RE: SPC - independent data

["Alson C H Look" <alook@pacbell.net>]

It is good to keep in mind that Shewhart's original 1931 book contained
an xbar and sigma chart that was autocorrelated.  Shewhart did not allow
the fact that there was correlation in his data to stop him from using
it.  In fact, a reading of this example indicates that Shewhart found
assignable causes (in spite of the autocorrelation).

Today, if our process has autocorrelation we need to think about what is
happening in the process.  If autocorrelation is a natural part of the
process then one simple method would be to spread the measurements apart
in a logical way.  Remember the goal of a Shewhart control chart is to
find special causes and not necessarily to adjust the process.

Alson 
alook@pacbell.net