Using a virtual reality role-playing game to enhance interaction dynamics and improve social attitudes toward transgender people.

Using a virtual reality role-playing game to enhance interaction dynamics and improve social attitudes toward transgender people - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Search Log in Primary site navigation Search Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice Sci Rep . 2026 Apr 10;16:16943. doi: 10.1038/s41598-026-46402-3 Using a virtual reality role-playing game to enhance interaction dynamics and improve social attitudes toward transgender people Cassandra L Crone 1, ✉ , Rachel W Kallen 1 , Michael J Richardson 1 Author information Article notes Copyright and License information 1 School of Psychological Sciences, Faculty of Medicine, Health and Human Sciences, Macquarie University, Sydney, NSW Australia ✉ Corresponding author. Received 2025 Jul 1; Accepted 2026 Mar 25; Collection date 2026. © The Author(s) 2026 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . PMC Copyright notice PMCID: PMC13230618  PMID: 41963407 Abstract While virtual reality (VR) interventions have demonstrated efficacy in reducing intergroup prejudice, they have yet to be applied to transprejudice. Accordingly, we developed a VR role-playing game (RPG) that integrates perspective-taking and intergroup-contact for transgender identities. We evaluated the RPG’s impact on cisgender individuals’ movement dynamics and social attitudes toward transgender people using a mixed experimental design, in which 162 participants embodied a self-customised transgender avatar and interacted with a cisgender or transgender virtual colleague. Empathy increased throughout the interaction, consistent with mediating the relationship between virtual embodiment and post-session gender essentialism, but only those partnered with a transgender colleague reported increased within-session interpersonal affiliation. Aligning with shifts in empathy, participants’ postural dynamics became less constrained and more complex over time, although women adapted more readily to the RPG than men. Taken together, our findings highlight the potential value of simulated interactions that promote rapport-building between cisgender and transgender individuals. Supplementary Information The online version contains supplementary material available at 10.1038/s41598-026-46402-3. Subject terms: Health care, Psychology, Psychology Introduction Against the backdrop of an evolving socio-political landscape, rising backlash to gender equality, and narratives that debate the legitimacy of transgender and gender diverse (TGD) identities, there is an increasingly urgent need to challenge the pervasiveness of gender essentialism (i.e., beliefs that gender is binary, discrete, immutable, and biologically based) 1 . Such beliefs contribute to gender inequities in education, healthcare, and the workplace 2 – 4 for both TGD people and cisgender women alike. Moreover, those who endorse gender essentialist beliefs are also more likely to favour cisgender individuals, enact delegitimising microaggressions, engage in gender identity denial, and stereotype TGD people based on their assigned sex at birth 5 . To reduce transprejudice within everyday interactions, there is a critical demand for interventions that go beyond traditional approaches. Recent research has found that virtual reality (VR) interventions—namely, VR perspective taking (VRPT) and VR-mediated contact—can effectively reduce intergroup prejudices 6 . The former, VRPT, adopts an individual strategy, whereby participants virtually embody an avatar representing a marginalised outgroup identity to simulate experiencing the world from another’s perspective 7 . The latter, VR-mediated contact, adopts an intergroup strategy, whereby participants synchronously interact with an outgroup member to simulate positive, direct intergroup contact 8 . Thus far, no research has applied either approach to TGD identities. In the current study, we investigate the impact of transgender VRPT and VR-mediated cisgender-transgender contact on gender essentialism, intergroup empathy, and interpersonal affiliation. Despite well-documented evidence of anti-transgender stigma in society, there is limited research on cisgender-transgender contact, overall. Much of the current body of knowledge relies on cross-sectional, survey-based, and correlational evidence, suggesting that self-reported contact with TGD people is associated with reduced levels of transphobia and transprejudice 9 . Some experimental paradigms have echoed these findings using imagined contact 10 , parasocial, vicarious contact 11 , and direct contact via teaching interventions 12 , finding that some contact with transgender outgroup targets is better than no contact at all for cisgender individuals. Although interventions based on direct contact experiences are expected to exhibit the strongest effects 13 , in-person encounters between cisgender and transgender people are limited and face several barriers (e.g., sparse minority population, intergroup anxiety and avoidance, potential for harm) 14 . Thus, we highlight a need to embrace more immersive simulations of contact when developing interventions aimed at reducing transprejudice among cisgender individuals. Only one study, to date, has employed an immersive, real-time simulation of contact, finding that online, text-based interaction with a transgender woman reduced transgender stigma, but only for cisgender men 15 . Text-based contact, however, lacks the visual and embodied cues that are present during real-world interpersonal interactions. Recent research has demonstrated that fully immersive interactions undertaken in VR result in greater empathy, prosocial behaviour, and prejudice reduction toward marginalised outgroup targets, in comparison to such simulations undertaken online 16 , 17 . Notably, this relies on embodied perspective taking, whereby participants experience a sense of ownership and agency over a virtual avatar representing a marginalised outgroup identity (i.e., VRPT) 18 , 19 . Mere exposure to such embodiment, however, is insufficient to induce attitudinal or behavioural changes 20 , 21 . Rather, successful VRPT requires that the content of the interaction be ecologically valid, more closely approximating real-world contexts in which intergroup prejudices arise 22 . Taken together, VR-mediated contact and VRPT may serve to improve cisgender-transgender relations by enhancing the affective (e.g., empathy) and cognitive (e.g., knowledge acquisition, familiarity) mechanisms associated with prejudice reduction 23 . Empathy, in particular, is thought to underlie the relationship between virtual immersion and positive change 24 . Thus, we designed a VR role-playing game (RPG), in which the narrative progressively increased rapport-building between cisgender participants and either a transgender or cisgender interaction partner. Not only does such an approach afford an ecologically valid simulation of positive contact during VRPT, but also it provides a novel method of capturing subtle information that an individual might reveal through non-verbal behaviours, such as postural activity, in real time. Postural activity, or more specifically postural sway , refers to the subtle shifts or fluctuations in movement that persist when standing upright, which are functionally critical to exploring the information in one’s environment 25 . This variability in postural sway systematically fluctuates over time, exhibiting meaningful structure that allows individuals to adapt to varying personal and task-relevant constraints 26 , 27 . An individual’s postural activity, thus, provides an embodied measure of the codependent, non-verbal, and linguistic processes that characterise direct, face-to-face social interactions 28 , 29 . Changes in situational constraints, as well as an individual’s intentional state, can significantly influence the structural complexity of postural fluctuations 30 , 31 . In the current study, we examine the dynamics of postural behaviours exhibited throughout the RPG simulation to explore how VR-mediated contact and VRPT are embodied. Given that such behaviour is not a static representation of mental states, but rather, an emergent property of the complex interactions between brain, body, and environment 32 , this approach can reveal which situational events, in particular, might contribute to adaptation as a result of undertaking the RPG. We expected that exposure to the RPG simulation would be associated with increased empathy and interpersonal affiliation with transgender outgroup targets, alongside lower endorsement of gender essentialist beliefs. Additionally, we hypothesised that empathy would play a mediating role in the relationships between virtual embodiment and these outcomes. However, no research, to date, has evaluated both VR-mediated contact and VRPT in the context of transprejudice among cisgender people. Using a mixed experimental design, we investigated the impact of our virtual RPG on the self-reported attitudes (i.e., gender essentialism, empathy, affiliation) and non-verbal behaviours (i.e., postural sway) of cisgender individuals, who embodied a self-customised transgender avatar while interacting with either a transgender or cisgender colleague. Results We designed the RPG as an after-work team-building event set in an arcade-style gaming centre to serve as a model of everyday interaction in which transprejudice may emerge. All participants undertook VRPT by embodying a transgender avatar within a structured series of scripted, rapport-building, narrative scenarios, while those who interacted with a transgender colleague additionally experienced VR-mediated contact (see Fig. 1 a). Postural activity timeseries were recorded across six discrete phases of the RPG (displayed in Fig. 1 b) and assessed within-subjects using detrended fluctuation analysis (DFA). We then used a generalised linear mixed model (GLMM) to assess whether changes in postural activity observed throughout the RPG differed based on the colleague’s gender (i.e., VR-mediated contact) as well as between women and men participants. Next, we examined differences by colleague and participant gender for all self-report measures, as recorded within- (i.e., state empathy, affiliation) and between-subjects (i.e., trait empathy, social desirability, embodiment, gender essentialism). Finally, we used a series of mediation analyses to assess indirect pathways from perceived virtual embodiment to post-session gender essentialism and within-session change in interpersonal affiliation via state empathy toward the virtual colleague. Further details on the research design are outlined in the Methods section. Fig. 1. Open in a new tab ( a ) Participants were randomly assigned to interact with either a transgender or cisgender virtual avatar, such that women were paired with either a cisgender woman or transgender man (left), whereas men were paired with a cisgender man or transgender woman (right). ( b ) Progression through the RPG narrative occurred across six discrete phases, as displayed here. Postural activity was recorded across each phase, while state empathy was recorded following the first, second, and third games, and following the debrief. Figure 1 a and b were composed by the authors using screenshots of the virtual environment, also developed by the authors and hosted in Unity (version 2020.3.42f1). Postural movement dynamics reveal reduced