The Proteus efect in virtual reality (VR) refers to the phenomenon where a user's behavior and perception are influenced by the characteristics of their avatar, often leading to changes aligned with the avatar's stereotypes. While this efect can enhance user experience, it may also result in unintended negative behaviors, such as increased aggression or bias. This study investigates methods to both enhance and mitigate the Proteus efect by introducing a Mental and Acting Protocol prior to avatar embodiment. The protocol consists of three stages: Introduction, Mental Imagery, and Acting. We hypothesized that providing users with information that contradicts the avatar's associated stereotypes would reduce the Proteus efect. A user study with 68 participants tested this hypothesis using avatars of elderly individuals, with walking speed as the primary behavioral measure. Contrary to expectations, the results showed no significant diferences in the Proteus efect across the diferent conditions. These findings contribute to the ongoing discussion on the variability of the Proteus efect in VR and provide guidelines for improving future research on this phenomenon.
tions of age and physical capability. ical responses, self-concept, and creativity [10]. Other fac
H1. The stereotype-contradicting (opposite) Mental tors, such as self-esteem and the illusion of embodiment, and Acting Protocol mitigates the Proteus efect of slower can modulate the Proteus efect in exergames featuring walking that appears when using elder avatars. Priming celebrity avatars [11]. An experiment using the avatar users with a protocol that contradicts age-related stereo- of the famous celebrity Kim Kardashian, known for her types can mitigate the typical Proteus efect seen with narcissistic behavior and materialistic purchases, showed elder avatars, leading to faster walking speeds than young that individuals might only adopt the desirable aspects avatars. of the avatars they embody while distancing themselves
H2. The stereotype-consistent Mental and Acting Pro- from the undesirable aspects [12]. tocol enhances the Proteus efect. Directly reinforcing It has also been shown that the average efect size of the stereotypes associated with aging (such as decreased mo- Proteus efect is small to moderate [ 13, 14]. In contrast, bility) might amplify the manifestation of those stereo- the Proteus efect of avatars with negative connotations, types, resulting in slower walking speeds. such as the black cloaks and dressing gowns of Ku Klux
H3. The stereotype-consistent Mental and Acting Pro- Klan (KKK), a secret society of white supremacy in the tocol increases the Sense of Embodiment (SoE), while USA, unconsciously teaches users negative beliefs, leadthe stereotype-contradicting Mental and Acting Protocol ing to antisocial intentions and aggressive thinking [10]. mitigates the SoE. Over-preparation or specific stereo- These negative aspects highlight the necessity of findtype reinforcement might detach the user from a sense of ing ways to not only enhance the Proteus efect but also natural avatar embodiment due to cognitive dissonance alleviate it. or overthinking the role.
Our contributions are outlined as follows:
• An experimental protocol to study the impact of Mental imagery exercises have been shown to guide imagthe Mental and Acting Protocol on mitigating the ination, promote perspective-taking abilities, and conProteus efect. tribute to empathy-related skills [38]. Engaging in just • A description of the lack of influence of the Men- ten minutes of mindfulness can significantly reduce bital and Acting Protocol on the Proteus efect. Non- ases towards diferent social groups, enhance empathy, significant diferences showed the following ef- and potentially improve SoE [39, 40]. Additionally, acting fect sizes: Young avatar condition = 0.415, activities can serve as efective methods for promoting Elder avatar condition = 0.0277, Elder avatar perspective-taking abilities [41]. Role-playing techniques with the stereotype-consistent Mental and Acting have been found useful in therapy and conflict resoluProtocol condition = 0.549, and for the Elder tion [42]. avatar with the stereotype-contradicting Mental A previous study proposed the Mental and Acting Proand Acting Protocol condition = 0.776. tocol for acting and mental imagery [7]. It was hypoth• Recommendations for conducting user studies on esized that providing prior knowledge about the avatar the efect of the Mental and Acting Protocol on would improve the SoE and the Proteus efect compared the Proteus efect. to not providing prior knowledge. This protocol was
Avatars can impact user behavior and perception through what is termed the Proteus efect, where users align their identity and behavior with the stereotypes linked to their avatars’ appearance [2]. A meta-analysis identified replications of the Proteus efect using various avatars and contexts [10]. These studies categorized avatars into two main groups: 1) bodily features, which encompass traits such as gender, race, or attractiveness, and 2) characters, which include the reference or use of famous individuals, uniforms, or even non-human figures Table 1.
These experiments have shown that the Proteus efect may influence psychological aspects such as users’ cognitive abilities, emotional states, motor behavior, physiologtested using the Hulk avatar from the Marvel series. Questionnaires were used to evaluate SoE, and punch speed, emotional state, and grip strength when destroying objects were measured as indicators of the Proteus efect. However, the protocol only enhanced the Proteus efect for participants familiar with the avatar before the experiment.
Given these findings, it is promising that by employing methods such as mental imagery and role-playing, we can efectively negate the stereotypes typically reinforced by the Proteus efect. These approaches are specifically used in our study to counteract the negative stereotypes associated with avatar characteristics, aiming to reduce biases and enhance empathy among users.
approximately ten minutes, and aids users in immersing themselves in their avatars. We modified the original Mental and Acting Protocol to make it either consistent or contradicting to the stereotype, depending on the desired efect of enhancing or diminishing the Proteus efect.
