Doppelganger-based training: Imitating our virtual self to accelerate interpersonal skills learning

Emmanuelle P Kleinlogel; Marion Curdy; João Rodrigues; Carmen Sandi; Marianne Schmid Mast

doi:10.1371/journal.pone.0245960

. 2021 Feb 10;16(2):e0245960. doi: 10.1371/journal.pone.0245960

Doppelganger-based training: Imitating our virtual self to accelerate interpersonal skills learning

Emmanuelle P Kleinlogel ^1,^*, Marion Curdy ², João Rodrigues ², Carmen Sandi ², Marianne Schmid Mast ¹

Editor: Doron Friedman³

PMCID: PMC7875421 PMID: 33566838

Abstract

Interpersonal skills require mastering a wide range of competencies such as communication and adaptation to different situations. Effective training includes the use of videos in which role models perform the desired behaviours such that trainees can learn through behavioural mimicry. However, new technologies allow new ways of designing training. In the present study, given that virtual reality is emerging as a valuable training setting, we compare two different demonstration conditions within virtual reality by investigating the extent to which the use of doppelgangers as role models can boost trainees’ interpersonal skills development as compared to a role model that does not resemble the trainees. We also assess trainees’ level of self-efficacy and gender as potential moderators in this relationship. Participants delivered a speech in front of a virtual audience twice. Before delivering their second speech, they watched a role model giving a speech in front of the same audience. The role model was either their doppelganger or an avatar of the same gender depending on the condition they were randomly assigned to. Results showed that the doppelganger-based training was the most beneficial for male trainees low in self-efficacy. These findings have important implications for training design, suggesting that doppelganger-based training might be effective only for a specific subset of trainees.

Introduction

Interpersonal skills are crucial in today’s society. At work or in our daily life, we are constantly confronted with different social interaction partners, such as recruiters, clients, colleagues, and supervisors at work, classmates and professors at school, and bankers, physicians, or childminders in our personal lives. Many of these interactions (e.g., with recruiters, superiors, or intimate partners) are high-stake in that our future depends on them. Being interpersonally skilled is thus an important asset to possess. However, developing interpersonal skills is not an easy task because it involves mastering several competencies at the same time while taking into account situational factors that are specific to each interaction situation. This complexity is reflected in the definition of interpersonal skills provided by Klein, DeRouin, and Salas [1], presenting these skills as “goal directed behaviours, including communication and relationship-building competencies, employed in interpersonal interaction episodes characterized by complex perceptual and cognitive processes, dynamic verbal and nonverbal interaction exchanges, diverse roles, motivations, and expectancies.”

How do we learn or improve our interpersonal skills? Emulating from a role model is an important aspect in interpersonal skills development and training. More generally, effective training is composed of four phases, namely information about the desired behaviours to be learnt, demonstration of the desired behaviours by a role model, trainees’ practice of the desired behaviours, and feedback to the trainees related to the practice phase [2, 3]. In the present research, we focus on the effect of role models for learning and training. Role models are part of the demonstration phase in which trainees observe individuals performing the desired behaviours so that they can learn through behavioural mimicry [4, 5]. Traditionally, demonstration involves the use of videos allowing trainees to watch examples of desired and undesired behaviours [2]. With the emergence of new technologies such as immersive Virtual Reality (VR) and the use of virtual humans for training, there are new possibilities available for training [6]. Past research has documented VR as an effective tool in interpersonal skills development and more specifically in public speaking training [7–11]. However, these studies have focused on the practice phase of training by showing positive outcomes after trainees’ participation to VR-based training sessions. Furthermore, these studies have mainly assessed VR as a tool to reduce public speaking anxiety, hence neglecting other interpersonal skills. To date, there is thus a lack of research on public speaking development per se involving the technology of VR. In the present research, given that VR is emerging as a valuable training setting, our goal is to compare two different demonstration conditions within VR. Specifically, we investigate the effectiveness of using a doppelganger as a role model in the development of interpersonal skills as compared to a role model that does not resemble the trainees. Doppelgangers are “virtual humans that highly resemble the real self but behave independently” [12].

Empirical evidence has shown that role models help individuals to learn desired behaviours [13–18]. Research has also revealed the effectiveness of using doppelgangers as role models in influencing individuals’ attitudes and behaviours [12, 19–21]. For instance, observing one’s doppelganger performing a certain physical act in VR can trigger behavioural changes such that individuals learn Tai Chi moves better [19] or report increased physical activity on the day following seeing one’s doppelganger in a training session on a treadmill [20]. In the field of public speaking, Aymerich-Franch and Bailenson [12] have demonstrated that the use of a doppelganger can reduce stress reactions in a public speaking task. In their study, before delivering a speech in front of an audience, participants took part in a relaxation exercise in which they listened to a voice-over describing that they were giving a successful speech. While listening to this voice, half of the participants watched their doppelganger delivering this speech in front of an audience (i.e., doppelganger condition), and the other half were asked to close their eyes and to visualise themselves delivering the speech (i.e., visualisation condition). Next, participants completed a questionnaire and then delivered their speech. Findings showed that participants did not react similarly to the relaxation exercise depending on their gender. In the doppelganger condition, male participants reported lower levels of anxiety and higher communicative competence than female participants. Inversely, in the visualisation condition female participants reported lower levels of anxiety and higher communicative competence than male participants.

