Start free trial
← Causal Inference 101

6 min read

Heterogeneous treatment effects & causal forests

Why average effects hide who benefits most — and how modern ML estimators find subgroups with different responses.

The average treatment effect (ATE) summarizes impact across everyone. Heterogeneous treatment effects (HTE) ask: does the effect differ by age, region, risk score, or behavior?

Knowing the ATE alone can mislead. A drug might be harmful on average but beneficial for a subgroup. A discount might lift conversions only for fence-sitters.

Causal forests and double ML

Causal forests split the covariate space to find regions with different estimated effects while controlling for confounding. Double machine learning (DML) combines flexible prediction with orthogonalization for high-dimensional confounders.

These methods shine when you have many covariates and want interpretable subgroup patterns, not just a single number.

From estimates to action

In marketing, HTE feeds targeting rules. In medicine, it supports personalized protocols within ethical bounds. In policy, it shows which communities gain most from an intervention.

Run this method on your data — no Python

CausalLens runs matching, DiD, causal forests, DoWhy refutation, and more — with balance tables, sensitivity checks, and PDF export.

Start 14-day Pro trialAll methodsPricing