Keywords
Avatar, unique avatar, humanized avatar, game, structural equation modeling
Abstract
Avatars are in-game characters that players use to represent themselves. Past research has clarified that players enjoy identifying with avatars, particularly those are good-looking and similar to themselves. However, avatars in online game possess many more features that can be viable but not yet examined, showing insufficiency in research. To address such insufficiency, our research took the perspective of Proteus effect and built the research model using three important avatar features and their effects on players’ identification with their in-game avatars. Avatar features were assessed using game experts’ responses, while avatar identification data were assessed using game players’ responses, forming two-sourced data and thereby negating common method concerns. We found that unique avatars are most influential for forming avatar identification, while interestingly, humanized avatars negatively affect the formation of avatar identification. Our results inform system designers that they should design game avatars as unique to each other, while not aiming to resemble real humans. We argue that the findings can be applied in other information system areas, e.g., the Metaverse and the gamified systems.
Recommended Citation
Pham, Thi Tuan Linh; Tai, Shih-I; Shih, Sheng-Pao; Chang, Yan-Cheng; Cheng, T.C.E.; and Teng, Ching-I, "Unique but Not Human-Like: Avatar Features and Responses" (2024). Digit 2024 Proceedings. 9.
https://aisel.aisnet.org/digit2024/9