In the stereotype-consistent version of the protocol, scenarios were created to instruct participants about declining motor abilities. Conversely, scenarios were designed to instruct participants about superior motor abilities in the stereotype-contradicting version of the protocol. In both cases, we used the following three stages: 1. Introduction: Participants are introduced to their avatars through detailed images and videos, illustrating their identity and inherent behavioral characteristics. These visual materials include representations of the avatar’s physical and behavioral traits, aiming to modify participants’ perceptions to either align with or counter the stereotypes typically associated with elderly individuals. For those in the stereotype-consistent protocol, videos showcased characteristics such as dificulty in moving one’s body and elderly individuals walking with canes, emphasizing tra- 3.2. System ditional stereotypes of aging, such as slower physical movements and dependence. Conversely, the An HTC Vive Pro connected to a high-performance PC stereotype-contradicting protocol ofered videos configuration was employed for our setup. This PC feafeaturing top elderly athletes exhibiting power- tured an NVIDIA GeForce RTX 2080 Ti graphics card, an ful, enthusiastic behaviors and exceptional motor Intel Core i9-9900K processor, and 32 GB of RAM. We skills, representing counter-stereotypical traits positioned HTC lighthouses within a room (Room B in such as vitality and active lifestyles. These Figure 1) measuring approximately two square meters at videos were sourced from free platforms like opposing corners to facilitate tracking and immersion. YouTube [43] and Japan Broadcasting Corpora- Participants were outfitted with the VR headset, two tion [44], ensuring a broad representation of the hand controllers, and two Vive trackers attached to their feet, enabling 5-point body tracking. The experimentraits described. One is a video of a 102-yearold man participating in track and field events, such as the javelin and shot put, and the other is a video released by the Japan Broadcasting Corporation, a Japanese television station, showing elderly Japanese people. 2. Mental Imagery: Participants engage in a guided mental imagery session designed to deepen their connection with their avatars. They are directed through an audio guide to envision themselves embodying the avatar, imagining nine scenarios that depict various aspects of a day in the life of an elderly person, such as “Having breakfast with the help of a caregiver” for those in the stereotype-consistent preparation group, or “Playing soccer with grandchildren, dribbling energetically and scoring goals” for those in the stereotype-contradicting group. Each scenario has an audio description and a 20-second imagination period, repeated twice to reinforce the imagery. Depending on their experimental group, participants either envision themselves with diminished capabilities that align with common age-related stereotypes or with enhanced abilities that challenge these stereotypes. 3. Acting: Brief audio instructions lead participants to perform five actions. For those in the stereotype-consistent group, these actions might include scenarios like “Climbing stairs, expressing the struggle but eventually reaching the top,” which simulate the challenges of aging, such as moving slowly or with dificulty. Conversely, for participants in the stereotype-contradicting group, the actions involve vigorous activities like “Engaging in exercise with energy and vigor,” representing tasks not typically associated with elderly avatars, such as jogging or dancing. Each action is accompanied by a 30-second acting period to deepen the embodiment experience, reinforcing the intended perception of the avatar’s capabilities or limitations.
Welcome and
Consent Hallway
33.7 m .518m .733m
Room A
Demographics
Walk
Preparation
VR : Free time
Walk
Survey VR : Visuomotor stimulation
VR : VEQ Walk tal software was developed using Unity 2020.1.4f1, integrating SteamVR and the FinalIK v2.1 plugin created by Pärtel Lang. Additionally, a C# script was employed to capture the motion of the hand controllers. To present questionnaire material within the VR environment and record user responses, we leveraged the VR Questionnaire Toolkit [45].
The VR environment was constructed using free room assets (See Figure 2). A mirror was placed in front of the user to observe his or her reflection in the virtual environment.
Preliminary checks were conducted to determine whether there was a diference in the perceived age of the elder avatars and the young avatars created. Fifteen people (five males, ten females), diferent from the participants in the main experiment, were asked to guess the ages of four types of avatars in a within-subjects 1 1 2 2 3 3 4 4 experiment. Shapiro-Wilk tests showed that the residuals were not normally distributed for Agency ( = 0.839, = 3.72 × 10− 4) and between female elderly and female young avatars ( = 0.866, = 1.40 × 10− 3). Wilcoxon rank sum tests were conducted to check the diference in the perceived age for each gender representation. The results of the pilot study showed significant diferences between male elder and male young avatars ( = 225, = 1.37 × 10− 6, 2 = 0.886) and between female elder and female young avatars ( = 222, = 4.02 × 10− 6, 2 = 0.846), confirming that they are avatars of diferent generations (See Figure 4).
The avatar movements underwent a two-step calibration process. First, the user was instructed to stand upright while wearing the VR equipment so that the experimenter could scale their avatar’s height using the headset’s tracked position. Subsequently, the VRIK Calibration script from the FinalIK plugin was used to calibrate arm and leg lengths and to position the user’s ***
*** Elder male
Young male
Elder female Avatar
Young female camera. Complementary adjustments were executed by manipulating the transforms of the target points associated with the head, hands, and feet, if necessary. Finally, the avatar’s height was set to match the average height of Japanese individuals, 170 cm for men and 157 cm for women.
