Body Image 11 (2014) 146–155 Contents lists available at ScienceDirect Body Image journal homepage: www.elsevier.com/locate/bodyimage Avatar body dimensions and men’s body image Jon-Paul Cacioli a,b,∗ , Alexander J. Mussap a a b School of Psychology, Deakin University, Melbourne, Australia School of Psychological and Clinical Sciences, Charles Darwin University, Darwin, Australia a r t i c l e i n f o a b s t r a c t Article history: Received 12 April 2013 Received in revised form 10 November 2013 Accepted 17 November 2013 Keywords: Avatars Men Body dissatisfaction, Drive for thinness, Drive for muscle Two online surveys examined the significance of the visual analogues, or ‘avatars’, men (total N = 266) create and use online. Two-dimensional (adiposity × muscle) somatomorphic matrices revealed that avatars are generally thinner than their creator’s actual body and similar to their ideal, but more muscular than either their actual or ideal. Men’s ratings of the importance of their avatar’s appearance correlated with their actual weight and muscle concerns, and disparity between their avatar and actual body dimensions predicted their offline context body change concerns additional to that accounted for by disparity between their ideal and actual bodies. Together with the observation that men also reported higher self-esteem, less social interaction anxiety and less social phobia while online (which correlated with the time they spent online), these results suggest that the physical dimensions of avatars used in social interactions online may serve a compensatory function. © 2013 Elsevier Ltd. All rights reserved. Introduction Sociocultural pressures on body image – from media, peers, and family (Frederick, Fessler, & Haselton, 2005; Thompson, Heinberg, Altabe, & Tantleff-Dunn, 1999) – have been implicated in appearance concerns and unhealthy body-change behaviours (Heywood & McCabe, 2006; Olivardia, Pope, Borowiecki, & Cohane, 2004; Pope, Phillips, & Olivardia, 2000). Although the focus of previous body image research has primarily been regarding women there is substantial evidence that some men and boys experience appearance concerns that can manifest as a drive for thinness and/or a drive for muscularity (Frederick et al., 2007; Grogan, 2008; Pope et al., 2000; Ricciardelli & McCabe, 2004). In the present article we examined men’s appearance concerns in the context of their social interactions online. The primary aim was to investigate the significance of men’s self-perceived bodies, their ideal bodies, and the bodies of the visual analogues, or avatars, they create and use to represent themselves online. Appearance Concerns, Online Social Interactions, and Avatar Use Appearance concerns in men have been linked to a range of negative psychosocial outcomes including depressed mood ∗ Corresponding author at: School of Psychological and Clinical Sciences, Charles Darwin University, Casuarina Campus, Ellengowan Drive, Darwin 0909, Northern Territory, Australia. Tel.: +61 08 8946 6818; fax: +61 08 8946 6151. E-mail addresses: [email protected], [email protected] (J.-P. Cacioli). 1740-1445/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.bodyim.2013.11.005 (Tiggemann & Kuring, 2004), low self-esteem (Wade, 2000), social anxiety (Cash & Fleming, 2002), problems with intimacy and impaired interpersonal relationships (Nezlek, 1999), and social phobia (Izgiç, Akyüz, Doğan, & Kuğu, 2004). The central premise of our research was that these psychosocial consequences may contribute to men’s reasons for interacting socially online and their motivation for creating and using avatars with particular physical dimensions during these interactions. This is a potentially important issue as time spent online has increased dramatically in the past decade (Miranda & Lima, 2012) with online social interactions becoming commonplace (Childress & Braswell, 2006; Cole & Griffiths, 2007; Martey & Stromer-Galley, 2007), more natural, and capable of combining anonymity with intimacy (Christopherson, 2007; Walther, 1996; Walther, Slovacek, & Tidwell, 2001; Whitty, 2008). Additionally, the use of avatars has become more common, allowing Internet users to project, and perhaps even identify with, a visual identity that is unconstrained by reality (Galanxhi & Nah, 2007; Trepte & Reinecke, 2010). According to Suler (1999), in an anonymous Internet environment avatars may be used to facilitate more controllable, safe, and potentially more rewarding social interactions. Galanxhi and Nah (2007) demonstrated this to be so. In their study participants were asked to deceive others in online interactions. When deceivers were permitted to represent themselves with an avatar they reported no increase in anxiety; when avatars were not permitted and the interaction remained entirely text-based, the deceivers reported increased anxiety. This difference was taken to confirm that avatars potentially enhance confidence and reduce anxiety in Internet users in difficult social interactions, potentially providing a distance from self. Similarly, Bessiere, Seay, and Kiesler (2007) proposed that social interactions online may be used to J.-P. Cacioli, A.J. Mussap / Body Image 11 (2014) 146–155 compensate for perceived personality deficiencies. They compared personality ratings of participants’ actual selves versus their avatars and discovered that avatars were rated as less neurotic, more extroverted and more conscientious. These studies lend support to the proposition that online avatar use has psychosocial implications. Given the everyday use of online interactions for both personal and business purposes, there is a growing emphasis on the social importance of these interactions which are taking place online, particularly in regard to the normalisation of this process and the unique ability to alter one’s appearance in the online context. In the context of body image research, the question of interest is whether the physical dimensions of the avatars used in these interactions are also relevant. We consider this issue in the next section in light of literature on body image disparity. Avatar Physical Dimensions and Self-Discrepancy Self-discrepancy theory (Higgins, 1987) proposes that negative and potentially unhealthy emotional and motivational states can result from disparity between one’s self-perceived body and one’s ideal body. Hence, the physical dimensions of avatars that men create and use online may be manifestations of this selfdiscrepancy. Low adiposity combined with muscularity is both a male appearance ideal as well as a symbol of masculinity and male competencies, such as physical strength and athletic prowess (Oehlhof, Musher-Eizenman, Neufeld, & Hauser, 2009). Therefore, discrepancy between physical dimensions of self and in the context of avatars may be aesthetic in nature and/or related to aspects of competency. The physical dimensions of avatars versus those of their creators may also reflect the avatars’ specific functions/roles online (Axelsson, 2002). The online physical dimensions may be realistic or at least self-referential depictions of oneself regardless of context (Vasalou & Joinson, 2009); depictions that conform to a particular context, function, role or character (Hussain & Griffiths, 2008; Kafai, Fields, & Cook, 2007; Taylor, 2002, 2003; Trepte & Reinecke, 2010); and/or depictions that represent a fantasy self, offering the creator the freedom to adopt a fundamentally different persona, perhaps one that reflects his/her ideal body (Konijn & Bijvank, 2009; Lawson, 2000; McKenna & Bargh, 2000). Although each of these contexts and functions is potentially relevant, the motivation behind valuing the avatar’s aesthetic properties over its functional/practical properties is potentially the most significant. To the extent that appearance objectification of one’s actual body is related to increased body concerns and unhealthy body change behaviours in the offline context (Olivardia et al., 2004; Schwartz & Tylka, 2008; Tylka, Bergeron, & Schwartz, 2005), an appearance focus in the context of one’s avatars online may reflect underlying body image disparity and appearance concerns (Lawson, 2000; McKenna & Bargh, 2000). Additionally the avatar could reflect the creators’ attempts to compensate for these offline world concerns (Trepte & Reinecke, 2010). These possibilities were tested in two studies with both quantitative and qualitative components. Our general aims were to identify the psychosocial factors that contribute to men’s use of avatars online and to explore the relevance of avatars’ physical dimensions – separately for weight and muscle – to their creators’ body concerns and body change drives. 