Experiment Registration for The Consumer Welfare Effects of Online Ads: Evidence from a 9-Year Experiment

How to cite

Garro, Haritz; Liaqat, Asad; Kutzman, Daley; Deisenroth, Daniel; Wernerfelt, Nils; Collis, Avinash; Brynjolfsson, Erik. 2025. "Experiment Registration for The Consumer Welfare Effects of Online Ads: Evidence from a 9-Year Experiment." AEA RCT Registry. AEARCTR-0016160. https://doi.org/10.1257/rct.16160-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Haritz Garro* , Asad Liaqat , Daley Kutzman , Daniel Deisenroth , Nils Wernerfelt , Avinash Collis , Erik Brynjolfsson

* Primary investigator

Registry ID
AEARCTR-0016160
DOI
https://doi.org/10.1257/rct.16160-1.0
First registered
Last updated
Status
Completed
Trial dates
2013-07-01 – 2025-06-01
Countries
Belgium (), Canada (), Germany (), Spain (), France (), United Kingdom of Great Britain and Northern Ireland (), Ireland (), Japan (), Korea (Democratic People's Republic of) (), Mexico (), Norway (), Romania (), United States of America ()
Registry record
http://www.socialscienceregistry.org/trials/16160

Abstract

Research on the effects of online advertising on consumer welfare is limited due to challenges in running large-scale field experiments. We analyze a long-running field experiment on Facebook in which a random subset of users received no ads in their newsfeeds. Using an incentive - compatible deactivation experiment, we find no significant differences in users’ valuation of Facebook across a representative sample of 53,083 Facebook users in the ads and no ads groups. Our sample size allows for precise estimates, suggesting that either the disutility of ads is relatively small or that there are offsetting benefits, such as product discovery.

Intervention

We analyze a long-running field experiment on Facebook in which a random subset of users received no ads in their newsfeeds experiment to understand how users in the ads and no ads groups value Facebook. We measure users' valuation of Facebook across a representative sample of 53,083 Facebook users using an incentive-compatible deactivation experiment.

Experimental design

We leveraged a long-running and pre-existing field experiment on Facebook in which a random subset of users received no ads in their newsfeeds experiment.

Design details

Randomization method
When a user joins Facebook, she is given a unique user ID. At the same time, a hash of that user ID will randomly determine whether that user is in the no ads or ads condition. Whatever the assignment is, the user will stay in that condition for the entire duration of her time on Facebook. This means that the users in our no ads condition have never seen ads in their newsfeed. Further, as the Facebook population grows, new users are continually added to both arms of the experiment.
Randomization unit
Facebook user
Clusters
1
Observations
53,083 Facebook users
Arms
39,647 in the ads arm, and 13,436 in the no-ads arm
Primary outcomes
User valuation of Facebook

Papers from this trial

  1. "The Consumer Welfare Effects of Online Ads: Evidence from a 9-Year Experiment" Erik Brynjolfsson, Avinash Collis, Daniel Deisenroth, Haritz Garro, Daley Kutzman, Asad Liaqat, Nils Wernerfelt - American Economic Review: Insights (Forthcoming) https://www.aeaweb.org/articles?id=10.1257/aeri.20240452&&from=f

Cited by

No citation data available for this registration yet.

Keywords