Standard deviation for population


Hello all,

I have a population of 100 samples that obviously do not meet a Ppk of 1.33 for my drawing tolerances. This dataset is from a prototype, so I can talk with my client and modify the tolerances according to my process.

My first though was the following:
  • New nominal value: average of my data set.
  • New tolerances: +/- 4 standard deviation (stdev.p from excel)
But comparing the standard deviation from excel and minitab they result different. Does anybody know why the standard deviation (sample) from excel is the same as the standard deviation (overall) from minitab?

Should I only calculate these formulas with minitab?

I know the difference is very little, but I am very curious about this. Any source of information is greatly appreciated.

Best regards,


Forum Moderator
Use stdev.s in Excel. Your 100 samples are just that, samples. They are not a population in the sense that you are using them as a sample to predict the performance of future runs.

Top Bottom