From 4eb756b1132d91cdfdbf8b58c3e131d1f2e25b24 Mon Sep 17 00:00:00 2001
From: NickCH-K <44816781+NickCH-K@users.noreply.github.com>
Date: Tue, 7 Sep 2021 22:10:26 -0700
Subject: [PATCH] Use only training data to train in Stata

---
 Machine_Learning/causal_forest.md | 1 +
 1 file changed, 1 insertion(+)

diff --git a/Machine_Learning/causal_forest.md b/Machine_Learning/causal_forest.md
index 233da72c..ece27c62 100644
--- a/Machine_Learning/causal_forest.md
+++ b/Machine_Learning/causal_forest.md
@@ -158,6 +158,7 @@ keep `r(varlist)'
 rcall: df.hold <- st.data()
 restore
 * Now go back to just the training data
+keep if split == 1
 
 * Run causal_forest, storing the effect predictions for the training data in the "effects" variable
 * the SEs of those effects in effectSE