Placebo Effect of Control Settings in Feeds Are Not Always Strong

要旨

Recent work has catalogued a variety of ``dark'' design patterns, including deception, that undermine user intent. We focus on deceptive ``placebo'' control settings for social media that do not work. While prior work reported that placebo controls increase feed satisfaction, we add to this body of knowledge by addressing possible placebo mechanisms, and potential side effects and confounds from the original study. Knowledge of these placebo mechanisms can help predict potential harms to users and prioritize the most problematic cases for regulators to pursue. In an online experiment, participants (N=762) browsed a Twitter feed with no control setting, a working control setting, or a placebo control setting. We found a placebo effect much smaller in magnitude than originally reported. This finding adds another objection to use of placebo controls in social media settings, while our methodology offers insights into finding confounds in placebo experiments in HCI.

受賞
Best Paper
著者
Silas Hsu
University of Illinois at Urbana-Champaign, Urbana, Illinois, United States
Vinay Koshy
University of Illinois at Urbana Champaign, Champaign, Illinois, United States
Kristen Vaccaro
University of California San Diego, San Diego, California, United States
Christian Sandvig
University of Michigan, Ann Arbor, Michigan, United States
Karrie Karahalios
University of Illinois at Urbana-Champaign, Urbana, Illinois, United States
DOI

10.1145/3706598.3714197

論文URL

https://dl.acm.org/doi/10.1145/3706598.3714197

動画

会議: CHI 2025

The ACM CHI Conference on Human Factors in Computing Systems (https://chi2025.acm.org/)

セッション: Dark Patterns and Content Moderation

G304
6 件の発表
2025-04-30 23:10:00
2025-05-01 00:40:00
日本語まとめ
読み込み中…