How do you estimate heterogeneous treatment effects (HTE)?

hard

Answer

$\mathrm{CATE}(x)\; = \;E[Y(1)\; - \;Y(0)\; \mid \;X\; = \;x]$
$Methods: (1) interaction terms in regression (Y ~ T + T\timesX); (2) causal forests / GRF (Wager & Athey); (3) meta-learners: T-learner (separate models), S-learner (single joint), X-learner (better in imbalanced treatment), R-learner; (4) doubly-robust CATE.$
$Uses: personalized recommendations, targeted policies (Uplift modeling in marketing).$
$Requires overlap + unconfoundedness (or RCT).$

Check yourself — multiple choice

Random
$\mathrm{CATE}\; = \;E[Y(1) - Y(0) \mid X]$ ; causal forests, T/S/X/R-learners, DR-learners; foundation of uplift modeling and personalized targeting
Same as ATE
Not real

HTE / CATE: causal forests + meta-learners; foundation of uplift modeling.

#causal-inference

215 interview questions in this topic.