NavList:

L.S.,

I do agree with Lars on this one. But that is because I take the standard deviation as a measure for the variance in the data, as texts like "Mathematical statistics and data analysis" by prof John A. Rice do.
Works with every distribution one can cook up.

Using the sample standard deviation as an (unbiased) estimator of the population standard deviation (all of Greg's lifetime observations) only works if the error (deviation between the observations and the actual - although unknown - measurand) is normally distributed. Which the Law of large numbers suggests one can do.

Regards,

Jaap

----------------------------------------------------------------
NavList message boards and member settings: www.fer3.com/NavList
Members may optionally receive posts by email.
To cancel email delivery, send a message to NoMail[at]fer3.com
----------------------------------------------------------------