar and R charts, we control the process average and range. In this 
>example probably, it could be appropriate to control the weekly 
>consumption rather than average/ day.

What is suggested here is possible. However, from the description of the 
problem, most likely it will increase the variance. One will measure the 
total variation which is composed of the variance between the days plus the 
variance within the days. In short, the subsample will be heterogeneous and 
so would not fit the "rational subsample" concept of Shewhart.

Bill
Latzko@att.net