behavioural constraint over time We used DFA to quantify the degree of self-similarity within participants’ postural sway timeseries across the six stages of the RPG (see Methods for a detailed description of data processing). DFA determines the average magnitude of residual variance across a range of window sizes within a given timeseries (i.e., segments of 8, 16, 32, 64, 128, 256, … data points) and then plots the average residual variance estimates (i.e., root mean square) as a function of window size in log–log form 32 . The slope of the regression line fitting this plot is denoted as α (alpha), which provides an estimate of the fractal, or self-similar structure within participants’ postural sway timeseries. Namely, α characterises the stochastic nature of this fractal structure, such that α values closer to 0.5 indicate more random, non-correlated, highly constrained behavioural variation (i.e., white noise); α values closer to 1.0 indicate moderately persistent, long-range correlated, more flexible, and less constrained behaviour (i.e., pink noise); and α values closer to 1.5 indicate highly persistent, overly controlled behaviour (i.e., brown noise) 33 , 34 . Of particular interest here, was the degree to which participants’ postural activity became more persistent or less constrained throughout the RPG (i.e., within-subjects), and whether these patterns of change differed by participant and colleague gender (i.e., between-subjects). Thus, we employed a 2 (between subjects; participant gender: woman, man) × 2 (between-subjects; colleague gender: transgender, cisgender) × 6 (within-subjects; phase) GLMM with α as the dependent variable and an unstructured covariance matrix. A sequential Bonferroni correction was applied to account for multiple comparisons. As described in Table 1 , results of the GLMM suggested no significant effects for colleague gender, nor for the two- and three-way interactions involving this factor. However, we observed a two-way interaction between phase and participant gender, as displayed in Fig. 2 . Although, α increased from baseline (Phase 1) to the conclusion of the RPG (Phase 6), and women exhibited higher values of α overall, women were significantly higher than men at Phase 2, Phase 4, and Phase 5. Table 1. GLMM Fixed Effects for the Overall Model. F df1 df2 p Corrected Model 32.76 23 217 < .001** Colleague 0.45 1 148 .502 Gender 11.36 1 148 < .001** Phase 128.71 5 145 < .001** Colleague x Gender 0.03 1 148 .860 Colleague x Phase 0.95 5 145 .453 Gender x Phase 2,52 5 145 .032* Colleague x Gender x Phase 1.29 5 145 .270 Open in a new tab ‘Colleague’ denotes the categorical variable colleague gender (cisgender = 0, transgender = 1), while ‘Gender’ denotes the categorical variable participant gender (women = 0, men = 1). * p < .05. ** p < .001. Fig. 2. Open in a new tab Change in Postural Activity across Phases by Participant Gender. The two-way interaction between phase and participant gender demonstrates differences between women and men over time. Overall, women exhibited higher levels of self-similarity as measured by α . However, all participants attained a statistically equivalent state of reduced constraint ( α closer to 1.0) at end-state compared to baseline ( α closer to 0.5). Additionally, women and men displayed similar patterns of change within-subjects across phases, with two exceptions. From Phase 2 to Phase 3, men’s α significantly increased, while women’s α showed no significant change. Conversely, from Phase 5 to Phase 6, women’s α significantly decreased, while men’s α showed no significant change. Pairwise comparisons supporting these effects are summarised in Table 2 . Taken together, these results indicate that both women and men attain a state of reduced constraint on their behaviour over Table 2. GLMM Pairwise Comparisons by Phase and Participant Gender. Between-subjects Phase Estimate SE t df Lower CI Upper CI 1 0.01 0.01 1.71 157 −0.01 0.03 2** 0.04 0.01 4.46 156 0.02 0.06 3 0.02 0.01 1.55 153 −0.01 0.03 4* 0.03 0.01 2.82 156 0.01 0.05 5* 0.03 0.01 2.56 157 0.01 0.05 6 0.01 0.01 0.16 122 −0.02 0.02 Estimate SE t df Lower CI Upper CI Within-subjects Women 1 vs 2** 0.10 0.01 12.28 137 0.08 0.11 2 vs 3 0.01 0.01 0.15 139 −0.02 0.02 3 vs 4** 0.04 0.01 4.44 105 0.02 0.06 4 vs 5 −0.02 0.01 −1.87 128 −0.04 0.01 5 vs 6* −0.03 0.01 −2.86 102 −0.05 −0.01 Men 1 vs 2** 0.07 0.01 9.43 151 0.05 0.08 2 vs 3* 0.03 0.01 2.92 150 0.01 0.04 3 vs 4* 0.03 0.01 3.06 185 0.01 0.04 4 vs 5 −0.01 0.01 −1.46 155 −0.03 0.01 5 vs 6 0.01 0.01 0.28 201 −0.01 0.02 Open in a new tab A significant two-way interaction between phase and participant gender was observed. Between-subjects effects summarise comparisons of α between women and men for each phase. Within-subjects effects summarise the change in α between each phase, in series, for women and men. CI = confidence interval. * p < .05. ** p < .001. the course of the interaction (i.e., exhibited more relaxed, less overtly controlled postural sway activity). However, women reach this state earlier in the RPG compared to men, implying that women adapted more readily to the interaction context. Such adaptation reflects behaviour that is not random or reactive from moment to moment 32 , 34 . Rather, emerging behaviour is shaped by an increasingly stable internal process, in which more coherent patterns of control support smoother and more responsive engagement with the environment or task. Colleague and participant gender differentially influence empathy and affiliation Prior to mediation analyses, we evaluated group means and variances among all self-report measures to describe the fundamental relationships within our data. Differences by participant and colleague gender for independent and repeated measures are presented in Tables 3 and 4 , respectively. Regarding between-subjects effects, we observed that women reported significantly higher levels of perceived avatar embodiment as well as higher affective, but not cognitive, trait empathy compared to men. Conversely, men endorsed gender essentialist attitudes and beliefs to a greater extent compared to women. However, social desirability did not differ based on participant or colleague gender. Table 3. Independent Measures Statistics by Colleague and Participant Gender. Women Participants Men Participants Participant Gender Colleague Gender Interaction Trans Colleague Cis Colleague Trans Colleague Cis Colleague F (1, 158) (η p 2 ) M ( SD ) QCAE-COG 62.15 (5.82) 61.10 (6.97) 59.43 (7.63) 59.85 (6.67) 3.44 (.021) 0.09 (.001) 0.47 (.003) QCAE-AFF 25.75 (3.84) 25.30 (3.41) 22.86 (3.52) 21.98 (3.72) 29.75** (.158) 1.37 (.009) 0.14 (.001) MCSDS 20.75 (3.14) 20.93 (3.22) 19.74 (2.56) 20.88 (3.30) 1.22 (.008) 1.86 (.012) 1.00 (.006) EQ 4.17 (0.55) 4.24 (0.68) 3.88 (0.73) 3.88 (0.78) 8.99* (.054) 0.13 (.001) 0.09 (.001) HABS-ESG 2.94 (1.45) 2.83 (1.32) 4.05 (1.60) 3.83 (1.44) 21.03** (.117) 0.51 (.003) 0.05 (.000) Open in a new tab QCAE = Questionnaire of Cognitive and Affective Empathy (COG = cognitive empathy dimension; AFF = affective empathy dimension). MCSDS = Marlowe Crowne Social Desirability Scale (social desirability). EQ = Embodiment Questionnaire (virtual embodiment). HABS-ESG = Heteronormative Attitudes and Beliefs Scale–Essential Sex and Gender (gender essentialism). Partial eta squared (η p 2 ) is reported as a measure of effect size for between-subjects comparisons using two-way ANOVA. * p < .05. ** p < .001. Table 4. Mixed-Design ANOVA for Repeated Measures. Empathy Within-subjects F (3, 474) η p 2 EM Means Empathy 230.73** .594 M SE 95% CI Empathy*Gender 1.65 .010 Within Empathy*Colleague 0.68 .004 Time 1 5.80 (0.13) 5.55, 6.04 Empathy*Gender*Colleague 0.85 .005 Time 2 5.85 (0.12) 5.61, 6.09 Between-subjects F (1, 158) Time 3 6.57 (0.13) 6.21, 6.73 Gender 19.09** .108 Time 4 7.76 (0.13) 7.51, 8.01 Colleague 0.08 .001 Gender*Colleague 0.13 .001 Between Pairwise Contrasts F (1, 158) Women 6.97 0.16 6.65, 7.29 1 vs 2 0.67 .004 Men 5.97 0.16 5.65, 6.29 2 vs 3 83.20** .345 3 vs 4 198.41** .557 Likeability Within-subjects F (1, 158) η p 2 EM Means Likeability 189.61** .545 M SE 95% CI Likeability*Gender 0.21 .001 Within Likeability*Colleague 0.69 .004 Pre 5.61 0.11 5.40, 5.83 Likeability*Gender*Colleague 1.74 .011 Post 7.22 0.13 6.97, 7.48 Between-subjects F (1, 158) Gender 23.04** .127 Between Colleague 0.67 .004 Women 6.91 0.15 6.62, 7.20 Gender*Colleague 0.48 .003 Men 5.92 0.15 5.63, 6.21 Similarity Within-subjects F (1, 158) η p 2 EM Means Similarity 53.15** .252 M SE 95% CI Similarity*Gender 0.01 .000 Within Similarity*Colleague 49.85** .240 Trans (pre) 3.08 0.19 2.71, 3.46 Similarity*Gender*Colleague 0.02 .000 Cis (pre) 5.51 0.19 5.13, 5.89 Between-subjects F (1, 158) Trans (post) 5.42 0.22 4.99, 5.86 Gender 5.89* .036 Cis (post) 5.56 0.22 5.11, 5.99 Colleague 27.57** .149 Gender*Colleague 0.40 .003 Between Women 5.19 .17 4.85, 5.53 Men 4.60 .17 4.26, 4.94 Open in a new tab Gender = participant gender (women = 0, men = 1). Colleague = colleague gender (cisgender = 0, transgender = 1). State empathy is measured across for timepoints (Time 1, Time 2, Time 3, Time 4), while likeability and similarity are measured pre- and post-RPG. Estimated marginal (EM) means are reported for significant effects within- and between-subjects. Partial eta squared (η p 2 ) is reported as a measure of effect size. ** p < .05. ** p < .001. Regarding repeated measures, empathy toward the colleague was recorded at four timepoints within the RPG, while interpersonal affiliation (i.e., likability, similarity) with the colleague was reported pre- and post-RPG (see Methods for further detail). Using a series of mixed-design ANOVAs, we observed that both empathy and affiliation significantly increased over time, and women provided higher ratings for empathy, likeability, and similarity compared to men. Further examination of the pairwise contrasts revealed that empathy increased from Time 2 to Time 3 (following discussions about workplace equity), and Time 3 to Time 4 (following the debrief), although there was no significant change between Time 1 and Time 2. Despite the presence of these main effects, the two- and three-way interactions involving participant gender were not significant. Similarly, neither state empathy nor likeability differed based on the colleague’s gender. However, the model evaluating similarity demonstrated an effect of colleague gender on perceptions of interpersonal similarity, such that only those who interacted with a transgender colleague reported increased feelings of similarity post-RPG. Specifically, participants reported lower feelings of similarity with a transgender colleague compared to a cisgender colleague before contact, whereas after contact, participants reported equivalent levels of similarity with either colleague. Empathy as a pathway from embodiment to gender essentialism and affiliation To test whether virtual embodiment was associated with post-session gender essentialism and within-session changes in interpersonal affiliation via state empathy, we first examined zero-order correlations among all self-report measures. Given that postural dynamics (indexed via DFA across phases of the simulation) were measured prior to the post-RPG self-reports, we also examined zero-order correlations among these variables (see Table 5 ). Correlations between virtual embodiment, state empathy, and the outcomes of interest (i.e., gender essentialism, likeability, similarity) indicated suitability for mediation analyses (moderate-to-large correlations). Further, the correlations among state empathy measured across four timepoints supported use of serial mediation (i.e., correlations > 0.70) 35 . Given that preceding results demonstrated differences by participant and colleague gender, we initially used Model 92 from Hayes 35 PROCESS macro (version 4.2) to assess for moderated mediation in the indirect pathways from embodiment (X) to gender essentialism (Y 1 ), likeability (Y 2 ), and similarity (Y 3 ) via empathy (M 1-4 ). We further included trait cognitive and affective empathy as covariates in each model to assess whether context-specific empathy was associated with outcomes over and above one’s dispositional and stable empathetic tendencies. Finally, as social desirability and DFA (Phases 4–6) were correlated with the outcomes of interest, these variables were examined as potential covariates. However, only DFA (Phase 6) remained statistically significant in the model for gender essentialism, and therefore the remaining variables were excluded from final analyses for the sake of model parsimony. Table 5. Zero-order Pearson Correlation Coefficients. