Tolerance Stacking

Home > Statistics Every Writer Should Know > The Stats Board > Discusssion

Tolerance Stacking
Message posted by Tommy on August 9, 2000 at 12:00 AM (ET)

In an assembly if parts (length of rod) each having normal distributions and being independent, the mean standard deviation can be calculated by using the variances of the individual parts. Why must the variances be used as opposed to the the standard deviations?

READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Tolerance Stacking
Message posted by nancy diehl on August 10, 2000 at 12:00 AM (ET)

Variances are additive, standard deviations are not. When we compare two groups of data we always compare
their variances, not their standard deviation (F-Test). When we pool the variability of two groups of data,
we pool their variances, not their standard deviation. Any combining is done via the variances. When all is done,
the standard deviation can be determined and is often preferred since it's back to the original units of measure.
But the bottom line is, you add variances not std dev., hence the term, Root Sum Square.

Your $5 contribution helps cover part the $500 annual cost of keeping this site online.