1. The first study focussed on documenting the diverse uses of avatars and the relative importance of these avatars – in terms of aesthetics versus competencies – to their creators’ offline body image concerns. 2. The second study focussed on the implications of men’s offline and online/avatar physical dimensions to psychosocial functioning (self-esteem, social phobia and social interaction anxiety), 147 and the relevance of these factors to men’s motivations for going online. Study 1 In Study 1 we examined the importance of men’s physical bodies and their online (avatar) bodies. An online questionnaire was completed by these men that assessed their body concerns separately for weight and muscle and the perceived importance of aesthetic versus competency factors in evaluating their own body and their avatar’s body. Additionally participants were invited to elaborate in writing on the meaning of their avatars to their lives. H1. Based on previous research it was hypothesised that a qualitative analysis would reveal that the physical attributes of men’s avatars would reflect either a representation of men’s offline self (Vasalou & Joinson, 2009), their ideal (Konijn & Bijvank, 2009; Lawson, 2000; McKenna & Bargh, 2000) or be independent of self, such as to fulfil a role in a particular context (Hussain & Griffiths, 2008; Kafai et al., 2007; Taylor, 2002, 2003; Trepte & Reinecke, 2010). H2. On the basis of men’s body image research in the offline or physical context, we hypothesised that men’s aesthetic concerns with their offline appearance would be higher than their concerns with their offline physical competencies (Arbour & Ginis, 2006; Grogan & Richards, 2002). H3. On the basis of literature on appearance self-objectification (Hebl, King, & Lin, 2004; Morry & Staska, 2001; Strelan & Hargreaves, 2005), we further hypothesised that men’s aesthetic, but not competency concerns, would be predictive of their level of their appearance concerns. Given the dearth of research on body image online, we could only speculate that men’s avatar concerns would mirror their concerns with their own bodies. This is an important element of the study as it allows for an analysis exploring whether the avatar is influenced by appearance objectification that may be evident in men’s perception of their own bodies, and if so, to what extent? Method Participants. An online community sample of 135 men aged from 18 years to 62 years (M = 34.17, SD = 11.02) who reported creating and regularly using one or more avatars online participated in Study 1. Respondents resided in Australia (31.6%), the United States of America (30.4%), Canada (9.1%) and the United Kingdom (6.1%). Participants reported currently using an average of 7.58 (SD = 12.53) avatars. In terms of the context in which avatars are used, a total of 63 software applications were cited. The majority of these were multiplayer role-playing games, the most popular being World of WarcraftTM (53.38%). The second most popular application of avatars was for social interactions such as Second LifeTM (24.1%). Almost half (45.9%) of the respondents reported using avatars in multiple contexts. Measures. In addition to questions about demographics, weight and height (to calculate Body Mass Index [BMI]), and avatar use, participants were asked that if they used multiple avatars, and which one was most important to them. They were instructed to think of this avatar when responding to avatar-specific items from the measures described below. Psychometric properties of the scales used in the Study 1 are reported in the results section (Table 1). Self-Objectification Questionnaire (SOQ). A modified version of the SOQ (Noll & Fredrickson, 1998) was used to assess the importance of physical appearance versus physical competency, separately for their actual body and their avatar body. The measure 148 J.-P. Cacioli, A.J. Mussap / Body Image 11 (2014) 146–155 Table 1 Pearson’s correlations of participant importance of appearance and competency for their real bodies and avatars, with concerns with weight and muscle. 1 2 3 4 5 6 1 2 3 4 SOQ Appearance (Self) SOQ Competency (Self) SOQ Appearance (Avatar) SOQ Competency (Avatar) MBAS Muscle MBAS Weight – .71** – .53** .27** – .27** .42** .49** – M SD Cronbach’s ˛ 6.12 2.24 .91 5 .33** .25** .31** .26** – 6.39 2.01 .87 4.64 3.11 .91 5.59 2.93 .92 1.58 1.01 .89 6 .22** .04 .26** 0.06 .47** – 2.27 1.22 .92 Note: SOQ Appearance, Self-Objectification Questionnaire Appearance subscale; SOQ Competency, Self-Objectification Questionnaire Competency subscale; MBAS Muscle, Male Body Attitudes Scale Muscle Dissatisfaction Subscale; MBAS Weight, Male Body Attitudes Scale Weight Dissatisfaction subscale. ** p < .01 (2-tailed). consists of five items relating to appearance (e.g., “How important to you is your physical attractiveness?”) and five relating to competence (e.g., “How important to you is your health and physical resilience?”). Although the original measure includes a ranking of the importance of these items out of 10, in the present study a Likert scale for each item was used anchored between “Not important” (0) to “Extremely important” (10). The SOQ has been shown to have good internal consistency (Fredrickson, Roberts, Noll, Quinn, & Twenge, 1998) and validity (Noll, as cited in Greive & Helmick, 2008, p. 292) and has previously been used in male body image studies (Greive & Helmick, 2008; Hallsworth, Wade, & Tiggemann, 2005). Male Body Attitudes Scale (MBAS). The body weight and body muscle subscales of the MBAS (Tylka et al., 2005) measure appearance on 6-point Likert scales rated from “Always” (0) to “Never” (6). Previous research has confirmed that the body weight and body mass subscales have high validity and reliability in a male sample (Tylka et al., 2005). Qualitative questions. At the completion of the questionnaire, participants were asked to respond to two open-ended questions: (i) “What does your avatar mean to you?” and (ii) “Is the appearance of your avatar important to you? Please elaborate.” Procedure. Permission to undertake the study was obtained from the Deakin University Ethics Committee. The study was advertised on Facebook, a social networking site, providing a brief study description and a link to the study’s Plain Language Statement. Webmaster’s and moderators of online forums and blogs that discussed avatar related issues such as online gaming, avatar fashion and current trends in avatar use online, such as www.themetaversjournal.com, were also contacted and asked for permission to display a brief description and link to the study. The advertisement stated that the study was recruiting male participants over the age of 18 years, who created and used avatars, in order to examine the relationship between the avatar and their physical selves. Participants were also invited to provide their email address, which was kept separately from the collected data to protect anonymity, and these were entered into a prize draw for a $100 Amazon voucher. Results Of the 135 men who completed the questionnaire, 2 were excluded from analyses for failing to meet the inclusion criteria. The remaining men (N = 133) ranged from 18 to 62 years in age (M = 32.09, SD = 11.02) with a mean BMI of 28.78 (SD = 7.69) ranging from 16.20 to 60.80. Quantitative responses were screened for missing values which were found to be randomly distributed across items. Less than .01% data points was missing from the dataset. No participant had more than 5% missing values. Multiple imputation using expectation maximisation was used to replace missing values. Variables were created by averaging summed responses to items. Variable descriptives, including Cronbach’s alphas, are included in Table 1. There were no outliers. Resultant variables also met criteria for normality (in relation to both skewness and kurtosis). As we had modified the SOQ, a factor analysis was conducted on the scale for both the offline and online contexts. Bartlett’s test of sphericity was significant for real world, 2 (45) = 