Due to its prominent role in online social networks, avatar creation has become an important research topic in the field of computer-mediated communication. One main motive for creating avatars is the representation of one’s own identity. Previous research indicates that avatar creation depends on the activity context for which the avatar is created. Though, studies comparing avatar creation for a wide variety of activity contexts are still missing. The present study addresses this empirical gap by examining (1) the self-representation of physical, demographical, and personality characteristics through avatars, (2) differences in self-representation between various online activity contexts, and (3) between-participant variance in ascribed personality traits. Participants attributed physical, demographical, and personality characteristics to their avatar for one of six online activity contexts and indicated the same characteristics for their actual and ideal selves. We analysed the data of 568 participants and found a high level of congruence between demographical/physical characteristics of the avatar, the actual self, and the ideal self. Furthermore, we found an idealised representation of the avatar’s personality traits, which was affected by the specific activity context. Last, the between-participant variance in the avatar’s personality traits was mainly larger than the variance in the ideal self but smaller than the variance in the actual self, indicating a link between avatar creation and social norms. These results deliver new insights into the strategies behind avatar creation in different contexts and build a basis for future research and practical implications for developers and designers of virtual worlds.
Avatars have become more and more prominent in social online networks and smartphone applications. Bailenson et al. (2008) defined avatars as digital representations of their users in digital environments. Avatars enable their users to interact and communicate in shared digital worlds. An increasingly large part of computer-based communication is mediated or at least flanked by avatars. The vision of a Metaverse focusing on avatar-based interactions (Lee et al., 2021) also shows the importance of avatars and their representation. However, relatively little research has been conducted on how avatars should be represented depending on the communication context. This question guides the present study.
Avatars are driven by humans in real-time and even can look and behave like them. Kromand (2007) differentiates between open and closed avatars: Closed avatars are pre-generated avatars that only can transform by pre-determined narrative progression. In contrast, their users define the appearance and personality of open avatars. The way open avatars are created is thereby linked to specific motives like exploring digital worlds, finding friends, and portraying one’s own identity (Lin & Wang, 2014), aiming at need fulfilment, including self-expression and self-esteem (for a review, see Sibilla & Mancini 2018). However, avatar creation and avatar identification can also be associated with adverse outcomes such as depression (Bessière et al., 2007) and problematic gaming behaviour (Lemenager et al., 2020). Consequently, avatars as communication tools and associated strategies of avatar creation have become an important research topic in the field of computer-mediated communication. Importantly, avatar creation seems to depend on the activity context for which the avatars are created. Some previous studies indicated differences in avatar creation between different activity contexts, such as different gaming scenarios (e.g., Trepte & Reinecke 2010; Trepte et al., 2009) or social networks (e.g., Triberti et al., 2017). However, studies that compare a wider variety of activity contexts that differ in primary communication goals are missing so far. The present study aims to fill this empirical gap by examining strategies of self-representation via avatars while considering different online activity contexts.
Self-representation through avatars
Self-representation is one central aspect of avatar creation and avatar usage. The theory of transformed social interaction (Bailenson et al., 2008) highlights that the transformation of self-representation is one of the main dimensions of avatar-mediated communication. This dimension describes the ability to strategically change the appearance and behaviour of the avatar regardless of the user’s actual characteristics. Bente et al. (2008) already pointed to the strategic emphasis and suppression of avatar features as one of the central factors of avatar-mediated communication.
Indications for self-representation through an avatar are often derived from impression management theory (Mummendey, 1995) and self-discrepancy theory (Higgins, 1987). According to impression management theory, people strive to represent themselves as idealised and positive as possible (Mummendey, 1995). Additionally, self-discrepancy theory differentiates between the actual self and the ideal self. While the actual self represents the current self-image, the ideal self includes the representation of how one would ideally like to be and the qualities one would like to have. Negative emotions arise from a sizeable discrepancy between the ideal self and the actual self (Higgins, 1987). Therefore, people strive to minimise this discrepancy via avatar creation (Bessière et al., 2007). The idealised virtual identity hypothesis (Manago et al., 2008) also states that people portray idealised characteristics of themselves in online networks and thus tend to describe how they would ideally see themselves rather than how they actually are. Hence, these theories advocate a general idealisation of avatars to achieve the most positive self-presentation possible. Most empirical findings are related to the context of video gaming: Sibilla & Mancini (2018) indicated in their review of previous research on avatar creation in massively multiplayer online games (MMOs) that users generally tend to actualise and idealise their avatars in terms of physical, demographic, and personality characteristics. Furthermore, especially people with a higher discrepancy between their actual self and ideal self seem to prefer idealised avatars (Loewen et al., 2021). However, this contrasts with the extended real-life hypothesis (Back et al., 2010). This hypothesis describes that users of online social networks use their profiles to portray their real personalities. Back et al. (2010) were able to show that the data provided on profiles can be used to predict the actual personality traits of users. Consequently, these contradictory hypotheses and results raise the question of the extent to which various feature levels of avatar creation are idealised or congruent with actual user characteristics and the extent to which the specific activity context influences avatar creation. The present study tackles this question.
Self-representation of physical, demographic, and personality characteristics
Previous findings are ambiguous in the context of the avatar’s physical and demographic characteristics. On the one hand, they are idealised by the users. For example, men and women tend to choose ideal male and female bodies, respectively, when creating an avatar (Cacioli & Mussap, 2014; Dunn & Guadagno, 2012). On the other hand, there is evidence that the avatar represents aspects of the actual self at least partially (Kafai et al., 2010) or comprehensively represent the actual self but with specific improvements in appearance (Ducheneaut et al., 2009; Messinger et al., 2008). Also, avatars can even be detached from the offline selves for aesthetical and functional reasons (Kafai et al., 2010), or users play with their identity by swapping gender (Hussain & Griffiths, 2008). Because of these mixed findings, it is unclear whether the avatar’s physical and demographical characteristics will be based on the actual self, the ideal self, or neither. Thus, we formulated an undirected hypothesis:
H1: Physical/demographic characteristics attributed to the avatar, the actual self, and the ideal self are interrelated
Regarding psychological characteristics, studies mainly examined the discrepancy between the user’s and the avatar’s personality traits (Sibilla & Mancini, 2018). Users seem to idealise the avatar’s personality (Ducheneaut et al., 2009; Bessière et al., 2007). This finding is especially true for negatively connoted personality traits (Dengah & Snodgrass, 2020). The present study focused on the Big Five personality traits encompassing extraversion, agreeableness, conscientiousness, neuroticism, and openness to experiences. These traits have been labelled as the basic dimensions of personality that cover many personality facets (Costa & McCrae, 1992). We hypothesised:
H2: Personality traits of the avatar are generally chosen by users in an idealised way
Self-representation in different online activity contexts
In addition to users’ actual and idealised self, the situational context seems to be an important determining factor of avatar creation. For example, the theory of transformed social interaction (Bailenson et al., 2008) highlights the transformation of the situational context of the digital world as one key factor. Bente et al. (2008) also underlined the situatedness and the co-presence of multiple users in shared virtual environments as one central aspect of avatar-mediated communication. Thus, the specific online activity context could influence avatar creation. Previous research mainly focused on video game contexts and found that avatars differed according to the demands of the specific games. Trepte & Reinecke (2010) compared avatar creation in terms of personality characteristics for video games which differed in the level of competitiveness. Their results showed that a video game’s competitiveness influenced the choice of avatar: while avatars with a similar personality to their users were created for non-competitive video games, dissimilar avatars were created for competitive video games. Also, Trepte et al. (2009) demonstrated that users tend to attribute predominantly male characteristics to their avatar for a video game which is perceived as masculine, and female characteristics to their avatar for a video game which is perceived as female. However, there was a preference for same-sex avatars as “participants rated those game descriptions and gaming scenarios more entertaining which required avatar features in line with their own sex role” (Trepte et al., 2009, p. 52). Users also seem to accentuate specific avatar characteristics depending on the activity context and associated requirements. For example, Vasalou & Joinson (2009) compared avatar creation regarding physical appearance for the online contexts of blogging, dating, and gaming. Avatars tended to be more physically attractive in an online dating context and more intellectual in video games. At the same time, they mainly reflected the user’s actual physical appearance in the blogging context (Vasalou & Joinson, 2009). Furthermore, Triberti et al. (2017) compared the creation of the appearance of an avatar (body, clothes, and accessories) for video games and a job-themed social network. The authors could show that the appearance, especially clothes, was changed when the activity context switched from gaming to social networks designed for job contacts, thus matching the appearance to the context’s requirements.
