Research:Understanding content moderation on English Wikipedia
The tasks of monitoring for and removal of content that violates Wikipedia’s guidelines is carried out by the community of volunteer editors. This study, conducted by researchers at the Berkman Klein Center, seeks to aid in the Wikimedia Foundation’s understanding and public discussion of Wikipedia's content moderation systems.
The study will provide an overview and assessment of English Wikipedia content moderation practices and policy implementation. The study will consist of several complementary components: a review of the regulatory context, a literature review of prior related research, development of a taxonomy of speech targeted for moderation, and two original research endeavors: a qualitative analysis based on interviews of volunteer Wikipedia content editors and a quantitative analysis of Wikipedia content removal, which will build upon prior efforts.
Ultimately, the findings from the project will help the Wikimedia Foundation to identify the strengths and weaknesses in the community’s efforts to moderate content. It will support Wikimedia’s efforts to educate lawmakers about the scale and efficacy of its community-driven model of removing harmful and illegal content. It will also provide a benchmark which can be returned to in order to better understand the effects of content removal laws and other regulatory trends on these community processes.
The goals of the project are as follows:
- Review, summarize, and document English Wikipedia guidelines and processes for content moderation.
- Review and summarize the current legal requirements for content moderation in the US, EU and India.
- Conduct a literature review of related scholarship.
- Develop a taxonomy of content categories and removal decisions, spanning illegal speech, harmful speech, vandalism, and other types of speech detrimental to Wikipedia and its community.
- Conduct a qualitative assessment of content moderation practices and policy implementation by Wikipedians based on interviews of volunteer Wikipedia content editors. This might also include informational interviews with NGO partners and collaborators.
- Building on prior efforts, conduct a quantitative content analysis of English Wikipedia articles and talk pages using the Perspective API. Sample Wikipedia across several dimensions (for example by topic, popularity, and editing protection levels) and assess the quantity of content removed across areas defined in the taxonomy. Note that some removed content of interest to this study (e.g. for legal reasons) is not publically available.
- Prepare a critical assessment of Wikipedia content moderation policies and structures and a comparative assessment of the same on a selection of other large platforms.
This project includes two original research endeavors:
We will conduct a qualitative assessment of content moderation practices by the Wikipedia community by interviewing Wikipedia content editors and administrators. We are seeking up to 16 Wikipedia editors and administrators with a range of experience to be interviewed about the processes and guidelines for content revision, content deletion, and quality control of English Wikipedia. The interviews will particularly focus on gaining an understanding of the community’s policies and decision making about handling harmful content both on articles and talk pages.
With the guidance of the Wikimedia Trust and Safety Team, we will post a call for volunteer interview subjects on the administrators’ noticeboard and are reaching out to individual administrators.
Quantitative content analysisEdit
Building on prior efforts, we will conduct a quantitative content analysis of English Wikipedia articles and talk pages using the Perspective API. We will sample Wikipedia across several dimensions (for example by topic, popularity, and editing protection levels) and assess the quantity of content removed across areas defined in the taxonomy.
The ultimate goal of the interviews and the quantitative analysis is to help the Wikimedia Foundation identify the strengths and weaknesses in the community’s efforts to moderate content and to improve the quality of content and positivity of conduct on the platform.
April - August 2019
Policy, Ethics and Human Subjects ResearchEdit
This study was reviewed by the Harvard University-Area IRB and approved on July 12, 2019, meeting criteria for exemption from IRB review per the regulations found at 45 CFR 46.104(d).
The researchers conclude that Wikipedia is largely successful at identifying and quickly removing a vast majority of harmful content despite the large scale of the project. The evidence suggests that efforts to remove malicious content are faster and more effective on Wikipedia articles compared to removal efforts on article talk and user talk pages. An ongoing challenge for Wikipedia is moderating and addressing harms in the interactions between community members. Wikipedia’s decentralized structure empowers editors to act quickly and decisively to interpret what constitutes damage and harm on the encyclopedia. Decentralization also results in inconsistency in the application of the guidelines and standards that guide moderation decisions, although this appears to be offset by the benefits and agility of autonomous context-guided decision making.
The full report was published on November 21, 2019 is available for download.