• Virtual Embodiment Questionnaire (VEQ): The VEQ is a 12-item questionnaire developed by Roth and Latoschik [46]. We used the Japanese translated version published by Roth, one of the authors [47]. The questionnaire provides a reliable and consistent way to assess the components of virtual embodiment, which are considered influencing factors of the Proteus effect [48, 49, 50, 21, 19]. The questionnaire was administered while in VR. • Custom Questionnaire: Supplementary data were collected through a 7-item custom questionnaire on a 7-point Likert scale. It included a measure of the avatar’s age perception. This questionnaire also included an optional comment section, allowing participants to provide openended feedback. It was administered after the session, outside of the virtual environment. The following is the detailed content.
– Do you feel that you are more athletic now than you usually are? Or do you feel like you are declining? 1:Feeling diminished, 7:Feeling superior – Compared to your own motor skills before the experiment, how much do you expect your motor skills to have changed now? 1:Decline diminished, 7:Improvement. – How do you feel the avatar you used affected your walking speed? Open-ended. – How good did the avatars you used look? Multiple-choice: under 10, teens, 20s, 30s, 40s, 50s, 60s, 70s, and 80s above. – Please describe any comments you have.
Open-ended. • Walking speed: To detect the Proteus Efect, we measured the 3D positions of the participants’ feet before and after the VR embodiment. The participants’ feet positions were recorded at a sampling rate of 30 Hz using mocopi sensors from Sony. Measurements were initially taken over a distance of 24 meters, as done in previous studies [21]. However, to ensure more consistent speed measurements across conditions and to better highlight the diferences between them, we focused on the longest straight segment of 18.5 m. Walking speed was measured by asking participants to walk through a corridor to another room (see Figure 1(left)). The participant was unaware that walking speed was being measured. We calculated the walking speed by dividing the distance moved by the time taken to move in meters per second. • Demographic information: We collected demographic information such as age, gender, prior exposure to VR technologies, and weekly video game playtime.
We calculated the required total sample size assuming an alpha error probability of 0.05 and a desired test power of 0.8, given an expected efect size of ( 2 = 0.25). The computed necessary sample size to achieve the desired power was 116 participants. However, recruiting that number in the time constraints we had was unrealistic, so we referred to previous research to determine the number of participants for each condition [7]. Sixty-eight volunteers aged 18 to 63 ( = 26.9, = 9.49) participated in the study. Each participant took part in a diferent condition, with each condition including ten male and seven female participants. Among these participants, 42.6% had used a VR headset once or twice, 14.7% used it several times a year, 1.5% used it frequently but not anymore, 32.4% had never used it, and 8.8% had frequent experience with a VR headset. Participants were recruited through the laboratory’s social media networks and a dedicated recruitment website [51] and were paid 1000 JPY in cash. All participants were unaware of the experiment’s objective and were randomly assigned to one of the four conditions. This experiment was conducted with the approval of the local ethics committee.
The experiment employed a between-subjects design with four conditions to explore the Proteus efect: • Conditions Y, E: Participants sat in a chair and engaged in solving puzzles for 12 minutes. Participants then started the VR experience without prior Mental and Acting Protocol about the avatars. • Condition EC: Participants received Mental and
Acting Protocol following the stereotype for 12 minutes via video instructions. • Condition EO: Participants received Mental and
Acting Protocol contrary to the stereotype for 12 minutes via video instructions. • Condition Y (Young avatar with puzzle): Participants embodied a young avatar and completed a neutral puzzle-solving task before the VR experience. • Condition E (Elder avatar with puzzle): Participants embodied an elder avatar and, like Condition Y, completed a puzzle-solving task prior to the VR experience. • Condition EC (Elder avatar with the The experimenter returned to the room at the end of stereotype-consistent protocol): Participants the task. Then, for all conditions, the participants were embodied an elder avatar and underwent a Men- immersed in VR and performed the same visuomotor tal and Acting Protocol reinforcing age-related stimulation Figure 1. During this time, the experimenter stereotypes before the VR session. remained in the same room with the participants, en• Condition EO (Elder avatar with the suring safety but facing away from them. Participants stereotype-contradicting protocol): Partici- were then asked to look at a virtual mirror and reproduce pants embodied an elder avatar and completed a gestures depicted in videos placed beside it. These gesMental and Acting Protocol that introduced in- tures involved simple arm and leg movements to induce formation contradicting age-related stereotypes visuomotor stimulation, such as raising arms or lifting before the VR session. legs, and were repeated for two minutes. After this induction phase, participants were instructed to spend the
The puzzle-solving task in Conditions Y and E served remaining time freely. This free time was provided to as a control activity, ensuring consistency in the duration ensure participants had ample time to experience the of the experimental procedure across all conditions. avatar, and the VR experience duration, including the induction phase, was set to ten minutes. Finally, partici4. Procedure pants responded to the VEQ using the questionnaire interface within the VR environment. Participants selected The procedure is summarized in Figure 1. In the be- responses using an activated handrail. The experimenter ginning, participants gathered in front of Room B and stopped the experimental software and removed the VR were guided along a walking path to Room A. The ex- equipment from the participants. perimenter followed them from behind. Upon arrival, Afterward, calibration of the mocopi sensors was perparticipants received an explanation of the experimental formed again for gait measurement. Participants were procedure and provided consent to participate. Subse- then instructed to return to Room A independently, folquently, participants completed a demographic question- lowing the path they took to Room B. Upon arrival, parnaire. ticipants removed the mocopi sensors and were asked to
Next, mocopi motion tracking sensors from Sony were respond to a post-experiment questionnaire. attached to the participants’ heads, wrists, ankles, and waist for gait measurement, followed by calibration. The 5. Analysis and Results purpose of the mocopi sensors was not disclosed to the participants; they were simply informed that they would 5.1. Proteus Efect need to wear the device throughout the entire experiment.
