If you compare the sigma value on this site with other values published in most six sigma literature, it’s important to know that a 1.5 σ “shift” is usually added. For example, if you see a sigma value of 1.08σ on measuringusability.com and want to compare it to other sources then add 1.5. The resulting process capability will be 2.58 σ.
The 1.5σ shift is a somewhat controversially calculated value derived from Michael Harry’s work observing manufacturing processes while at Motorola. The shift is best incorporated for processes that last over a “long” period of time. There is no data supporting a shift, much less one with the magnitude of a 1.5σ shift for software processes. It may exist, but until it’s well documented I leave the shift out unless I’m comparing the sigma value across industries.
Most sigma value charts from Six Sigma consulting firms include the shift in their calculation of DPMO (defects per million opportunities) so be sure to look for it if you’re comparing sigma values after using the NORMSINV formula from Excel which doesn’t include a shift.
Finally, people have been using the standard normal z table for decades. It’s become second nature to anyone using a control chart or is used to analyzing data that 99.7% of data in a normal distribution falls within + or – 3 standard deviations of the mean.
For a lengthy discussion of the controvery of the 1.5σ shift you can read some of the debate at isixsigma.com
|Denver UX Boot Camp: August 16-18 3 Days of Hands-On Training on UX Methods, Metrics and Analysis|