This literature has revealed promising results regarding the use of doppelgangers as role models. Nonetheless, additional research is needed in the field of public speaking. First, empirical evidence is needed to further document the effectiveness of the use of doppelgangers as a training tool [6]. Second, past research has suggested that the use of doppelgangers has a different effect on trainees’ learning depending on whether trainees are women or men [12]. Third, existing studies have focused almost exclusively on self-report data but not on performance data. Self-report data might be affected by demand effects much more so than observational data of actual performance.

In the present study, we investigate whether the use of doppelgangers can improve interpersonal skills, measured as public speaking performance. Drawing on past findings [12, 19, 20] and research in marketing on the role played by identification in shaping individuals’ attitudes and behaviours [22, 23], we expect that using the doppelganger as the role model will increase training effectiveness. For instance, Ahn and Bailenson [22] have shown across three experiments that participants had more positive attitudes towards a brand and higher purchase intentions when the brand they were exposed to was associated to themselves (either through a photograph or through their avatar) as compared to when it was associated to another person or with only a text-based advertisement. The theoretical background of these findings lies in the self-referencing effect stating that individuals learn better (i.e., learn faster and remember longer) when the new information is delivered in association with the self [24, 25]. In our study, we expect this self-referencing effect to occur through the use of a role model that is maximally similar to the participant: their doppelganger. Using an exploratory lens, we also assess whether participant gender matters. Drawing on past research [12], it is plausible to expect that male participants benefit the most from the use of a doppelganger as the role model.

Research has shown that individual differences, such as cognitive ability [3, 26], conscientiousness [3, 27], or anxiety [27–29] play a role in the learning process. In the present study, we are interested in the influence that individual levels of self-efficacy might have on the effectiveness of doppelganger-based training. Self-efficacy is crucial in the learning process; individuals are motivated to learn a desired behaviour and to put effort into it only if they believe that they can achieve their goal [30–32]. Individuals low in self-efficacy believe that they are not able to achieve specific goals. Hence, they are easily discouraged and are more likely to give up than individuals high on self-efficacy who have this inner motivation to learn and to persevere in order to achieve desired outcomes. Accordingly, research has shown that self-efficacy is positively associated with individuals’ motivation to learn and with their performance [33–38]. For the development of interpersonal skills, considered to be a complex task [1], research has revealed that self-efficacy is positively associated with skill acquisition and maintenance [39, 40]. We hypothesise that individuals low in self-efficacy benefit more from training using a doppelganger as the role model because using the doppelganger closes the gap between one’s own behaviour and the role model’s behaviour and thus less self-efficacy is “necessary” to achieve the performance of the role model. Therefore, individuals low in self-efficacy might have an extra benefit from seeing themselves–their doppelganger–already perform the desired interpersonal skills behaviour.

Materials and methods

Experimental design

We conducted a quasi-experiment composed of a between-subject variable and two quasi-experimental variables. The between-subject variable had two levels capturing our manipulation of the type of role model that participants watched during the training in VR. The role model was either the participant’s doppelganger (DG condition) or an unknown avatar of the same gender (UA condition). Participants watched the virtual human (either the DG or the UA) give a charismatic speech in front of a virtual audience (the speech was identical in the DG and the UA conditions). The two quasi-experimental variables were participant level of self-efficacy and gender. Before and after the training session, participants delivered a speech in VR in front of the same virtual audience. We tested whether the type of role model led to greater performance in terms of body language persuasiveness (while controlling for their performance prior to the training session), and whether participant level of self-efficacy and gender moderate this relationship.

Participants and recruitment

We recruited 76 students from the subject pool of a Swiss university (M_age = 21.24, SD = 2.49, 37% women). Participants were mainly undergraduate students (80%) and the majority of them were enrolled either in a Business and Economics program (29%) or in an engineering program (39%). Almost all participants indicated having low to moderate experience in public speaking (96%) and they were willing to improve their public speaking skills (99%). Participants received 70 Swiss Francs (about $78) as compensation for their participation. Swiss Ethics Committee on research involving humans approved our research (project ID: 2018–00156) in a written form.

Materials

The immersive VR equipment was an HTC Vive Pro headset and we ran the study on the VR development platform WorldViz Vizard. To create the virtual human, we used 3ds Max toolkit. Finally, we collected data on participant level of self-efficacy, social anxiety, trait anxiety, and demographic information through online questionnaires using Qualtrics software. We performed our statistical analyses using SPSS, Version 25.

Procedure

Participants came to the laboratory twice. Session 1 lasted about 30 minutes and Session 2 lasted about 1h45.

Session 1

Participants obtained information about the study and signed an informed consent form. They then filled in an online questionnaire (15 minutes) including measures of social anxiety (SIAS) [41, 42], trait anxiety (STAI-T) [43, 44], and sociodemographic questions (e.g., participant age, gender). At this stage, we also collected data on the participants’ prior experience in public speaking and the extent to which they would like to improve their public speaking skills. Finally, we took photos of each participant (i.e., face and right profile) for the creation of the doppelganger.

After Session 1, we randomly assigned participants to the DG (N = 39) and the UA (N = 37) conditions while verifying that the sample was balanced in both conditions in terms of gender [X² (1, N = 76) = 0.031, p = .861], age [t(74) = -0.714, p = .477], social anxiety [t(74) = -0.406, p = .686], and trait anxiety [t(74) = -0.623, p = .535]. For participants assigned to the DG condition, we created their virtual doppelganger using their photos. The UA was created in the same way as the doppelgangers from photos of a male and a female lab member (unfamiliar to the participants). This was to ensure that the quality of the virtual humans was comparable in both conditions. Participants in the UA condition always encountered a virtual human of the same gender.

