I've found I've lost a handy table that gave the 5% and 1% limits for
declaring an observation and outlier, depending on the sample size.
Obviously, a standard deviation is going to be influenced by an outlier
(especially for n=3), and this table accommodated that, providing limits
based on the distance from the mean using the sd of all the
observations.
Can anybody provide guidance in this area?
Thanks, Todd
|