Understanding the relationship between the Proteus effect, immersion, and gender among World of Warcraft players: An empirical survey study

ABSTRACT

Online gaming is part of contemporary life, with a range of influences on gamer’s behaviour. The convergence and alignment of an individual’s attitude and behaviour with features and characteristics of their in-game representation (i.e. their avatar), is conceptualised as the ‘Proteus effect’ (PE). In the massively multiplayer online role-playing game World of Warcraft (WoW), the Draenei (‘exiled ones’) are one type of in-game character faction. In the present study, the interplay between choosing the class of Draenei, experiencing game-related immersion, along with the gamer’s biological gender was examined in relation to PE behaviours among 404 WoW gamers (males = 299; 74%; females = 104; 26%). Participants completed the Proteus Effect Scale and the Immersion sub-scale of the User-Avatar Questionnaire. A hierarchical regression analysis indicated a significant positive association between immersion and PE behaviours. Furthermore, mediation and moderated mediation analyses suggested that being a Draenei was positively associated with the level of game-related immersion, which in turn associated to higher PE behaviours offline, independent of participants’ gender. Results indicate that gamers who are more immersed within the gaming world tend to exhibit elevated PE behaviours. Moreover, being a Draenei in WoW appeared to relate with higher game-immersion experiences, which resulted in higher PE behaviours offline.

KEYWORDS:

• Proteus effect
• immersion
• avatar
• online gaming
• world of Warcraft
• gaming gender

Disclosure statement

No potential conflict of interest was reported by the author(s).

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Notes

1 Analyses Assumptions: The assumption of independent errors was upheld following recommendations of conservative Durbin-Watson benchmarks proposed by Field (2013). The assumption of normality was violated, as the histogram displayed a positive skewness, furthermore, the normal P–P plot reinforces this as it showed systematic deviation of residuals. At this point, and before referring to further analyses’ assumptions, it should be noted that the deviation from the normality assumption was addressed with the use of bootstrapping, in line with past literature recommendations (De Boer et al. 2009). The assumption of homoscedasticity was upheld as the residuals on the scatterplot roughly showed that variance was equally spread across the entire level of the predictor. The assumption of linearity was upheld as the scatterplot showed a comparable spread of residuals with little systematic deviation around the zero point. The assumption of multicollinearity was upheld as tolerance values did exceed 0.2 and Variance Inflation Factor statistics were below 10. Furthermore, the Condition Index value was below 15 (if above 15 possible collinearity problems may exist; greater than 30 indicates serious problems; Thompson et al. 2017). This suggests that there were no significant suppression effects (due to multicollinearity) that could have confounded the analysis. Using a case-wise cut-off value of 2.5 a single univariate outlier (2.529) was discovered. Multivariate outliers were detected as Mahalanobis distance (57.11) surpassed the chi-square distribution value (p = .05, based on three predictors [namely age and gender as the controlled covariates and Immersion as the predictor of interest] was 7.81; Tabachick & Fidell, 2012; as cited in Field 2013). Furthermore, Cook’s distance (.038) was assessed to uncover whether outliers influenced the model. The value was below one which suggests that outliers held no significant influence over the model, therefore the analysis retained them (Cook & Weisberg, 1982 as cited in Field 2013). Please see appendix graphs and tables.