Dynamics of Disagreement: Large-Scale Temporal Network Analysis Reveals Negative Interactions in Online Collaboration

TL;DR

This study analyzes the timing and patterns of negative interactions in online collaboration, revealing systematic attack behaviors and social dynamics using large-scale temporal network analysis of Wikipedia reverts.

Contribution

It introduces a method to detect and analyze negative social interactions in large-scale online collaboration networks, focusing on temporal patterns and social status effects.

Findings

Individuals attack the same person and retaliate faster than random chance.

High-status contributors tend to attack many others serially.

Lower-status contributors are more likely to defend and initiate attacks.

Abstract

Disagreement and conflict are a fact of social life and considerably affect our well-being and productivity. Such negative interactions are rarely explicitly declared and recorded and this makes them hard for scientists to study. We overcome this challenge by investigating the patterns in the timing and configuration of contributions to a large online collaboration community. We analyze sequences of reverts of contributions to Wikipedia, the largest online encyclopedia, and investigate how often and how fast they occur compared to a null model that randomizes the order of actions to remove any systematic clustering. We find evidence that individuals systematically attack the same person and attack back their attacker; both of these interactions occur at a faster response rate than expected. We also establish that individuals come to defend an attack victim but we do not find evidence that attack victims "pay it forward" or that attackers collude to attack the same individual. We further find that high-status contributors are more likely to attack many others serially, status equals are more likely to revenge attacks back, while attacks by lower-status contributors trigger attacks forward; yet, it is the lower-status contributors who also come forward to defend third parties. The method we use can be applied to other large-scale temporal communication and collaboration networks to identify the existence of negative social interactions and other social processes.