971.95, p < .01, and online world, 2 (45) = 1085.04, p < .01, contexts. The Kaiser–Meyer–Olkin measures of sampling adequacy were .88 for the offline context and .85 for the online context. Communalities were all above .3. For the offline context, two factors were found. Initial eigenvalues explained 59.28% of the variance on the first factor and 12.25% of the variance on the second factor. After Oblimin rotation, one item cross loaded on both factors above .3 while the remaining nine items loaded above .68 on their respective factor and below .25 on the other factor. The online world condition also showed two factors with initial eigenvalues showing that the first factor explained 56.20% of the variance and the second factor explained 19.77%. After Oblimin rotation, all items loaded above .68 on their respective factor and below .25 on the other factor. The factor analysis supports the factorial validity of the scale. Analyses were conducted initially on the qualitative responses using thematic analysis followed by inferential analyses of the quantitative data. Thematic analysis involved a theoretical approach utilising an essentialist paradigm (Braun & Clarke, 2006). The thematic analysis was conducted based on the guidelines suggested by Braun and Clarke (2006), including the generation of initial codes and the collation of themes. The predominant themes, identified across the dataset generated by each of the qualitative questions, related to the theoretical framework of the study; specifically, the relationship between the avatar and how/if it represents the avatar’s creator in the online world. Identified themes were reviewed in regard to the coded extracts then defined and classified based on the primary function of the avatar as identified from the responses given. A second rater reviewed the results of the thematic analysis against the participants’ transcripts; discrepancies in interpretation were identified, discussed, and resolved to yield final classifications. The significance of avatar physical dimensions. Approximately 99.5% of participants responded to the first open-ended question concerning the relevance/meaning of the avatar to them, with the mean response length being 20.73 words (SD = 22.67, ME = 15). In regard to the second qualitative question concerning the significance of their avatar’s appearance, 99.98% of participants responded with a mean response length of 20.95 words (SD = 22.72, ME = 15). Thematic analysis of these responses supported our J.-P. Cacioli, A.J. Mussap / Body Image 11 (2014) 146–155 hypothesis (H1) and identified the following appearance-relevant typology: • Utilitarian/independent of self – avatars that visually conform to the functional demands imposed by the context, character or role independent of the user (e.g., the avatar “has to look ‘the part’ to fulfil its role” and “The avatar is a separate person, which I ‘direct’ from time to time”) • Actual self – avatars that visually correspond to the user’s selfperceived physical dimensions (e.g., “I identify with my avatar in any game. I try to make him look as similar to myself as possible to immerse myself in the game”) and/or their personal characteristics (e.g., “A representation of my personality” and “it’s my digital self.”) and • Idealised self – avatars that visually correspond to the appearance ideals held by the user (e.g., My SL (Second Life) avatar is an idealised form of myself, and so it’s an important way of representing myself online. As I am more often in contact with people online than in person, my avatar is crucial to my interactions with others”) and/or their idealised personal characteristics (e.g., “It is a visual representation of how I feel about myself on the inside, strong, capable, and appealing to others.”). 58.02% of respondents indicated that their avatar’s appearance was predominantly ‘utilitarian’; 21.37% indicated that their avatar corresponded to their ‘actual self’; and 12.98% indicated that their avatar corresponded to their ‘idealised self’. The remaining respondents were unable to be classified in the above categories but their avatar still appeared to have a personal meaning with strong attachment themes; for example, the avatar may represent a significant person in their life or even a has become a significant person in its own right through the participant’s interactions with it. The importance of appearance versus competence. Regardless of avatar type, the importance of the avatar’s appearance (including its adornments and equipment) was a central theme (“It’s a digital representation of the player, regardless of how physically similar it is. How other players perceive you (me) depends largely on the avatar appearance”). To test hypothesis H2, that participants would rate aesthetic concerns as more important than competency in the offline and online world and distinguish these aesthetic aspects of the avatar’s physical dimensions from its functional aspects, a 2 × 2 repeated measures ANOVA was conducted on importance ratings (‘appearance’ v ‘competence’; corresponding to the two subcomponents of the SOQ) and domain (oneself in the ‘offline context’ v one’s avatar in the ‘online context’). A significant main effect of domain, F(1, 132) = 29.52, p < .01, indicated that the appearance (M = 6.11, SD = 2.24) and competence (M = 6.39, SD = 2.01) of men’s bodies in the offline context are more important to them than their avatar’s appearance (M = 4.64, SD = 3.11, d = 0.54) and competence (M = 5.59, SD = 2.93, d = 0.32) in the online world. Furthermore, a significant main effect of importance rating, F(1, 132) = 11.62, p < .01, indicated that physical competence is more important to men than appearance. A significant interaction between these factors was also obtained, F(1, 132) = 8.90, p < .01, with the difference between competence and appearance being most pronounced in the online world. A comparison of offline importance of competency and appearance was borderline (p = .047) while the same comparison in the online world revealed that men rated their competency as more important than their appearance (p < .001) in the online context. This pattern of results did not support our hypotheses (H2) that aesthetic concerns would be more important than competency ones both in the offline and online world. To examine whether a focus on appearance rather than competence – that is, the tendency to appearance self-objectify (Oehlhof et al., 2009) – is relevant to men’s weight and muscle concerns, 149 correlations were conducted. The results are shown in Table 1, and reveal the presence of strong to moderate correlations between appearance and competence ratings for self and avatar respectively. Regardless of the dimension (appearance or competence) explored, we have interpreted these correlations as reflecting the importance participants place on their offline and online selves. The results further reveal that the importance of men’s own appearance was highly correlated with that of their avatar as was the importance of men’s physical competency in relation to their avatar’s competency. To this extent, the results supported our hypothesis that men’s physical, offline body and those of their avatar are influenced by similar values. Table 1 reveals that men’s ratings of the importance of appearance were more strongly correlated with their body concerns, particularly weight concerns, than were their ratings of the importance of physical competency. More importantly, the same was also found to be the case for ratings of avatar appearance versus competency. In order to fully explore hypothesis H3, multiple regression analyses were conducted examining whether the importance of appearance was a stronger predictor than competency for men’s dissatisfaction with weight. There was no evidence of multicollinearity or singularity amongst the IVs (Tolerance > .1 and VIF < 10). The results of the regression (Table 2), when controlling for importance of appearance in the offline context, indicated that participant rated importance of competency did not explain any additional variance for dissatisfaction with weight. This result was also found in the online context. Regressions were not conducted with muscle dissatisfaction due to a lack of significant correlation between importance place on competency and muscle dissatisfaction. The results support