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 1. QCAE-COG – 2. QCAE-AFF .08 – 3. MCSDS .12 −.03 – 4. EQ .02 .25* .07 – 5. HABS-ESG −.06 −.42* .10 −.26* – 6. Likeability pre .07 .25* .19* .22* −.32* – 7. Likeability post .14 .32* .07 .46* −.43* .57* – 8. Similarity pre .05 .09 .05 .19* −.17* .44* .39* – 9. Similarity post .15 .20* .08 .37* −.28* .39* .61* .35* – 10. Empathy 1 .13 .32* .15 .30* −.30* .52* .62* .32* .44* – 11. Empathy 2 .05 .29* .10 .35* −.35* .48* .60* .28* .43* .87* – 12. Empathy 3 .11 .35* .16* .35* −.37* .53* .64* .29* .46* .81* .87* – 13. Empathy 4 .22* .40* .16* .36* −.43* .55* .75* .30* .48* .72* .73* .77* – 14. DFA 1 .01 .04 −.01 .07 −.04 .07 .05 .17* .02 −.02 −.01 .02 .03 – 15. DFA 2 .06 .15 .18* .08 −.05 .09 .01 .06 .07 .01 −.09 −.07 .02 .23* – 16. DFA 3 −.08 .03 .05 −.03 0.01 .05 .05 .01 .05 .09 .01 .02 .01 .25* .17* – 17. DFA 4 −.06 .18* .00 .16* −.15 .02 .09 .08 −.02 .07 .06 .03 .09 .27* .28* .32* – 18. DFA 5 −.05 .22* −.01 .08 −.05 .11 .09 −.03 −.01 .18* .18* .15 .15 .23* .27* .32* .34* – 19. DFA 6 −.03 −.08 −.07 .00 .17* −.04 −.03 −.01 .02 −.02 −.02 −.04 .03 .22* .13 .37* .21* .36* – Open in a new tab DFA = detrended fluctuation analysis. * p < .05. Cognitive and affective empathy were included as covariates in all models. Shading denotes correlations between the key predicting, mediating, and outcome variables evaluated. Small, positive, significant correlations with these key variables (underlined, bold) were examined as potential covariates in the mediation models. As neither social desirability nor postural activity at Phases 4 and 5 were significantly associated with the mediators (state empathy) or outcome variables (gender essentialism, likeability post , similarity post ), these variables were excluded from the final mediation models. Postural activity at Phase 6 was retained only in the model assessing gender essentialism. Virtual embodiment and gender essentialism via state empathy over time Although DFA was not conceptualised as an outcome of the mediation models, exploratory correlations and model testing indicated that end-state DFA (Phase 6) was significantly associated with post-session gender essentialism. Thus, we included this factor as an additional covariate here to assess the robustness of the mediation model to end-state postural dynamics. As analyses with Model 92 suggested no effects consistent with moderated mediation, we used Model 6 for the final analysis while retaining participant and colleague gender as covariates. Figure 3 summarises the serial mediation model where virtual embodiment is negatively associated with endorsement of post-session gender essentialism via state empathy ( F (10,151) = 7.30, p < 0.001, R 2 = 0.33). Fig. 3. Open in a new tab Serial Mediation Model of Gender Essentialism. Increases in empathy over time were consistent with serial mediation for the relationship between virtual embodiment and post-session gender essentialism, with participant and colleague gender, cognitive and affective empathy, and postural activity ( α ) at Phase 6 included as covariates. The total effect is displayed below the horizontal line, while the direct effect is displayed above the line. Despite the previously observed zero-order relationships between virtual embodiment and gender essentialism, neither the direct nor total effects were significant when accounting for covariates and mediators. However, we observed significant indirect effects for the pathway from embodiment to gender essentialism via state empathy at all timepoints, as well as the pathway excluding Time 1 empathy. Given the minimal changes in empathy from Time 1 to Time 2, this latter effect is not overly surprising. Regarding covariates, affective empathy was positively associated with state empathy at Time 1 and negatively associated with gender essentialism, while cognitive empathy was positively associated with state empathy at Time 4. These covariate associations not only indicate that the affective and cognitive dimensions of trait empathy are differentially related to state empathy across the RPG, but also they support the independence of the association between state empathy and gender essentialism when controlling for empathetic disposition. Additionally, end-state DFA (Phase 6) remained negatively associated with gender essentialism, suggesting that end-state postural dynamics explained variance in gender essentialism over and above embodiment and empathy. The indirect effects and covariate statistics supporting these findings are summarised in Table 6 . Taken together, these effects suggest complexity in the relationships between embodiment, empathy, and gender essentialism. Neither the sense of embodiment nor empathy alone were associated with the endorsement of gender essentialist attitudes and beliefs after the RPG. Instead, embodiment is first associated with empathy, which in turn increases throughout the RPG, where end-state empathy, in turn, is negatively associated with gender essentialism. Table 6. Indirect Effects and Covariates for Model 6 Predicting Gender Essentialism. Indirect Effect Coefficient ( SE ) 95% CI Effect Size ( SE ) 95% CI Total* −0.18 (0.08) −0.35,  −0.04 −0.08 (0.04) −0.16,  −0.12 a 1 b 1 0.09 (0.07) −0.03, 0.26 0.04 (0.03) −0.01, 0.12 a 2 b 2 −0.03 (0.04) −0.12, 0.03 −0.02 (0.02) −0.06, 0.01 a 3 b 3 −0.01 (0.02) −0.04, 0.03 −0.01 (0.01) −0.02, 0.01 a 4 b 4 −0.05 (0.05) −0.16, 0.03 −0.02 (0.02) −0.07, 0.01 a 1 d 1 b 2 −0.06 (0.07) −0.21, 0.07 −0.03 (0.03) −0.10, 0.03 a 1 d 4 b 3 −0.01 (0.01) −0.03, 0.03 −0.01 (0.01) −0.01, 0.02 a 1 d 6 b 4 −0.03 (0.02) −0.08, 0.00 −0.01 (0.01) −0.04, 0.00 a 2 d 2 b 3 −0.01 (0.02) −0.05, 0.05 −0.01 (0.01) −0.02, 0.02 a 2 d 5 b 4 −0.01 (0.01) −0.03, 0.01 −0.01 (0.01) −0.01, 0.01 a 3 d 3 b 4 −0.01 (0.01) −0.04, 0.01 −0.01 (0.01) −0.02, 0.01 a 1 d 1 d 2 b 3 −0.01 (0.04) −0.08, 0.08 −0.01 (0.02) −0.04, 0.04 a 1 d 1 d 5 b 4 −0.02 (0.02) −0.05, 0.02 −0.01 (0.01) −0.03, 0.01 a 1 d 4 d 3 b 4 −0.01 (0.01) −0.04, 0.01 −0.01 (0.01) −0.02, 0.01 a 2 d 2 d 3 b 4 * −0.02 (0.02) −0.06,  −0.01 −0.01 (0.01) −0.03,  −0.01 a 1 d 1 d 2 d 3 b 4 * −0.03 (0.03) −0.10,  −0.01 −0.02 (0.01) −0.05,  −0.01 Covariates Coefficient SE Lower CI Upper CI Time 1 Empathy (M1) Cognitive 0.02 0.02 −0.02 0.06 Affective* 0.08 0.03 0.02 0.15 Gender −0.48 0.27 −1.01 0.04 Colleague 0.03 0.24 −0.44 0.51 End-state DFA 0.01 1.64 −3.22 3.25 Time 2 Empathy (M2) Cognitive −0.01 0.01 −0.03 0.01 Affective 0.01 0.02 −0.04 0.04 Gender −0.02 0.14 −0.30 0.27 Colleague −0.11 0.13 −0.37 0.14 End −state DFA −0.12 0.88 −1.85 1.61 Time 3 Empathy (M3) Cognitive 0.01 0.01 −0.01 0.03 Affective 0.03 0.02 −0.01 0.07 Gender −0.08 0.15 −0.37 0.21 Colleague 0.10 0.13 −0.16 0.36 End-state DFA −0.42 0.90 −2.20 1.36 Time 4 Empathy (M4) Cognitive* 0.03 0.01 0.01 0.06 Affective 0.05 0.02 −0.01 0.09 Gender −0.25 0.18 −0.60 0.10 Colleague 0.09 0.16 −0.22 0.41 End-state DFA 1.63 1.09 −0.51 3.78 Gender Essentialism (Y) Cognitive 0.01 0.02 −0.02 0.04 Affective* −0.09 0.03 −0.15 −0.03 Gender 0.44 0.23 −0.03 0.89 Colleague 0.23 0.21 −0.18 0.64 End-state DFA* 3.39 1.43 0.58 6.21 Open in a new tab * p < .05. Completely standardised indirect effects are reported as a measure of effect size. CI = confidence interval. Gender: 0 = women, 1 = men. Colleague: 0 = cisgender, 1 = transgender. Virtual embodiment and interpersonal affiliation with the transgender colleague Figure 4 a summarises the moderated serial mediation model for perceived likeability toward the colleague at post-RPG, holding constant perceived likeability at baseline ( F (9, 152) = 30.14, p < 0.001, R 2 = 0.64). Our exploratory analyses using Model 92 suggested that colleague gender served to moderate only the serial mediators. Thus, Model 91 was used for the final analysis, retaining participant gender as a covariate. Here, a significant indirect effect was observed only for those who interacted with the transgender colleague, such that the relationship between embodiment and likeability was consistent with mediation for Time 1 followed by Time 4 empathy. As illustrated in Fig. 4 a, the interaction between colleague gender and state empathy at Time 1 was significantly associated with state empathy at Time 4. However, this effect was significant for the transgender ( b = 0.39, SE = 0.15, p = 0.012, 95% CI [0.09, 0.69]), but not the cisgender interaction partner ( b =  −0.04, SE = 0.13, p = 0.765, 95% CI [ −0.29, 0.22]). In other words, the pattern consistent with moderated mediation observed here suggests that although all participants reported increased feelings of likeability toward their interaction partner following the RPG simulation, it is the increase from baseline to end-state empathy that appears to matter for more positive attitudes toward a transgender outgroup target. Fig. 4. Open in a new tab ( a ) Increases in empathy over time were consistent with mediating the relationship between virtual embodiment and likeability ratings following the RPG, but only for those who interacted with the transgender colleague (moderated d-path). Participant gender, cognitive and affective empathy, and baseline likeability were included as covariates. The total effect is displayed below the horizontal line, while the direct effect is displayed above the line. ( b ) Increases in empathy over time were consistent with mediating the relationship between virtual embodiment and similarity ratings following the RPG, but only for those who interacted with the transgender colleague (moderated b-path) Participant gender, cognitive and affective empathy, and baseline similarity were included as covariates. The total effect is displayed below the horizontal line, while the conditional (moderated) direct effects are displayed above the line for the transgender and cisgender colleagues. The full indirect effects and covariates table for this model can be viewed in Table 7 . Similar to the previous model, affective empathy was positively associated with state empathy at Time 1, while both affective and cognitive empathy were positively associated with state empathy at Time 4. Thus, we again see support for the differential relationship of both trait and state empathy across the progression of the RPG. Additionally, baseline likeability ratings were associated with post-RPG likeability and empathy at all timepoints except Time 2, where instead, empathy toward the cisgender colleague was significantly higher. This