Sigma versus sigma

The originators of the Six Sigma initiative use a unit of measurement, a ‘Sigma’, to measure performance, the higher the value of Sigma the better the performance. For example, a 6 Sigma process is rated better than a 5 Sigma one. This meaning contrasts with the standard statistical meaning and usage of the term, sigma. The symbol, sigma, a Greek letter symbolizing a statistical measure of variation, termed the standard deviation is in universal use. The lower the value of the standard deviation, ‘sigma’, the less the variation present and the better the performance of a characteristic. The opposing meanings inevitably cause confusion. This confusion can be compounded as the internationally accepted standard ‘sigma’ provides the derivation of the Sigma metric used in Six Sigma projects and also the relationship between non-conformities per million opportunities and the Six Sigma, Sigma.

Thoughts
An advantage of the Sigma measure is its simplicity and practicality. This appeals to all those who do not wish to get too embroiled in statistical niceties but just want a simple readily understandable scale of performance measurement. The fact that 6 Sigma denotes something better than 5 Sigma and that 5 Sigma is better than 4 Sigma, and so on makes good practical sense to them and they are quite prepared to run with it. Others, who are already actively engaged in the application of statistical process control, will readily understand the use of Sigmas in a different way to sigmas but may feel uncomfortable and uneasy with both the frailty of the underlying concept and the outcome. This is because an alternative soundly based approach is available with the use of the internationally defined measured data process capability indices such as Cp. A Cp of 3 indicates less variation than a Cp of 2, and so on. Such measures require the pre-establishment of both process stability and a knowledge of the underlying pattern of variation.