Besides, avatar creation depends on the gender of the user and the audience of self-representation: Female users tended to change their avatar more often than male users when they expected to communicate with friends rather than with strangers. The authors concluded that women tend to express themselves and their gender more through avatars than men, especially when interacting with friends (Triberti et al., 2017). Hence, previous research revealed different strategies for avatar creation depending on the activity context and its demands. Despite this evidence, however, a comprehensive comparison of avatar creation strategies across the most common online contexts is lacking, and the question of avatar idealisation in different online contexts has, to our best knowledge, not yet been explored and compared for a wide variety of online contexts. Thus, the present study scrutinised such activity context effects concerning avatar idealisation. Specifically, we examined avatar creation for six activity contexts: dating, competitive and non-competitive gaming, and social networks for different audiences (friends, strangers, and job contacts). Besides analysing the relation between avatar, actual self, and ideal self within each activity context (H3), we hypothesised that avatar idealisation regarding personality traits would significantly differ across activity contexts (H4). Furthermore, as gender also seems to affect avatar creation, we examined if the user’s gender influences avatar idealisation regarding personality traits (H5):
H3: The physical/demographic characteristics and personality traits of the avatar, the actual self, and the ideal self are chosen by users differently depending on the specific activity context
H4: There is a significant main effect of the activity context on the extent to which personality traits of the avatar are chosen by users in an idealised way
H5: There is a significant main effect of the user’s gender on the extent to which personality traits of the avatar are chosen by users in an idealised way
Between-participant variance in ascribed personality traits
Notably, the personality traits of an avatar do not necessarily have to be in the range of the actual self and the ideal self. For example, the avatar’s personality could even be worse than the user’s actual personality, allowing users to explore the effect of characteristics that are impossible or socially undesirable in the offline world (Mancini & Sibilla, 2017). Vasalou and Joinson (2009) found high variance in self-representation through avatars. However, it is still unclear whether users are taking advantage of these extensive possibilities. So far, research examining the personality aspects of avatar creation has mainly focused on differences between online self and offline self. However, this approach does not answer whether the diversity of possibilities offered by digital worlds is reflected in a wide variety of avatars. If this is the case, the personalities of avatars should vary more than those of offline selves, which are bound to the real possibilities and norms of the offline world. Therefore, we compared the dispersion of personality traits between avatar, actual self, and ideal self to address this issue. We hypothesised that personality traits of the avatar show higher between-participant variance than those of the actual self and the ideal self (H6). We also explored the variance patterns within each specific activity context (H7).
H6: The personality traits of the avatars chosen by users show a higher between-participant variance than those of the ideal self and actual self
H7: The between-participant variance in personality traits of the avatar, the actual self, and the ideal self, indicated by users, differ between online activity contexts
We conducted an online experiment with the software Unipark (Tivian, 2017). Participants were recruited through a combination of convenience and snowball sampling methods via social media and e-mail distribution lists of several German universities. Inclusion criteria were a minimum age of 18 years, understanding of the German language, and informed consent. All procedures performed in the study were in accordance with the ethical guidelines of the German Psychological Society (DGPs) and with the 1964 Helsinki declaration. At the beginning of the study, participants were informed that the data of this study will be used for research purposes only and that all data would be collected anonymously. Thus, no identifying information was collected. We also highlighted that participation is voluntary and that participants can terminate participation in this study at any time and without giving reasons. Thus, participants who prematurely stopped the survey were not included in the analyses and all of their data were deleted from the dataset. An email address of a contact person for queries of any kind was also provided. None of the participants reported technical problems or significant stress while conducting the study. Informed consent to participate in this study was provided by clicking a corresponding box.
Overall, 594 participants completed the study, but 25 of them were excluded because they interrupted the experiment in the meantime or had an implausibly long processing time, both of which counteracted the experimental manipulation. We also excluded one further participant due to implausible information about their own age. No further exclusion criteria were applied. Thus, the data of 568 participants (64.1% female, Mage = 28.3, SDage = 11.63) were included in the analyses. The participants were randomly assigned to one of the six activity contexts by the survey software: dating (n = 89, 62.9% female; Mage = 28.45, SDage = 10.86), competitive gaming (n = 94, 67% female; Mage = 28.50, SDage = 11.93), non-competitive gaming (n = 93, 69.9% female, Mage = 28.10, SDage = 11.53), social network with friends (n = 98, 57.1% female, Mage = 28.83, SDage = 12.97), social network with strangers (n = 92, 62% female, Mage = 28.26, SDage = 11.85), and social network with job contacts (n = 102, 65.7% female; Mage = 27.72, SDage = 10.77).
All participants reported (desirable) physical and demographic characteristics and personality traits of their avatar, actual self, and ideal self.
The physical and demographic characteristics included one’s gender (single-choice: male, female, or diverse), height (in centimetres), age (in years), weight (in kilogram), hair colour (single-choice: brown, blonde, black, red, grey, bald head, or others), and physical attractiveness (10-point scale ranging from 0 = “not attractive at all” to 9 = “very attractive”).