pattern might suggest that one’s initial perceptions of another—at least in terms of favourability, pleasantness, or agreeability—could be meaningfully related to how empathy develops across the course of an interaction. Unlike the previous model, however, the total and direct effects of embodiment on likeability were significant, indicating that embodiment remained associated with likeability when holding constant only the covariates, as well as all mediators and covariates, respectively. Table 7. Indirect Effects and Covariates for Model 91 Predicting Likeability. Unconditional Indirect Effects b SE Lower CI Upper CI a 1 b 1 0.05 0.05 −0.03 0.16 a 2 b 2 −0.02 0.03 −0.09 0.04 a 3 b 3 0.01 0.01 −0.02 0.04 a 4 b 4 0.09 0.08 −0.04 0.27 Cisgender Colleague Transgender Colleague Index b ( SE ) 95% CI b ( SE ) 95% CI b ( SE ) 95% CI Conditional Indirect Effects a 1 d 1 b 2 −0.02 (0.04) −0.12, 0.05 −0.03 (0.05) −0.13, 0.06 −0.01 (0.01) −0.02, 0.01 a 1 d 4 b 3 0.01 (0.01) −0.01, 0.03 0.01 (0.01) −0.01, 0.02 −0.01 (0.01) −0.03, 0.01 a 1 d 6 b 4 −0.01 (0.02) −0.06, 0.03 0.07 (0.05) 0.01, 0.20* 0.08 (0.06) 0.01, 0.22* a 2 d 2 b 3 0.01 (0.02) −0.03, 0.05 0.01 (0.02) −0.03, 0.06 0.01 (0.01) −0.01, 0.02 a 2 d 5 b 4 0.02 (0.02) −0.01, 0.07 0.01 (0.03) −0.05, 0.07 −0.01 (0.03) −0.09, 0.05 a 3 d 3 b 4 0.02 (0.03) −0.02, 0.08 0.01 (0.02) −0.02, 0.05 −0.01 (0.02) −0.06, 0.02 a 1 d 1 d 2 b 3 0.01 (0.02) −0.03, 0.06 0.01 (0.03) −0.04, 0.08 0.01 (0.01) −0.01, 0.03 a 1 d 1 d 5 b 4 0.03 (0.03) −0.02, 0.09 0.01 (0.04) −0.08, 0.10 −0.01 (0.05) −0.12, 0.07 a 1 d 4 d 3 b 4 0.01 (0.02) −0.01, 0.05 0.01 (0.01) −0.01, 0.03 −0.01 (0.02) −0.05, 0.03 a 2 d 2 d 3 b 4 0.03 (0.03) −0.01, 0.10 0.02 (0.02) −0.01, 0.07 −0.01 (0.03) −0.08, 0.03 a 1 d 1 d 2 d 3 b 4 0.04 (0.03) −0.01, 0.11 0.03 (0.03) −0.01, 0.10 −0.01 (0.03) −0.08, 0.05 Covariates b SE Lower CI Upper CI Time 1 Empathy (M1) Cognitive 0.02 0.02 −0.01 0.05 Affective* 0.07 0.03 0.01 0.13 Gender −0.08 0.25 −0.56 0.41 Likeability (pre)* 0.48 0.08 0.33 0.64 Time 2 Empathy (M2) Cognitive −0.01 0.01 −0.03 0.01 Affective 0.01 0.02 −0.03 0.04 Gender 0.01 0.15 −0.28 0.30 Likeability (pre) 0.02 0.05 −0.08 0.13 Time 3 Empathy (M3) Cognitive 0.01 0.01 −0.01 0.03 Affective 0.03 0.02 −0.01 0.07 Gender −0.01 0.15 −0.30 0.29 Likeability (pre)* 0.13 0.05 0.03 0.24 Time 4 Empathy (M4) Cognitive* 0.03 0.01 0.01 0.06 Affective* 0.05 0.02 0.01 0.10 Gender −0.10 0.18 −0.45 0.26 Likeability (pre)* 0.18 0.07 0.05 0.31 Likeability (Y) Cognitive −0.01 0.01 −0.03 0.03 Affective 0.01 0.02 −0.05 0.05 Gender 0.18 0.19 −0.19 0.56 Likeability (pre)* 0.24 0.07 0.10 0.38 Open in a new tab * p < .05. A statistically significant index of moderated mediation suggests that the indirect effect is conditional on the values of the moderator. CI = confidence interval. Gender: 0 = women, 1 = men. Pre = baseline likeability. Figure 4 b displays the moderated mediation model for perceived similarity with the colleague at post-RPG, holding constant perceived similarity at baseline ( F (15, 146) = 6.57, p < 0.001, R 2 = 0.40). Here, Model 92 suggested that colleague gender served to moderate only the b-path. Thus, we used Model 89 for the final analysis, retaining participant gender as a covariate. Reflecting our earlier mixed-design ANOVA, the direct effect was significant only for those who interacted with the transgender colleague. Likewise, a significant indirect effect was observed only for the transgender colleague, such that embodiment was associated with increased similarity via state empathy across all timepoints. Figure 4 b denotes two significant interactions between colleague gender and empathy at Time 1 and Time 4. An examination of the conditional simple effects revealed that the strength of the relationship between Time 1 empathy and similarity did not significantly differ between transgender ( b =  −0.46, SE = 0.27, p = 0.086, 95% CI [ −0.99, 0.07]) and cisgender colleagues ( b = 0.37, SE = 0.22, p = 0.094, 95% CI [ −0.06, 0.80]), although the direction of these effects were in opposition to each other. However, the strength of the relationship between Time 4 empathy and similarity was significant only for the transgender (( b = 0.57, SE = 0.17, p = 0.001, 95% CI [0.23, 0.91]), and not the cisgender colleague ( b =  −0.15, SE = 0.21, p = 0.470, 95% CI [ −0.56, 0.26]). The pattern consistent with moderated mediation observed here again suggests that end-state empathy is important for more positive attitudes, in particular for greater feelings of closeness and lower psychological distance with those who are initially perceived as dissimilar or ‘other’. Further aligning with the previous models, affective empathy was positively associated with state empathy at Time 1, while cognitive empathy was positively associated with state empathy at Time 4. However, baseline similarity ratings were only associated with empathy at Time 1 and similarity at post-RPG, indicating a more limited role of one’s initial interpersonal perceptions. Table 8 summarises the indirect effects and covariates for this model. Table 8. Indirect Effects and Covariates for Model 89 Predicting Similarity. Cisgender Colleague Transgender Colleague Index Effect b ( SE ) 95% CI b ( SE ) 95% CI b ( SE ) 95% CI a 1 b 1 0.15 (0.12) −0.07, 0.41 −0.19 (0.16) −0.55, 0.07 −0.34 (0.22) −0.81, 0.02 a 2 b 2 −0.04 (0.09) −0.27, 0.10 0.09 (0.10) −0.06, 0.32 0.13 (0.14) −0.06, 0.49 a 3 b 3 0.04 (0.06) −0.06, 0.19 −0.01 (0.03) −0.08, 0.06 −0.05 (0.07) −0.22, 0.08 a 4 b 4 −0.02 (0.08) −0.26, 0.06 0.09 (0.09) −0.07, 0.27 0.12 (0.14) −0.84, 0.46 a 1 d 1 b 2 −0.06 (0.12) −0.29, 0.20 0.13 (0.14) −0.10, 0.46 0.19 (0.18) −0.14, 0.57 a 1 d 4 b 3 0.03 (0.04) −0.01, 0.13 −0.01 (0.02) −0.06, 0.04 −0.04 (0.05) −0.16, 0.02 a 1 d 6 b 4 −0.01 (0.03) −0.09, 0.03 0.04 (0.04) −0.01, 0.13 0.05 (0.05) −0.01, 0.18 a 2 d 2 b 3 0.07 (0.08) −0.01, 0.28 −0.02 (0.05) −0.12, 0.08 −0.09 (0.10) −0.35, 0.02 a 2 d 5 b 4 −0.01 (0.01) −0.04, 0.02 0.02 (0.03) −0.01, 0.09 0.02 (0.03) −0.02, 0.10 a 3 d 3 b 4 −0.01 (0.02) −0.05, 0.03 0.02 (0.03) −0.02, 0.08 0.03 (0.04) −0.04, 0.11 a 1 d 1 d 2 b 3 0.11 (0.09) −0.01, 0.35 −0.03 (0.07) −0.18, 0.12 −0.13 (0.12) −0.44, 0.04 a 1 d 1 d 5 b 4 −0.01 (0.02) −0.05, 0.03 0.03 (0.04) −0.02, 0.12 0.03 (0.04) −0.04, 0.13 a 1 d 4 d 3 b 4 −0.01 (0.01) −0.05, 0.01 0.02 (0.02) −0.01, 0.06 0.02 (0.03) −0.01, 0.09 a 2 d 2 d 3 b 4 −0.01 (0.03) −0.10, 0.02 0.04 (0.03) −0.01, 0.11 0.05 (0.05) −0.01, 0.19 a 1 d 1 d 2 d 3 b 4 −0.02 (0.04) −0.14, 0.03 0.06 (0.04) 0.01, 0.15* 0.07 (0.06) 0.01, 0.24* Covariates b SE Lower CI Upper CI Time 1 Empathy (M1) Cognitive 0.02 0.02 −0.02 0.05 Affective* 0.08 0.03 0.01 0.14 Gender −0.41 0.26 −0.92 0.09 Similarity (pre) 0.18 0.06 0.07 0.29 Time 2 Empathy (M2) Cognitive −0.02 0.01 −0.03 0.01 Affective 0.01 0.02 −0.04 0.04 Gender −0.02 0.14 −0.30 0.26 Similarity (pre) −0.05 0.03 −0.07 0.06 Time 3 Empathy (M3) Cognitive 0.01 0.01 −0.01 0.03 Affective 0.03 0.02 −0.01 0.07 Gender −0.07 0.15 −0.36 0.22 Similarity (pre) 0.03 0.03 −0.04 0.09 Time 4 Empathy (M4) Cognitive* 0.03 0.01 0.01 0.06 Affective 0.04 0.02 −0.01 0.09 Gender −0.25 0.18 −0.60 0.10 Similarity (pre) 0.03 0.04 −0.05 0.11 Similarity (Y) Cognitive 0.03 0.02 −0.01 0.07 Affective −0.03 0.04 −0.10 0.05 Gender 0.24 0.29 −0.34 0.82 Similarity (pre)* 0.28 0.09 0.11 0.45 Open in a new tab * p < .05. A statistically significant index of moderated mediation suggests that the indirect effect is conditional on the values of the moderator. CI = confidence interval. Gender: 0 = women, 1 = men. Pre = baseline similarity. Discussion In this study, we evaluated a virtual RPG designed to support cisgender-transgender interactions and address transprejudice. Our findings suggest that state empathy may play a unique role in shaping attitudinal responses among cisgender majority group members following embodied interactions in VR. These findings advance recent work that has demonstrated the potential of VRPT and VR-mediated contact as independent approaches to prejudice reduction in other identity domains 7 , 8 by employing both approaches simultaneously and providing preliminary evidence for their impact on gender essentialist beliefs and cisgender-transgender affiliation. By capturing non-verbal postural dynamics across the interaction, we offer novel insight into how embodied processes shift across key perturbations of the VR interaction. Extending previous work on binary gender 17 , we found evidence of an indirect association between perceived virtual embodiment and post-session gender essentialism via state empathy for all participants, such that higher embodiment was associated with greater empathy, which in turn was associated with lower endorsement of gender essentialism. We also observed a novel pattern for within-session affiliation, whereby virtual embodiment was associated with increased empathy, which in turn was associated with increased feelings of likeability and similarity, but only among those who interacted with a transgender colleague. Taken together, these findings suggest that VRPT (e.g., embodying a transgender avatar) may be a promising approach for engaging with internalised cognitions about gendered norms that precede transprejudice and discrimination 36 . However, it should be noted that our mediation analyses tested conditional associations consistent with, but not wholly demonstrating, the proposed empathy-based mechanism linking embodiment to essentialism and affiliation. To further evaluate this initial result, future work should experimentally manipulate virtual embodiment (e.g., comparing transgender and cisgender embodiment) to determine whether induced differences in embodiment produce corresponding changes in state empathy and downstream shifts in essentialist beliefs or intergroup affiliation. Alongside VRPT, VR-mediated contact may offer an additional benefit for facilitating opportunities to develop a sense of shared identity. As described above, we observed within-session increases in likeability and similarity only when cisgender individuals both embody a transgender outgroup identity and interact with a transgender outgroup member. Thus, our findings suggest that combining perspective taking and contact within VR contexts may be a promising direction for future prejudice reduction research. In addressing transprejudice, the roles of likeability and similarity should not be overlooked. Fostering recognition of shared similarities can reduce anxiety about and avoidance of those who deviate from traditional, binary gender norms 5 . Moreover, enhancing these factors contributes to increased readiness for future intergroup interactions 37 , counteracting pervasive negative stereotypes that construe transgender individuals as ‘other’. However, it should be noted that within-session affiliation was measured via single likeability and similarity items in the current study. To extend this initial finding, future research could evaluate the multidimensional nature of intergroup perceptions using more comprehensive measures of interpersonal affiliation, such as multi-item scales 38 , semantic differential ratings 39 , or social distance metrics 40 . Using non-linear fractal analysis (i.e., DFA) to examine postural activity dynamics, we also found that undertaking the RPG simulation allowed participants to achieve a state of reduced constraint on their behaviour. In other words, the increased complexity of participants’ postural behaviours indicated an emerging pattern toward a balanced state of predictability and variability, in which past