Session 2

Six months after Session 1, participants returned to the VR laboratory. They were instructed to deliver a 3 minute speech in front of a large virtual audience in a conference room on the topic of university fees. We chose this topic so that it is familiar to our sample of university students. To standardise the quality of the arguments or any initial differences in knowledge about the topic of university fees, we provided a list of 7 arguments and participants chose 3 of them to justify why university fees should not increase. Furthermore, they were instructed to be both convincing and charismatic during their speech. Specifically, they were asked to pay attention to their nonverbal behaviour, such as making sure that their gesture and posture were appropriate for the public speaking situation (to be not too static, to not move around too much, and to adopt an open body posture) and in terms of vocal qualities (to not speak too fast and to vary their voice tone). Participants had 5 minutes of preparation time. Once ready, participants were equipped with the headset and immersed in the VR environment. In this VR environment, they were standing on a stage in a large conference room with a computer on a desk displaying the 3 minute countdown (see Fig 1). Facing the participants, virtual humans who looked like students and members of the school management team followed participants with their gaze. Moreover, a typical conference room background sound was played in the headset’s earphones. Participants were video-recorded during their speech.

Fig 1 — Virtual conference room in which participants delivered their speech. Top right of the Figure shows the view of the desk with the 3 minute countdown on the computer screen.

After delivering their speech, participants were asked to watch a virtual person (DG or UA) giving an impressive charismatic speech on the same topic. To do so, the VR environment changed such that participants were outside the conference room, looking at the presenter from behind a door through a window (see Fig 2). The picture of the virtual person was displayed on a whiteboard behind the stage during the speech to ensure that participants correctly saw the face of the virtual person delivering the speech. Participants watched the virtual person’s speech twice. Before displaying the speech for a second time, participants were instructed to focus their attention on the body language of the virtual person. Participants watched the speech outside of the conference room because we wanted to muffle the sound so that participants in the DG condition would not be irritated to see themselves speaking with a different voice. Furthermore, prior to watching the speech, participants met the virtual person for 2 minutes in a flat (see Fig 3). We did this in order to ensure that participants in the DG condition realised that the virtual person was a virtual representation of themselves, and also to avoid any surprise effect in the DG condition when watching the speech so that participants could fully concentrate on the virtual presenter and their nonverbal behaviour.

Fig 2 — Example of a participant’s view of the virtual human giving a charismatic speech.

Fig 3 — Example of a female participant’s view in the virtual person’s flat and who is in the unknown avatar condition.

After having watched the virtual person giving her/his speech twice, participants delivered their speech a second time in front of the virtual audience (after 3 minutes of preparation time). Beforehand, they completed a short online questionnaire designed to assess their level of self-efficacy in delivering their speech a second time. Immediately after the preparation, participants were equipped with the headset and immersed in the same conference room as before for their 3 minute speech. After delivering their speech, participants completed an online questionnaire assessing the extent to which they identified with the virtual human.

Measures

Body language persuasiveness

We measured participant body language persuasiveness by asking an independent coder to watch the videotapes of participants’ first and second speeches without listening to the sound. Her task was to rate the appropriateness of the participants’ gestures and body openness and the expressiveness of their body language on a scale ranging from 1 (not at all) to 5 (completely). The three items were then averaged and higher values indicate more body language persuasiveness (M_speech1 = 2.63, SD_speech1 = 1.01, α_speech1 = .94; M_speech2 = 2.89, SD_speech2 = .98, α_speech2 = .94). Interrater reliability was established with an additional coder on a subset of 40 videos (i.e., the two videos of a randomly selected sample of 20 participants). Result showed a good reliability, ICC = .61 (two-way mixed, consistency, single measure intra-class correlation) [45].

Self-efficacy

We measured participant level of self-efficacy with regard to the second speech right before giving this speech using 23 items. We adapted 10 items from the General Self-Efficacy (GSE) scale [46] and we self-developed 13 items. Sample items are “During my speech, it will be easy for me to stick to my aims and deliver a good speech” and “I feel that I will give a good speech.” Participants indicated the extent to which each statement was true for them on a 4-point Likert-type scale ranging from 1 (not at all true) to 4 (totally true) for the 10 adapted items and they indicated the extent to which they agreed with each statement on a 5-point Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree) for the 13 self-developed items. To create our measure of self-efficacy, the 4-point scale items were converted to a 5-point scale and then the 23 items were averaged (M = 3.30, SD = .77, α = .96).

In order to better understand how training affects participants with different levels of self-efficacy, we created three categories, that is, a low, medium, and high level of self-efficacy. We opted for categories instead of a continuous variable because we wanted to assess the effect of our manipulation on the two extreme groups (i.e., low and high self-efficacy individuals). Prior to creating these categories, we ran a t-test to assess whether there was a gender difference. Results showed that male participants had significantly higher scores of self-efficacy (M = 3.45, SD = .69) than female participants (M = 3.04, SD = .84), t(74) = -2.33, p = .023. To avoid having gender as a confound of the self-efficacy categorisation, we created the three categories for each gender separately. We categorised the participants according to terciles. For female participants, a low score corresponds to a value below 2.80 (N = 10), a medium score corresponds to a value between 2.80 and 3.53 (N = 9), and a high score corresponds to a value equal to or above 3.54 (N = 9). For male participants, a low score corresponds to a value below 3.22 (N = 16), a medium score corresponds to a value between 3.23 and 3.88 (N = 16), and a high score corresponds to a value equal to or above 3.89 (N = 16).

