Gang Wrote:
-------------------------------------------------------
> Michael,
>
> > This is referring to standard deviation,
> correct? Additionally, How would you calculate the
> standard error?
>
> Regarding standard deviation vs. standard error:
> standard deviation usually refers to the extent of
> how spread or concentrated the data (or samples)
> are, while standard error is more used to
> described the accuracy or reliability of an
> estimator (e.g., estimated coefficient in a
> model).
>

To calculate the standard devation you said the formula is:

beta divided by t (or z)

The beta is the subbrik before the z subbrik in the output of 3dLME?

What is the calculation for the standard error? Does it differ from the traditional calculation of standard-deviation/sqrt(n).



> > Why is it an approximation? Because it is
> derived from a mixed model as opposed to a
> standard glm?
>
> Yes, the statistics from a linear mixed-effects
> model are mostly approximate based on asymptotic
> properties. In other words, they are more accurate
> if the sample size is reasonably large enough.


Great. Thanks!