behaviour has a measurable influence on future behaviour 41 . These behavioural states are characterised by adaptability and resilience in response to internal disruptions or environmental changes 26 , 27 , such as those inherent in VRPT (e.g., embodied cognition, simulated context, virtual realism). While the colleague’s gender identity did not appear to influence postural activity, we observed that women achieved a more adaptable and resilient state earlier in the RPG than men. However, men exhibited a significant increase in complexity following discussions that specifically focused on allyship, gender identity, and equity in the workplace. Given that women are often subject to persistent experiences of prejudice and discrimination at work 42 , they may shift more readily into the VRPT experience of embodying a transgender identity. Aligning with previous work 43 , 44 , men may require exposure to specific content aimed at highlighting the impact of privilege on marginalised groups before they begin to benefit from the VRPT experience. These non-verbal behavioural dynamics appear to co-vary with self-reported measures of empathy. First, empathy peaked at the conclusion of the simulation, and both empathy and DFA at this end-state uniquely contributed to variance in post-session gender essentialism. Second, women reported greater empathy than men at all timepoints and exhibited earlier increases in the persistence and flexibility of their postural activity. Third, men exhibited increases in postural complexity following discussions regarding workplace allyship, gender identity, and equity, which coincided with a significant increase in state empathy at Time 3. By triangulating postural activity data with self-reported, context-specific empathy, we offer an initial descriptive account of how the affective and cognitive processes implicated in theoretical models of prejudice reduction may fluctuate during simulated interactions. Indeed, previous work has identified empathy as a significant contributor to positive outcomes following VRPT 24 , and as a mediator of intergroup contact and prejudice reduction more broadly 23 . Future work could experimentally manipulate empathetic engagement to test whether the coupling between state empathy and non-verbal postural dynamics reflects a functional relationship. Extending previous work using VRPT and VR-mediated contact in other intergroup contexts 7 , 8 , the current study offers two key considerations for the development of future interventions targeting transprejudice. First, simulated interactions should aim to be synchronous, rather than asynchronous (e.g., imagined) 10 or observational (e.g., parasocial media representations) 11 . While these latter approaches are often preferred for feasibility, their reduced social presence can, in turn, impact interaction quality, interpersonal trust, and outgroup favourability 45 . However, immersive VR interactions remain parasocial in nature, as participants engage with virtual avatars and other computer-generated entities. Accordingly, the present findings should not be interpreted as directly equivalent to real-world, human–human interaction, which involves features such as reciprocal spontaneity, emotional complexity, and genuine unpredictability 46 , 47 . Rather, VR may represent a higher social presence form of parasocial interaction that more closely approximates certain features of direct contact (e.g., synchrony, contingency, embodied interaction) 48 . Thus, a key direction for future research is to systematically compare human–human and VR-mediated interactions to determine whether and under what conditions observed effects may have practical significance beyond simulated contexts. Second, simulated cisgender-transgender interactions should aim to enable rapport building, rather than simply exposing majority group individuals to marginalised identities to enhance familiarity. While rapport and familiarity are related, rapport involves a higher degree of emotional connection or interpersonal closeness 49 . Indeed, feelings of closeness with outgroup members are stronger predictors of reduced prejudice and increased acceptance 50 , with efforts to foster such relationships being most effective when they involve personal and meaningful connections 51 . Although our results highlight the potential utility of simulated exposures to gender identity microaggressions (e.g., bathroom access), it is critical for intervention designers to engage in community consultation when representing other aspects of transgender lived experience (e.g., navigating gender-exclusive spaces, accessing medical care, securing workplace or community benefits) in future interventions targeting cisgender individuals. As a tool for research, VR balances experimental control with ecological validity, more closely approximating direct contact and enabling visceral perspective-taking without relying on the labour of vulnerable populations. However, there are limitations to this work that must be considered. First, certain methodological constraints should be noted. As virtual embodiment was not experimentally manipulated, the serial mediation models are observational with respect to this variable. Additionally, the current study did not include a cisgender avatar embodiment condition or a non-intervention control group, and gender essentialism was assessed only at post-RPG. Accordingly, effects attributable to intervention cannot be definitively established. Certain self-report constructs were also assessed using single-items indicators (i.e., likeability, similarity), and internal consistency values for virtual embodiment, social desirability, and affective trait empathy fell below conventional thresholds (α < 0.70), which may have attenuated or inflated the observed associations. We encourage future research to extend the lines of enquiry initiated in this study by pursuing the trajectories outlined in the preceding discussion and developing extended (e.g., fully crossed) designs that allow for robust tests of intervention effects. Second, our data was collected among an Australian sample, which may not be representative of individuals on a global scale. Although the Australian context aligns with cultural, legal, and social tensions surrounding gender identity in Western and Global North locations, future work should continue to investigate how interventions can be applied in contexts where transgender identities are criminalised 52 , 53 and in conflict or post-conflict settings 54 , 55 . Third, we did not investigate the impact of multiple exposures. While research suggests that initial exposures may be the strongest predictors of future intentions 37 , 56 , significantly more longitudinal research is needed to understand the degree to which benefits are sustained in the long term. Finally, participants interacted with avatars that were designed from Asian and White models, created to represent a majority of the Australian university population. Thus, participants were not explicitly matched with an interaction partner based on their own racial or ethnic identities. While the current study targeted gender identity, holding additional marginalised identities may further impact experiences of perspective taking and intergroup contact 57 . As interventions frequently address identity categories independently, this highlights a need for future work that targets intersectional identities, both in VR and in the intervention space more broadly, to better represent the dynamics of real-world diversity. Methods Participants Cisgender undergraduates ( N = 162; Women = 80, Men = 82; M age = 20.28, SD = 3.17) were recruited from the university’s research participant pool and randomly assigned to interact with either the transgender (Women = 40, Men = 42) or cisgender colleague (Women = 40, Men = 40). A priori power analyses employed a moderate effect size and suggested that 136 participants were needed to achieve 80% power for mixed models with one within- and two between-subjects factors (G*Power), while a range of 150–155 participants was needed for models with four serial mediators (Monte Carlo simulation). We initially recruited 173 participants, although 11 were excluded from the analyses due to incomplete data (e.g., inattention, data recording error, technical difficulties). The final sample was comprised largely of Asian (17.9%) and White (34%) participants, followed by European (16.7%), Middle Eastern (8.6%), Indian (8.6%), Black (1.2%), and Hispanic (1.9%) participants, while 11.1% of participants reported mixed backgrounds. Previous experience with virtual reality was reported by 58% of participants (Limited: n = 72, Moderate: n = 18; Frequent: n = 4). No other demographic data were collected (e.g., religious beliefs, political affiliation), as the current study was not designed to examine individual differences in these variables, although future work may wish to consider their potential moderating role. Procedure In exchange for course credit, undergraduate students volunteered to complete an RPG simulation of an after-work team-building event undertaken within VR. Participants were required to be at least 18 years of age, with normal or corrected-to-normal vision, and no history of motion sickness or neurological, neuromuscular, or skeletal disorders. On arrival at the lab, participants provided voluntary, written, informed consent to undertake the following procedures, as approved by the Macquarie University Human Research Ethics Committee (Project ID: 10,903) in accordance with the requirements set out in the National Statement on Ethical Conduct in Human Research 2007. First, participants customised one cisgender and one transgender avatar to represent themselves in VR. Guided by the experimenter, participants were allotted 15 min to create each avatar (i.e., 30 min total; see Participant Avatars for further detail). Although we initially informed participants that they would be randomly assigned to embody one of their customised avatars in VR, all participants were required to embody their transgender avatar to enable VRPT. Instead, they were randomly assigned to interact with either a transgender or cisgender virtual colleague (Fig. 1 a), such that women were paired with either a cisgender woman or transgender man, whereas men were paired with a cisgender man or transgender woman. The purpose of this deception was to minimise demand characteristics by obscuring the study’s focus on transgender embodiment. Following avatar customisation, participants were presented with information about their assigned colleague (e.g., work history, pronouns, gender identity; see Supplementary Material A) before providing baseline ratings of perceived interpersonal affiliation. Notably, the colleague’s work history indicated that they worked in the information technology field. Replicating previous work 17 , this setting was intentionally selected to situate the simulation within a STEM context, where gender disparities in representation and workplace climate are well-documented 58 , 59 , making it a relevant backdrop for the present investigation. Prior to entering VR, participants viewed an instructional video outlining directives for VR gameplay, teleportation within the virtual environment, and how to interact with their colleague using the dialogue engine. The instructional video can be viewed at https://youtu.be/HJRGRy53EsA . Participants then entered VR using the HTC Vive system, where they first appeared in the lobby of