minmse performs treatment assignment for (field) experiments based on pre-treatment information, i.e., it assigns treatment group number(s) to observations based on covariates specified by the user. It is suited for experiments with one or several treatment groups. The created treatment groups are balanced with respect to the specified covariates according to the minMSE-method proposed by Schneider and Schlather (2017) ; the package implements their treatment assignment method (based on work by Kasy, 2016). Pre-treatment information to be considered can be continuous and multivariate (i.e., several variables). Optimization is performed using the stochastic simulated annealing algorithm (Kirkpatrick, Gelatt, and Vecchi, 1983).
We provide R and Stata implementation for treatment assignment using the minMSE method as proposed by Schneider and Schlather (2017).
Stata implementation: Get the Stata ado package ‘minMSE’. You can also install the package directly in Stata:
ssc install minmse
Need help with automatic installing? Try this brief explanation or this more extended explanation.
Need help with manual installing the ado- and help-files into the appropriate folder? Have a look at this FAQ entry.
If all else fails: Contact me, I’m happy to help.
Looking for the R implementation? See R Implementation of the minMSE treatment assignment method for one or multiple treatment groups.