Laboratory Statistical Methods - 4.10.6.6 in QS-9000

[Zanzi]

laboratory Statistical Methods

Can anyone help me with understanding the requirements of clause 4.10.6.6 Laboratory Statistical Methods. It states that:

Appropriate statistical techniques should be applied to verification activities whose deliverables are date.

Any ideas would be most appreciated.

Thanks

 

[KenK - 2009]

Are you sure the phrase isn't

"...whose deliverables are data."???

I would assume it means that you can and do use appropriate statistics to provide summary information. This can be as simple as providing the mean, standard deviation, median, percentiles, etc...

Also, if you are verifying an instrument, you might need to use confidence interals or comparison tests (t-tests) as part of your decision criteria.

 

[Atul Khandekar]

Since this is about lab, it may include measurement of uncertainty
( or uncertainty of measurement ?), Calibration related data etc...

- Atul.

 

[Marc]

Measurement Uncertainty

KenK and Atul Khandekar pretty well summed things up for you.

Yup - most labs should be able to explain how they address uncertainty of instruments and as early as the design phase, but at least well into the 'quality Planning' stage, uncertainty of the measurement system should be addressed and understood in the context of the product. This is to say in a company injection molding 1 inch plastic 'dolls' for Cracker Jacks boxes uncertainty and some other aspects of the Measurement System(s) will not be as important as in a situation where the company is measuring optics for telescopes or crank shafts for cars.

In QS, measurment as a system was really pushed hard and is being pushed harder. The big push started in 1996-1997. Prior to that few folks had heard of measurement uncertainty, much less knew how to determine it. Also see the AIAG's MSA - Measurement Systems Analysis 'manual'.

Don't forget you must understand uncertainty principles and be able to state the uncertainty of the instrument(s) used to make the measurement as well as the system as a whole.