hypothesis H3, with the importance placed on appearance being stronger predictors of muscle and weight concerns. Study 2 In Study 2 we examined the relevance of men’s body image disparity to their body change drives. The novel aspect of this study was that we considered disparity not only between men’s actual and ideal bodies, but also between their actual and avatar bodies, and we did this separately for body adiposity and body muscle. We also examined the extent to which poor psychosocial functioning – low self-esteem, social interaction anxiety and social phobia – are related to body image disparity and contribute to men’s motivations for going online. Study 2 consisted of 157 men who did not participate in Study 1, and who reported creating and regularly using one or more avatars, completed an online questionnaire in which they used a 2-dimensional (2D) somatomorph matrix to estimate the physical dimensions (in terms of adiposity and muscle) of their actual, ideal, and avatar bodies. They also completed measures of self-esteem, social phobia and social interaction anxiety in the context of their social interactions in the offline and the online contexts; measures of body comparison and internalisation of the thin/muscular male ideal, again, separately for the offline context and the online context; and measures of their drive for thinness and muscle in the offline context. H4. In light of previous research showing the prevalence of the lean/muscular male ideal (Grogan, 2008; Pope et al., 2000; Ricciardelli & McCabe, 2004), it was hypothesised that men’s ideal body would be both thinner and more muscular than their actual body. H5. Furthermore, based on Higgins’ (1987) self-discrepancy theory which proposes that the disparity which is perceived to exist between an individual’s actual and ideal body image leads to unhealthy body change behaviours and poor body image, it was 150 J.-P. Cacioli, A.J. Mussap / Body Image 11 (2014) 146–155 Table 2 Hierarchical regression analyses of participant ratings of importance of competency in the offline (or online) context predicting dissatisfaction with weight and controlling for participant ratings of importance of appearance in the offline (or online) context. IV R2 R2 change * ˇ t * r sr2 Step 1 Importance of Appearance (offline) .05 .21 2.56 .22 .05 Step 2 Importance of Appearance (offline) Importance of Competency (offline) .07 .37** −.22 3.13 −1.84 .22 .04 .07 .02 Importance of Appearance (online) .07** .26** 3.07 .26 .07 ** 3.11 −.90 .26 .06 .07 .01 Step 1 Step 2 * ** Importance of Appearance (online) Importance of Competency (online) .07 .02 .01 .30 −.09 p < .05. p < .01. hypothesised that the disparity between these bodies would predict body concern and drive for thinness and muscle. We speculated that men’s avatars would combine physical elements of their ideal body as well as elements that are dependent on the online context. However, the question of primary interest was whether the disparity between actual and avatar physical dimensions would explain variance in body concern and drive for thinness/muscle in addition to that explained by the disparity between men’s actual and ideal body. Body image has been shown to have a close association with self-esteem in men suggesting that dissatisfaction with one’s body can lead to lower levels of self-esteem (Davison & McCabe, 2005) as well as other psychosocial outcomes, such as depression (Olivardia et al., 2004). H6. In relation to psychosocial functioning, on the basis of previous research (Bessiere et al., 2007) it was hypothesised that men will exhibit higher levels of self-esteem in an online context compared to offline context self-esteem. In addition to self-esteem, poor body image has also been linked to issues with social functioning amongst men (Davison & McCabe, 2005) as well as greater social anxiety (Cash & Fleming, 2002). H7. However, due to the nature of the Internet, such as the control it provides to users and the anonymity it affords (Galanxhi & Nah, 2007; Suler, 1999), it was further hypothesised that men would demonstrate lower levels of social interaction anxiety and social phobia in the online context, compared to the offline context. H8. Furthermore, it was hypothesised that men with poor social interactions – those with high levels of social interaction anxiety and phobia – would be more likely to embrace online social interactions and spend time online. Method Participants. Study 2 included 131 male adults with a mean age of 33.12 years, ranging from 18 years to 68 years (M = 33.12, SD = 12.06). Participants had a mean BMI of 28.16 (SD = 6.87) ranging from 14.20 to 51.70. The majority of participants were from the United States of America (33.5%), Australia (16.8%), Canada (8.2%) and the United Kingdom (7.2%). Participants were asked to record their Internet use based on frequency. 0.8% of respondents indicated that they spent less than 1 h a day on the Internet; 2.3% between 1 and 2 h a day; 24.4% between 2 and 4 h a day; 45.8% between 6 and 8 h a day; and 26% more than 8 h a day. Measures. The Internet Use Questionnaire (IUQ). The IUQ (Campbell, Cumming, & Hughes, 2006) is divided into two parts and asks participants the activities and applications they use the Internet for, such as emailing, online gaming and gambling. It also asks participants to rank these activities according to use. The second part of the IUQ queries participants in regard to the time they spend using the Internet and the division of time between the activities and applications used. Psychometrics properties are not available for the IUQ as its primary purpose is to collect information regarding to Internet use types and frequency. The Social Interaction Anxiety Scale (SIAS) and Social Phobia Scale (SPS). The SIAS and SPS (Mattick & Clarke, 1998) are 20 item self-report scales which use 5-point Likert measures anchored between “Not at all” (1) and “Extremely” (5) and assess selfreported difficulties in various social contexts. Both scales have high reported internal consistency, validity and test–retest reliability (Orsillo, 2001; Rodebaugh, Woods, Heimberg, Liebowitz, & Schneier, 2006). In the present study participants were asked to complete these measures twice, first in reference to their offline social interactions, then in reference to their social interactions online using their ‘most important’ avatar. The SIAS items yielded a Cronbach’s ˛ of .89 in the offline context, and .87 in the online context, while the SPS yielded a Cronbach’s ˛ of .89 in the offline context and .95 in the online context. The Coopersmith Self-Esteem Inventory Adult Short Form (CSEI). The CSEI is a 25 item questionnaire designed to measure self-esteem in an adult population (Coopersmith, 1989). Participants are asked to rate whether each statement is “Like me” or “Unlike me” in response to statements such as “I wish I were someone else” and “My family understands me”. The CSEI has found to be a reliable and valid measure of self-esteem in the adult population (Francis, 1997). In the present study participants were asked to complete these measures twice, first in reference to the offline context, then in reference to the online context. Cronbach’s ˛ in the offline context was .87 and in the online context it was .79. The Somatomorphic Matrix (Modified). The somatomorphic matrix (modified; Cafri & Thompson, 2004), provides participants with 34 images of men, arranged in a 10 × 10 grid represents 10 incremental levels of adiposity and 10 of muscularity. The adiposity dimension begins at 4% adiposity and increases by increments of 4%; the muscle dimension begins at a fat free muscle mass index of 16.5 kg/m2 and increases by increments of 1.5 kg/m2 (Cafri & Thompson, 2004). Note, however, that test–retest reliability of this measure remains in question (Cafri, Roehrig, & Thompson, 2004), and validity against more objective anthropometric techniques has not been established. However, the somatomorphic matrix has demonstrated good construct validity (Cafri & Thompson, 2004). An example of the 2D matrix, along with superimposed group means for actual, ideal, and avatar bodies is shown in Fig. 1. In the present study participants were asked to move their mouse cursor to the 2D location on the modified somatomorphic matrix which best represented their own body, their ideal body, and their primary avatar’s body. This resulted in a pixel location that represented muscularity (x) and adiposity (y), and allowed comparisons to be made between actual, ideal and online muscularity and adiposity. J.