Quantile treatment effect (QTE) vs ATE.
hardAnswer
- ATE captures mean effect.
- ^(-1)^(-1)(τ) captures effect on the τ-th quantile — useful when the mean is misleading (heavy tails, revenue metrics dominated by top 1%).
- Example: an A/B test with +1 at the median → most users hurt, a few whales lift the average.
- Estimation: quantile regression, IPW-quantile, causal forests.
- Standard tail-metric analysis at big tech.
Check yourself — multiple choice
- Same as ATE
- QTE(τ) = effect on τ-th quantile; matters for heavy-tail / revenue metrics where mean is misleading; via quantile regression or IPW-quantile
- Random
- Not real
QTE: effect on quantile τ; useful under heavy tails / revenue metrics.
#ab-testing#causal-inference
Practise Statistics Fundamentals
215 interview questions in this topic.