Question

I have a normal process (Anderson-Darling pvalue=.54   and skew=.3. The median

The R chart is out of control (1 outlier). This means that the process variability is unstable, and the control limits on the xbar chart are unreliable. The process is not stable over time. What if anything does this have to do with choosing the center of the distribution? Use the median in this instance. Outliers will not affect median as much as the mean. That is my best guess.

Answer 1

Mean and R-chart used to describe observed central value and spread with respect to time respectively.

It is given that the R chart is out of control which shows that process variability is not stable and the control limits of the x-bar chart is not reliable. In short, the process is not stable with respect to the time.

Median chart is one of the special chart used to determine the central location as well as the variation. Median is suitable for the skewed data as it does not get affected by the outliers in the data set. For each of the subgroup, compute the median value and after that arranged in sorted order. This chart plots all the values instead of only subgroup statistics. This chart is especially useful when the ranges of the subgroups shows a great variation in the distribution of the data. It is useful to explain that central value lies within the limits of the control chart but it may possible that individual values may lies outside the limits of the control chart.

Therefore, it can be said that if the R-chart is out of control and x-bar chart limits are not reliable, then median chart is the best possible choice to find the central value of the distribution.