Identification with the virtual person

We measured participant identification with the virtual person using the three following items: “The virtual person resembled me”, “The face of the virtual person resembled mine”, and “I identified with the virtual person.” Participants indicated the extent to which they agreed with each statement on a 5-point Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree). We averaged these items to create our measure of identification with the virtual person (M = 2.50, SD = 1.18, α = .75). Results of a t-test showed that, as expected, participants in the DG condition (M = 3.05, SD = 1.07) identified to a higher extent with the virtual person than participants in the UA condition (M = 1.92, SD = 1.01), t(74) = -4.73, p < .001. Results also showed that in the DG condition, male (M = 3.20, SD = 1.00) and female participants (M = 2.79, SD = 1.19) similarly identified with their doppelganger, t(37) = -1.16, p = .253.

Results

We performed a 2 (type of role model: DG—vs. UA) by 2 (participant gender: female vs. male) by 3 (levels of self-efficacy: low vs. medium vs. high) analysis of covariance (ANCOVA) using body language persuasiveness related to the second speech as a dependent variable, and we entered participants’ body language persuasiveness resulting from their first speech as the covariate. Beforehand, we checked whether our dependent variable was normally distributed. Results revealed that the distribution of our dependent variable differs from a normal distribution, D(76) = 0.17, p < .001. However, the Kolmogorov-Smirnov test can be significant even if the scores slightly differ from a normal distribution if the sample exceeds 50 participants [47]. Hence, we followed recommended procedures [47, 48] by further exploring the data through a visual inspection (histogram and Probability-Probability plot) and by converting the skew and kurtosis values to z-scores. Both the visual check and the z-scores revealed a normal distribution. Specifically, both z-scores are lower than 1.96 and thus are non-significant (p < .05). We also performed our ANCOVA while including participant level of identification with the virtual person as a covariate and results showed similar results. Hence, we decided to remove this variable from our model for parsimonious reasons.

Results revealed neither a significant effect of type of role model, nor a significant main effect of gender and level of self-efficacy, all Fs < 2.21, all ps > .10. There was no significant 2-way interaction involving type of role model, all Fs < .40, all ps > .10. However, there was a significant 3-way interaction effect, F(2, 63) = 3.30, p = .043. To further investigate this interaction effect, we ran additional ANCOVAs separately for female and male participants to test how self-efficacy affects body language persuasiveness depending on the type of role model.

These follow-up analyses showed a significant interaction effect for male participants, F(2, 41) = 3.34, p = .045, but not for female participants, F(2, 21) = 0.84, p = .444. Simple main effect analyses for male participants revealed that those who are relatively low in self-efficacy benefited the most from seeing their doppelganger; they were more persuasive in their body language (M = 3.55, SE = .20, N = 6) as compared to those with medium (M = 2.56, SE = .16, N = 9, p < .001) or relatively high levels of self-efficacy (M = 2.89, SE = .15, N = 10, p = .016). Furthermore, we found that male participants relatively low in self-efficacy that watched their doppelganger were more persuasive in their body language (M = 3.55, SE = .20, N = 6) than those that watched the unknown avatar (M = 2.91, SE = .15, N = 10, p = .014). Fig 4 reports these results in a bar graph.

Fig 4 — Bar chart reporting the estimated means and standard errors of participant body language persuasiveness related to the second speech as a function of participant gender, level of self-efficacy, and the type of virtual human. Participant body language persuasiveness related to the first speech was entered as the covariate.

Discussion

We investigated the effectiveness of using a doppelganger as the role model in training in the specific context of interpersonal skills development. We tested whether watching one’s doppelganger delivering a charismatic speech in front of an audience led participants to perform better in a subsequent speech than watching an unknown avatar. We captured participant performance in terms of body language persuasiveness. Furthermore, we also investigated whether participant level of self-efficacy and gender moderate this relationship. Results went in the expected direction by showing that the use of a doppelganger helped improving performance as compared to the use of a same gender avatar role model. However, findings suggested this effect to be specific to male participants that were relatively low in self-efficacy, as measured following the VR training. These results provide some evidence towards our expectation that individuals relatively low in self-efficacy would benefit the most from doppelganger-based training than individuals with relatively higher levels of self-efficacy. We argue that, by watching their virtual self delivering the charismatic speech, participants relatively low in self-efficacy had a greater motivation to put effort into the task and to persevere than those who watched an unknown avatar performing the same charismatic speech.

