Mean adjusted rand index score across the random seed runs for each algorithm in each trial (an adjusted rand index score approaching 1 would represent almost complete agreement between cluster composition in the 10 random seed changes runs, whereas, a score of 0 represents almost no agreement). PAM = partitioning around medoids, HC = Hierarchical clustering, MOB = model based recursive partitioning, CF = Causal forest, XL-RF = X-learner with Random Forest (RF); XL-BART = Bayesian Additive Regression Trees (BART).