Antisocial Behavior in Online Discussion Communities

By Justin Cheng, Cristian Danescu-Niculescu-Mizil, Jure Leskovec. 2015

344 citations recorded by the source corpus.

User contributions in the form of posts, comments, and votes are essential to the success of online communities. However, allowing user participation also invites undesirable behavior such as trolling. In this paper, we characterize antisocial behavior in three large online discussion communities by analyzing users who were banned from these communities. We find that such users tend to concentrate their efforts in a small number of threads, are more likely to post irrelevantly, and are more successful at garnering responses from other users. Studying the evolution of these users from the moment they join a community up to when they get banned, we find that not only do they write worse than other users over time, but they also become increasingly less tolerated by the community. Further, we discover that antisocial behavior is exacerbated when community feedback is overly harsh. Our analysis also reveals distinct groups of users with different levels of antisocial behavior that can change over time. We use these insights to identify antisocial users early on, a task of high practical importance to community maintainers.

Subfields: Artificial Intelligence, Communication, Information Systems

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Communication

In Information Systems

Citations in the corpus

Deep Learning for User Comment Moderation (2017, background): “Cheng et al. (2015) predict which users would be banned from on-line communities.”
Comparison of Naive Bayes, Random Forest, Decision Tree, Support Vector Machines, and Logistic Regression Classifiers for Text Reviews Classification (2017, background): “…reasons: firstly, for detecting antisocial online behavior, antisocial users in a community, or that which act strangely or even appear dangerous (Cheng et al., 2014); secondly, classification allows the investigation of global social and information networks to gather special knowledge derived…”
Hate Me, Hate Me Not: Hate Speech Detection on Facebook (2017, background): “A behavioral analysis of banned users is in [7], showing a certain degree of similarity in their texts, containing often irrelevant content too.”
Can Large Language Models Transform Computational Social Science? (2023, background): “3.2.5 Toxicity Prediction Toxicity is a major area of social research in online communities, as online disinhibition (Suler, 2005) makes antisocial behaviour especially prevalent (Cheng et al., 2015).”
Emoticons vs. Emojis on Twitter: A Causal Inference Approach (2015, methodology): “There has been some recent work employing causal inference techniques such as matching in large scale quantitative studies using observational social media data (Reis and Culotta 2015; Cheng, Danescu-Niculescu-Mizil, and Leskovec 2015).”
A Survey on Troll Detection (2020, methodology): “In particular, [37] presents a study on anti-social behaviour in online discussion communities, focusing on those users that will eventually be banned.”
Deeper Attention to Abusive User Content Moderation (2017, background): “1 User comments, however, and more generally user content can also be abusive (e.g., bullying, profanity, hate speech) (Cheng et al., 2015).”
Detecting harassment in real-time as conversations develop (2019, background): “Focusing on users instead of utterances, Cheng et al. (2015) aim to detect ’antisocial users’ in online communities over a longer period of time.”
A Bad Workman Blames His Tweets? The Consequences of Citizens’ Uncivil Twitter Use When Interacting with Party Candidates (2016, background): “…for democratic discourse (Mutz and Reeves, 2005; Papacharissi, 2004) and the anonymity behind which many users choose to hide themselves allows for limitless abuse (Cheng et al., 2015; Davis, 2009) which can ultimately influence the motivations behind the communication style of candidates.”
Conversations Gone Awry: Detecting Early Signs of Conversational Failure (2018, background): “between conversations that derail and those that stay on track, and that accounts for trivial topical confounds (Kittur et al., 2009; Cheng et al., 2015).”
A Measurement Study of Hate Speech in Social Media (2017, neutral): “Aside from leveraging text based features researchers also explored other features like leveraging user history [8] or even community detection [6]; These techniques can be used in addition to the text based features.”
Anyone Can Become a Troll: Causes of Trolling Behavior in Online Discussions (2017, methodology): “Further, such behavior comprises a substantial fraction of user activity on many web sites [18, 24, 30] – 40% of internet users were victims of online harassment [27]; on CNN.”
False Information on Web and Social Media: A Survey (2018, methodology): “Feature-based algorithms have been used to identify various different types of malicious users and activities, such as identifying bots [97], trolls [20], vandals [49], sockpuppets [47], and many more.”
Community Interaction and Conflict on the Web (2018, background): “This finding echoes those on anti-social behavior, which show that troll-like behaviors are concentrated in small number of communities [12], and that taking precautionary measures, such as banning particularly egregious communities, can be effective in curbing this behavior [10].”
You Can't Stay Here (2017, methodology): “Research on automatic approaches to moderating online antisocial behavior has shown that textual cyberbullying [15, 49] and undesirable posting [6, 8, 10, 43] can be identified based on topic models, presence of insults and user behavior.”