The official site of Braun and Schwartz (2021) “The A/B Test Deception…”

News

The working paper is available on SSRN

A/B Test Deception refers to bias generated by divergent delivery and response heterogeneity in results of “randomized” experiments conducted in a targeted advertising environment. A/B tests, as described in Braun and Schwartz (2022).

Now, with a Shiny App!!!!

The purpose of the app is to illustrate the bias under response heterogeneity and different targeting policies, as described in Braun and Schwartz (2022).

To avoid lots of issues with hosting, authentication, credentials, etc., we are sharing the app as an R package. Once installed, you can run the app with just one command.

Installation

You can install the released version of ABTestDeception from github

install_github("braunm/ABTestDeception")

Running the app

Once you have installed the package, launch the app.

ABTestDeception::launch()

That’s it.

Understanding the app

At this time, you probably need to read the paper to understand what all the symbols mean. We plan to write some instructions… eventually.