Why is ANOVA so slow for performing a variance decomposition on a sample with a large number of groups?

Is xtsum an interesting alternative? Are estimates from xtsum unbiased in the case of strongly unbalanced panel?

Similar questions and discussions