-P. Cacioli, A.J. Mussap / Body Image 11 (2014) 146–155 Fig. 1. Participants’ mean subjective actual, ideal and avatar adiposity and muscle superimposed on the 2D somatomorphic matrix (Cafri & Thompson, 2004) used in Study 2. ±1 standard error bars are included. Drive for Thinness Scale (DTS). The DTS (Garner, Olmstead, & Polivy, 1983) is a subscale from the Eating Disorder Inventory. It consists of seven items that use 6-point Likert scales rated from “Always” (1) to “Never” (5) to measure participant dieting concerns, fear of becoming overweight and preoccupation with weight. Internal consistency in a male sample has been found to be high (Spillane, Boerner, Anderson, & Smith, 2004). Cronbach’s ˛ for the measure in the current study was .82. Drive for Muscularity Scale (DMS). The DMS (McCreary & Sasse, 2000) uses 15 6-point Likert scales anchored between “Always” (1) and “Never” (5) to measure participants’ preoccupation with gaining muscle mass. High internal consistency has been found in a male sample (McCreary & Sasse, 2000). The DMS had a Cronbach’s ˛ of .87 in the current study. Procedure. The procedure and recruitment of participants for Study 2 followed that specified in Study 1 with the exception that advertisements were not placed on Facebook. Results Of the 157 men who completed the questionnaire, 12 were removed due to failing to meet the inclusion criteria of having or using an avatar, and an additional 14 were removed due to having >5% missing values. For the remaining 131 participants, missing values were found to be randomly distributed across items and replaced with multiple imputation using expectation maximisation. Less than .01% data points was missing from the dataset. Five items were removed from the SPS and one from the SIAS due to high rates of non-response in the context of participant avatar use. Variables were created by averaging the summed responses to items. The criteria for normality, in terms of skew and kurtosis, were met by all variables. 151 Body image disparity. A Pearson’s correlation between participants’ BMI and actual body adiposity and muscle from the somatoform matrix was conducted, revealing a significant strong correlation between BMI and actual body adiposity (r = .63, p < .01). The correlation between BMI and actual body muscle was not significant (r = .01, p > .05). To partially test hypothesis H4, that men would have an ideal body that was thinner than their actual body, a repeated measures ANOVA (with Greenhouse–Geisser correction due to mild non-sphericity) revealed that the adiposity of men’s actual, ideal, and “most important” avatar differed significantly, FG-G (1.74, 226.43) = 39.78, p < .001. Paired comparisons confirmed that men’s ideal is thinner than their actual body (342.90 ± 112.32 v 417.76 ± 136.30; p < .001, d = −0.60), but not significantly different from their avatar (320.27 ± 127.69 v 342.90 ± 112.32; p = .08, d = −0.19). Means are shown in Fig. 1, superimposed on the 2D somatomorphic matrix used in the study. A similar repeated measures ANOVA on body muscle was also conducted to examine if the men’s ideal body was more muscular than their actual body, as stipulated by hypothesis H4. The ANOVA revealed that men’s actual, ideal, and avatar bodies also differed significantly in terms of muscle, FG-G (1.49, 193.80) = 28.75, p < .001, with paired comparisons confirming that their ideal is more muscular than their actual body (452.70 ± 160.84 v 551.57 ± 182.85; p < .001, d = −0.57), and also that their avatar is more muscular than their actual or ideal (398.04 ± 247.73 v 452.70 ± 160.84; p < .05, d = −0.26). The increment in muscle from actual, to ideal, to avatar body was found to be quasi-linear, F(1, 130) = 39.42, p < .001, h2 = .23. That is, although the avatar is similar in adiposity to men’s ideal, in terms of muscle it appears to be an extrapolated (i.e., exaggerated) version of their ideal. Body image disparity and drive for thinness/muscle. To test hypothesis H5, in which we predicted that the disparity between men’s actual, ideal and avatar body would lead to unhealthy body change behaviours, body image disparity was calculated in two ways: as the difference between actual versus ideal body (separately for adiposity and muscle), and as the difference between actual and avatar body (again, separately for adiposity and muscle). The question of interest was whether disparity in relation to the avatar’s physical dimensions would predict variance in men’s body change drives and, if so, if this was variance additional to that predicted by disparity in relation to the ideal body. To test this question two-step hierarchical multiple regressions were conducted in which drive for thinness (or drive for muscle) was regressed on ideal-actual adiposity (or muscle) in Step 1, followed by avataractual adiposity (or muscle) in Step 2. Prior to running the multiple regression, an evaluation of the correlation matrix confirmed that the independent variables correlated with the dependent variable and there was no evidence of multicollinearity or singularity amongst the IVs (Tolerance > .1 and VIF < 10). The results, summarised in Table 3, reveal that body image disparity between actual and ideal adiposity and muscle significantly predicted variance in drive for thinness and drive for muscle. As expected the thinner or more muscular one’s ideal body, the higher the drive to lose weight and gain muscle. Furthermore, disparity between actual and avatar adiposity and muscle significantly predicted variance in drive for thinness, but not drive for muscle. That is, the creation of an avatar that is thinner than oneself is associated with the pursuit of thinness in one’s offline life. A third step in the hierarchical regression included moderating for the interaction term of actual body adiposity, or muscle, and avatar adiposity, or muscle. No significant interaction was found. Possible reasons for this, as well as applications of the effect, are explored further in the Discussion. Psychosocial factors relevant to online social interactions. It was hypothesised (H6) that the use of avatars to engage in 152 J.-P. Cacioli, A.J. Mussap / Body Image 11 (2014) 146–155 Table 3 Hierarchical regression analyses of avatar adiposity (or muscle) predicting drive for thinness (or muscle) controlling for participant ideal body adiposity (or muscle) and moderating for the interaction term of real body adiposity (or muscle), and avatar adiposity (or muscle). DV Drive for Thinness Step 1 Step 2 Step 3 Drive for Muscle Step 1 Step 2 Step 3 * ** R2 change ˇ t .05* .21* 2.48 .21 .05 .08* .03* .06 .23* 0.49 2.01 .21 .27 .00 .03 .01 .08 .23* .09 0.68 2.03 1.05 .21 .27 .03 .00 .03 .01 IV R2 Actual-Ideal Adiposity Actual-Ideal Adiposity Actual-Avatar Adiposity sr2 r Actual-Ideal Adiposity Actual-Avatar Adiposity Actual-Ideal Adiposity × Actual-Avatar Adiposity .08 Actual-Ideal Muscle .15* .39** 4.75 .39 .15 .15 .00 .36** .07 3.90 0.71 .39 .23 .10 .00 .01 .37** .07 .07 3.99 0.76 0.88 .39 .23 −.01 .11 .00 .01 Actual-Ideal Muscle Actual-Avatar Muscle Actual-Ideal Muscle Actual-Avatar Muscle Actual-Ideal Muscle × Actual-Avatar Muscle .16 p < .05. p < .01. safe, self-affirming, and/or anonymous social interactions online would be appealing to men who are challenged by offline social interactions. We first tested if self-esteem, social anxiety and social phobia improved online when using avatars compared to the offline context. Paired t-tests confirmed that men reported higher self-esteem (1.67 ± .22 v 1.79 ± .15), t(130) = −7.96, p < .001, d = −0.63, less social interaction anxiety (1.40 ± .81 v 1.02 ± .59), t(130) = 6.63, p < .001, d = 0.54, and less social phobia (.70 ± .67 v .40 ± .47), t(130) = 9.43, p < .001, d = 0.52, while online compared to the offline world, supporting hypothesis H7. In support of the proposition that these differences motivate some men to go online and use avatars, offline world social phobia and anxiety correlated with time spent online (r = −.22, p < .05 and r = −.21, p < .05 respectively), supporting hypothesis H8. Discussion The majority of men surveyed in Study 1 indicated that they created and used