Results related to participant gender are in line with findings from Aymerich-Franch and Bailenson [12]. One plausible explanation of this gender effect is the specific context of the studies. Public speaking skills are traditionally associated with male prototypes. Research has shown that individuals’ prototypical representations of leaders, which refer to managerial positions requiring self-presentation, assertiveness, and persuasion, are commonly associated with men [49]. This skills-gender association is detrimental to women, leading for instance to discrimination against women aspiring to high status positions. It is possible that female performance was reduced due to the phenomenon of stereotype threat [50, 51]. Stereotype threat occurs when individuals feel at risk of conforming to a negative stereotype that applies to their social group in a specific situation. To illustrate, a female manager might feel threatened to be perceived by her subordinates as incompetent in her leading position, which then induces a reaction of stress and negatively impacts her performance, thus confirming the stereotype that female managers are less competent than male managers. Indeed, research has revealed that female role models help women overcome this stereotype threat [52]. However, Latu, Schmid Mast, Bombari, Lammers, and Hoyt [53] have shown that the effect of female role models is limited to famous female role models (as compared to unknown female role models). In our study, none of the doppelganger-based or the unknown-avatar-based training had an effect on female participant performance. These results are in line with findings from Latu et al. [53] such that none of our role models were famous, and thus did not help women in developing their skills. If stereotype threat is at work for women, then seeing their doppelganger might not enhance their performance because they might still fear that they are unable to reproduce the performance of the doppelganger.

Training to use immersive VR as well as training to use doppelgangers as role models is a growing field. Hence, future research is needed to investigate the effectiveness of the use of doppelgangers to develop interpersonal skills. Past research has shown promising results [19, 20]. However, the present study reveals that the use of doppelgangers might benefit only a subset of the population, namely male trainees low in self-efficacy. Furthermore, in our work, we tested the effectiveness of the use of doppelgangers in one single VR session, whereas literature on training recommends a stepwise procedure based on four phases to maximize trainees’ learning, namely information, demonstration, practice, and feedback [2, 3]. Specifically, research has shown that the practice and feedback phases are of particular importance [54]. This literature hence suggests several rounds of trials through which, after receiving information and watching demonstration related to the desired behaviour, trainees participate in several practice sessions, followed by feedback, such that, step-by-step, they can improve their performance by addressing in each session the points raised in the previous feedback session. We consider our work as a first step to investigate the effectiveness of doppelganger-based training. Nonetheless, future research is needed to assess the extent to which the effect of one unique VR doppelganger-based session lasts in time as well as whether the effectiveness of this type of training would be strengthened as the number of VR sessions increases.

Our findings should be interpreted with caution due to our relatively small sample size for testing a 3-way interaction and, therefore, further empirical evidence is needed before drawing definite conclusions about the effectiveness of using doppelgangers as a training tool. It has to be noted also that our sample is not representative of the entire population; it is relatively homogenous. Therefore, we do not know whether the people classified as high or low in self-efficacy are really the extremes in the population or whether they represent all rather high, low, or medium levels of self-efficacy. The low self-efficacy group in our study is thus low relative to the other participants in our sample. This is why we talk about relatively low (or high) self-efficacy, relatively meaning with respect to the participants in our sample. If we wanted to preselect trainees based on their level of self-efficacy in order to decide which training is best for them, we would first have to benchmark the self-efficacy score on a representative sample. Another discussion point is the reliability of our dependent variable (body language persuasiveness). Although it is considered as good [45], having a higher reliability would be preferable to avoid type-II errors: the lower the interrater reliability, the larger the amount of measurement error and thus the amount of noise in the data [55]. Ideally, future research should reach higher levels of interrater reliability (i.e., ICC ≥ .75).

As a research agenda, we suggest to first conduct additional (preferably longitudinal) studies investigating the role played (a) by doppelganger-based training as compared to unknown avatar-based training and (b) by participant gender in the development of interpersonal skills. Finally, research should further investigate the moderating effect of individual differences such as individual levels of self-efficacy. Furthermore, it would be interesting to study whether trainees’ level of self-efficacy varies throughout the training and if yes, to assess how and the extent to which this variation is affected by the performance of trainees as well as the extent to which it affects their subsequent performance. Understanding the profile of trainees who benefit the most from doppelganger-based training will allow managers to make more precise recommendations for how to develop interpersonal skills for different collaborators.

Acknowledgments

We are grateful to Erik Studer for his help in designing the study and collecting the data.

Data Availability

The datasets analysed during the current study, the data dictionary, and the syntax are available in the OSF repository, osf.io/8uedt.

Funding Statement

This study was funded by the Collaborative Research on Science and Society (https://www.epfl.ch/schools/cdh/research-2/cross-collaborative-research-on-science-and-society/) grant designed to foster collaboration between the Swiss Federal Institute of Technology (EPFL) in Lausanne and the University of Lausanne (UNIL) to MSM and CS. This research was also supported by a grant from the Swiss Science Foundation (http://www.snf.ch/en/Pages/default.aspx) to MSM and CS (CRSII5_183564/1). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