After the sensors were fitted, participants were instructed We extracted the straight 18.5m section immediately after to walk back to Room B, following the same path as exiting Room B from the measured 3D feet positions before, without guidance. This step was taken to measure (see Figure 1) and calculated the walking speed. The their normal walking speed before embodiment. mid-point of both feet is used as the user’s position. To
During the preparation phase, the experimenter left account for the persistence of the Proteus efect, a straight Room B and the participant completed the task alone to line was measured immediately after exit [21]. As a result minimize external influences and distractions. For the of excluding participants with incomplete measurements, rest of the time, the experiment was conducted with the one subject was excluded from the E condition. experimenter present with the participants. We conducted a two-factor analysis with four exper
In Room B, without a VR headset, participants per- imental conditions (Y, E, EC, and EO) and two meaformed one of the following four preparation tasks: surement timings at which walking speed was mea
Y-b Y-a E-b E-a EC-b EC-a EO-b EO-a sured (before and after VR exposure) as independent variables and walking speed as a dependent variable (see Figure 5). A Shapiro-Wilk test confirmed that the residuals significantly deviated from the normal distribution ( = 0.842, < 1.17 × 10− 10). Therefore, we used an aligned rank transform (ART) ANOVA. This test found a significant efect of the two measurement timings at which walking speed was measured, (1, 126) = 14.2, = 2.47 × 10− 4, 2 = 0.101. On the other hand, the main efect of the four conditions of the experiment was not confirmed, (3, 126) = 0.383, = 0.765, 2 = 9.04 × 10− 3. No significant interaction between the timing at which walking speed was measured and the four conditions of the experiment was found, (3, 126) = 0.882, = 0.882, 2 = 5.23 × 10− 3.
Since the Proteus efect was not observed in the E and EC conditions, the fundamental premise for hypothesis verification has been undermined. Thus, H0, H1, and H2 were not confirmed.
culate three distinct scores: Ownership, Agency, and Change [46] (see Figure 6). Shapiro-Wilk tests showed that the residuals were not normally distributed for Agency ( = 0.927, = 6.92 × 10− 4), while they were for Ownership ( = 0.978, = 0.283) and Change ( = 0.977, = 0.249). Levene’s tests showed that the variances are homogeneous ( > 0.05 for all dimensions). Ownership and Change were analyzed using ANOVA, while Agency, which did not meet the assumption of normality, was analyzed using ART ANOVA. However, no significant diferences were found in the main efects or in the interaction of any of the three scores. The main efect of Ownership was insignificant ( (3, 64) = 0.102, = 6 y c n4 e g A 2 0 8 6 e g n4 a h C 2 0
Y Y Y
E E E
EC EC EC
EO EO EO
In comparing the walking speed before and after VR exposure for each condition, the Proteus efect did not appear, and the assumptions were violated. Additionally, no significant diferences were observed in SoE. In this section, we discuss the reasons why no significant results were obtained and ofer guidelines for future studies.
Proteus efect was not observed. We examined how participants perceived the age of their avatars. Data on perceived age was collected via a questionnaire at the end of the experiment. Given the subjective feedback, this perception might have varied among participants.
The perceived age of the avatar was gathered through a post-experiment questionnaire. Figures 7 and 8 show the relationship between the perceived age of their avatars and their relative change in walking speed before and after VR exposure.
A regression slope test was performed on the change
As to why the Proteus efect did not appear during the experiment, the age of the avatars used may not have appeared as intended. According to post-hoc interviews on the perceived age of avatars, the average age of the elder female and male avatars appeared 6.48 and 14 years younger in VR than the average age in the pilot study, respectively. On the other hand, the average age of the young female and male avatars appeared only 3.81 and 1.67 years younger in VR than the average age in the pilot study, respectively. Thus, the appearance of the avatars may have difered significantly between the monitor and the VR HMD, potentially afecting the perceived age. This possibility was dificult to predict in advance. Note that, however, the pilot study used a within-subjects design, while the main experiment employed a between-subjects design, which prevented direct comparison through statistical tests.
Additionally, the cross-race efect, which is the tendency to more easily recognize faces of one’s own racial group [52], may have influenced age recognition in this study. The avatars, while created to look as Asian as possible, were based on European-looking assets, which may have impacted their perceived age.