multiple avatars within various contexts ranging from games to social simulations and social networking. When asked to respond in relation to the appearance of their most important avatar, three types of avatar were revealed: avatars whose appearance is utilitarian and independent of self; avatars whose appearance reflects actual self; and avatars whose appearance represents an idealised version of self, supporting hypothesis H1. This typology, along with men’s qualitative responses indicated men can create avatars which are similar but enhanced versions of themselves to increase game enjoyment (Axelsson, 2002; Hsu, Lee, & Wu, 2005); they can role-play and influence social interactions online using alternate visual identities (Hussain & Griffiths, 2008; Kafai et al., 2007; Taylor, 2002; Trepte & Reinecke, 2010); and that they can use their avatars to express their ideals, identities, etc., online (Lawson, 2000; McKenna & Bargh, 2000). In terms of the factors that shaped avatar physical dimensions, utilitarian considerations dominated over aesthetic ones, in the online world more so than in the offline world supporting hypothesis H2. This is consistent with the proposition that online game competitiveness is a major influence on avatar creation. As reported by Trepte and Reinecke (2010), avatar users who had high trait competitiveness or played competitive games online were more likely to create avatars with attributes designed to achieve game goals. However, the importance of competency factors online did not preclude the possibility that avatar physical dimensions are relevant to men’s body image in the offline context. In support of hypothesis H3, not only did the importance given by men to their appearance in the offline context correlate with the importance of their avatar’s appearance, their concerns with their own weight and muscle also correlated. This was taken to suggest that an avatar’s physical dimensions reflect its creator’s concerns with their own appearance. To this extent, online visual representations may be of psychosocial relevance to men’s body image and body change drives. In Study 2 we explored this possibility by estimating men’s body image disparity and relating it to their body change drives and their psychological and psychosocial functioning. Not surprisingly, and as predicted in hypothesis H4, the results showed that men’s ideal body is of lower adiposity and greater muscularity than their actual body. This is consistent with previous research which suggests that males desire more muscle and less fat (Demarest & Allen, 2000; Fallon & Rozin, 1985; Tiggemann, Martins, & Kirkbride, 2007). However while actual BMI correlated significantly with participants’ actual body adiposity as derived from the modified somatomorphic matrix, it did not correlate with participants’ actual body muscle. This may be due to inadequate validity of the matrix itself, and/or the fact that the ‘weight’ component of the BMI formula confounds adiposity and muscle and thus serves as a poor proxy for body mass. More surprising was the finding that men’s avatars were not only thinner and more muscular than their actual body, but also more muscular than their ideal body. Is this because men’s avatars are merely a caricature of the male muscular ideal, a functional requirement of the hyper-masculine characters that populate online game environments, or is this of psychosocial significance? The results supported hypothesis H5 and a third and most interesting possibility, at least in one important sense: disparity between men’s actual and avatar body image was a better predictor of their drive for thinness than disparity between actual and ideal body image. Apparently, men who lose weight in an attempt to achieve greater muscle tone in the offline world, manifest this aspiration in terms of the avatars they create and/or use online. We consider this to be the most important implication of our study, and one that is counterintuitive from the perspective of selfdiscrepancy theory, for it suggests that the avatar is a better proxy for what a man aspires to physically than their self-reported ideal! We speculate that this is due to limitations inherent in attempting to quantify the ideal body. Perhaps responses to questions about one’s ideal body are inherently constrained by reality; by one’s knowledge of the range of body shapes that are possible for humans; and ultimately by one’s knowledge of what is feasible for J.-P. Cacioli, A.J. Mussap / Body Image 11 (2014) 146–155 oneself. That is, men’s descriptions of their ideal body may actually be more like an aspirational or goal body than an ideal in the literal sense. Avatars are likely to be less constrained by reality, by what is feasible for humans, and what is feasible for oneself. In this way avatar physical dimensions might serve to better quantify the aspirations underlying men’s body change drives. Avatar physical dimensions might be particularly salient for men whose social interactions online are more rewarding. Study 2 provided some insight into men’s motivations for creating such avatars and going online in the first place. It was found that men’s self-reported self-esteem, social interaction anxiety and social phobia, were all superior online compared to the offline context, supporting hypotheses H6 and H7. Moreover, in support of hypothesis H8, the magnitude of their social anxiety and phobia in the offline context predicted the amount of time they spent online each day. These results suggest that online social interactions, as well as the physical dimensions of avatars used online, serve a social compensatory function at least for some men. In light of these results, we suggest that future research investigate the possibility of including men’s avatar images into discussions of their body image concerns and behaviours. The avatars might provide researchers and clinicians with insights into individual men’s motivations for body change, and with the opportunity to raise with them the issue of unrealistic body image ideals, particularly amongst men who use online interactions to maintain interpersonal distance. Furthermore, unlike other self-report measures of appearance, avatars possess strong ecological validity: they are three-dimensional, dynamic, and most importantly, they are created by men voluntarily. Limitations Although previous research (Taylor, 2003) and our own findings confirmed the importance of utilitarian factors in determining how avatars are created and used, we could not control for the influence of these factors on the results. Most problematic were the limitations imposed on users by the software applications they use to create their avatars. As an example, consider that in the most popular game application sampled in our research – World of WarcraftTM – users are unable to create obese avatars. In addition, we did not examine whether there was a relationship between participants’ application or game choice and their body image, nor the impact of this on the results. This is a limitation of the study because, for example, combat-focussed applications might attract individuals who desire more masculine attributes compared to socially-focussed applications. And the differing foci of these applications may lead to the creation of avatars whose physical dimensions reflect these foci. Our studies also relied entirely on self-report measures of appearance. While practical and easy to implement online, such measures are prone to social desirability (DelPrete, Caldwell, English, Banspach, & Lefebvre, 1992; Gorber, Tremblay, Moher, & Gorber, 2007), and it is unknown to what extent these biases exist in relation to participants’ avatars. Perhaps future research could request actual images of avatars from participants. We focussed exclusively on men. Given the increased take-up of online games by women (they now represent approximately 47% of players, [Entertainment Software Association, 2012]), it is important to extend our research to women and their body image online. As noted in our concluding remarks, we were surprised by the responsiveness of our all-male sample to the open-ended questions of Study 2. This raises suspicions of a self-selection bias whereby men who value their avatars were more likely to agree to participate. 