an arcade-style gaming centre facing a virtual mirror, in which they could see the reflection of their customised transgender avatar. Here, the experimenter verbally confirmed and reinforced participants’ transgender identity and initiated a movement integration exercise prior to beginning the RPG simulation (see Supplementary Material B). The RPG then began with participants being introduced to their virtual colleague and practicing navigation via teleportation. Participants progressed through the RPG by playing three games with their colleague, which took place between scripted dialogues designed to build rapport with the virtual colleague, gradually increasing in intimacy (see Dialogue Engine and Scripting for further detail). The first and third games involved competitive (i.e., air hockey) and cooperative tasks (i.e., sheep herding), with presentation of task type counterbalanced across participants, while the second game served to maintain engagement and progress the RPG (i.e., whack-a-mole, neither competitive nor cooperative). Following the completion of gameplay, participants were guided to a virtual bar where the final scripted event sequence commenced, which was designed to serve as an exposure to real-world prejudice and discrimination commonly faced by transgender individuals. First, a drink was spilled, and participants were instructed to wash their hands in the bathroom. Next, participants were required to decide which of two bathrooms to enter (i.e., women’s, men’s). Once inside, they triggered a faucet using their controller to simulate washing their hands, while viewing themselves in a virtual bathroom mirror. Upon exiting, participants encountered another patron in the arcade who challenged their use of the ‘wrong’ bathroom. After the confrontation, participants returned to their virtual colleague, who engaged them in a debrief about the incident and concluded the RPG simulation. Empathy toward the virtual colleague was recorded at four timepoints during the simulation: after the first, second, and third games, as well as following the final debrief. Participants’ postural activity was recorded throughout the RPG (see Data Processing for further detail). After exiting VR, participants completed a battery of self-report questionnaires, first completing state measures of interpersonal affiliation, perceived sense of embodiment, and gender essentialism, followed by trait measures of empathy and social desirability. Participants were then debriefed about the true purpose of study (i.e., VRPT, VR-mediated contact) and the deception used (i.e., transgender avatar embodiment, random assignment to colleague). Materials Apparatus Participants used the HTC Vive Pro Eye system to view and interact within the virtual environment. This system is comprised of a VR headset with dual-OLED display and a combined resolution of 2880 × 1600 pixels and 615 PPI, two hand-controllers, and wall-mounted base stations. Three Vive 2.0 base stations were affixed to the uppermost portion of the wall to maximise coverage and minimise obstructions. These base stations consistently track and determine the precise position and orientation of the Vive headset and controllers, allowing for real-time motion tracking and avatar motor control. Participant avatars Participants customised their avatars using Adobe Fuse (version 1.3). To do this, participants first selected their preferred body type from a selection of premade, height-scaled, feminine and masculine models. The model that best matched the participant’s height, within a range of 153–182 cm, was then imported into Adobe Fuse, where sliders were available to adjust size and musculature of the body (e.g., arms, legs, torso, hips, breasts), facial structure (e.g., cheek bones, eye shape, lips, jawline), and features such as skin tone and eye colour. After finalising adjustments to the body, participants were guided through options to choose their avatar’s dynamic features (e.g., clothing, hair style, makeup, facial hair). The same baseline model was used for customisation of both the transgender and cisgender avatars. The customised transgender avatar was then rigged for animation using ActorCore AccuRIG (version 1.2.1) before being imported into the virtual environment. A selection of participant avatars is displayed in Supplementary Material C. Pre-programmed avatars Figure 1 a displays the transgender and cisgender colleague avatars, while the two cisgender aggressor avatars can be viewed in Supplementary Material C. Cisgender avatars were selected from a larger pool of models created by the research team for use in human–computer interaction research, while transgender avatars were specifically designed to minimise stereotypical representations of transgender identities. These avatars were created using Adobe Fuse and Blender 3D modelling software and rigged for animation using Mixamo ( https://www.mixamo.com ). Avatars were pre-programmed into the virtual environment and synced with pre-recorded voiceover, conducted by transgender and cisgender voice actors, which was integrated with SALSA LipSync software to enable facial animations using an approximation technique. Virtual environment The virtual environment was created and hosted in Unity (version 2020.3.42f1). Participants first arrived in the lobby of the arcade-style gaming centre, where they viewed their avatar in a virtual mirror. Once the RPG was initiated, this mirror was removed, revealing the full arcade environment. Relative to the participant’s initial position, a welcome desk was located on the right side of the lobby. The air hockey and herding game tables were located at the centre of the gaming area, while whack-a-mole and additional arcade machines lined each side of the room. The bar and dining area was located at the far right, adjacent to the bathroom doors. To further enhance the realism of the environment, additional non-player avatars were placed throughout the environment at key locations (i.e., welcome desk, arcade machines, dining area), and directional background noise was integrated within the environment (e.g., arcade game music, air hockey puck, ambient bar clatter). Virtual mirrors lined the full length of the right-hand wall and the bathroom sink, allowing participants to view their avatars throughout the RPG. Additional scenes illustrating the virtual environment are displayed in Supplementary Material D. Notably, all non-player avatars were drawn from the broader pool of models developed by the research team for use in human–computer interaction research and were designed to represent cisgender individuals. Transgender characters were not included among the non-player avatars as the simulation was intended to reflect a predominantly cisgender environment, thereby positioning the transgender participant and colleague avatars as socially distinct within the interaction context. To allow participants to move through the environment, blue teleportation rings were placed at key locations. The colleague avatar was programmed to walk to the next location before text-based instructions appeared to guide participants to each location in sequence (e.g., welcome desk, gaming tables, bar and dining area, bathrooms, lobby). Participants then navigated to the specified location by pointing their controller at a blue ring and selecting it. Using teleportation as the locomotion mechanic allowed participants to be instantly moved to the selected point in virtual space, without physically walking in the laboratory while wearing the VR headset. Dialogue engine and scripting In an RPG, a dialogue engine is a computational framework that facilitates interactive communication, narrative progression, and player agency by using dialogue trees and scripting logic . Dialogue trees refer to the hierarchical structure of predefined player dialogue options and resulting interaction partner responses, such that the player is permitted to choose their conversational path from a predetermined set of options, while the interaction partner’s response may differ depending on the option selected (i.e., branching narratives). Scripting logic then refers to the rules and conditions that dynamically adjust conversations based on the player’s choices. In the current study, participants were enabled to choose how they would like to respond to their colleague from the options presented within VR during conversational turn-taking. To enhance realism of the interaction, participants were instructed to speak their chosen dialogue aloud before selecting their response with their controller, which then cued the next scripted action (e.g., colleague dialogue, environment navigation, gameplay). The content of the dialogue was designed to progressively increase in intimacy, such that participant-colleague interactions began with discussions surrounding the workplace and information technology skills, followed by discussions focused on likes and dislikes (e.g., food, pets), and then workplace equity, diversity, and inclusion. During the confrontation, participants were enabled to choose from a selection of responses to the aggressor that were modelled as passive or assertive. In addition to allowing the colleague to respond specifically to participant dialogue selections, our scripting logic adjusted the dialogue depending on the participant’s avatar gender (i.e., transgender woman/man) as well as the colleague’s gender (transgender/cisgender, woman/man). In particular, the colleague’s dialogue during the debrief was adjusted to represent lived experience for the transgender colleague versus allyship for the cisgender colleague. The full script and branching logic can be viewed in Supplementary Material E. Data processing Postural sway timeseries were recorded across the six phases of the RPG by tracking the position of the head using the Vive headset, which captures linear motion trajectories along the X-, Y-, and Z-axes at 90 Hz (i.e., 90 captures per second). As postural sway was not recorded during gameplay, timeseries captured postural activity across the following periods of participant-colleague interaction: (1) from the start of the experiment to the start of the first game, (2) from the end of the first game to the start of the second game, (3) from the end of the second game to the start of the third game, (4) from the end of the third game to bathroom choice, (5) during the confrontation, (6) when debriefing with the colleague. To prepare timeseries for DFA, we calculated a two-dimensional vector using data from the X- and Z-axes of linear motion. At baseline, the X-axis captures medial–lateral (M-L) sway, while the Z-axis captures anterior–posterior (A-P) sway. However, M-L and A-P sway were not independently constrained to these axes, as participants were permitted to rotate on the spot within the experimental space. That is, each phase of the RPG occurred at a different location within the virtual environment via teleportation, with the participant’s orientation rotated accordingly. To ensure that teleportation behaviours were excluded from postural activity data, each timeseries was trimmed by 1024 frames (i.e., 11.11 s) from the initial and final timestamps. Data processing was completed using Python (version 3.9.12), while DFA was completed using MATLAB (version R2024b). Example code can be viewed at: https://github.com/xkiwilabs/MATLAB-Toolboxes/tree/master/Fractal_Analysis_Toolbox . Detrended fluctuation analysis (DFA) DFA is a non-linear timeseries method that determines the degree