Avatars appearing younger than 70 years old in VR likely contributed to the insignificant results of the Proteus efect. Studies on the relationship between age and walking speed indicate that it decreases after the age of 70 [53]. This efect may have been further influenced by Japan’s longer life expectancy. In Japan, studies show no reduction in walking speed even at 70 years old [54], and this higher life expectancy compared to other countries may shape perceptions of elderly individuals as relatively active. Previous studies on the Proteus efect using elderly avatars have targeted European and American participants [8, 9]. This suggests that cultural diferences in perceptions of elderly capabilities between Japan and Western countries could have influenced the manifestation of the Proteus efect. in walking speed before and after VR exposure and the perceived age based on the avatar’s appearance. A significant diference in the slope was observed across all 6.2. Experimental Procedure data (Figure 7), indicating a negative correlation ( = 1.24 × 10− 2, = 0.633). These results suggest that the In the Mental and Acting Protocol scenario, the term perceived age did indeed afect walking speed. However, “elderly” was mentioned for avatars in the EC and EO no significant diference in the slope was found in each conditions. In the EO condition, elements contradictof the four conditions (Figure 8). The results for the Y, ing to the stereotype were explained after mentioning E, EC and EO conditions are ( = 0.09, = 0.913), this. While the participants in the EO condition were ( = 0.158, = 0.771), ( = 0.121, = 0.821), and expected to walk faster, they actually walked slower af( = 0.122, = 0.820), respectively. Sample size based ter VR exposure (Figure 5). Perhaps the participants in on a power analysis was 116 participants. After consider- the EO condition were more conscious about the coning the feasibility of recruiting participants and previous trast between the young and the elderly, which may have studies, data was collected from 68 participants in this caused cognitive dissonance. Cognitive dissonance is the experiment. The lack of significant diferences may be phenomenon of attempting to eliminate discomfort or due to the small sample size (16 or 17 participants per anxiety when contradictory information or feelings exist condition). within oneself by justifying or glossing over one or the other [55]. Under the EO conditions, participants used 6. Longitudinal and Detailed Evaluations
Conduct detailed evaluations on how perceived avatar age influences behaviors and follow up long-term to assess persistence of efects, deepening understanding of VR’s impact on behavior.
avatars perceived as “elderly” while introducing elements that contradicted the stereotype (e.g., youthful features). This contrast may have caused cognitive dissonance and discomfort, leading participants to act unconsciously by slowing down their walking speed.
The fact that the experimenter followed the participants during walking speed measurement and that the measurement was conducted after removing the HMD may have also influenced the walking speed. We have no precise models of how the Proteus efect evolves after VR exposure, so the timing and method of measuring walking speed might not have been optimal.
age perception, the measurement environment, and cultural background on the Proteus efect. Incorporating the following guidelines into future studies can significantly enhance the quality and impact of research on the Proteus efect in virtual reality.
the Proteus efect by employing the Mental and Acting Protocol. Our study compared walking speed and Sense of Embodiment (SoE) across three conditions: the Elderly avatar (E) condition, the Elderly avatar with stereotypecontradicting Mental and Acting Protocol (EO) condition, and the Elderly avatar with stereotype-consistent Mental and Acting Protocol (EC) condition. The results did not demonstrate significant diferences, suggesting that the method of providing prior knowledge may not substantially impact the Proteus efect or the Sense of Embodiment.
Additionally, our research has provided insights into 1. Detailed Assessment of Avatar Perception factors that must be considered when designing and conConduct detailed surveys before and after experi- ducting experiments in this field. While this study did not ments to accurately capture variations in partici- succeed in developing efective methods to counteract pant perception of avatar age, ensuring nuanced the Proteus efect, it underscored the importance of conunderstanding and reliable conclusions. sidering various experimental aspects, such as the precise 2. Considering the Visual Medium setup of the avatar appearance and the careful monitorEvaluate avatars using the same hardware for ing of participants’ interactions with their avatars. These both pilot and main studies to ensure consistency insights pave the way for future research to propose in avatar representation across diferent viewing innovative solutions to the challenges posed by the Promediums and to address potential biases. For ex- teus efect. It is necessary that subsequent studies build ample, a pilot study involving an avatar viewed upon these findings, employing refined methodologies on a desktop monitor may yield significantly dif- and expanded theoretical frameworks. By doing so, we ferent results compared to a main study using a can enhance the understanding of avatar-mediated exVR headset. periences in virtual reality and possibly develop more 3. Improvement of Measurement Environment efective strategies to manage the Proteus efect.