We assessed personality traits in terms of the Big Five personality traits via the German short version of the Big Five Inventory (BFI-K; Rammstedt & John 2005). The factorial structure of this standardised and economical instrument designed for online contexts features good psychometric properties and has been validated in homogenous student samples and larger heterogeneous samples (Kovaleva et al., 2013). The BFI-K includes 21 Likert-type items (1 = “disagree strongly” to 5 = “agree strongly”), four to five items for each of the personality traits, namely extraversion (Cronbach’s α on the level of the whole sample = 0.64−0.84), agreeableness (α = 0.66−0.69), conscientiousness (α = 0.67−0.73), neuroticism (α = 0.59−0.73), and openness (α = 0.67−0.74). The present range of the scales’ internal consistencies has already been found in previous studies (e.g., Kaspar & Fuchs 2021; Meier et al., 2021; Steiner et al., 2012). A more detailed analysis of Cronbach’s α at the level of specific online activity contexts can be found in the Supplementary Information File, Table 1.
Research design and procedure
We employed a 6 (online activity context: dating, competitive gaming, non-competitive gaming, social network with friends, social network with strangers, social network with job contacts) × 3 (evaluation subject: avatar, actual self, ideal self) design with online activity context as between-participant factor and evaluation subject as within-participant factor. After providing informed consent, participants were randomly assigned to one of the six online activity contexts. A context-specific instruction was presented, serving as the experimental manipulation: First, it was explained to all participants that the purpose of this study was to investigate the criteria for creating avatars in online worlds and networks, defined as artificial persons or graphic figures. Participants were then instructed to create an avatar which should represent themselves for one of the following activity contexts: (1) a dating website where they can make contact with strangers in order to meet them in real life (dating); (2) an online videogame where they compete with other players in order to win the game on their own (competitive gaming); (3) an online videogame where they cooperate with other players in order to win the game as a team (non-competitive gaming); (4) a digital social network where they can connect and chat with friends as well as comment on their posts in order to stay in contact (social network with friends); (5) a digital social network where they can connect and chat with strangers as well as comment on their posts in order to make new contacts (social network with strangers); (6) a digital social job network where they make new and maintain current business contacts in order to improve networking and to get job offers (social network with job contacts). Furthermore, we specified that the avatar should represent a whole human with specific physical and demographic characteristics, respectively, and personality traits, which would also be displayed like skill points and, in turn, would be viewable by other users. For example, the participants who were assigned to the dating context received the following instruction:
In online worlds and networks, avatars (i.e., artificial persons or graphical characters) are increasingly used to represent users instead of actual photos of the respective persons (e.g., Second Life, BitMoji, MeMojis). Therefore, we would like to capture the criteria by which users create avatars in online environments.
Imagine you are signing up for a dating website where you can contact and write to strangers to meet them in real life subsequently. At the beginning of the sign-up process, you must create a personal avatar that represents yourself to the other users of the dating site. This avatar should represent a person holistically. Therefore, you should now assign characteristics and physical features to your avatar that will be visible to the other users. For this purpose, you will be asked to answer some questions about your avatar.
Only the descriptions of the activity context and the purpose of registration differed between the experimental groups. After the instruction, the participants filled out several questionnaires in the following order: We initially asked them to think of their avatar and its characteristics. Participants provided their virtual avatar’s physical characteristics, demographic characteristics, and personality traits. In the second step, we asked participants to think about how they really are, thus indicating their actual physical characteristics, demographic characteristics, and personality traits (actual self). Finally, we asked the participants to think about how they ideally would like to be if they could freely choose, thus indicating their ideal self’s physical characteristics, demographic characteristics, and personality traits. After that, participants were discharged.
All analyses were conducted using SPSS 27. To analyse the relations between avatar, actual self, and ideal self concerning physical and demographic characteristics (H1, H3), we calculated Spearman correlations regarding height, weight, age, and physical attractiveness. We quantitatively examined gender and hair colour via cross tables.
To examine avatar idealisation (i.e., avatar characteristics are closer to the ideal self than to the actual self) on the level of personality traits, we first computed differences of the Big Five trait scores between the ideal self and the avatar (= avatar-discrepancy, ideal self minus avatar) and between the ideal self and the actual self (= self-discrepancy, ideal self minus actual self) as proposed by Bessière et al. (2007). We then computed the difference between avatar-discrepancy and self-discrepancy (avatar-discrepancy minus self-discrepancy) as the extent of avatar idealisation. We then used this extent of avatar idealisation as the dependent variable for one-sample t-tests (H2, H3) and a two-way ANOVA (H4, H5).
Finally, we analysed the between-participant variance in the Big Five trait scores. We statistically compared this variance between the avatar, the actual self, and the ideal self through Pitman-Morgan tests (Kauttonen, 2021). This analysis was done for the whole sample (H6) and individually for each online activity context (H7).
Physical and demographic characteristics of avatar, actual self, and avatar (H1, H3)
At the level of the total sample, we consistently found significant positive correlations between avatar, actual self, and ideal self with respect to age (all rs ≥ 0.685, ps < 0.001), height (all rs ≥ 0.773, ps < 0.001), weight (all rs ≥ 0.810, ps < 0.001), and physical attractiveness (all rs ≥ 0.345, ps < 0.001). Results are depicted in Table 1. More fine-grained correlation analyses for each activity context revealed that this result pattern was comparable across activity contexts (see Supplementary Information File). Furthermore, most participants attributed the identical gender to their avatar, actual self, and ideal self (Table 2). Regarding hair colour, Table 3 shows that the reported hair colour mostly matched between avatar, actual self, and ideal self. This analysis was limited to blonde and brown hair colour as additional hair colours showed overall very low frequencies (3.2% black, 3.0% red, 3.3% grey, 1.1% bald head, 1.9% others). However, most participants generally selected their actual hair colour for their avatar and ideal self. A more detailed analysis of attributed gender and hair colour at the level of specific online activity contexts can be found in the Supplementary Information File. Overall, and as expected (H1), we found high congruence between avatar, actual self, and ideal self for physical and demographic characteristics. In contrast to our expectation (H3), this pattern was relatively stable in all online activity contexts.
Idealisation of personality traits in general (H2) and within activity contexts (H3)
We initially analysed the extent of the self-idealisation of personality traits. For this purpose, one-sample t-tests compared the mean extent of avatar idealisation with zero for the whole sample and each activity context separately. In the case of neuroticism, it must be considered that a significant positive difference between avatar-discrepancy (ideal self minus avatar) and self-discrepancy (ideal self minus actual self) indicates avatar idealisation due to the negative connotation of neuroticism. In contrast, avatar idealisation is reflected by a negative difference between avatar-discrepancy and self-discrepancy regarding extraversion, agreeableness, conscientiousness, and openness. For a more intuitive presentation, we therefore reversed the signs of the latter four personality traits. Thus, positive values always indicate avatar idealisation, and if the avatar idealisation was significantly above zero, we considered the avatar idealised.
With respect to the whole sample, we found that avatars were idealised regarding extraversion, t(567) = 8.26, p < 0.001, d = 0.35, agreeableness, t(567) = 6.65, p < 0.001, d = 0.28, conscientiousness, t(567) = 9.13, p < 0.001, d = 0.38, and neuroticism, t(567) = 18.70, p < 0.001, d = 0.78, but not regarding openness, t(567) = 0.07, p = 0.943, d = 0.003. Overall, this result supports H2.
