Re: Shewhart's Maximum Control

["John G. Wendt" <jgwendt@earthlink.net>]

I am not a statistician, and typically use x/moving range or xbar/R charts,
depending on how the data is arrayed, meaning where the test sites which produce
the data are physically located in relation to each other.  I do a lot of work
with non-woven, rolled material webs, and use Don Wheeler's books a lot.

In general, I rely on the robustness of the charts, and don't worry much about
it.  However, if the data is known to be skewed  sufficiently to truly cause
concern, might this be a place to use the median range rather than the mean
range, since it would be affected less by the distribution shape?

John Wendt
=========================================================================