My team is conducting an analysis using a large dataset of health data from >150 demographic and health surveys with the aim of identifying optimal metrics to quantify early childhood growth at the population level.
We are interested in identifying robust metrics, using model predictions or regression approaches, that can more accurately/reliably inform between- and within-country comparative analyses of child health status. After much investigation of the literature and various statistical packages, we believe that robust regression (e.g. deriving MM-estimators) could be a suitable approach for identifying candidate metrics; however, we are unsure whether this can be implemented in R, given our complex survey sample design variance estimation (using stratum/cluster variables plus a weight).
I would appreciate any guidance!