to which a behaviour or signal changes over time and across different timescales. DFA was conducted on each of the six trimmed timeseries, across all participants, by dividing each timeseries into segments of equal length varying by window size (i.e., length of non-overlapping segments comprised of 8, 16, 32, 64, 128, 256 data points). Within each segment, the local trend is identified and extracted from the data (i.e., detrended), isolating the residuals (i.e., differences between the data points and the local trend). The residual variance is then calculated for each segment, quantifying the degree to which the data fluctuates around the trend within that segment. Root mean square is provided as an estimate of residual variances averaged across segments for each window size. These residual variance estimates, representing the average magnitude of fluctuations in postural activity, are plotted as a function of window size in log–log form, with the slope of the regression line fitting this plot denoted as α (alpha). The scaling exponent α characterises the complexity of the fractal, or self-similar, structure within each timeseries 33 , 34 . This factor was extracted for each of the six postural sway timeseries and evaluated against standardised values representing white noise ( α = 0.5; random, highly constrained variation), pink noise ( α = 1.0; moderately persistent, highly flexible), and brown noise ( α = 1.5; highly persistent, overly controlled). Measures Embodiment Perceived sense of virtual avatar embodiment was measured using the Embodiment Questionnaire 60 . Participants rated their experience of virtual avatar embodiment during the RPG on perceived body ownership, agency, self-location, and external appearance across 15 items from 1 ( strongly disagree ) to 7 ( strongly agree ). The Embodiment Questionnaire has previously demonstrated convergent, discriminant, and construct validity, with good internal consistency at α = 0.69-0.75 61 . In the current study, the embodiment questionnaire achieved an α of 0.65. Gender essentialism Endorsement of gender essentialism was measured using the Essential Sex and Gender subscale of the Heteronormative Attitudes and Beliefs Scale (HABS) 36 . Participants indicated the degree to which they endorsed eight statements concerning binary conceptualisations of gender from 1 ( strongly disagree ) to 7 ( strongly agree ). The HABS has previously demonstrated convergent and concurrent validity, with excellent internal consistency for this subscale at α = 0.85-0.92 36 . In the current study, the HABS achieved an α of 0.90. Social desirability The degree to which participants tended to respond to self-report measures in a favourable light was measured using the Marlowe-Crowne Social Desirability Scale (MCSDS) 62 . Participants responded to 33 true/false items concerning common, undesirable and unlikely, desirable behaviours. The MCSDS has previously demonstrated convergent and discriminant validity and good internal consistency at α = 0.70-0.80 63 . In the current study, the MCSDS achieved an α of 0.68. Trait empathy Trait-level empathetic tendencies were measured using the Questionnaire of Cognitive and Affective Empathy (QCAE) 64 . Participants responded to 19 items for cognitive empathy and eight items for affective empathy, excluding the Peripheral Responsivity subscale 65 from 1 ( strongly disagree ) to 4 ( strongly agree ). The QCAE has previously demonstrated external, differential, and convergent validity, with good to excellent internal consistency (Cognitive: α = 0.86, Affective: α = 0.73) 65 . In the current study, we observed α = 0.82 for the Cognitive subscale and α = 0.68 for the Affective subscale. State empathy Context-specific empathy was measured at four timepoints during the RPG, using three items from Batson and colleagues 66 . Within VR, participants rated their feelings of compassion, warmth, and sympathy toward their colleague from 0 ( not at all ) to 10 ( very much ) following each game and after the debrief. This three-item measure has previously demonstrated convergent, discriminant, and predictive validity and excellent internal consistency at α = 0.85-0.88 67 . In the current study, we observed α = 0.81 at time 1, α = 0.85 at time 2, α = 0.86 at time 3, and α = 0.87 at time 4. Interpersonal affiliation Perceived likeability of and similarity with the colleague was measured pre- and post-RPG, using single item statements adapted from the Reysen Likeability Scale 38 , 68 . Participants were shown an image of their randomly assigned colleague’s avatar and provided ratings from 1 ( not at all ) to 9 ( very much ) for likeability ( How much do you like your colleague? ) and similarity ( How similar to you is your colleague? ). Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 (768.6KB, pdf) Acknowledgements The authors would like to thank Jessica Liang, Polina Kurbanova, Gianna Edirimanne, and Justin Biceski for assistance with data collection, as well as Meredith Porte for assistance with programming in VR. Author contributions CLC and RWK conceived and designed the study. CLC and MJR analysed and interpreted the data. CLC wrote the main manuscript text and prepared all display elements, with oversight and supervision provided by RWK and MJR. All authors reviewed the manuscript and approved their contributions to the work. Funding The current research was supported by an Australian Government Research Training Program (RTP) Scholarship, granted to the first author. Data availability The de-identified dataset generated and analysed during this study is available at: https://doi.org/10.6084/m9.figshare.29288978 . Declarations Competing interests The authors have no relevant financial or non-financial conflicts of interest to disclose. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References 1. Saguy, T., Reifen-Tagar, M. & Joel, D. The gender-binary cycle: the perpetual relations between a biological-essentialist view of gender, gender ideology, and gender-labelling and sorting. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 376 , 20200141. 10.1098/rstb.2020.0141 (2021). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 2. Morgenroth, T., Kirby, T. A. & van der Toorn, J. Heteroprofessionalism: The power of the gender/sex binary in the workplace. Curr. Opin. Psychol. 60 , 101908. 10.1016/j.copsyc.2024.101908 (2024). [ DOI ] [ PubMed ] [ Google Scholar ] 3. Skewes, L., Fine, C. & Haslam, N. Beyond Mars and Venus: The role of gender essentialism in support for gender inequality and backlash. PLoS ONE 13 , e0200921. 10.1371/journal.pone.0200921 (2018). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 4. Bartels, M. S., Le Forestier, J. M., Hug, A. M. O., Morgenroth, T. & Roselló-Peñaloza, M. The influence of essentialist and social constructionist notions on perceptions of “realness”: Implications for LGBTIQ+ experiences. J. Soc. Issues. 80 , 843–870. 10.1111/josi.12642 (2024). [ Google Scholar ] 5. Gallagher, N. M. & Bodenhausen, G. V. Gender essentialism and the mental representation of transgender women and men: A multimethod investigation of stereotype content. Cognition 217 , 104887. 10.1016/j.cognition.2021.104887 (2021). [ DOI ] [ PubMed ] [ Google Scholar ] 6. Chen, R. & White, F. A. The future of prejudice reduction research: A critical review of the role of virtual reality (VR). Comput. Hum. Behav. 152 , 108073. 10.1016/j.chb.2023.108073 (2024). [ Google Scholar ] 7. Nikolaou, A., Schwabe, A. & Boomgaarden, H. Changing social attitudes with virtual reality: A systematic review and meta-analysis. Ann. Int. Commun. Assoc. 46 , 30–61. 10.1080/23808985.2022.2064324 (2022). [ Google Scholar ] 8. Tassinari, M., Aulbach, M. B. & Jasinskaja-Lahti, I. The use of virtual reality in studying prejudice and its reduction: A systematic review. PLoS ONE 17 , e0270748. 10.1371/journal.pone.0270748 (2022). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 9. Michelson, M. R. & Harrison, B. F. Ties that bind: The effects of transgender contact on transphobia. J. Homosex. 70 , 2848–2900. 10.1080/00918369.2022.2081524 (2023). [ DOI ] [ PubMed ] [ Google Scholar ] 10. Moss-Racusin, C. A. & Rabasco, H. Reducing gender identity bias through imagined intergroup contact. J. Appl. Soc. Psychol. 48 , 457–474. 10.1111/jasp.12525 (2018). [ Google Scholar ] 11. Massey, Z. B., Wong, N. C. H. & Barbati, J. L. Meeting the (trans)parent: Test of poarasocial contact with transgender characters on reducing stigma toward transgender people. Commun. Stud. 72 , 232–250. 10.1080/10510974.2021.1876125 (2021). [ Google Scholar ] 12. Ruprecht, K. A., Dunlop, W. A., Wah, E., Phillips, C. & Martin, S. J. Intergroup contact improves medical student attitudes and skill in transgender health care. Transgend. Health 9 , 162–173. 10.1089/trgh.2021.0203 (2024). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 13. MacInnis, C. C. & Page-Gould, E. How can intergroup interaction be bad if intergroup contact is good? Exploring and reconciling an apparent paradox in the science of intergroup relations. Perspect. Psychol. Sci. 10 , 307–327. 10.1177/1745691614568482 (2015). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 14. White, F. A. et al. Beyond direct contact: The theoretical and societal relevance of indirect contact for improving intergroup relations. J. Soc. Issues 77 , 132–153. 10.1111/josi.12400 (2021). [ Google Scholar ] 15. Boccanfuso, E., White, F. A. & Maunder, R. D. Reducing transgender stigma via an e-contact intervention. Sex Roles 84 , 326–336. 10.1007/s11199-020-01171-9 (2021). [ Google Scholar ] 16. Herrera, F., Bailenson, J., Weisz, E., Ogle, E. & Zaki, J. Building long-term empathy: A large-scale comparison of traditional and virtual reality perspective-taking. PLoS ONE 13 , e0204494. 10.1371/journal.pone.0204494 (2018). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 17. Crone, C. L. & Kallen, R. W. Interview with an avatar: Comparing online and virtual reality perspective taking for gender bias in STEM hiring decisions. PLoS ONE 17 , e0269430. 10.1371/journal.pone.0269430 (2022). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 18. Kilteni, K., Groten, R. & Slater, M. The sense of embodiment in virtual reality. Presence 21 , 373–387. 10.1162/PRES_a_00124 (2012). [ Google Scholar ] 19. Kilteni, K., Maselli, A., Kording, K. P. & Slater, M. Over my fake body: Body ownership illusions for studying the multisensory basis of own-body perception. Front. Hum. Neurosci. 9 , 141. 10.3389/fnhum.2015.00141 (2015). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 20. Lopez, S. et al. Investigating implicit gender bias and embodiment of White males in virtual reality with full body visuomotor synchrony. CHI Conference on Human Factors in Computing Systems Proceedings , 1–12, 10.1145/3290605.3300787 (2019). 21. Schulze, S., Pence, T., Irvine, N. & Guinn, C. The effects of embodiment in virtual reality on implicit gender bias. HCII: International Conference on Human-Computer Interaction: Virtual, Augmented and Mixed Reality. Multimodal Interaction , 361–374 (2019). 22. Khanna, J., Lee, A. S. & Kalantari, S. The mediating role of technology affordances in virtual reality perspective-taking: A systematic literature review. Virtual Real. 29 , 66. 10.1007/s10055-024-01095-9 (2025). [ Google Scholar ] 23. Pettigrew, T. F. & Tropp, L. R. How does intergroup contact reduce prejudice? Meta-analytic tests of three mediators. Eur. J. Soc. Psychol. 38 , 922–934. 