Figure 1 shows the extent of avatar idealisation regarding the Big Five personality traits for each activity context (H3). In all activity contexts, avatars were idealised regarding extraversion, all ts ≥ 2.08, ps ≤ 0.040, dmin = 0.22, dmax = 0.55, and neuroticism, all ts ≥ 6.27, ps < 0.001, dmin = 0.66, dmax = 1.00. Avatars were also idealised regarding agreeableness in most activity contexts, all ts ≥ 2.61, ps ≤ 0.011, dmin = 0.27, dmax = 0.50, except for competitive gaming, t = 0.20, p = 0.842, d = –0.02. Moreover, avatars were idealised regarding conscientiousness in most contexts, all ts ≥ 2.94, ps ≤ 0.004, dmin = 0.31, dmax = 0.69, apart from dating, t = 1.85, p = 0.067, d = 0.20, and social network with friends, t = 1.39, p = 0.169, d = 0.14. Surprisingly, avatars were significantly closer to the actual self than to the ideal self regarding openness in the context of competitive gaming, t = − 3.07, p = 0.003, d = –0.32. There were no other significant results regarding openness, all ts ≤ 1.73, ps ≥ 0.086, dmin = –0.06, dmax = 0.18. In summary, we found the expected idealised representation of avatars in almost every activity context, but also some notable exceptions, partly supporting our expectations (H3).
Idealisation of avatar’s personality traits across activity contexts (H4) and as a function of participant’s gender (H5)
Next, for each personality trait, we calculated a two-way ANOVA with activity context and the participants’ gender as independent variables, and the extent of avatar idealisation as dependent variable to test for differences between activity contexts and gender effects. In case of a significant main effect of the activity context, we further performed pairwise comparisons. We found a main effect of the activity context on the extent of avatar idealisation for conscientiousness, F(5, 556) = 4.43, p < 0.001, ηp2 = 0.038, and openness, F(5, 556) = 3.45, p = 0.004, ηp2 = 0.030. There were no other significant main effects of the activity context on personality traits, all Fs(5, 556) ≤ 1.94, ps ≥ 0.086, ηp2 ≤ 0.017. Thus, H4 was partly supported by the data. Interestingly, only few pairwise comparisons met the Bonferroni-adjusted significance level (see Fig. 1): The extent of avatar idealisation regarding conscientiousness was lower in dating compared to competitive gaming, p = 0.010, MDiff = 0.34, 95%-CI [0.05, 0.63], and non-competitive gaming, p = 0.004, MDiff = 0.36, 95%-CI [0.07, 0.66], and lower in social network with friends than in competitive gaming, p = 0.010, MDiff = 0.33, 95%-CI [0.47, 0.62], and non-competitive gaming, p = 0.004, MDiff = 0.36, 95%-CI [0.07, 0.64]. Besides, avatar idealisation regarding openness was lower in competitive gaming than in social network with friends, p = 0.005, MDiff = 0.30, 95%-CI [0.06, 0.55]. Furthermore, we found significant main effects of gender on the avatar idealisation regarding conscientiousness, F(1, 556) = 6.347, p = 0.012, ηp2 = 0.011, and neuroticism, F(1, 556) = 6.347, p < 0.001, ηp2 = 0.027. Males (Mideal = 0.34, SDideal = 0.74) idealised their avatar more strongly than females (Mideal = 0.22, SDideal = 0.56) regarding conscientiousness, females (Mideal = 0.88, SDideal = 0.97) idealised their avatars more strongly regarding neuroticism than males (Mideal = 0.53, SDideal = 0.89). So, H5 was partly supported. There were no significant interaction effects on the extent of avatar creation for any of the personality traits, all Fs(5, 556) ≤ 1.199, ps ≥ 0.308, ηp2 ≤ 0.011.
Between-participant variance in personality traits of avatar, actual self, and ideal self (H6, H7)
Next, we analysed the between-subject variance in personality ratings attributed to the avatar, the ideal self, and the actual self. This analysis was conducted at the level of the whole sample and separately for the six activity contexts. Significance testing was done via Pitman-Morgan tests (Kauttonen, 2021).
Figure 2 shows the variances of the whole sample regarding personality traits. For all personality traits, we found that the variance in the actual self scores was higher than the variance in the avatar scores, all Pitman’s Ts(566) ≥ 2.35, ps ≤ 0.019, except for openness, Pitman’s T(566) = 0.86, p = 0.388. Also, for all personality traits, the variance in actual self scores was higher than the variance in ideal self scores, all Pitman’s Ts(566) ≥ 2.85, ps ≤ 0.005. Finally, we found higher variances in avatar scores compared to ideal self scores for all personality traits, all Pitman’s Ts(566) ≥ 4.88, ps < 0.001, except agreeableness, Pitman’s T(566) = 0.86, p = 0.388. Thus, personality scores generally varied most when participants rated their actual selves, personality scores varied less when they rated the avatar, and variance in personality scores was again significantly lower for the ideal self. This result pattern partially contradicts our H6, stating that avatar scores should show a higher variance than the ideal self and the actual self.
Table 4 shows the results of the Pitman-Morgan tests separately for each activity context (H7). Regarding differences in variance between the avatar’s and the actual self’s personality scores, the variance in the actual self was significantly higher than the avatar’s variance regarding several activity contexts: This was true for openness in the context of dating, Pitman’s T(87) = 2.24, p = 0.028, neuroticism in competitive gaming, Pitman’s T(92) = 3.40, p = 0.001, conscientiousness in non-competitive gaming, Pitman’s T(91) = 2.42, p = 0.017, and neuroticism in non-competitive gaming, Pitman’s T(91) = 2.21, p = 0.030, neuroticism in social networking with friends, Pitman’s T(96) = 2.15, p = 0.034, as well as with job contacts, Pitman’s T(100) = 3.17, p = 0.002, extraversion in social network with job contacts, Pitman’s T(100) = 3.00, p = 0.003, and agreeableness in social network with job contacts, Pitman’s T(100) = 2.93, p = 0.005.
Next, we compared the variances in the avatar’s and the ideal self’s personality scores. The variance in the avatar was significantly higher than the variance in the ideal self for conscientiousness, all Pitman’s Ts ≥ 2.65, ps ≤ 0.009, and neuroticism, all Pitman’s Ts ≥ 3.08, ps ≤ 0.003, in all activity contexts. This was also true for extraversion in the activity contexts dating, Pitman’s T(87) = 2.20, p = 0.030, competitive gaming, Pitman’s T(92) = 4.52, p < 0.001, and non-competitive gaming, Pitman’s T(91) = 4.85, p < 0.001, as well as for openness in the activity contexts competitive gaming, Pitman’s T(92) = 2.28, p = 0.025, non-competitive gaming, Pitman’s T(91) = 2.14, p = 0.035, and social network with friends, Pitman’s T(96) = 4.11, p < 0.001. No significant difference between avatar and ideal self was found for agreeableness, all Pitman’s Ts ≤ 1.84, ps ≥ 0.069.
