Ariel Linden
Linden Consulting Group, LLC
In randomized controlled trials (RCT) of sufficient size, we expect the treatment and control groups to be balanced on both observed and unobserved characteristics, and any imbalances are considered to be due to chance. CTA can be used to determine whether treatment assignment can be predicted by observed pre-intervention covariates–separately or interacted with other covariates. In this paper, I describe how to assess the quality of the randomization process in RCTs using the new Stata package for implementing CTA.
odajournal.com 