Maximize control over experimental environments, ensuring participants perform tasks in- References dependently to minimize experimenter influence. 4. Cultural Considerations in Research [1] A. S. Praetorius, D. Görlich, How avatars influence Assess diferences in perceptions of elderly ac- user behavior: A review on the proteus efect in tivity levels between cultures to understand cul- virtual environments and video games, in: Proceedtural impacts on the Proteus efect, broadening ings of the 15th International Conference on the the applicability of findings. As the perception of Foundations of Digital Games, 2020, pp. 1–9. elderly avatars used in this study difers between [2] N. Yee, J. Bailenson, The proteus efect: The efJapan and Western countries, cultural diferences fect of transformed self-representation on behavior, may influence the manifestation of the Proteus Human communication research 33 (2007) 271–290. efect. [3] D. A. De Vries, J. Peter, Women on display: The 5. Optimizing Data Collection Timing efect of portraying the self online on women’s selfDevelop models to understand how the Proteus objectification, Computers in Human Behavior 29 efect evolves post-VR exposure for optimal tim- (2013) 1483–1489. ing of data collection, enhancing measurement [4] J. Fox, J. N. Bailenson, L. Tricase, The embodiment precision. of sexualized virtual selves: The proteus efect and experiences of self-objectification via avatars, Com- 1093.
puters in Human Behavior 29 (2013) 930–938. [17] R. Ratan, Y. J. Sah, Leveling up on stereotype threat: [5] L. K. Kaye, C. R. Pennington, J. J. McCann, Do ca- The role of avatar customization and avatar embodsual gaming environments evoke stereotype threat? iment, Computers in Human Behavior 50 (2015) examining the efects of explicit priming and avatar 367–374. gender, Computers in Human Behavior 78 (2018) [18] N. Yee, J. N. Bailenson, N. Ducheneaut, The pro142–150. teus efect: Implications of transformed digital self[6] Z. Sylvia, T. K. King, B. J. Morse, Virtual ideals: representation on online and ofline behavior, ComThe efect of video game play on male body image, munication Research 36 (2009) 285–312.
Computers in Human Behavior 37 (2014) 183–188. [19] N. Yee, J. N. Bailenson, The diference between [7] A. Genay, E. Kimura, M. Hachet, A. Lécuyer, Y. Hi- being and seeing: The relative contribution of selfrao, M. Perusquía-Hernández, U. H., K. Kiyokawa, perception and priming to behavioral changes via Preparing users to embody their avatar in vr: In- digital self-representation, Media Psychology 12 sights on the efects of priming, mental imagery, (2009) 195–209.
and acting on embodiment experiences, 2024. [20] Y. Bian, L. Han, C. Zhou, Y. Chen, F. Gao, The pro[8] S.-C. Yoo, J. F. Peña, M. E. Drumwright, Virtual teus efect in virtual reality social environments: shopping and unconscious persuasion: The prim- Influence of situation and shyness, Acta Psychologing efects of avatar age and consumers’ age dis- ica Sinica 47 (2015) 363. crimination on purchasing and prosocial behaviors, [21] R. Reinhard, K. G. Shah, C. A. Faust-Christmann, Computers in Human Behavior 48 (2015) 62–71. T. Lachmann, Acting your avatar’s age: efects [9] J. A. Bargh, M. Chen, L. Burrows, Automaticity of virtual reality avatar embodiment on real life of social behavior: Direct efects of trait construct walking speed, Media Psychology 23 (2020) 293– and stereotype activation on action., Journal of 315.
personality and social psychology 71 (1996) 230. [22] E. Ash, Priming or proteus efect? examining the [10] A. Martin Coesel, B. Biancardi, S. Buisine, A theo- efects of avatar race on in-game behavior and postretical review of the proteus efect: understanding play aggressive cognition and afect in video games, the underlying processes, Frontiers in Psychol- Games and Culture 11 (2016) 422–440. ogy 15 (2024) 1379599. doi:10.3389/fpsyg.2024. [23] J. Peña, S. Khan, C. Alexopoulos, I am what i see: 1379599. How avatar and opponent agent body size afects [11] N. Sadek, P. Elagroudy, A. Khalil, S. Abdennad- physical activity among men playing exergames, her, The superhero pose: Enhancing physical per- Journal of Computer-Mediated Communication 21 formance in exergames by embodying celebrity (2016) 195–209. avatars in virtual reality, in: Nordic Human- [24] Y. K. Joo, K. Kim, When you exercise your avatar Computer Interaction Conference, 2022, pp. 1–11. in a virtual game: The role of avatars’ body shape [12] J. McCain, S. J. Ahn, W. K. Campbell, Is desirability and behavior in users’ health behavior, Interacting of the trait a boundary condition of the proteus with Computers 29 (2017) 455–466. efect? a pilot study, Communication Research [25] M. F. G. Ferrer, B. Porras-García, M. Moreno, Reports 35 (2018) 445–455. P. Bertomeu Panisello, J. Gutiérrez Maldonado, Em[13] R. Ratan, D. Beyea, B. J. Li, L. Graciano, Avatar bodiment in diferent size virtual bodies produces characteristics induce users’ behavioral conformity changes in women’s body image distortion and diswith small-to-medium efect sizes: a meta-analysis satisfaction, Annual Review of CyberTherapy and of the proteus efect, Media Psychology 23 (2020) Telemedicine, 2018, vol. 16, p. 111-117 (2018). 651–675. [26] A. S. Won, J. Bailenson, J. Lee, J. Lanier, Homuncular [14] D. Beyea, R. Ratan, Y. Lei, H. Liu, G. E. Hales, C. Lim, lfexibility in virtual reality, Journal of ComputerA new meta-analysis of the proteus efect: Studies Mediated Communication 20 (2015) 241–259. in vr find stronger efect sizes, Presence: Virtual [27] L. Hoyet, F. Argelaguet, C. Nicole, A. Lécuyer, and augmented reality 31 (2022) 189–202. “wow! i have six fingers!”: Would you accept struc[15] N. Yee, N. Ducheneaut, M. Yao, L. Nelson, Do men tural changes of your hand in vr?, Frontiers in heal more when in drag? conflicting identity cues Robotics and AI 3 (2016) 27. between user and avatar, in: Proceedings of the [28] W. Steptoe, A. Steed, M. Slater, Human tails: ownSIGCHI conference on Human factors in computing ership and control of extended humanoid avatars, systems, 2011, pp. 773–776. IEEE transactions on visualization and computer [16] R. A. Ratan, M. Dawson, When mii is me: A graphics 19 (2013) 583–590.