Looking at the differences in variances between the actual self and the ideal self, the variance in the actual self was significantly higher than the variance in the ideal self regarding extraversion, all Pitman’s Ts ≥ 2.16, ps ≤ 0.034, conscientiousness, all Pitman’s Ts ≥ 4.21, ps < 0.001, and neuroticism, all Pitman’s Ts ≥ 4.54, ps < 0.001, in all activity contexts. This was also true for agreeableness in the activity contexts social network with friends, Pitman’s T(96) = 2.14, p = 0.035, and social network with strangers, Pitman’s T(90) = 2.32, p = 0.023, as well as for openness in the activity contexts dating, Pitman’s T(87) = 4.22, p < 0.001, competitive gaming, Pitman’s T(92) = 2.99, p = 0.004, social network with friends, Pitman’s T(96) = 4.62, p < 0.001, and social network with job contacts, Pitman’s T(100) = 2.57, p = 0.012.
The present study investigated the preferences of self-presentation through avatars compared to actual self-presentation and an idealised self-presentation depending on several online activity contexts.
First, we examined the relations between the physical/demographical characteristics attributed to the avatar, the actual self, and the ideal self. We consistently found high positive correlations or high degrees of congruence between the respective characteristics of the avatar, the actual self, and the ideal self. These findings were confirmed for the whole sample (i.e., context-independent) and each online activity context. Hence, the avatar’s physical and demographic characteristics generally appear to be based on the characteristics of the actual self, with some enhancements and improvements, as also indicated by previous research (Cacioli & Mussap, 2014; Ducheneaut et al., 2009; Messinger et al., 2008) and the extended real-life hypothesis (Back et al., 2010). The present data do not support a complete decoupling of avatar characteristics from the offline self or a tendency to try out different identities oriented towards fictional characters.
Second, we found an idealised representation of avatar personality traits in almost every activity context. An idealised avatar was confirmed for extraversion, agreeableness, neuroticism, and conscientiousness regarding the whole sample. Interestingly, the extent of idealisation was most pronounced for neuroticism, extending previous findings that especially negatively connoted personality traits are presented in an improved way (e.g., Dengah & Snodgrass 2020). In the case of openness, no significant extent of idealisation was found. These results can be generalised to the individual activity contexts with few exceptions. When focusing on a direct comparison of activity contexts, we found significant differences regarding the idealisation of the avatar’s personality traits for conscientiousness and openness. So, avatar creation partly depends on the activity context and the corresponding purpose of the avatar (Triberti et al., 2017; Trepte & Reinecke, 2010; Trepte et al., 2009; Vasalou & Joinson, 2009). Additionally, we found significant gender effects on avatar idealisation regarding conscientiousness and neuroticism. For conscientiousness, male participants idealised their avatar more strongly than female participants, whereas female participants idealised their avatars more than males concerning neuroticism. Our results align with Triberti et al. (2017) and Trepte et al. (2009), who also found gender effects on avatar design, and underline the importance of considering differential effects when investigating avatar creation.
Last, we hypothesised that the personality traits of the avatar would vary more than those of the actual self and the ideal self. At the level of the whole sample, the results only partly support this expectation. The variance in the avatar’s personality traits in terms of the Big Five was larger than the variance in trait scores attributed to the ideal self, except for agreeableness. Contrary to our expectation, however, the variance in the avatar’s personality scores was mainly lower than the variance found for the actual self, with the exception of openness. Thus, the comparison of variances could shed new light on the strategies of avatar creation, underlining the potential benefits of this analysis method.
Implications for research and practice
In general, our results regarding the idealisation of the avatar’s personality traits are in line with impression management theory (Mummendey, 1995), self-discrepancy theory (Higgins, 1987), and idealised virtual identity hypothesis (Manago et al., 2008), according to which people tend to present themselves as positively as possible. This tendency also seems to be remarkably robust across different activity contexts for which the avatar is created. In this sense, our results also replicated previous research stating that personality traits are mainly actualised or idealised (Sibilla & Mancini, 2018; Ducheneaut et al., 2009; Bessière et al., 2007). Our relatively robust results on avatar idealisation support the assumption that people create avatars with the intention of minimizing their self-discrepancy (cf. Bessière et al., 2007). In practice, it remains to be shown whether the overreporting of positive personality characteristics leads to far-reaching consequences in the respective activity context. On the one hand, research regarding avatar creation in video games found short-term positive effects via reduction of self-discrepancy by identifying with the avatar and by adopting its desired characteristics during gameplay (Klimmt et al., 2009). On the other hand, previous studies could show that the identification with and the use of an excessively idealised avatar can be related to low self-esteem and game addiction (Lemenager et al., 2020), depression and depressive symptoms (You et al., 2017; Bessière et al., 2007), and internet gaming disorder (T’ng & Pau, 2020). Furthermore, gaming disorder is linked to negative self-concepts, which might be compensated through the excessive use of and identification with an idealised avatar (Lemenager et al., 2020). These relationships between low self-esteem, self-discrepancy, identification with an idealised avatar, and excessive media use have mainly been studied in the context of video games and gaming disorder. The present results now also bring other contexts into focus such as dating platforms and social networking sites. In this context, Brunskill (2013) already highlighted the effects of avatars and their idealisation on social media on users’ psychopathology. Furthermore, idealised avatars on social media could negatively affect upward social comparisons by providing unrealistic comparative images. Social comparisons can negatively affect self-image and mental health, as previous research has already shown (e.g., Robinson et al., 2019; Fardouly et al., 2015). Some work also stressed the negative effect of idealised avatars on body image and body satisfaction (Park, 2018; Cacioli & Mussap, 2014). Consequently, further research on the impact of creating and using idealised avatars on users’ mental health in digital contexts other than gaming is generally essential.
Interestingly, avatars seem to be more idealised in gaming contexts than in dating or some social network contexts regarding the personality trait conscientiousness. This result exemplarily highlights the strategic properties in avatar creation. People can represent themselves as more idealised in more anonymous video gaming contexts, whereas in the contexts of dating or social networking, where personal contact outside the virtual world is more likely, a more balanced self-representation appears to be appropriate. On the one hand, this underlines that avatars can be idealised in video games more extensively, shifting the research focus to potential side effects on players’ well-being and mental health as well as potential transfer effects from the online to the offline world (cf. Kaspar, 2017). On the other hand, it becomes clear that findings about the extent of avatar idealisation in one activity context cannot simply be generalised to other contexts. In general, however, we found that participants portrayed idealised personality traits via the avatar, supporting the virtual identity hypothesis (Manago et al., 2008). In contrast, most participants portrayed their actual physical and demographic characteristics via the avatar, supporting the extended real-life hypothesis (Back et al., 2010). Hence, avatar creation seems to be highly variable across online activity contexts and it strongly depends on the type of avatar characteristics.
