the robust scale is normally outputed by the routine you used to estimate the MM. For example in R: library(robustbase) data(coleman) set.seed(0) RlmST <- lmrob(log.light ~ log.Te, data = starsCYG) RlmST$scale It's formula is simply: $$(\sum_{i=1}^n w_i)^{-1}(\sum_{i=1}e_i^2w_i)$$ where $e_i$ ($w_i$) is the residual (weight) (from the MM fit) associated with the $i^{th}$ observation.