@inproceedings{10dacd57ee7d402daf11a7600dd444f2,
title = "Placebo Effect of Control Settings in Feeds Are Not Always Strong",
abstract = "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.",
keywords = "control settings, dark pattern, deception, placebo, social media, Twitter",
author = "Silas Hsu and Vinay Koshy and Kristen Vaccaro and Christian Sandvig and Karrie Karahalios",
note = "A big thanks to everybody that participated in our study. Additional thanks to the people in the Social Spaces group that gave feedback on this paper. This work was supported by NSF grant CHS-1564041 and The Center for Just Infrastructures at UIUC.; 2025 CHI Conference on Human Factors in Computing Systems, CHI 2025 ; Conference date: 26-04-2025 Through 01-05-2025",
year = "2025",
month = apr,
day = "26",
doi = "10.1145/3706598.3714197",
language = "English (US)",
series = "Conference on Human Factors in Computing Systems - Proceedings",
publisher = "Association for Computing Machinery",
booktitle = "CHI 2025 - Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems",
address = "United States",
}