MD is absolutely correct on the notion that a sample of five or so is much too small to give precise estimates of long term SD. A 1" rifle will routinely print 1/2" to 1 1/2" groups with no change whatever in ammunition, technique, or the firearm. Five piece samples from a large population of ammunition with an SD of 25 will routinely produce SD estimates between about 15 and 75.
I'm not so sure about the notion that the uncertainty in screen spacing affects SD. I think that is more likely a constant error, which will not affect SD. Otherwise, I think it would not be possible to get SD readings in the single digits.
Last edited by denton; 11/26/15.