Regarding the found effects of the participant’s gender on avatar idealisation, we can only speculate why men idealised their avatars in terms of conscientiousness and women idealised their avatars in terms of neuroticism more compared to the respective opposite gender. This finding may be related to different motives of men and women for using internet-based communication environments. Previous research has shown that men tend to be more task- and achievement-oriented in social media and video games, while women tend to use the internet for relationship-oriented goals instead (Guadagno et al., 2011; Williams et al., 2009). Since conscientiousness is positively related to task performance (Bakker et al., 2012) and neuroticism is negatively related to relationship commitment (Kurdek, 1997), the respective idealisation of these personality traits might reflect different communication goals of men and women. This result opens exciting perspectives for future research on the different goal-oriented aspects of avatar usage in virtual environments.
Last, the analysis of dispersion measures, which has rarely been done, suggest interesting implications. The lower variance in avatar’s personality scores compared to those of the actual self pinpoints that users do not fully explore all possibilities of avatar creation across online activity contexts. Instead, users all seem to have a relatively similar idea of what a suitable avatar for self-representation should look like. This could be based on social ideas that do not reflect the variance in the actual selves. At the same time, however, we observed more variance in the avatar’s personality traits compared to the ideal self. Although our results suggest a general idealisation of avatars, some traits are allowed to vary more than others and are somewhat closer to the ideal self. As some previous studies indicated (Vasalou & Joinson, 2009; Vasalou et al., 2008), this could present a partly balanced and strategic representation of users’ own personalities through avatars. Remarkably, variances in the avatar’s and the actual self’s personality traits did not differ significantly in most cases when analysing individual activity contexts, which was paralleled by a decrease of sample size and consequently test power. If they did differ significantly, the variance in the avatar was always smaller than the variance in the actual self. In contrast, the variance in the avatar’s personality traits was larger than the variance in the ideal self’s personality traits in most activity contexts. Interestingly, the activity contexts differed in which of the avatar’s personality traits varied more and which varied less. This could indicate that, depending on the activity context and the associated purpose of avatar creation, there is a more or less shared understanding of the permissible scope for avatar design. From a practical point of view, these results provide preliminary information for digital media developers. Examining the variances in avatar creation can give developers a sense of the scope of possible avatar characteristics that a video game or social networking site has to offer to be inclusive. This study is only a first step in this direction, in any case. Future research should investigate what other features of an avatar, including those specific to the platform, should be customisable and to what extent.
Strengths and limitations
This study has a number of strengths and innovations: First, we studied the creation of avatars in a wide range of different online contexts. To our knowledge, such a wide range has not been realised before. Second, by comparing the variances between the avatar, the actual self, and the ideal self, we used a statistical approach that is novel in the field. This approach has provided new insights and a deeper understanding of avatar creation strategies. Third, we obtained a large sample with a reasonably high test power to determine the existence of relevant effects. Finally, although we could not consider all potential characteristics of an avatar in our study, we captured both key physical/demographic and psychological variables, whereas previous studies have often focused on one of the two dimensions.
In terms of limitations, we asked about hypothetical avatar creation for different activity contexts via vignettes. Most previous studies of avatar creation used avatars created for a specific situation (e.g., for specific video games, cf. Trepte & Reinecke, 2010), allowing participants to see their avatar and interact with them in the specific activity context. So, seeing how the avatar is built could influence the physical characteristics attributed to the avatar, just as the possibility of interacting with the avatar could influence the evaluation of the psychological variables. It has been shown that, for example, the style of visualisation (Yoon et al., 2019) and the visual similarity of the avatar with the user (Mansour et al., 2006) can have effects on behaviour in and perception of social interactions in collaborative digital environments. Schuurink & Toet (2010) also found that acting with an avatar in a third-person perspective in which the avatar is visible (versus first-person perspective from the avatar’s point of view) has positive effects on engagement and viewing strategies in digital environments. Therefore, the visualisation of avatars can impact people’s behaviour and perceptions in digital environments. However, since our study had a strong focus on the psychological characteristics of avatars, the lack of visualisation may have had only a minor influence on the corresponding results. The effect of (non-)visualisation on physiological and demographic characteristics may be more pronounced. Indeed, we would argue that the hypothesised engagement with an avatar’s personality traits realised here was an ideal first step, as a concrete form of visualisation could severely limit the generalizability of trait attribution.
Furthermore, we did not track actual behaviour in the different activity contexts. In fact, actions via an avatar can influence real-life behaviour and behaviour change (Yoon & Vargas, 2014; Yee & Bailenson 2007). It would be interesting to investigate whether and to what extent avatar idealisation and differences in avatar idealisation across activity contexts are associated with concrete behaviour and how the users’ personality traits are manifested in their behaviour in the digital environments. For example, video games such as Little Big Planet (Media Molecule, 2008) already allow different facial expressions and gestures by pressing a button and tilting the joystick. So, different levels of the Big Five could be represented by the avatar’s non-verbal behaviour, which the real person could control. Of course, the possibilities also depend on technological progress. It is generally possible to track avatar and user behaviour in digital environments. For example, a high level of extraversion in an online dating context could be related to the frequency of initiating contacts, or a high level of openness could be related to the amount and type of personal information one discloses in specific digital environments. There is also the option to use virtual reality headsets, special suits, and cameras to directly map users’ gestures and facial expressions onto avatars. It is already feasible to study users’ behaviour in the digital context and its relationships to their personality profiles based on nonverbal behaviour, text- and audio-based conversation protocols, and via analyses of avatar-avatar-interactions within virtual environments.
Although we covered the most common avatar characteristics, others may still be idealised to a different extent or would be even more appropriate for identity play. Video games provide extensive opportunities to specify even more specific details in the avatar’s appearance (e.g., tattoos), ethnicity, abilities, skills, status (e.g., feared or famous), and even species (human and non-human). Other extensive options might be more motivating for trying different identities than the characteristics examined in the present study, which would lead to a higher variance in avatars than that found here. The study of such extensive possibilities would be fascinating in other contexts.
All in all, we have begun to fill a major empirical gap with the present study. With the emergence and discussion of the avatar-based Metaverse, avatar creation will become an increasingly important topic in the context of diverse virtual environments that focus on tools for self-representation. In this context, we found a strong tendency to create avatars with an idealised personality. We also found significant differences in the extent of idealisation between online activity contexts. This and differences in avatar variances across contexts also point to strategic properties of avatar creation and the role of social norms. Furthermore, there were gender effects on the idealisation of the avatar’s conscientiousness and neuroticism, highlighting the relevance of differential effects on avatar creation. Moreover, the demographic and physical characteristics of the avatar, the actual self, and the ideal self showed high congruence. Thus, the strategies of avatar creation also seem to differ between the characteristics to be assigned. Hence, this study helps to understand basic strategies behind avatar creation in different contexts and provides a fruitful basis for future research and practice, particularly research on potential problems with self-image and excessive media use as well as practical implications for developers and designers of virtual worlds.