1.Klein C, DeRouin RE, Salas E. Uncovering workplace interpersonal skills: A review, framework, and research agenda In: Hodgkinson GP, Ford JK, editors. International review of industrial and organizational psychology. New York: Wiley and Sons, Ltd; 2006. p. 80–126. [Google Scholar]
2.Bedwell W. L., Fiore SM, Salas E. Developing the future workforce: An approach for integrating interpersonal skills into the MBA classroom. Academy of Management Learning & Education. 2014;13(2):171–86. [Google Scholar]
3.Salas E, Cannon-Bowers JA. The science of training: A decade of progress. Annual Review of Psychology. 2001;52:471–99. 10.1146/annurev.psych.52.1.471 [DOI] [PubMed] [Google Scholar]
4.Bandura A. Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall; 1986. [Google Scholar]
5.Bandura A. Social cognitive theory: An agentic perspective. Annual Review of Psychology. 2001 2001;52:1–26. 10.1146/annurev.psych.52.1.1 [DOI] [PubMed] [Google Scholar]
6.Schmid Mast M, Kleinlogel EP, Tur B, Bachmann M. The future of interpersonal skills development: Immersive virtual reality training with virtual humans. Human Resource Development Quarterly. 2018;29:125–41. [Google Scholar]
7.Anderson PL, Zimand E, Hodges LF, Rothbaum BO. Cognitive behavioral therapy for public-speaking anxiety using virtual reality for exposure. Depression and Anxiety. 2005;22:156–8. 10.1002/da.20090 [DOI] [PubMed] [Google Scholar]
8.Gorini A, Riva G. Virtual reality in anxiety disorders: The past and the future. Expert Review of Neurotherapeutics. 2008;8(2):215–33. 10.1586/14737175.8.2.215 [DOI] [PubMed] [Google Scholar]
9.Harris SR, Kemmerling RL, North MM. Brief virtual reality therapy for public speaking anxiety. Cyberpsychology & Behavior. 2002;5(6):543–57. [DOI] [PubMed] [Google Scholar]
10.Lindner P, Miloff A, Fagernäs S, Andersen J, Sigeman M, Andersson G, et al. Therapist-led and self-led one-session virtual reality exposure therapy for public speaking anxiety with consumer hardware and software: A randomized controlled trial. Journal of Anxiety Disorders. 2019;61:45–54. [DOI] [PubMed] [Google Scholar]
11.Stupar-Rutenfrans S, Ketelaars LEH, van Gisbergen MS. Beat the fear of public speaking: Mobile 360° video virtual reality exposure training in home environment reduces public speaking anxiety. Cyberpsychology, Behavior, and Social Networking. 2017;20(10):624–33. 10.1089/cyber.2017.0174 [DOI] [PubMed] [Google Scholar]
12.Aymerich-Franch L, Bailenson J. The use of doppelgangers in virtual reality to treat public speaking anxiety: A gender comparison. Proceedings of the International Society for Presence Research 2014. p. 173–86.
13.Collins RL. For better or worse: The impact of upward social comparison on self-evaluations. Psychological Bulletin. 1996;119(1):51–69. [Google Scholar]
14.Lockwood P, Kunda Z. Superstars and me: Predicting the impact of role models on the self. Journal of Personality and Social Psychology. 1997;73(1):91–103. [Google Scholar]
15.Marx DM, Roman JS. Female role models: Protecting women’s math test performance. Personality and Social Psychology Bulletin. 2002;28(9):1183–93. [Google Scholar]
16.Marx DM, Ko SJ, Friedman RA. The ‘‘Obama Effect”: How a salient role model reduces race-based performance differences. Journal of Experimental Social Psychology. 2009;45:953–6. [Google Scholar]
17.Rich GA. The sales manager as a role model: Effects on trust, job satisfaction, and performance of salespeople. Journal of the Academy of Marketing Science. 1997;25(4):319–28. [Google Scholar]
18.Taylor SE, Lobel M. Social comparison activity under threat: Downward evaluation and upwards contacts. Psychological Review. 1989;96(4):569–75. 10.1037/0033-295x.96.4.569 [DOI] [PubMed] [Google Scholar]
19.Bailenson J, Patel K, Nielsen A, Bajscy R, Jung S-H, Kurillo G. The effect of interactivity on learning physical actions in virtual reality. Media Psychology. 2008;11:354–76. [Google Scholar]
20.Fox J, Bailenson JN. Virtual self-modeling: The effects of vicarious reinforcement and identification on exercise behaviors. Media Psychology. 2009;12:1–25. [Google Scholar]
21.Fox J, Bailenson JN, Binney J. Virtual experiences, physical behaviors: The effect of presence on imitation of an eating avatar. Presence: Teleoperators and Virtual Environments. 2009;18(4):294–303. [Google Scholar]
22.Ahn SJ, Bailenson JN. Self-endorsing versus other-endorsing in virtual environments. Journal of Advertising. 2011;40(2):93–106. [Google Scholar]
23.Martin BAS, Kwai-Choi Lee C, Yang F. The influence of ad model ethnicity and self-referencing on attitudes: Evidence from New Zealand. Journal of Advertising. 2004 2004/12/1;33(4):27–37. [Google Scholar]
24.Ahn SJ, Bailenson J. Self-endorsed advertisements: When the self persuades the self. Journal of Marketing Theory and Practice. 2014;22(2):135–6. [Google Scholar]
25.Kuiper NA, Rogers TB. Encoding of personal information: Self-other differences. Journal of Personality and Social Psychology. 1979;37(4):499–514. [DOI] [PubMed] [Google Scholar]
26.Greco LM, Charlier SD, Brown KG. Trading off learning and performance: Exploration and exploitation at work. Human Resource Management Review. 2019;29(2):179–95. [Google Scholar]
27.Colquitt JA, LePine JA, Noe RA. Toward an integrative theory of training motivation: A meta-analytic path analysis of 20 years of research. Journal of Applied Psychology. 2000;85(5):678–707. 10.1037/0021-9010.85.5.678 [DOI] [PubMed] [Google Scholar]
