English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We propose strategies to estimate and make inference on key features ofheterogeneous effects in randomized experiments. These key features includebest linear predictors of the effects using machine learning proxies, averageeffects sorted by impact groups, and average characteristics of most and leastimpacted units. The approach is valid in high dimensional settings, where theeffects are proxied (but not necessarily consistently estimated) by predictiveand causal machine learning methods. We post-process these proxies intoestimates of the key features. Our approach is generic, it can be used inconjunction with penalized methods, neural networks, random forests, boostedtrees, and ensemble methods, both predictive and causal. Estimation andinference are based on repeated data splitting to avoid overfitting and achievevalidity. We use quantile aggregation of the results across many potentialsplits, in particular taking medians of p-values and medians and otherquantiles of confidence intervals. We show that quantile aggregation lowersestimation risks over a single split procedure, and establish its principalinferential properties. Finally, our analysis reveals ways to build provablybetter machine learning proxies through causal learning: we can use theobjective functions that we develop to construct the best linear predictors ofthe effects, to obtain better machine learning proxies in the initial step. Weillustrate the use of both inferential tools and causal learners with arandomized field experiment that evaluates a combination of nudges to stimulatedemand for immunization in India.

Fisher-Schultz Lecture: Generic Machine Learning Inference on  Heterogenous Treatment Effects in Randomized Experiments, with an Application  to Immunization in India