Anyone Can Become a Troll: Causes of Trolling Behavior in Online Discussions

By Justin Cheng, Michael S. Bernstein, Cristian Danescu-Niculescu-Mizil, Jure Leskovec. 2017

542 citations recorded by the source corpus.

In online communities, antisocial behavior such as trolling disrupts constructive discussion. While prior work suggests that trolling behavior is confined to a vocal and antisocial minority, we demonstrate that ordinary people can engage in such behavior as well. We propose two primary trigger mechanisms: the individual's mood, and the surrounding context of a discussion (e.g., exposure to prior trolling behavior). Through an experiment simulating an online discussion, we find that both negative mood and seeing troll posts by others significantly increases the probability of a user trolling, and together double this probability. To support and extend these results, we study how these same mechanisms play out in the wild via a data-driven, longitudinal analysis of a large online news discussion community. This analysis exposes temporal mood effects, and explores long range patterns of repeated exposure to trolling. A predictive model of trolling behavior reveals that mood and discussion context together can explain trolling behavior better than an individual's history of trolling. These results combine to suggest that ordinary people can, under the right circumstances, behave like trolls.

Subfields: Artificial Intelligence, Communication, Statistical and Nonlinear Physics

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 Statistical and Nonlinear Physics

Citations in the corpus

Fake News: A Survey of Research, Detection Methods, and Opportunities (2018, unknown): “Recent studies can be seen in, for example, in [Cheng et al. 2017; Shao et al. 2017], that analyze behavioral patterns of bots and trolls in fake news propagation, in [Chu et al. 2012], that automatically classifies humans, bots and cyborgs in terms of their profiles, tweeting behavior and content…”
Fake News Detection on Social Media: A Data Mining Perspective (2017, background): “Trolling behaviors are highly aﬀected by people’s mood and the context of online discussions, which enables the easy dissemination of fake news among otherwise “normal” online communities [14].”
Can Large Language Models Transform Computational Social Science? (2023, background): “Predictive models can be used to understand the early signs of later toxicity (Cheng et al., 2017) for downstream causal analysis on the evolution of toxicity (Mathew et al., 2020) and the effectiveness of intervention methods (Kwak et al., 2015).”
The Effect of Extremist Violence on Hateful Speech Online (2018, background): “This has led to efforts to detect, understand, and quantify such harmful speech online, with goals such as modeling socially deviant behavior (Cheng et al. 2017), building better content ﬁltering and moderation tools (Matias et al. 2015), and informing policy makers (Faris et al. 2016).”
Conversations Gone Awry: Detecting Early Signs of Conversational Failure (2018, background): “, 2017), annotators are asked to label personal attacks in the context of the conversations in which they occur, since antisocial behavior can often be contextdependent (Cheng et al., 2017).”
Critical Thinking for Language Models (2020, background): “…reasoning (Kahneman, 2011; Sunstein and Hastie, 2015), their limited critical thinking skills (Paglieri, 2017), the resulting omnipresence of fallacies and biases in texts and the frequently low argumentative quality of online debates (Hansson, 2004; Guiaşu and Tindale, 2018; Cheng et al., 2017).”
Community Interaction and Conflict on the Web (2018, background): “…interactions in online platforms has focused primarily on controversies [4, 13, 20, 21, 45, 52] and anti-social behavior, in the form of trolling [11, 12, 42], sockpup-petry [39], harassment and cyberbullying [8, 23, 32, 35, 55, 66, 79], vandalism [43], hate speech [9, 15, 51, 56, 72], and…”
Predicting Dynamic Embedding Trajectory in Temporal Interaction Networks (2019, unknown): “and Wikipedia, ensures platform integrity [9, 25, 26].”
Hate Lingo: A Target-based Linguistic Analysis of Hate Speech in Social Media (2018, background): “In (Cheng et al. 2017), the authors observe that negative mood increased a user’s probability to engage in trolling, and that anger begets more anger.”
Interpretable Multi-Modal Hate Speech Detection (2021, background): “…a breeding ground for abuse in the form of misuse of private user details (Yang & Srinivasan, 2014; Kontaxis et al., 2011), hate speech and cyberbullying (Djuric et al., 2015; Sood et al., 2012; Singh et al., 2017), and false or malicious information (Cheng et al., 2017; Aro, 2016; Weimann, 2010).”