Submit a question to our community and get an answer from real people.
Submit

When calculating control limits, should values outside of two/three sigma be used even if their is no special cause?

We retest based on anything out of two sigma, so I was wondering when we recalculate our limits, should we use these data points.

Report as

Outliers represent data points that may be compromised, but unless you can show that they are errors they are still a valid part of the data set. Therefore, I would use all data points to calculate the standard deviation. Otherwise, you will narrow the control limits with each iteration, throwing out more and more outliers each time until the control limits are basically equal to the mean.

Helpful Fun Thanks for voting Comments (0)
Report as
Add a comment...

When calculating control limits, values outside of two/three sigma should not be used if there is no special cause. This is because a special cause is necessary for the variation of results to be achieved. Learn more from http://www.au.af.mil/au/awc/awcgate/navy/bpi_manual/mod10-control.pdf.

Helpful Fun Thanks for voting Comments (1)
Report as
You have no idea what this question is about, do you???
Report as
Add a comment...
Do you have an answer?
Answer this question...
Did you mean?
Login or Join the Community to answer
Popular Searches