Statistical methods in ISO 9001:2000

[Jase Eyre]

Hi y'all

A company here in Australia is sending out promotional material that states the following:

"Inevitable changes to ISO/QS9000 make the use of statistical methods compulsory to ALL organisations that wish to meet the standard. Don't let this become a burden, attend our seminars..."

I wasn't aware that statistical process control methods were 'compulsory' in the new standard. That certainly wasn't how I read it. Am I wrong?

------------------
JasE

 

[David Mullins]

How else can you scare up some attendees to a seminar on stats? (rhetorical question - sorry)

9001:1991 Requires us to identify the need for statistical techniques required.
9001:2000 (draft) Requires us to determine the need for statistical techniques.
Identify vs Determine.
I believe ISO is merely re-enforcing the stance that if you don't apply statistical techniques then you must be able to prove (objective evidence) that such a determination took place, and that it is appropriate to your business.

Do you have measuring and monitoring systems?
Should you have measuring and monitoring systems?
If not, prove to me how you DETERMINED that there would be no benefit to the organisation or the customer in applying statistical techniques!


------------------

 

[Andy Bassett]

Im not sure of the detail, but after my brief reading for the new standard i beleive that they are seeking more statistical or objective evidence that the system is working and that your customers are satisfied.

With this in mind i wont be lauching extra SPC activities, but i will encourage customers to gather more objective evidence(stats if possible) that all is well ie Customer Satisfaction Surveys, Measurements for each Process etc.

Regards



------------------
Andy B

 

[Marc]

Originally posted by David Mullins:

Identify vs Determine.

I believe ISO is merely re-enforcing the stance that if you don't apply statistical techniques then you must be able to prove (objective evidence) that such a determination took place, and that it is appropriate to your business.

Do you have measuring and monitoring systems?
Should you have measuring and monitoring systems?
If not, prove to me how you DETERMINED that there would be no benefit to the organisation or the customer in applying statistical techniques!

I agree. Misinformation is rampany, especially when some lamer is trying to hawk their software, seminars, canned procedures and such.

 

[Marc]

Originally posted by Andy Bassett:

....i beleive that they are seeking more statistical or objective evidence that the system is working and that your customers are satisfied.

With this in mind i wont be lauching extra SPC activities, but i will encourage customers to gather more objective evidence(stats if possible) that all is well ie Customer Satisfaction Surveys, Measurements for each Process etc.

I separated SPC from Statistical Techniques long ago. SPC is only 1 of many, many possible statistical techniques. I think you're making a wise decision.

 

[Laura M]

Let me qualify my statement by saying I haven't read ISO9000:y2k. Having said that, I wonder if the emphasis on continuous improvement is of the same vain as in QS9000...page 19...."the supplier shall develop a prioritized action plan for continuous improvement in processes that have demonstrated stability, acceptable capability and performance" Note above provides further clarrification. This would imply uses of statistical analysis.

Just a thought...again, I don't know the CI language in ISO9000:y2k

Laura

 

[Marc]

Originally posted by Laura M:

....This would imply uses of statistical analysis.

I have only had 2 clients who got away with no statistical techniques. Both were small ( <20 people) and could show they considered possible places to use statistical techniques but did not feel (believe or whatever) that statistics would benefit them. Auditors have a hard time with this (as do I to some degree) however the requirement was not, nor do I believe it is in the new version, outright required. Reality is larger companies have little choice. Like the US Census, to 'measure' the country (quite a few people and since everyone won't stand in one place for a year or two so they can be counted 1 by 1) requires (despite republicans requiring adhereance to an outdated interpretation of 'Talk to every single sole') statistical analysis techniques. On the other hand, to inventory your house and the houses of 2 of your neighbors ONLY really doesn't lend its self to statistical analysis unless you plan to interpolate the findings to predict what you will find in the country as a whole (population prediction).