10.1002/ejsp.504 (2008). [ Google Scholar ] 24. Chen, V. H. H. & Ibasco, G. C. All it takes is empathy: How virtual reality perspective-taking influences intergroup attitudes and stereotypes. Front. Psychol. 14 , 1265284. 10.3389/fpsyg.2023.1265284 (2023). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 25. Carpenter, M. G., Murnaghan, C. D. & Inglis, J. T. Shifting the balance: Evidence of an exploratory role for postural sway. Neuroscience 171 , 196–204. 10.1016/j.neuroscience.2010.08.030 (2010). [ DOI ] [ PubMed ] [ Google Scholar ] 26. Bardy, B. G., Oullier, O., Lagarde, J. & Stoffregen, T. A. On perturbation and pattern coexistence in postural coordination dynamics. J. Mot. Behav. 39 , 326–336. 10.3200/jmbr.39.4.326-336 (2007). [ DOI ] [ PubMed ] [ Google Scholar ] 27. Manor, B. et al. Physiological complexity and system adaptability: Evidence from postural control dynamics of older adults. J. Appl. Physiol. 109 , 1786–1791. 10.1152/japplphysiol.00390.2010 (2010). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 28. Shockley, K., Santana, M. V. & Fowler, C. A. Mutual interpersonal postural constraints are involved in cooperative conversation. J. Exp. Psychol. Hum. Percept. Perform. 29 , 326–332. 10.1037/0096-1523.29.2.326 (2003). [ DOI ] [ PubMed ] [ Google Scholar ] 29. Richardson, D. C., Dale, R. & Shockley, K. Synchrony and swing in conversation: Coordination, temporal dynamics, and communication. in Embodied communication in humans and machines (eds Wachsmuth, I., Lenzen, M., & Knoblich, G.) 75–94 (Oxford University Press, 2008). 30. Delignières, D., Fortes, M. & Ninot, G. The fractal dynamics of self-esteem and physical self. Nonlinear. Dyn. Psychol. Life. Sci. 8 , 479–510 (2004). [ PubMed ] [ Google Scholar ] 31. Correll, J. 1/f noise and effort on implicit measures of bias. J. Pers. Soc. Psychol. 94 , 48–59. 10.1037/0022-3514.94.1.48 (2008). [ DOI ] [ PubMed ] [ Google Scholar ] 32. Richardson, M. J., Dale, R. & Marsh, K. L. Complex dynamical systems in social and personality psychology: Theory, modeling, and analysis. in Handbook of research methods in social and personality psychology, 2nd ed. (eds Reis, H. T. & Judd, C. M.) 253–282 (Cambridge University Press, 2014). 33. Ihlen, E. A. Introduction to multifractal detrended fluctuation analysis in MATLAB. Front. Physiol. 3 , 141. 10.3389/fphys.2012.00141 (2012). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 34. Eke, A. et al. Physiological time series: Distinguishing fractal noises from motions. Pflügers Arch. Eur. J. Physiol. 439 , 403–415. 10.1007/s004249900135 (2000). [ DOI ] [ PubMed ] [ Google Scholar ] 35. Hayes, A. F. Introduction to mediation, moderation, and conditional process analysis: A regression-based approach, 3rd ed. , (Guildford Press, 2022). 36. Habarth, J. M. Development of the heteronormative attitudes and beliefs scale. Psychol. Sex. 6 , 166–188. 10.1080/19419899.2013.876444 (2015). [ Google Scholar ] 37. Kauff, M. et al. How do we get people into contact? Predictors of intergroup contact and drivers of contact seeking. J. Soc. Issues. 77 , 38–63. 10.1111/josi.12398 (2021). [ Google Scholar ] 38. Reysen, S. Construction of a new scale: The Reysen likability scale. Soc. Behav. Pers. 33 , 201–208. 10.2224/sbp.2005.33.2.201 (2005). [ Google Scholar ] 39. Monahan, J. L. I don’t know it but I like you: The influence of non-conscious affect on person perception.. Hum. Commun. Res. 24 , 480–500. 10.1111/j.1468-2958.1998.tb00428.x (1998). [ Google Scholar ] 40. Liviatan, I., Trope, Y. & Liberman, N. Interpersonal similarity as a social distance dimension: Implications for perception of others’ actions. J. Exp. Soc. Psychol. 44 , 1256–1269. 10.1016/j.jesp.2008.04.007 (2008). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 41. Riley, M. A. & Turvey, M. T. Variability and determinism in motor behavior.. J. Mot. Behav. 34 , 99–125. 10.1080/00222890209601934 (2002). [ DOI ] [ PubMed ] [ Google Scholar ] 42. Son Hing, L. S. et al. Gender inequities in the workplace: A holistic review of organizational processes and practices.. Hum. Resour. Manag. Rev. 33 , 100968. 10.1016/j.hrmr.2023.100968 (2023). [ Google Scholar ] 43. Van Laar, C., Van Rossum, A., Kosakowska-Berezecka, N., Bongiorno, R. & Block, K. Mandatory: Why men need (and are needed for) gender equality progress.. Front. Psychol. 15 , 1263313. 10.3389/fpsyg.2024.1263313 (2024). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 44. Fath, S., Ma, A. & Rosette, A. S. Self-views of disadvantage and success impact perceptions of privilege among White men.. Organ. Behav. Hum. Decis. Process. 169 , 104114. 10.1016/j.obhdp.2021.104114 (2022). [ Google Scholar ] 45. Schumann, S., Klein, O., Douglas, K. & Hewstone, M. When is computer-mediated intergroup contact most promising? Examining the effect of out-group members’ anonymity on prejudice. Comput. Hum. Behav. 77 , 198–210. 10.1016/j.chb.2017.08.006 (2017). [ Google Scholar ] 46. Gilam, G. & Hendler, T. With love, from me to you: Embedding social interactions in affective neuroscience. Neurosci. Biobehav. Rev. 68 , 590–601. 10.1016/j.neubiorev.2016.06.027 (2016). [ DOI ] [ PubMed ] [ Google Scholar ] 47. Madhavan, P. & Wiegmann, D. A. Similarities and differences between human–human and human–automation trust: An integrative review. Theor. Issues Ergon. Sci. 8 , 277–301. 10.1080/14639220500337708 (2007). [ Google Scholar ] 48. O’Donnell, A. W. et al. Technological and analytical advancements in intergroup contact research.. J. Soc. Issues 77 , 171–196. 10.1111/josi.12424 (2021). [ Google Scholar ] 49. Moore, S. M. & Geuss, M. N. Familiarity with teammate’s attitudes improves team performance in virtual reality. PLoS ONE 15 , e0241011. 10.1371/journal.pone.0241011 (2020). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 50. Wright, S., Brody, S. & Aron, A. Intergroup contact: Still our best hope for reducing prejudice. in The social psychology of prejudice: Historical perspectives (eds Crandall, C. S. & Schaller, M.) 115–142 (Lewinian Press, 2005). 51. Davies, K., Tropp, L. R., Aron, A., Pettigrew, T. F. & Wright, S. C. Cross-group friendships and intergroup attitudes: A meta-analytic review. Pers. Soc. Psychol. Rev. 15 , 332–351. 10.1177/1088868311411103 (2011). [ DOI ] [ PubMed ] [ Google Scholar ] 52. Mogotsi, I. et al. Psychosocial correlates of LGBTIQ+ experiences in selected African countries: Reimagining LGBTIQ+ research. J. Soc. Issues. 80 , 1079–1111. 10.1111/josi.12640 (2024). [ Google Scholar ] 53. Ahuwalia, D. C., Singh, P., Jami, H., Ummak, E. & Osin, E. From oppressive to affirmative: Situating the health and well-being of LGBTIQ+ people as impacted by systemic and structural transitions in Russia, Turkey, Pakistan, and India. J. Soc. Issues. 80 , 1056–1078. 10.1111/josi.12644 (2024). [ Google Scholar ] 54. Borinca, I., Koc, Y. & Mustafa, S. Fostering social cohesion in post-conflict societies: The power of normative apologies in reducing competitive victimhood and enhancing reconciliation and intergroup negotiation. Eur. J. Soc. Psychol. 0 , 1–12. 10.1002/ejsp.3116 (2024). [ Google Scholar ] 55. Van Assche, J. et al. Intergroup contact is reliably associated with reduced prejudice, even in the face of group threat and discrimination. Am. Psychol. 78 , 761–774. 10.1037/amp0001144 (2023). [ DOI ] [ PubMed ] [ Google Scholar ] 56. Brauer, M. Stuck on intergroup attitudes: The need to shift gears to change intergroup behaviors. Perspect. Psychol. Sci. 19 , 280–294. 10.1177/17456916231185775 (2024). [ DOI ] [ PubMed ] [ Google Scholar ] 57. McConnell, E. A., Janulis, P., Phillips, G., Truong, R. & Birkett, M. Multiple minority stress and LGBT community resilience among sexual minority men. Psychol. Sex. Orientat. Gend. Divers. 5 , 1–12. 10.1037/sgd0000265 (2018). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 58. Cross, K. J., Farrell, S., Chavela, R. & Tsugawa, M. Queerness in science, technology, engineering, and mathematics (STEM): Insights and foresights from experienced lesbian, gay, bisexual, transgender, queer, and plus (LGBTQ+) advocates. J. Div. Higher Edu. 10.1037/dhe0000528 (2023). [ Google Scholar ] 59. Cech, E. A. & Waidzunas, T. J. Systemic inequalities for LGBTQ professionals in STEM. Sci. Adv. 7 , eabe0933. 10.1126/sciadv.abe0933 (2021). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 60. Gonzalez-Franco, M. & Peck, T. C. Avatar embodiment. Towards a standardized questionnaire. Front. Robot. AI 5 , 74. 10.3389/frobt.2018.00074 (2018). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 61. Crone, C. L. & Kallen, R. W. Measuring virtual embodiment: A psychometric investigation of a standardised questionnaire for the psychological sciences. Comput. Hum. Behav. Rep. 14 , 100422. 10.1016/j.chbr.2024.100422 (2024). [ Google Scholar ] 62. Crowne, D. P. & Marlowe, D. A new scale of social desirability independent of psychopathology. J. Consult. Psychol. 24 , 349–354. 10.1037/h0047358 (1960). [ DOI ] [ PubMed ] [ Google Scholar ] 63. Helmes, E., Holden, R. R. & Ziegler, M. Response bias, malingering, and impression management. In Measures of Personality and Social Psychological Constructs (eds Boyle, G. J. et al.) 16–43 (Academic Press, 2015). [ Google Scholar ] 64. Reniers, R. L., Corcoran, R., Drake, R., Shryane, N. M. & Völlm, B. A. The QCAE: A questionnaire of cognitive and affective empathy. J. Pers. Assess. 93 , 84–95. 10.1080/00223891.2010.528484 (2011). [ DOI ] [ PubMed ] [ Google Scholar ] 65. Gomez, R., Brown, T., Watson, S. & Stavropoulos, V. Confirmatory factor analysis and exploratory structural equation modeling of the factor structure of the questionnaire of cognitive and affective empathy (QCAE). PLoS ONE 17 , e0261914. 10.1371/journal.pone.0261914 (2022). [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 66. Batson, C. D., Early, S. & Salvarani, G. Perspective taking: Imagining how another feels versus imaging how you would feel. Pers. Soc. Psychol. Bull. 23 , 751–758. 10.1177/0146167297237008 (1997). [ Google Scholar ] 67. Moss-Racusin, C. A. et al. Reducing STEM gender bias with VIDS (video interventions for diversity in STEM). J. Exp. Psychol. Appl. 24 , 236–260. 10.1037/xap0000144 (2018). [ DOI ] [ PubMed ] [ Google Scholar ] 68. Douglas, H. M., Toohey, S., Richardson, M. J. & Kallen, R. W. Concealable stigmatized identity disclosure as a possible perturbation to complex social systems. Proc. Annu. Meet. Cog. Sci. Soc. 5 , 454–460 (2020). [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Supplementary Material 1 (768.6KB, pdf) Data Availability Statement The de-identified dataset generated and analysed during this study is available at: https://doi.org/10.6084/m9.figshare.29288978 . Articles from Scientific Reports are provided here courtesy of Nature Publishing Group ACTIONS View on publisher site PDF (8.1 MB) Cite Collections Permalink RESOURCES Similar articles Cited by other articles Links to NCBI Databases Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top

Related documents

• A queer duoethnography of Final Fantasy VII#86449
• Virtual Reality Device for Rehabilitation of Stroke Patients#979235
• Virtual Reality Distraction for Pain and Anxiety in Hemodialysis Patients#879924
• Exergames for women with fibromyalgia: A randomised controlled trial to evaluate the effects on mobility skills, balance and fear of falling#947515
• Video Gaming and Psychological Well-Being Among Sexual and Gender Diverse Youth in Canada, the United States, Mexico, the United Kingdom, and Australia: Protocol for a LEVEL UP! Cross-Sectional Survey and Screenshot Elicitation Interview Study.#1091448