153 The study employed a cross-sectional design and correlational analyses which prevent interpretation of the presence or direction of causality between variables. Additionally, the study used a modified version of the SOQ which has not been validated. Although note that a factor analysis suggested that it was suitable for use in the current study. Additionally the high level of correlation between competency and men’s muscle/weight concerns could potentially be due to the SOQ and the MBAS measuring similar constructs and therefore these results need to be interpreted carefully. Concluding Remarks As the online world becomes more important to people’s work and social lives (Bailenson & Beall, 2006; McKenna & Bargh, 2000), the psychological relevance of online interactions is likely to increase and require closer empirical attention. Viewed in this context, the present study can be taken as an early attempt to explore the potential of avatar physical dimensions as a proxy for underlying body image concerns, particularly amongst men who value their avatar’s appearance and turn to the Internet for relatively ‘safe’ online social interactions. We are encouraged in this endeavour by what we consider to be the most surprising finding of our research: the remarkable willingness of our male participants (with only one exception) to respond thoughtfully and in a self-reflective manner to the open-ended questions in Study 1 about the significance of their avatars. We suggest that these men have invested time and energy in creating their avatars; that their avatars hold special significance to them; and that they are uncharacteristically open to discussing the significance of these avatars because they are not as self-conscious as they would be talking about their own appearance or feelings in person. For these reasons, and given the ecological validity inherent in using a measure of body image that a person has willingly created by themselves and for their own use, we propose that, where relevant, consideration be given to the appearance of men’s avatars in therapeutic discussions of their body image. Acknowledgements The authors would like to thank all the blogs and forums which participated in recruitment for this study, in particular the Metaverse Journal (http://www.metaversejournal.com/). References Arbour, K., & Ginis, K. M. (2006). Effects of exposure to muscular and hypermuscular media images on young men’s muscularity dissatisfaction and body dissatisfaction. Body Image, 3, 153–161. http://dx.doi.org/10.1016/j.bodyim.2006.03. 004 Axelsson, A. (2002). The digital divide: Status differences in virtual environments. In R. Schroeder (Ed.), Social life of avatars: Presence and interaction in shared virtual environments (pp. 188–204). New York: Springer. http://dx.doi.org/10.1007/978-1-4471-0277-9 11 Bailenson, J. N., & Beall, A. C. (2006). Transformed social interaction: Exploring the digital plasticity of avatars. In R. Schroeder & A. Axelsson (Eds.), Avatars at work and play: Collaboration and interaction in shared virtual environments (pp. 1–16). Berlin: Springer-Verlag. http://dx.doi.org/10.1007/1-4020-3898-4 1 Bessiere, K., Seay, A. F., & Kiesler, S. (2007). The ideal elf: Identity exploration in the World of Warcraft. Cyberpsychology & Behavior, 10, 530–535. http://dx.doi.org/10.1089/cpb.2007.9994 Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3, 77–101. http://dx.doi.org/10.1191/1478088706qp063oa Cafri, G., Roehrig, M., & Thompson, J. K. (2004). Reliability assessment of the Somatomorphic Matrix. International Journal of Eating Disorders, 35, 596–600. http://dx.doi.org/10.1002/eat.10270 Cafri, G., & Thompson, J. K. (2004). Measuring male body image: A review of the current methodology. Psychology of Men and Masculinity, 5, 18–29. http://dx.doi.org/10.1037/1524-9220.5.1.18 154 J.-P. Cacioli, A.J. Mussap / Body Image 11 (2014) 146–155 Campbell, A. J., Cumming, S. R., & Hughes, I. (2006). Internet use by the socially fearful: Addiction or therapy? Cyberpsychology & Behavior, 9, 69–81. http://dx.doi.org/10.1089/cpb.2006.9.69 Cash, T. F., & Fleming, E. C. (2002). The impact of body-image experiences: Development of the Body Image Quality of Life Inventory. International Journal of Eating Disorders, 31, 455–460. http://dx.doi.org/10.1002/eat.10033 Childress, M. D., & Braswell, R. (2006). Using massively multiplayer online roleplaying games for online learning. Distance Education, 27, 187–196. http://dx.doi.org/10.1080/01587910600789522 Christopherson, K. M. (2007). The positive and negative implications of anonymity in Internet social interactions: “On the Internet, nobody knows you’re a dog”. Computers in Human Behaviour, 23, 3038–3056. http://dx.doi.org/10.1016/j.chb.2006.09.001 Cole, H., & Griffiths, M. D. (2007). Social interactions in massively multiplayer online role-playing gamers. Cyberpsychology & Behavior, 10, 575–583. http://dx.doi.org/10.1089/cpb.2007.9988 Coopersmith, S. (1989). Manual of self-esteem inventories. Palo Alto, CA: Consulting Psychologists Press. Davison, T. E., & McCabe, M. P. (2005). Relationships between men’s and women’s body image and their psychological, social and sexual functioning. Sex Roles, 52, 463–475. http://dx.doi.org/10.1007/s11199-005-3712-z DelPrete, L. R., Caldwell, M., English, C., Banspach, S. W., & Lefebvre, C. (1992). Self-reported and measured weights and heights of participants in communitybased weight loss programs. Journal of the American Diet Association, 92, 1483–1486. Demarest, J., & Allen, R. (2000). Body Image: Gender, ethnic and age differences. The Journal of Social Psychology, 140, 465–472. http://dx.doi.org/10.1080/00224540009600485 Entertainment Software Association. (2012). Essential facts about the computer and video game industry. Washington, DC: Entertainment Software Association. Retrieved from http://theesa.com/facts/ Fallon, A. E., & Rozin, P. (1985). Sex differences in perceptions of desirable body shape. Journal of Abnormal Psychology, 94, 102–105. http://dx.doi.org/10.1037/0021-843X.94.1.102 Francis, L. (1997). Coopersmith’s model of self-esteem: Bias towards the stable extrovert? The Journal of Social Psychology, 137, 139–141. http://dx.doi.org/10.1080/00224549709595425 Frederick, D. A., Buchanan, G. M., Sadehgi-Azar, L., Peplau, L. A., Haselton, M. G., Berezovaskaya, A., & Lipinski, R. E. (2007). Desiring the muscular ideal: Men’s body satisfaction in the United States, Ukraine, and Ghana. Psychology of Men and Masculinity, 8, 103–117. http://dx.doi.org/10.1037/1524-9220.8.2.103 Frederick, D. A., Fessler, D. M. T., & Haselton, M. G. (2005). Do representations of male muscularity differ in men’s and women’s magazines? Body Image, 2, 81–86. http://dx.doi.org/10.1016/j.bodyim.2004.12.002 Fredrickson, B. L., Roberts, T., Noll, S. M., Quinn, D. M., & Twenge, J. M. (1998). That swimsuit becomes you: Sex differences in self-objectification, restrained eating, and math performance. Journal of Personality and Social Psychology, 75, 269–284. http://dx.doi.org/10.1037/0022-3514.75.1.269 Galanxhi, H., & Nah, F. F. (2007). Deception in cyberspace: A comparison of text-only vs. avatar-supported medium. International Journal of Human-Computer Studies, 65, 770–783. http://dx.doi.org/10.1016/j.ijhcs.2007.04.005 Garner, D. M., Olmstead, M. P., & Polivy, J. (1983). Development and validation of a multidimensional eating disorder inventory for anorexia nervosa and bulimia. International Journal of Eating Disorders, 2, 15–34. http://dx.doi.org/ 10.1002/1098-108X(198321)2:2<15::AID-EAT2260020203>3.0.CO;2-6 Gorber, S. C., Tremblay, M., Moher, D., & Gorber, B. (2007). A comparison of direct vs. self-report measures for assessing height, weight and body mass index: A systematic review. Obesity Reviews, 8, 307–326. http://dx.doi.org/10.1111/j.1467-789X.2007.00347.x Greive, R., & Helmick, A. (2008). The influence of men’s self-objectification on the drive for muscularity: Self-esteem, body satisfaction and muscle dysmorphia. International Journal of Men’s Health, 7, 288–298. http://dx.doi.org/10.3149/jmh.0703.288 Grogan, S. (2008). Body Image. Understanding body dissatisfaction in men, women, and children (2nd ed.). New York: Routledge. Grogan, S., & Richards, H. (2002). Body image: Focus groups with boys and men. Men and Masculinities, 4, 219–232. http://dx.doi.org/10.1177/1097184X02004003001 Hallsworth, L., Wade, T., & Tiggemann, M. (2005). Individual differences in