What do you think of the Cohen's f² measure of "local" effect size ? It has been recently proposed (by Selya et al., 2012, "A Practical Guide to Calculating Cohen’s f², a Measure of Local Effect Size, from PROC MIXED") to evaluate the practical significance of individual effect sizes of particular variables between two nested regression models (including multilevel regression, but only if it is the fixed effects component that is concerned). This is what is meant by local effect size. See the link below for the appropriate Stata syntax.
http://www.ats.ucla.edu/stat/s...
I hope you'll write a post that covers this topic.