psychophysiological examination of avatar self- [29] D. Banakou, S. Kishore, M. Slater, Virtually being relevance, Communication Research 43 (2016) 1065– einstein results in an improvement in cognitive task performance and a decrease in age bias, Frontiers [42] P. Bertrand, J. Guegan, L. Robieux, C. A. McCall, in psychology 9 (2018) 917. F. Zenasni, Learning empathy through virtual real[30] G. Gorisse, S. Wellenreiter, S. Fleury, S. Richir, ity: multiple strategies for training empathy-related O. Christmann, et al., I am a genius! influence of abilities using body ownership illusions in embodvirtually embodying leonardo da vinci on creative ied virtual reality, Frontiers in Robotics and AI performance, IEEE Transactions on Visualization (2018) 26.
and Computer Graphics (2023). [43] 102-Year-Old is a Champion Sprinter, https://www. [31] J. Guegan, S. Buisine, F. Mantelet, N. Maranzana, youtube.com/watch?v=pEPdU3m1tk4, ???? F. Segonds, Avatar-mediated creativity: When em- [44] Japan Broadcasting Corporation, https: bodying inventors makes engineers more creative, //www2.nhk.or.jp/archives/movies/?id= Computers in Human Behavior 61 (2016) 165–175. D0002160910_00000, ???? [32] J. Guegan, F. Segonds, J. Barré, N. Maranzana, [45] M. Feick, N. Kleer, A. Tang, A. Krüger, The virtual F. Mantelet, S. Buisine, Social identity cues to im- reality questionnaire toolkit, in: Adjunct Proceedprove creativity and identification in face-to-face ings of the 33rd Annual ACM Symposium on User and virtual groups, Computers in Human behavior Interface Software and Technology, 2020, pp. 68–69. 77 (2017) 140–147. [46] D. Roth, M. E. Latoschik, Construction of the [33] J. Peña, J. T. Hancock, N. A. Merola, The priming ef- virtual embodiment questionnaire (veq), IEEE fects of avatars in virtual settings, Communication Transactions on Visualization and Computer Graphresearch 36 (2009) 838–856. ics 26 (2020) 3546–3556. doi:10.1109/TVCG.2020. [34] S. Buisine, J. Guegan, J. Barré, F. Segonds, A. Aous- 3023603.
sat, Using avatars to tailor ideation process to in- [47] The Virtual Embodiment Questionnovation strategy, Cognition, Technology & Work naire, https://sites.google.com/view/ 18 (2016) 583–594. virtualembodimentquestionnaire/startseite, [35] S. J. Ahn, J. Bostick, E. Ogle, K. L. Nowak, K. T. ????
McGillicuddy, J. N. Bailenson, Experiencing na- [48] K. Kilteni, R. Groten, M. Slater, The sense of emture: Embodying animals in immersive virtual envi- bodiment in virtual reality, Presence: Teleoperators ronments increases inclusion of nature in self and and Virtual Environments 21 (2012) 373–387. involvement with nature, Journal of Computer- [49] K. Kilteni, I. Bergstrom, M. Slater, Drumming in Mediated Communication 21 (2016) 399–419. immersive virtual reality: the body shapes the way [36] A. Frisanco, M. Schepisi, G. Tieri, S. M. Aglioti, In we play, IEEE transactions on visualization and god’s virtual shoes: embodying the avatar of an om- computer graphics 19 (2013) 597–605. nipotent agent modulates the perception of one’s [50] S. A. Osimo, R. Pizarro, B. Spanlang, M. Slater, own abilities and enhances feelings of invulnerabil- Conversations between self and self as sigmund ity (2022). freud—a virtual body ownership paradigm for self [37] C. Christou, D. Michael, Aliens versus humans: counselling, Scientific reports 5 (2015) 13899. do avatars make a diference in how we play the [51] Collecting participants for the experiment, https: game?, in: 2014 6th International Conference on //www.jikken-baito.com, ???? Games and Virtual Worlds for Serious Applications [52] C. Feingold, The influence of environment on (VS-GAMES), IEEE, 2014, pp. 1–7. identification of persons and things, Journal of [38] L. K. Hildebrandt, C. McCall, T. Singer, Diferen- Criminal Law and Police Science 5 (1914) 39–51. tial efects of attention-, compassion-, and socio- URL: https://www.jstor.org/stable/1133283. doi:10. cognitively based mental practices on self-reports 2307/1133283. of mindfulness and compassion, Mindfulness 8 [53] R. Kozakai, S. Tsuzuku, K. Yabe, F. Ando, N. Niino, (2017) 1488–1512. H. Shimokata, Age-related changes in gait velocity [39] A. Lueke, B. Gibson, Mindfulness meditation re- and leg extension power in middle-aged and elderly duces implicit age and race bias: The role of reduced people, Journal of epidemiology 10 (2000) 77–81. automaticity of responding, Social Psychological [54] H. Kawai, Y. Taniguchi, S. Seino, R. Sakurai, and Personality Science 6 (2015) 284–291. Y. Osuka, S. Obuchi, Y. Watanabe, H. Kim, H. Ina[40] D. Romano, A. Maravita, M. Perugini, Psychometric gaki, A. Kitamura, et al., Reference values of gait properties of the embodiment scale for the rubber parameters measured with a plantar pressure plathand illusion and its relation with individual difer- form in community-dwelling older japanese adults, ences, Scientific reports 11 (2021) 5029. Clinical Interventions in Aging (2019) 1265–1276. [41] D. Whitebread, M. Basilio, M. Kuvalja, M. Verma, [55] L. Festinger, A theory of cognitive dissonance row, The importance of play, Brussels: Toy Industries of Peterson and company (1957).