28.Thoresen JC, Francelet R, Coltekin A, Richter K-F, Fabrikant SI, Sandi C. Not all anxious individuals get lost: Trait anxiety and mental rotation ability interact to explain performance in map-based route learning in men. Neurobiology of Learning and Memory. 2016;132:1–8. 10.1016/j.nlm.2016.04.008 [DOI] [PubMed] [Google Scholar]
29.Herrero AI, Sandi C, Venero C. Individual differences in anxiety trait are related to spatial learning abilities and hippocampal expression of mineralocorticoid receptors. Neurobiology of Learning and Memory. 2006;86(2):150–9. 10.1016/j.nlm.2006.02.001 [DOI] [PubMed] [Google Scholar]
30.Bandura A. Self-efficacy: Toward a unifying theory of behavioral chance. Psychological Review. 1977;84(2):191–215. 10.1037//0033-295x.84.2.191 [DOI] [PubMed] [Google Scholar]
31.Bandura A. Social cognitive theory: An agentic perspective. Asian Journal of Social Psychology. 1999;2:21–41. [DOI] [PubMed] [Google Scholar]
32.Bandura A, Wood R. Social cognitive theory of organizational management. Academy of Management Review. 1989;14(3):361–84. [Google Scholar]
33.Ackerman PL, Kanfer R, Goff M. Cognitive and noncognitive determinants and consequences of complex skill acquisition. Journal of Experimental Psychology: Applied. 1995;1(4):270–304. [Google Scholar]
34.Chemers MM, Hu L-T, Garcia BF. Academic self-efficacy and first-year college student performance and adjustment. Journal of Educational Psychology. 2001;93(1):55–64. [Google Scholar]
35.Locke EA, Frederick E, Lee C, Bobko P. Effect of self-efficacy, goals, and task strategies on task performance. Journal of Applied Psychology. 1984;69(2):241–51. [Google Scholar]
36.Multon KD, Brown SD, Lent RW. Relation of self-efficacy beliefs to academic outcomes: A meta-analytic investigation. Journal of Counselling Psychology. 1991;38(1):30–8. [Google Scholar]
37.Stajkovic AD, Luthans F. Self-efficacy and work-related performance: A meta-analysis. Psychological Bulletin. 1998;124(2):240–61. [Google Scholar]
38.Zimmerman BJ. Self-efficacy: An essential motive to learn. Contemporary Educational Psychology. 2000;25:82–91. 10.1006/ceps.1999.1016 [DOI] [PubMed] [Google Scholar]
39.Gist ME, Stevens CK, Bavetta AG. Effects of self-efficacy and post-training intervention on the acquisition and maintenance of complex interpersonal skills. Personnel Psychology. 1991;44(4):837–61. [Google Scholar]
40.Stevens CK, Gist ME. Effects of self-efficacy and goal-orientation training on negotiation skill maintenance: What are the mechanisms?. Personnel Psychology. 1997;50:955–78. [Google Scholar]
41.Mattick RP, Clarke JC. Development and validation of measures of social phobia scrutiny fear and social interaction anxiety. Behaviour Research and Therapy. 1998;36:455–70. 10.1016/s0005-7967(97)10031-6 [DOI] [PubMed] [Google Scholar]
42.Lacroix F, Zhou B. Échelles traduites en français: Anxiété sociale et Taijin Kyofusho: Concordia University; Université de Montréal2015.
43.Bergeron J. State-trait anxiety in French-English bilinguals: Cross-cultural considerations In: Spielberger CD, Diaz-Guerrero R, editors. Cross-cultural anxiety Washington, DC: Hemisphere/Wiley; 1983. p. 157–76. [Google Scholar]
44.Spielberger CD, Gorsuch RL, Lushene R, Vagg PR, Jacobs GA. Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press; 1983. [Google Scholar]
45.McGraw KO, Wong SP. Forming inferences about some intraclass correlation coefficients. Psychological Methods. 1996;1(1):30–46. [Google Scholar]
46.Schwarzer R, Jerusalem M. Generalized Self-Efficacy scale In: Weinman J, Wright S, Johnston M, editors. Measures in health psychology: A user’s portfolio Causal and control beliefs. Windsor, England: Nfer-Nelson; 1995. p. 35–7. [Google Scholar]
47.Elliott AC, Woodward WA. Statistical analysis quick reference guidebook: With SPSS examples. Thousand Oaks, CA: Sage; 2007. [Google Scholar]
48.Field A. Discovering statistics using SPSS. London: Sage; 2009. [Google Scholar]
49.Koenig AM, Eagly AH, Mitchell AA, Ristikari T. Are leader stereotypes masculine? A meta-analysis of three research paradigms. Psychological Bulletin. 2011;137(4):616–42. 10.1037/a0023557 [DOI] [PubMed] [Google Scholar]
50.Spencer SJ, Steele CM, Quinn DM. Stereotype threat and women’s math performance. Journal of Experimental Social Psychology. 1999;35:4–28. [Google Scholar]
51.Steele CM. A threat in the air: How stereotypes shape intellectual identity and performance. American Psychologist. 1997;52:613–29. 10.1037//0003-066x.52.6.613 [DOI] [PubMed] [Google Scholar]
52.Latu IM, Schmid Mast M, Lammers J, Bombari D. Successful female leaders empower women’s behavior in leadership tasks. Journal of Experimental Social Psychology. 2013;49:444–8. [Google Scholar]
53.Latu IM, Schmid Mast M, Bombari D, Lammers J, Hoyt CL. Empowering mimicry: Female leader role models empower women in leadership tasks through body posture mimicry. Sex Roles. 2019;80:11–24. 10.1007/s11199-018-0911-y [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Smith-Jentsch KA, Salas E, Baker DP. Training team performance-related assertiveness. Personnel Psychology. 1996;49(4):909–36. [Google Scholar]
55.Hallgren KA. Computing inter-rater reliability for observational data: An overview and tutorial. Tutorials in Quantitative Methods for Psychology. 2012;8(1):23–34. 10.20982/tqmp.08.1.p023 [DOI] [PMC free article] [PubMed] [Google Scholar]

