Six Sigma in terms of Statistics - An Explanation and Discussion

[ralphsulser]

OK, thanks Al, Isee it now
Sorry to get

 

[wmarhel]

OK, thanks Al, Isee it now
Sorry to get

C'mon, you gotta cover that bad eye up and use the good one.

Wayne

 

[Jim Wynne]

Thank you, I stand corrected, it is 3.4 ppm I typed ppb. I am not a SS person so I did not remember if it was 1 or 1.5 shift.

It really doesn't make any difference because the 1.5 shift is a theoretical construct that doesn't hold water anyway.

 

[Statistical Steven]

It really doesn't make any difference because the 1.5 shift is a theoretical construct that doesn't hold water anyway.

But it does matter, because it justifies the 3.4PPM. Many of the metrics for SS are based on this concept (DPMO, etc.).

 

[Jim Wynne]

But it does matter, because it justifies the 3.4PPM. Many of the metrics for SS are based on this concept (DPMO, etc.).

I understand that, but the fact is that there is considerable controversy within the SS community itself on this subject--it's far from being an accepted idea. My own disdain for SS is based in part on the fact that SS practioners can't agree on the subject, which means that there's something lacking in the data.

It's pretty simple, really. Proponents of the 1.5-sigma shift just don't understand what really goes on in a production environment, and assume that changes in the mean (driven, of course, by changes in the individuals) will go undetected until some undetermined time in the future when the data is charted. The fact is that most operators make adjustments on the fly, in reaction to what they're seeing as they measure things, especially when tolerances are tight. The idea of the 1.5 shift is based on faulty reasoning--that operators must be hit over the head by a shift in the mean before actually doing anything, or that undetected shifts in the mean will result in faulty PPM calculations and inaccurate predictions.

 

[Statistical Steven]

I understand that, but the fact is that there is considerable controversy within the SS community itself on this subject--it's far from being an accepted idea. My own disdain for SS is based in part on the fact that SS practioners can't agree on the subject, which means that there's something lacking in the data.

It's pretty simple, really. Proponents of the 1.5-sigma shift just don't understand what really goes on in a production environment, and assume that changes in the mean (driven, of course, by changes in the individuals) will go undetected until some undetermined time in the future when the data is charted. The fact is that most operators make adjustments on the fly, in reaction to what they're seeing as they measure things, especially when tolerances are tight. The idea of the 1.5 shift is based on faulty reasoning--that operators must be hit over the head by a shift in the mean before actually doing anything, or that undetected shifts in the mean will result in faulty PPM calculations and inaccurate predictions.

I agree wholeheartedly. My issue is that the shift is not just arbitrary, but does not account for asymmetrical tolerances, where the process is shifted to account for drift.

So with all the issues within the SS community, why is it that it still has the clout to keep regular QE's out of jobs?

 

[e006823]

The following link is a good explantion about the origins of the 1.5 Sigma shift :

http://www.isixsigma.com/forum/showmessage.asp?messageID=39713

 

[Steve Prevette]

I agree wholeheartedly. My issue is that the shift is not just arbitrary, but does not account for asymmetrical tolerances, where the process is shifted to account for drift.

So with all the issues within the SS community, why is it that it still has the clout to keep regular QE's out of jobs?

And ASQ keeps religiously pushing Six Sigma, making things all the worse.

http://www.asq.org/media-room/press-releases/2006/20060221-milwaukee-county.html

But, oh well. Six Sigma bashing doesn't seem to be getting us anywhere. Just got to keep delivering good product. Idaho National Lab has requested I give them a day's training on control charts in March. . .

 

[Bev D]

guys - it's really simple: those who believe in the 1.5 sigma shift are hacks. they were the hacks for SPC in the eighties and ISO in the nineties and wil be the hacks for whatever the 'next big thing' is.
those who truly practice 'six sigma' don't talk about the shift because it means NOTHING to six sigma it was a marketing gimmick - a campaign slogan; it had nothing to do with solving problems. and only the hacks talk about 3.4ppm.

If you make fun of or deride the 1.5 sigma shift and 3.4 ppm you are doing the same thing as any good six sigma practioner did years ago. Even GE abandoned it shortly after starting their six sigma initiative because it was non value add, silly and irrelevent to the cause.

it doesn't matter. we need to move on. If we wish to discuss Six Sigma, let's discuss what it is not what it isn't.

 

[Statistical Steven]

guys - it's really simple: those who believe in the 1.5 sigma shift are hacks. they were the hacks for SPC in the eighties and ISO in the nineties and wil be the hacks for whatever the 'next big thing' is.
those who truly practice 'six sigma' don't talk about the shift because it means NOTHING to six sigma it was a marketing gimmick - a campaign slogan; it had nothing to do with solving problems. and only the hacks talk about 3.4ppm.

If you make fun of or deride the 1.5 sigma shift and 3.4 ppm you are doing the same thing as any good six sigma practioner did years ago. Even GE abandoned it shortly after starting their six sigma initiative because it was non value add, silly and irrelevent to the cause.

it doesn't matter. we need to move on. If we wish to discuss Six Sigma, let's discuss what it is not what it isn't.

I took some time to look into what Six Sigma is...and unfortunately, I found that it is just TQM rebundled.