The data that support the findings of this study, as well as the used materials, are available from the corresponding author upon reasonable request.
Back, M. D., Stopfer, J. M., Vazire, S., Gaddis, S., Schmukle, S. C., Egloff, B., & Gosling, S. D. (2010). Facebook profiles reflect actual personality, not self-idealization. Psychological Science, 21(3), 372–374
Bailenson, J. N., Yee, N., Blascovich, J., Beall, A. C., Lundblad, N., & Jin, M. (2008). The use of immersive virtual reality in the learning sciences: Digital transformations of teachers, students, and social context. The Journal of the Learning Sciences, 17(1), 102–141. https://doi.org/10.1080/10508400701793141
Bakker, A. B., Demerouti, E., & Lieke, L. (2012). Work engagement, performance, and active learning: The role of conscientiousness. Journal of vocational behavior, 80(2), 555–564. https://doi.org/10.1016/j.jvb.2011.08.008
Bente, G., Krämer, N. C., & Eschenburg, F. (2008). Is there anybody out there? Analyzing the effects of embodiment and nonverbal behavior in avatar-mediated communication. In E. Konijn, S. Utz, M. Tanis, & S. Barnes(Hrsg.) Mediated interpersonal communication (pp. 131–157). Routledge
Bessière, K., Seay, A. F., & Kiesler, S. (2007). The ideal elf: Identity exploration in World of Warcraft. Cyberpsychology & Behavior, 10(4), 530–535. https://doi.org/10.1089/cpb.2007.9994
Brunskill, D. (2013). Social media, social avatars and the psyche: Is Facebook good for us? Australasian Psychiatry, 21(6), 527–532
Cacioli, J. P., & Mussap, A. J. (2014). Avatar body dimensions and men’s body image. Body Image, 11(2), 146–155. https://doi.org/10.1016/j.bodyim.2013.11.005
Costa, P. T., Jr., & McCrae, R. R. (1992). Four ways five factors are basic. Personality and Individual Differences, 13(6), 653–665. https://doi.org/10.1016/0191-8869(92)90236-I
Dengah, H. F., & Snodgrass, J. G. (2020). Avatar creation in videogaming: Between compensation and constraint. Games for Health Journal, 9(4), 265–272. https://doi.org/10.1089/g4h.2019.0118
Ducheneaut, N., Wen, M. H., Yee, N., & Wadley, G. (2009). Body and mind: A study of avatar personalization in three virtual worlds. In Proceedings of the SIGCHI conference on human factors in computing systems (pp. 1151–1160). https://doi.org/10.1145/1518701.1518877
Dunn, R. A., & Guadagno, R. E. (2012). My avatar and me–Gender and personality predictors of avatar-self discrepancy. Computers in Human Behavior, 28(1), 97–106. https://doi.org/10.1016/j.chb.2011.08.015
Fardouly, J., Diedrichs, P. C., Vartanian, L. R., & Halliwell, E. (2015). Social comparisons on social media: The impact of Facebook on young women’s body image concerns and mood. Body Image, 13, 38–45. https://doi.org/10.1016/j.bodyim.2014.12.002
Guadagno, R. E., Muscanell, N. L., Okdie, B. M., Burk, N. M., & Ward, T. B. (2011). Even in virtual environments women shop and men build: A social role perspective on Second Life. Computers in Human Behavior, 27(1), 304–308. https://doi.org/10.1016/j.chb.2010.08.008
Higgins, E. T. (1987). Self-discrepancy: A theory relating self and affect. Psychological Review, 94(3), 319–340. https://doi.org/10.1037/0033-295X.94.3.319
Hussain, Z., & Griffiths, M. D. (2008). Gender swapping and socializing in cyberspace: An exploratory study. CyberPsychology & Behavior, 11(1), 47–53. https://doi.org/10.1089/cpb.2007.0020
Kafai, Y. B., Fields, D. A., & Cook, M. S. (2010). Your second selves: Player-designed avatars. Games and Culture, 5(1), 23–42
Kaspar, K. (2017). Lernen durch Computerspielen: Erwünschte und unerwünschte Nebeneffekte. In: Zielinski, W., Aßmann, S., Kaspar, K. & Moormann, P. (Hrsg.), Spielend lernen! Computerspiele(n) in Schule und Unterricht (S.27–37). Kopaed. https://doi.org/10.1007/978-3-658-25090-4_68-1
Kaspar, K., & Fuchs, L. A. M. (2021). Who likes what kind of news? The Relationship between characteristics of media consumers and news interest. SAGE Open, 11(1), 1–12. https://doi.org/10.1177/21582440211003089
Kauttonen, J. (2021). PitmanMorganTest(x,y). In MATLAB central file exchange. Retrieved March 24, 2021, from https://www.mathworks.com/matlabcentral/fileexchange/67910-pitmanmorgantest-x-y
Klimmt, C., Hefner, D., & Vorderer, P. (2009). The video game experience as “true” identification: A theory of enjoyable alterations of players’ self-perception. Communication theory, 19(4), 351–373. https://doi.org/10.1111/j.1468-2885.2009.01347.x
Kovaleva, A., Beierlein, C., Kemper, C. J., & Rammstedt, B. (2013). Psychometric properties of the BFI-K: A cross-validation study. The International Journal of Educational and Psychological Assessment, 13(1), 34–50
Kromand, D. (2007). Avatar Categorization. In Situated Play, Proceedings of DiGRA 2007 Conference. University of Tokyo, 400–406
Kurdek, L. A. (1997). Relation between neuroticism and dimensions of relationship commitment: Evidence from gay, lesbian, and heterosexual couples. Journal of Family Psychology, 11(1), 109–124. https://doi.org/10.1037/0893-3220.127.116.11
Lee, L. H., Braud, T., Zhou, P., Wang, L., Xu, D., Lin, Z., Kumar, A., Bermejo, C., & Hui, P. (2021). All one needs to know about metaverse: A complete survey on technological singularity, virtual ecosystem, and research agenda. arXiv preprint arXiv:2110.05352. https://doi.org/10.48550/arXiv.2110.05352
Lemenager, T., Neissner, M., Sabo, T., Mann, K., & Kiefer, F. (2020). “Who Am I” and “How Should I Be”: A Systematic review on self-concept and avatar identification in gaming disorder. Current Addiction Reports, 7(2), 166–193. https://doi.org/10.1007/s40429-020-00307-x
Lin, H., & Wang, H. (2014). Avatar creation in virtual worlds: Behaviors and motivations. Computers in Human Behavior, 34, 213–218. https://doi.org/10.1016/j.chb.2013.10.005
Little Big Planet [Software] (2008). Guildford, GB:Media Molecule
Loewen, M. G., Burris, C. T., & Nacke, L. E. (2021). Me, Myself, and Not-I: Self-discrepancy type predicts avatar creation style. Frontiers in Psychology, 11, 1902. https://doi.org/10.3389/fpsyg.2020.01902
Mancini, T., & Sibilla, F. (2017). Offline personality and avatar customisation. Discrepancy profiles and avatar identification in a sample of MMORPG players. Computers in Human Behavior, 69, 275–283. https://doi.org/10.1016/j.chb.2016.12.031