male body-image: An examination of self-objectification in recreational body builders. British Journal of Health Psychology, 10, 453–465. http://dx.doi.org/10.1348/135910705X26966 Hebl, M. R., King, E. B., & Lin, J. (2004). The swimsuit becomes us all: Ethnicity, gender, and vulnerability to self-objectification. Personality and Social Psychology Bulletin, 30, 1322–1331. http://dx.doi.org/10.1177/0146167204264052 Heywood, S., & McCabe, M. P. (2006). Negative affect as a mediator between body dissatisfaction and extreme weight loss and muscle gain behaviors. Journal of Health Psychology, 11, 833–844. http://dx.doi.org/10.1177/1359105306069077 Higgins, E. T. (1987). Self-discrepancy: A theory relating self and affect. Psychological Review, 94, 319–340. http://dx.doi.org/10.1037/0033-295X.94.3.319 Hsu, S. H., Lee, F. L., & Wu, M. C. (2005). Designing action games for appealing to buyers. Cyberpsychology & Behavior, 8, 585–591. http://dx.doi.org/10.1089/cpb.2005.8.585 Hussain, Z., & Griffiths, M. D. (2008). Gender swapping and socializing in cyberspace: An exploratory study. Cyberpsychology & Behavior, 11, 47–53. http://dx.doi.org/10.1089/cpb.2007.0020 Izgiç, F., Akyüz, G., Doğan, O., & Kuğu, N. (2004). Social phobia among university students and its relation to self-esteem and body image. Canadian Journal of Psychiatry, 49, 630–634. Kafai, Y. B., Fields, D. A., & Cook, M. (2007). Your second selves: Resources, agency, and constraints in avatar designs and identity play in a teen virtual world. In Paper presented at the Situated Play, DiGRA 2007 Conference Tokyo. Konijn, A., & Bijvank, M. N. (2009). Door to another me. Identity construction through digital game play. In U. Ritterfeld, M. Cody, & P. Vorderer (Eds.), Serious games. Mechanisms and effects (pp. 179–203). New York: Routledge. Lawson, K. (2000). Beyond corporeality: The virtual self in postmodern times. Journal of Psychological Practice, 6, 35–43. Martey, R. M., & Stromer-Galley, J. (2007). The digital dollhouse: Context and social norms in The Sims. Games and Culture, 2, 314–334. http://dx.doi.org/10.1177/1555412007309583 Mattick, R. P., & Clarke, J. C. (1998). Development and validation of measures of social phobia scrutiny fear and social interaction anxiety. Behavior Research and Therapy, 36, 455–470. http://dx.doi.org/10.1016/S0005-7967(97)10031-6 McCreary, D. R., & Sasse, D. K. (2000). An exploration of the drive for muscularity in adolescent boys and girls. Journal of American College Health, 48, 297–304. http://dx.doi.org/10.1080/07448480009596271 McKenna, K. Y. A., & Bargh, J. A. (2000). Plan 9 from cyberspace: The implications of the Internet for personality and social psychology. Personality and Social Psychology Review, 4, 57–75. http://dx.doi.org/10.1207/S15327957PSPR0401 6 Miranda, L. C. M., & Lima, C. A. S. (2012). Trends and cycles of the Internet evolution and worldwide impacts. Technological Forecasting and Social Change, 79, 744–765. http://dx.doi.org/10.1016/j.techfore.2011.09.001 Morry, M. M., & Staska, S. L. (2001). Magazine exposure: Internalization, selfobjectification, eating attitudes, and body satisfaction in male and female university students. Canadian Journal of Behavioural Science, 33, 269–279. http://dx.doi.org/10.1037/h0087148 Nezlek, J. (1999). Body image and day-to-day social interaction. Journal of Personality, 67, 793–817. http://dx.doi.org/10.1111/1467-6494.00073 Noll, S. M., & Fredrickson, B. L. (1998). A mediational model linking selfobjectification, body shame, and disordered eating. Psychology of Women Quarterly, 22, 623–636. http://dx.doi.org/10.1111/j.1471-6402.1998.tb00181.x Oehlhof, M. E. W., Musher-Eizenman, D. R., Neufeld, J. M., & Hauser, J. C. (2009). Self-objectification and ideal body shape for men and women. Body Image, 6, 308–310. http://dx.doi.org/10.1016/j.bodyim.2009.05.002 Olivardia, R., Pope, H. G., Jr., Borowiecki, J. J., & Cohane, G. H. (2004). Biceps and body image: The relationship between muscularity and self-esteem, depression, and eating disorder symptoms. Psychology of Men and Masculinity, 5, 112–120. http://dx.doi.org/10.1037/1524-9220.5.2.112 Orsillo, S. M. (2001). Measures for social phobia. In M. M. Antony, S. M. Orsillo, & L. Roemer (Eds.), Practitioner’s guide to empirically based measures of anxiety (pp. 165–187). New York, NY: Springer. http://dx.doi.org/10.1007/0-306-47628-2 14 Pope, H. G., Phillips, K. A., & Olivardia, R. (2000). The Adonis complex: The secret crisis of male body obsession. New York: The Free Press. Ricciardelli, L., & McCabe, M. (2004). A biopsychosocial model of disordered eating and the pursuit of muscularity in adolescent boys. Psychological Bulletin, 130, 179–205. http://dx.doi.org/10.1037/0033-2909.130.2.179 Rodebaugh, T. L., Woods, C. M., Heimberg, R. G., Liebowitz, M. R., & Schneier, F. R. (2006). The factor structure and screening utility of the social interaction anxiety scale. Psychological Assessment, 18, 231–237. http://dx.doi.org/10.1037/1040-3590.18.2.231 Schwartz, J. P., & Tylka, T. L. (2008). Exploring entitlement as a moderator and mediator of the relationship between masculine gender role conflict and men’s body esteem. Psychology of Men and Masculinity, 9, 67–81. http://dx.doi.org/10.1037/1524-9220.9.2.67 Spillane, N. S., Boerner, L. M., Anderson, K. G., & Smith, G. T. (2004). Comparability of the Eating Disorder Inventory-2 between women and men. Assessment, 11, 85–93. http://dx.doi.org/10.1177/1073191103260623 Strelan, P., & Hargreaves, D. (2005). Reasons for exercise and body esteem: Men’s response to self-objectification. Sex Roles, 53, 495–503. http://dx.doi.org/10.1007/s11199-005-7137-5 Suler, J. (1999, April). Gender switching in cyberspace. The psychology of cyberspace. Retrieved from http://www.rider.edu/∼suler/psycyber/genderswap.html Taylor, T. L. (2002). Living digitally: Embodiment in virtual worlds. In R. Schroeder (Ed.), The social life of avatars: Presence and interaction in shared virtual environments (pp. 40–62). London: Springer-Verlag. http://dx.doi.org/10.1007/978-1-4471-0277-9 3 Taylor, T. L. (2003). Intentional bodies: Virtual environments and the designers who shape them. International Journal of Engineering Education, 19, 25–34. Thompson, J. K., Heinberg, L. J., Altabe, M., & Tantleff-Dunn, S. (1999). Exacting beauty: Theory, assessment, and treatment of body image disturbance. Washington, DC: American Psychological Association. Tiggemann, M., & Kuring, J. K. (2004). The role of body objectification in disordered eating and depressed mood. British Journal of Clinical Psychology, 43, 299–311. http://dx.doi.org/10.1348/0144665031752925 Tiggemann, M., Martins, Y., & Kirkbride, A. (2007). Oh to be lean and muscular: Body image ideals in gay and heterosexual men. Psychology of Men and Masculinity, 8, 15–24. http://dx.doi.org/10.1037/1524-9220.8.1.15 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 Psychology, 22, 171–184. http://dx.doi.org/10.1027/1864-1105/a000022 J.-P. Cacioli, A.J. Mussap / Body Image 11 (2014) 146–155 Tylka, T. L., Bergeron, D., & Schwartz, J. P. (2005). Development and psychometric evaluation of the Male Body Attitudes Scale (MBAS). Body Image, 2, 161–175. http://dx.doi.org/10.1016/j.bodyim.2005.03.001 Vasalou, A., & Joinson, A. N. (2009). Me, myself and I: The role of interactional context on self-presentation through avatars. Computers in Human Behaviour, 25, 510–520. http://dx.doi.org/10.1016/j.chb.2008.11.007 Wade, J. T. (2000). Evolutionary theory and self-perception: Sex differences in body esteem predictors of self-perceived physical and sexual attractiveness and self-esteem. International Journal of Psychology, 35, 36–45. http://dx.doi.org/10.1080/002075900399501 155 Walther, J. B. (1996). Computer-mediated communication: Impersonal, interpersonal and hyperpersonal interaction. Communication Research, 39, 274–279. http://dx.doi.org/10.1177/009365096023001001 Walther, J. B., Slovacek, C., & Tidwell, L. (2001). Is a picture worth a thousand words? Photographic images in long-term and short-term computer-mediated communication. Communication Research, 28, 105–134. http://dx.doi.org/10.1177/009365001028001004 Whitty, M. T. (2008). Liberating or debilitating? An examination of romantic relationships, sexual relationships and friendships on the Net. Computers in Human Behaviour, 24, 1837–1850. http://dx.doi.org/10.1016/j.chb.2008.02.009