A1 Refereed original research article in a scientific journal
Do players communicate differently depending on the champion played? Exploring the Proteus effect in League of Legends
Authors: Şengün Sercan, Santos Joao M, Salminen Joni, Jung Soon-gyo, Jansen Bernard J
Publisher: Elsevier
Publication year: 2022
Journal: Technological Forecasting and Social Change
Journal name in source: TECHNOLOGICAL FORECASTING AND SOCIAL CHANGE
Journal acronym: TECHNOL FORECAST SOC
Article number: 121556
Volume: 177
Number of pages: 15
ISSN: 0040-1625
eISSN: 1873-5509
DOI: https://doi.org/10.1016/j.techfore.2022.121556
Web address : https://doi.org/10.1016/j.techfore.2022.121556
Abstract
We investigate how the Proteus effect, which is players changing their way of communication based on characters with which they play, is associated with players' champion usage in the popular online game League of Legends, where champions are the characters that the players control. First, we create two sets of variables: (a) objective champion characteristics based on information from the game developer, which we further enrich by semiotic coding, and (b) subjective champion characteristics based on crowdsourced opinions about the champions. Then, we analyze 13.6 million in-game chat messages to measure whether the players' vocality (character counts of messages), valence (negative versus positive scores of language use), and toxicity (frequency of toxic word usage) change depending on the characteristics of the champions they employ. We find that champions' body type, role, and gender are associated with players' higher vocality, toxicity, and negative valence. We also find that the players' communication significantly changes in toxicity and valence when they play using different champions. We discuss our methodology and results in detail and propose design directions and other implications based on them.
We investigate how the Proteus effect, which is players changing their way of communication based on characters with which they play, is associated with players' champion usage in the popular online game League of Legends, where champions are the characters that the players control. First, we create two sets of variables: (a) objective champion characteristics based on information from the game developer, which we further enrich by semiotic coding, and (b) subjective champion characteristics based on crowdsourced opinions about the champions. Then, we analyze 13.6 million in-game chat messages to measure whether the players' vocality (character counts of messages), valence (negative versus positive scores of language use), and toxicity (frequency of toxic word usage) change depending on the characteristics of the champions they employ. We find that champions' body type, role, and gender are associated with players' higher vocality, toxicity, and negative valence. We also find that the players' communication significantly changes in toxicity and valence when they play using different champions. We discuss our methodology and results in detail and propose design directions and other implications based on them.
UTU Research Portal