RQL/AQL from Attribute Data

P

PaulHurley

Hi all - a stats novice and my first post so it's all happening!!! :D
I have some historical attribute data from our process and need to establish the process performance.

I've read that I cannot calculate a Ppk/Cpk from attribute data - but if anyone can help me to do that it would be ideal.

How do I use pass/fail data (297 fails from 7947 inspected) to give statement of process performance.

:thanks:
 

Bev D

Heretical Statistician
Leader
Super Moderator
Hi Paul and welcome to the cove! :bigwave:

What you can say about the process performance is the defect rate just divide the number of failures by the number inspected.

A better approach is to plot the defect rate over time (by lot or day or whichever makes most sense) on a p chart - this will show the variation of the process over time and let you know if it is stable or if there were times when you a lot better or a lot worse. This might provide clues as to causes to help you improve the performance.

You can calculate the Cpk value by 'back calculating' from the defect rate but why would want to do that? It really provides no more insight than simply knowing the defect rate and not as much insight as the control chart...

May I ask what your ultimate goal is?
 
Top Bottom