2021/11/10 by Galen Weld, Amy X. Zhang, Weld, Galen +3 · 7 citations
Physics and Astronomy · Social Sciences · #Complex Network Analysis Techniques #Computers and Society (cs.CY) #FOS: Computer and information sciences #Human-Computer Interaction (cs.HC) #Knowledge Management and Sharing #Misinformation and Its Impacts #Social Media and Politics #Social and Information Networks (cs.SI)
paper · pdf · doi:10.48550/arxiv.2111.05835
openalex publication_date 2021/11/10 · openalex created_date 2022/10/29 · openalex updated_date 2026/07/28
Making online social communities 'better' is a challenging undertaking, as\nonline communities are extraordinarily varied in their size, topical focus, and\ngovernance. As such, what is valued by one community may not be valued by\nanother. However, community values are challenging to measure as they are\nrarely explicitly stated. In this work, we measure community values through the\nfirst large-scale survey of community values, including 2,769 reddit users in\n2,151 unique subreddits. Through a combination of survey responses and a\nquantitative analysis of public reddit data, we characterize how these values\nvary within and across communities. Amongst other findings, we show that\ncommunity members disagree about how safe their communities are, that\nlongstanding communities place 30.1% more importance on trustworthiness than\nnewer communities, and that community moderators want their communities to be\n56.7% less democratic than non-moderator community members. These findings have\nimportant implications, including suggesting that care must be taken to protect\nvulnerable community members, and that participatory governance strategies may\nbe difficult to implement. Accurate and scalable modeling of community values\nenables research and governance which is tuned to each community's different\nvalues. To this end, we demonstrate that a small number of automatically\nquantifiable features capture a significant yet limited amount of the variation\nin values between communities with a ROC AUC of 0.667 on a binary\nclassification task. However, substantial variation remains, and modeling\ncommunity values remains an important topic for future work. We make our models\nand data public to inform community design and governance.\n