From Niche to Mainstream: Community Size and Engagement in Social Media Conversations

Jacopo Nudo, Matteo Cinelli, Andrea Baronchelli, and Walter Quattrociocchi · 2025

Read From Niche to Mainstream: Community Size and Engagement in Social Media Conversations by Jacopo Nudo, Matteo Cinelli, Andrea Baronchelli, and Walter Quattrociocchi (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-12076.txt.

Introduction

Jacopo Nudo, Matteo Cinelli, Andrea Baronchelli, and Walter Quattrociocchi study how community size shapes talk on social media in a 2025 paper. Platforms now host public debate at a huge scale. The team asks whether the size of a platform or a thread changes how people join in and how long they stay.

They tracked user behavior across six platforms over more than three decades. Smaller platforms held richer, longer exchanges, while larger ones drew wider but briefer input. As communities grew, users were less likely to return to a thread. Niche spaces broke that rule, keeping strong repeat input. The authors link these patterns to how each platform is built.

Catalog record

CREATORJacopo Nudo, Matteo Cinelli, Andrea Baronchelli, and Walter Quattrociocchi
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE29 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-12076

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

From Niche to Mainstream: Community Size and Engagement in Social Media Conversations Jacopo Nudo, Matteo Cinelli, Andrea Baronchelli, and Walter Quattrociocchi Abstract The architecture of public discourse has been profoundly reshaped by social media platforms, which mediate interactions at an unprecedented scale and complexity. This study analyzes user behavior across six platforms over 33 years, exploring how the size of conversations and communities influences dialogue dynamics. Our findings reveal that smaller platforms foster richer, more sustained interactions, while larger platforms drive broader but shorter participation. Moreover, we observe that the propensity for users to re-engage in a conversation decreases as community size grows, with niche environments as a notable exception, where participation remains robust. These findings show an interdependence between platform architecture, user engagement, and community dynamics, shedding light on how digital ecosystems shape the structure and quality of public discourse. 1 Center of Data Science and Complexity for Society, Department of Computer Science, Sapienza University of Rome, Rome, Italy 2 Department of Mathematics, City St George’s, University of London, and The Alan Turing Institute, London, UK 1 Introduction Social media platforms have transformed online participation, becoming integral to daily life as primary sources of information, entertainment, and communication [ 1 , 2 , 3 ] . Although these platforms provide unprecedented opportunities for connectivity and information sharing, their integration with complex social dynamics has raised significant concerns about their social impact. A major challenge in studying how platforms influence human behavior is the lack of accessible data [ 4 ] , and even when researchers obtain data through special agreements with companies such as Meta, it may still be insufficient to distinguish inherent behaviors from the effects of platform design [ 5 , 6 , 7 , 8 ] . This difficulty arises because the data, deeply embedded in platform interactions, complicate separating intrinsic human behavior from the influences exerted by the platform’s design and algorithms. Research has delved deeply into issues such as polarization, misinformation, and antisocial behaviors in digital spaces [ 9 , 10 , 11 , 12 , 13 ] , highlighting the intricate and multifaceted effects of social networks on public discourse. In this context, a key area of research focuses on identifying differences in dialogue dynamics across platforms, exploring whether these are shaped by algorithmic or design choices, or by community-driven factors [ 14 , 15 , 16 ] . The attention a piece of content gets on a platform helps measure its reach, impact, and spread. When users comment repeatedly on a post, it shows they are deeply engaged and willing to share their thoughts [ 17 , 18 ] . This study compares public conversations, combining a long-term perspective with a multi-platform dataset to examine their dynamics across different digital environments. We define an interaction as the comments a user posts within a thread. To measure user engagement—and, by extension, their attention—we analyze the number of comments left in a thread and the probability of re-entering the conversation after posting, following the methods outlined in [ 19 ] . We investigate how these aspects are influenced by the platform where the conversation takes place and by size of the context, defined either as the number of people commenting in a thread or as the number of people that belong to a community or a page where the conversation is hosted. Looking at the probability re-entry in a thread, that is similar to the probability of leaving more than one comment in a conversation, we notice its sensitiveness to size of the context, analogous to quorum sensing in bacteria, where population density governs collective behaviors through the exchange of signaling molecules [ 20 ] ; much like how the behavior of individual particles in a system often depends critically on the size of the system itself, as observed in phenomena like finite-size effects in phase transitions and critical phenomena, where collective behavior dominates individual dynamics [ 21 ] . So the main research questions are: RQ1 Do users maintain consistent participation and re-entry patterns across various platforms? RQ2 How does the size of the crowd participating in a conversation affect the probability of user re-entry? RQ3 How does the outreach size of the community hosting the conversation impact the probability of user re-entry? Using data from [ 14 ] , we systematically compare threads (or conversations) linked to posts on Usenet, Gab, Reddit, Voat, Twitter, and Facebook to answer these questions. The conversations are from a wide range of topics (vaccines, conspiracies, politics, news, Brexit, climate change) and cover 33 years. This research provides actionable insights for platform designers and policymakers seeking to foster healthier online conversations. 2 Related works The academic literature has explored various aspects of user behavior in online communities. Studies such as [ 19 ] analyzed the likelihood of users re-entering a conversation after leaving a comment, using specific participation patterns as predictors. Researchers studied the structure of the conversation on Reddit, intended as a tree graph, and analyzed the local and global features, like post sentiment and Subreddit popularity [ 22 ] . While [ 23 , 24 ] investigated how the structure is connected to the toxicity level of the comments produced. Additionally, [ 18 ] suggests that persuasiveness increases as comments become more deeply nested in a thread structure. Works like [ 25 ] analyze how users explore diverse options before settling into specific communities, highlighting their persistence within them and their consistency in discussing similar topics despite volatile engagement across different communities. Studies such as [ 26 ] and [ 27 ] analyze loyalty and turnover among user or consumer in groups or subscription programs, looking for factors that influence it. In [ 28 ] , researchers demonstrated how variability in user activity changes significantly with the size of the community on Reddit, suggesting implications for platform design. Relatedly, literature on the attention economy investigated human attention [ 29 , 30 , 31 ] as limited resource that platforms and communities compete to capture and retain. A comprehensive multi-platform analysis of vocabulary usage is provided by [ 16 ] , exploring linguistic patterns and variations across various social media platforms. Finally, recent studies such as [ 14 ] have conducted multi-platform analysis to identify consistent patterns in toxic discourse dynamics, highlighting persistent patterns in toxic behavior across various platforms. Despite the limited material, these studies provide a solid starting point for analyzing attention dynamics, offering a comprehensive overview of user behavior in online communities that can be used a solid background for further analysis. 3 Materials and methods This section outlines the data sources for all analyzed social media platforms, detailing the data acquisition process, the preprocessing steps, and the measures employed in the analysis. 3.1 Data Collection Usenet: Usenet is one of the oldest communication networks on the internet, established in 1980. It is organized into a decentralized system of newsgroups, each dedicated to a specific topic or theme. Users post messages or “articles” to these newsgroups, creating public threads for others to view, read, and reply to. Due to its unmoderated and decentralized nature, Usenet is known for hosting a variety of niche discussions, some of which predate modern social networks. We collected

Read the complete text