Please could someone explain the following,
if i defien a reliability criterion as the rsd of a particular quantity,
e.g. a regression coefficient, so we have b(mean)/b(std). If the
number of samples from which this was gathered was small, less
than 30 I would assume that the rsd would be distributed as
Students t wityh n degrees of freedom, correct? What i need to
statistically find are those rsd's where b(mean) is significantly
greater then b(std), is it ok to simply apply a critical value and
hence set up a rejection region, and if so how does this work?
Thankyou for helping with this somewhat mundain problem
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|