This is a very basic description, so please, all the stat nerds out there cut me some slack (I’m one too). There are much longer responses. Sounds like you are fairly new to 6sigma, so I don’t want to get into issues like sigma shift or other various distribution/probability explanations. Please search the isxsigma site for these if you want more – it provides more detailed information.
When a process is being measured, we can normally expect to see values fall within 3 standard deviations of the mean and a control limit. For six sigma, you are improving a process to meet a target more often, either by shifting the mean toward the target and/or reducing the standard deviation to meet the target more often. Sigma in Six Sigma is a measure of probability that you will hit a target/goal. When we talk about sigma level, we are talking about target and spec, not mean and control limit. A common mistake is to confuse the sigma level to the engineering specs of the normal curve (e.g., x-bar +/- (3*std dev)).
If you can fit 6 standard deviations between the target and an upper specification (one side), you are operating at a 6 sigma level. At this level, you would expect only 3.4 defects per million opportunities to occur outside of one of your spec limits.
For example, a process has a target of 10, an upper spec limit of 16, and a std deviation of 2. Say your process mean is meeting your target of 10. In this instance, you could fit 3 std deviations between 10 and 16 (i.e., 10+(3*2)=16). The process is operating at 3 Sigma. If you reduce the std deviation to 1 (and the mean stays the same), you would be operating at a 6 sigma level (10+(6*1)= 16).
Again – very basic explanation. My guess is that you will get a few responses to my text, arguing how I’m off the mark a bit. The explanation above is the simplest I can make. Good Luck!
.
