@inproceedings{manzoor-etal-2022-status,
    title = "Status Biases in Deliberation Online: Evidence from a Randomized Experiment on {C}hange{M}y{V}iew",
    author = "Manzoor, Emaad  and
      Jo, Yohan  and
      Montgomery, Alan",
    editor = "Goldberg, Yoav  and
      Kozareva, Zornitsa  and
      Zhang, Yue",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2022",
    month = dec,
    year = "2022",
    address = "Abu Dhabi, United Arab Emirates",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/fix-sig-urls/2022.findings-emnlp.474/",
    doi = "10.18653/v1/2022.findings-emnlp.474",
    pages = "6351--6363",
    abstract = "Status is widely used to incentivize user engagement online. However, visible status indicators could inadvertently bias online deliberation to favor high-status users. In this work, we design and deploy a randomized experiment on the ChangeMyView platform to quantify status biases in deliberation online. We find strong evidence of status bias: hiding status on ChangeMyView increases the persuasion rate of moderate-status users by 84{\%} and decreases the persuasion rate of high-status users by 41{\%} relative to the control group. We also find that the persuasive power of status is moderated by verbosity, suggesting that status is used as an information-processing heuristic under cognitive load. Finally, we find that a user{'}s status influences the argumentation behavior of other users they interact with in a manner that disadvantages low and moderate-status users."
}