Hi,

We have 2 sets of data, one is the mean and the other its corresponding Std. deviation. We have independent targets for both against which they are tracked. Can you let us know if there is any way of combining the two into an equivalent single metric and project.

Thanks

Sridhar

You can use Z-value, Hypothesis testing control limits along with actual range of deviation both mean and std.deviation observed to see whether process is within 3 sigma or 6 sigma limits. It is not tough to combine and set sigma limits and look for opportunities to improve.

if these data sets are independent then the below two METRIC can be used.

1- First is sigma level (Z)which uses mean & stdev both together along with specifications

2- CV (coefficient of variation) can also be used which is stdev divided by the eman of the data set. Its preferably used to compare the variation of two processes when their means are significantly different.

would like to be enlightened if further value additions are made by other members

Regards

SanjayBishnoi

MBB-Mumbai

Sridhar,

If you want to define a single metric to measure the combined improvement of the mean and the standard deviation, you must use the Process Capability measurement.

Cpk is the metric which you would look at.

Cp would tell you how godd or bad is your variation, and Cpk will tell you how far you are away from the target.

Also, please remember, though your business goals might force you to optimize BOTH the mean and the variation at the SAME time, please note, six sigma method will tell you to minimize the variation first and then the mean.

And even before you look at variation and mean, please also make sure that your process is under control and that the variation is not caused by any special causes - Clue to find it out.... is look at the outliers in a histogram or a box plot...

Hope these hints help you...

I realize you had posted it long back - Hopefully, you would get back and check this answer.