Europe (2012).
were used to deliver the Mental and Acting Protocol. The participants were briefed about the process before going through it. All procedures were carried out by video. To alleviate embarrassment, the experimenter left the room, and the participants proceeded with the procedure themselves.
A.2.1. Introduction
The experimenter showed pictures of the avatar model.
“This is what you look like in VR. You have the appearance of an elderly person, but you are energetic and active.
The following video introduces such an elderly person.”
An excerpt of the following video was shown: https://www.youtube.com/watch?v=pEPdU3m1tk4 A.2.2. Mental imagery A.1.1. Introduction This script was delivered to participants: “Your appearance is that of an elderly person. Wrinkles This script was delivered to participants: and sagging whitened hair are noticeable. And you are The experimenter showed pictures of the avatar model. powerful and motivated. What I’m about to tell you is a “This is what you look like in VR. You are physically day in your life. In the morning we go out to the park and do disabled. The following video introduces such an elderly pull-ups on the bars. Find a child drowning in a pond and person. ” swam to save him. Drink protein smoothies for breakfast.
An excerpt of the following video was shown: Go to the gym and do full-body workouts and yoga to https://www2.nhk.or.jp/archives/movies strengthen their bodies. Unintentionally bending the spoon /?id=D0002160910_00000 during lunch. Play football with my grandchildren. You dribble through and shoot without any adult supervision.
A.1.2. Mental imagery Carry your tired grandson home on your back. Dinner is served with your favourite steak. You sweat it out in the sauna before bed.” This script was delivered to participants: “Your appearance is that of an elderly person. Wrinkles and sagging whitened hair are noticeable. And you have a slight limp A.2.3. Acting in your movements. What I’m about to tell you is a day in your life. Wake up slowly and have a carer help you to This script was delivered to participants: get up. Breakfast is taken with the help of a carer. Enjoy “You are an energetic elderly person. For acting, change reading while listening to music. With the help of a carer, your mind to be a healthy elderly person. For the act, you they eat a nutritionally balanced diet. Relax and enjoy will do a morning exercise in the park. First, stand with a the view outside. Family members come to visit and enjoy smile on your face. First, perform a spirited appearance. conversation. Light exercises are performed with the help Your fists are clenched tightly, and your back is arched. of a carer. Dinner is taken with the help of a carer. TV You are so fired up that you feel like shouting. Next, the viewing before going to bed .” body warm-up is performed. Unwind your body by doing some light stretching and large arm movements. In doing A.1.3. Acting so, try to make it feel fun and lively. Finally, act out the ifgure engaging in the exercise. Act it out with an energetic image.”
“You are an energetic elderly person. For acting, change your mind to be a healthy elderly person. For the act, you will do a morning exercise in the park. First, stand with a smile on your face. First, perform a spirited appearance. Your fists are clenched tightly, and your back is arched. You are so fired up that you feel like shouting. Next, the body warm-up is performed. Unwind your body by doing some light stretching and large arm movements. In doing so, try to make it feel fun and lively. Finally, act out the ifgure engaging in the exercise. Act it out with an energetic image.” B. Additional feedback questionnaire • Do you feel that you are more athletic now than you usually are? Or do you feel like you are declining? 1:Feeling diminished, 7:Feeling superior • Compared to your own motor skills before the experiment, how much do you expect your motor skills to have changed now? 1:Decline diminished, 7:Improvement • How do you feel the avatar you used afected your walking speed? Open-ended response • How good did the avatars you used look? Multiple-choice: under 10, teens, 20s, 30s, 40s, 50s, 60s, 70s, and 80s above • Please describe any comments you have. Openended response