From 73b37d5a2fd81c9258400b5bc10d9a1bbc733d36 Mon Sep 17 00:00:00 2001 From: Bertil Hatt Date: Wed, 13 Nov 2024 12:22:03 +0100 Subject: [PATCH] Include mention of the Delta method in the Clustered Analysis page --- docs/experiment-analysis/clustered-analysis.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/experiment-analysis/clustered-analysis.md b/docs/experiment-analysis/clustered-analysis.md index e862de71..1ad25222 100644 --- a/docs/experiment-analysis/clustered-analysis.md +++ b/docs/experiment-analysis/clustered-analysis.md @@ -20,10 +20,10 @@ When creating an experiment, choose the randomization entity as usual. Then in ` ![Create clustered analysis experiment](/img/experiments/clustered-analysis/clustered-analysis-experiment.png) -Once created and running, there are a few differences on how experiment results are calculated. Notice that assignment numbers will refer to the entity but subentity numbers will also be available. Metrics will be calculated by using the subentity as the denominator. +Once created and running, there are a few differences on how experiment results are calculated. Notice that assignment numbers will refer to the entity but subentity numbers will also be available. Metrics will be calculated by using the subentity as the denominator. For that, we leverage the Delta method described in [Deng et al.](https://alexdeng.github.io/public/files/kdd2018-dm.pdf) and recommended in Chapter 18 of _Trustworthy Online Controlled Experiments_. ![Subentities in experiment results](/img/experiments/clustered-analysis/metrics-with-subentities.png) Explore charts behave the same way. Assignment numbers are reported, but metrics and their split bys will use the subentity numbers for the denominator. -![Subentities in Explore charts](/img/experiments/clustered-analysis/explore-with-subentities.png) \ No newline at end of file +![Subentities in Explore charts](/img/experiments/clustered-analysis/explore-with-subentities.png)