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

Causal decision making (CDM) based on machine learning has become a routinepart of business. Businesses algorithmically target offers, incentives, andrecommendations to affect consumer behavior. Recently, we have seen anacceleration of research related to CDM and causal effect estimation (CEE)using machine-learned models. This article highlights an important perspective:CDM is not the same as CEE, and counterintuitively, accurate CEE is notnecessary for accurate CDM. Our experience is that this is not well understoodby practitioners or most researchers. Technically, the estimand of interest isdifferent, and this has important implications both for modeling and for theuse of statistical models for CDM. We draw on prior research to highlight threeimplications. (1) We should consider carefully the objective function of thecausal machine learning, and if possible, optimize for accurate treatmentassignment rather than for accurate effect-size estimation. (2) Confoundingdoes not have the same effect on CDM as it does on CEE. The upshot is that forsupporting CDM it may be just as good or even better to learn with confoundeddata as with unconfounded data. Finally, (3) causal statistical modeling maynot be necessary to support CDM because a proxy target for statistical modelingmight do as well or better. This third observation helps to explain at leastone broad common CDM practice that seems wrong at first blush: the widespreaduse of non-causal models for targeting interventions. The last two implicationsare particularly important in practice, as acquiring (unconfounded) data on allcounterfactuals can be costly and often impracticable. These observations opensubstantial research ground. We hope to facilitate research in this area bypointing to related articles from multiple contributing fields, including twodozen articles published the last three to four years.

Causal Decision Making and Causal Effect Estimation Are Not the Same... and Why It Matters