Please use this identifier to cite or link to this item:
http://arks.princeton.edu/ark:/88435/dsp0173666786w
Title: | Investigating the Effectiveness of Ad Controls for Targeted Advertising on Facebook |
Authors: | Castleman, Jane |
Advisors: | Korolova, Aleksandra |
Department: | Computer Science |
Class Year: | 2024 |
Abstract: | Targeting advertising controls, such as marking "See less" to certain ads, are touted by platforms as a mechanism for users to express their preferences and exert control over the ad delivery algorithm. Our study aims to investigate the effectiveness of the "See less" targeting advertising control on Facebook and the analyze the targeting explanations attached to each ad. First, we conduct an anecdotal analysis of the misalignment between user-facing ad controls and advertiser-facing ad campaign settings, and test marking "See less" to Body Weight Control}on our own Facebook account. Next, we recruit 201 participants and randomly assign them the intervention of marking "See less" to the topics of Body Weight Control or Parenting, collecting data prior to the intervention, one hour post-intervention, and more than 72 hours post-intervention. Our results suggest that the "See less" ad control is ineffective, finding no evidence of a significant decrease in the proportion of related ads delivered pre-intervention versus more than 72 hours post-intervention, and the Body Weight Control topic saw a slight increase in the proportion of related ads over time. Participants in correlated demographics with their topic of interest saw significantly more related ads post-intervention than participants in non-correlated demographics. In our analysis of ad targeting criteria, the majority of ad targeting explanations were not aligned with local ad content and did not provide users with sufficient information to understand why related ads were delivered post-intervention. To remedy these issues, Facebook must adjust their ad delivery algorithm to penalize ads relating to blocked topics, align their ad controls with the shift to machine learning-based ad audience selection, and increase targeting transparency in ad targeting explanations. Overall, sociotechnical audits of platform algorithms are crucial for validating platform policies and upholding legal standards. |
URI: | http://arks.princeton.edu/ark:/88435/dsp0173666786w |
Type of Material: | Princeton University Senior Theses |
Language: | en |
Appears in Collections: | Computer Science, 1987-2024 |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
CASTLEMAN-JANE-THESIS.pdf | 3.58 MB | Adobe PDF | Request a copy |
Items in Dataspace are protected by copyright, with all rights reserved, unless otherwise indicated.