Manago, A. M., Graham, M. B., Greenfield, P. M., & Salimkhan, G. (2008). Self-presentation and gender on MySpace. Journal of Applied Developmental Psychology, 29(6), 446–458. https://doi.org/10.1016/j.appdev.2008.07.001
Mansour, S., El-Said, M., Rude-Parkins, C., & Nandigam, J. (2006). The interactive effect of avatar visual fidelity and behavioral fidelity in the collaborative virtual reality environment on the perception of social interaction. WSEAS Transactions on Communications, 5(8), 1501–1509
Meier, J. V., Noel, J. A., & Kaspar, K. (2021). Alone together: computer-mediated communication in leisure time during and after the COVID-19 pandemic. Frontiers in Psychology, 12, 2040. https://doi.org/10.3389/fpsyg.2021.666655
Messinger, P. R., Ge, X., Stroulia, E., Lyons, K., Smirnov, K., & Bone, M. (2008). On the relationship between my avatar and myself. Journal For Virtual Worlds Research, 1(2), https://doi.org/10.4101/jvwr.v1i2.352
Mummendey, H. D. (1995). Psychologie der Selbstdarstellung [Psychology of Self-Representation], Hogrefe
Park, J. (2018). The effect of virtual avatar experience on body image discrepancy, body satisfaction and weight regulation intention. Cyberpsychology: Journal of Psychosocial Research on Cyberspace, 12(1), https://doi.org/10.5817/CP2018-1-3
Rammstedt, B., & John, O. P. (2005). Kurzversion des Big Five Inventory (BFI-K). Diagnostica, 51(4), 195–206. https://doi.org/10.1026/0012-1918.104.22.168
Robinson, A., Bonnette, A., Howard, K., Ceballos, N., Dailey, S., Lu, Y., & Grimes, T. (2019). Social comparisons, social media addiction, and social interaction: An examination of specific social media behaviors related to major depressive disorder in a millennial population. Journal of Applied Biobehavioral Research, 24(1), e12158. https://doi.org/10.1111/jabr.12158
Schuurink, E. L., & Toet, A. (2010). Effects of third person perspective on affective appraisal and engagement: Findings from SECOND LIFE. Simulation & Gaming, 41(5), 724–742
Sibilla, F., & Mancini, T. (2018). I am (not) my avatar: A review of the user-avatar relationships in massively multiplayer online worlds. Cyberpsychology: Journal of Psychosocial Research on Cyberspace, 12(3), https://doi.org/10.5817/CP2018-3-4
Steiner, M., Allemand, M., & McCullough, M. E. (2012). Do agreeableness and neuroticism explain age differences in the tendency to forgive others? Personality and Social Psychology Bulletin, 38(4), 441–453. https://doi.org/10.1177/0146167211427923
Tivian. (2017). EFS Survey, Version Summer 2017. Tivian XI GmbH
T’ng, S. T., & Pau, K. (2020). Identification of avatar mediates the associations between motivations of gaming and internet gaming disorder among the Malaysian youth. International Journal of Mental Health and Addiction, 1–16. https://doi.org/10.1007/s11469-020-00229-9
Trepte, S., & Reinecke, L. (2010). Avatar creation and video game enjoyment. Effects of life satisfaction, game competitiveness, and identification with the avatar. Journal of Media Psychologie, 22(4), 171–184. https://doi.org/10.1027/1864-1105/a000022
Trepte, S., Reinecke, L., & Behr, K. M. (2009). Creating virtual alter egos or superheroines? Gamers’ strategies of avatar creation in terms of gender and sex. International Journal of Gaming and Computer-Mediated Simulations, 1(2), 52–76. https://doi.org/10.4018/jgcms.2009040104
Triberti, S., Durosini, I., Aschieri, F., Villani, D., & Riva, G. (2017). Changing avatars, changing selves? The influence of social and contextual expectations on digital rendition of identity. Cyberpsychology Behavior and Social Networking, 20(8), 501–507. https://doi.org/10.1089/cyber.2016.0424
Vasalou, A., & Joinson, A. N. (2009). Me, myself and I: The role of interactional context on self-presentation through avatars. Computers in human behavior, 25(2), 510–520. https://doi.org/10.1016/j.chb.2008.11.007
Vasalou, A., Joinson, A., Bänziger, T., Goldie, P., & Pitt, J. (2008). Avatars in social media: Balancing accuracy, playfulness and embodied messages. International Journal of Human-Computer Studies, 66(11), 801–811. https://doi.org/10.1016/j.ijhcs.2008.08.002
Williams, D., Consalvo, M., Caplan, S., & Yee, N. (2009). Looking for gender: Gender roles and behaviors among online gamers. Journal of Communication, 59(4), 700–725. https://doi.org/10.1111/j.1460-2466.2009.01453.x
Yee, N., & Bailenson, J. (2007). The Proteus effect: The effect of transformed self-representation on behavior. Human communication research, 33(3), 271–290. https://doi.org/10.1111/j.1468-2958.2007.00299.x
Yoon, G., & Vargas, P. T. (2014). Know thy avatar: The unintended effect of virtual-self representation on behavior. Psychological Science, 25(4), 1043–1045
Yoon, B., Kim, H. I., Lee, G. A., Billinghurst, M., & Woo, W. (2019). The effect of avatar appearance on social presence in an augmented reality remote collaboration. In 2019 IEEE Conference on Virtual Reality and 3D User Interfaces (VR) (pp. 547–556). IEEE. https://doi.org/10.1109/VR.2019.8797719
You, S., Kim, E., & Lee, D. (2017). Virtually real: Exploring avatar identification in game addiction among massively multiplayer online role-playing games (MMORPG) players. Games and Culture, 12(1), 56–71
Open Access funding enabled and organized by Projekt DEAL.
Conflict of interest
The authors declare that they have no conflict of interest.
All procedures performed in the study were in accordance with the ethical guidelines of the German Psychological Society (DGPs) and with the 1964 Helsinki declaration. According to the guidelines of the German Research Association, no ethical approval was needed because the research did not pose any threats or risks to the respondents, it was not associated with high physical or emotional stress, and the respondents were informed about the objectives of the study (http://www.dfg.de/foerderung/faq/geistes_sozialwissenschaften/index.html).
Consent to participate/for publication
Written informed consent to participate in this study and for publication of their data was provided online by the participants.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Below is the link to the electronic supplementary material.
(PDF 272 KB)
About this article
Cite this article
Zimmermann, D., Wehler, A. & Kaspar, K. Self-representation through avatars in digital environments. Curr Psychol (2022). https://doi.org/10.1007/s12144-022-03232-6
- Avatar creation
- Virtual world
- Activity context
- Big five