PLoS One. doi: 10.1371/journal.pone.0245960.r001

Decision Letter 0

Doron Friedman

1 Sep 2020

PONE-D-20-19196

Doppelganger-based training: Imitating our virtual self to accelerate interpersonal skills learning

PLOS ONE

Dear Dr. Kleinlogel,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please address comments by the reviewers and especially reviewer #2.

Please submit your revised manuscript by Oct 16 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Doron Friedman

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: On line 116, I believe you want to say "learn and perform" rather than "learn and performance"

On line 280, saying "Results did not reveal and difference in results" is a bit clumsy. I'd suggest "Outcomes did not reveal any differences in results" or some other synonym for results.

On line 319, I'm not sure why you capitalized the S in "Self." That is inconsistent with other uses in the paper.

Reviewer #2: The current manuscript examined the use of virtual doppelgangers as role models in training interpersonal skills in VR. In particular, comparing the use of doppelgangers vs a gender-matched ‘unknown’ avatar meant to represent a given participant for training the interpersonal skill of giving a convincing speech. The authors didn’t find a main effect for the use of doppelgangers across all participants, but they did find a slight effect of doppelgangers for males with low self efficacy.

Overall I found the manuscript to be well-written and clearly communicated their ideas and methodologies. I thought the general approach to capturing and analyzing data was useful and clearly stated. To test the idea that there may be differences in efficacy of doppelgangers based on sex or self-efficay was important to test. However, I did find some fundamental theoretical issues with the manuscript that should be addressed before publication, stated below:

While the current work does reference relevant published research investigations (Balienson’s work examining the use of VR for public speaking, in particular), no theoretical justification is provided as to why a virtual model looking like yourself (i.e., doppelganger) should provide a more effective model in order to learn. It seems to be based on intuition and one other result (that should be replicated before taken as a solid finding). If this were designed as a replication of that work, I still think it’s important to . Intuitively, I believe there is reason for participants to accept their doppelganger less than an unknown avatar since the dissonance between doppelganger and themselves may be more apparent. In any case, I think the author’s need to resolve the theoretical justification for their work instead of relying on just one study. Without this I don’t see the merit in publishing this work. Perhaps there is work from the research areas focusing on training/modeling for improving public speaking.

Is the idea that public speaking can be taught in the one VR session? Why? Is there a methodology regarding public speaking effectiveness that supports this methodology? If yes then this needs to be reconciled with the current methodology. If not then it’s worth stating as much.

It seems that splitting groups up into terciles in order to examine differences at a more ‘meaningful’ group level artificially induces differences that may not be meaningful. In other words, If the sample were much larger and incorporated a wide variety of individuals I would believe that the current sample distribution is representative of the population. However, with a small sample such as this and with such a homogenous sample it's likely we are only viewing one relatively small part of the distribution of scores. By splitting into terciles, there is an assumption that the lowest score from this sample is representative of the population, which is unlikely to be true given the reasons stated above. Therefore, I advise declaring 'good' or 'poor' performers, as all of the scores captured here may be of 'good' performers. If there is reason to believe that this is not the case, then this needs to be stated.

Minor Comments:

IRR = 0.6 seems pretty weak, particularly with a seemingly small effect size. That is the lack of IRR may be substantially contributing to the relatively small differences in the sample

“Results did not reveal any difference in results between these analyses and those performed using the non-transformed variable” - is it appropriate to run ANCOVA on skewed distribution? I believe one of the assumptions of the ANCOVA model is symmetry, which is violated with skewed distributions

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Charles E. Hughes

Reviewer #2: Yes: Michael Casale

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2021 Feb 10;16(2):e0245960. doi: 10.1371/journal.pone.0245960.r002

Author response to Decision Letter 0

25 Sep 2020

Reviewer #1

1. On line 116, I believe you want to say "learn and perform" rather than "learn and performance"

Response: Research has shown that self-efficacy is related to both people motivation to learn and people performance. To avoid confusion, we clarify the sentence as followed (page 5, line 119):

“Accordingly, research has shown that self-efficacy is positively associated with individuals’ motivation to learn and with their performance [28-33].”

2. On line 280, saying "Results did not reveal and difference in results" is a bit clumsy. I'd suggest "Outcomes did not reveal any differences in results" or some other synonym for results.

Response: Thank you for pointing out the clumsiness of this sentence. We edited this section, leading to the deletion of this sentence from the manuscript.

3. On line 319, I'm not sure why you capitalized the S in "Self." That is inconsistent with other uses in the paper.

Response: Thank you. We fixed this typo (page 14, line 326).

Reviewer #2

1. The current manuscript examined the use of virtual doppelgangers as role models in training interpersonal skills in VR. In particular, comparing the use of doppelgangers vs a gender-matched ‘unknown’ avatar meant to represent a given participant for training the interpersonal skill of giving a convincing speech. The authors didn’t find a main effect for the use of doppelgangers across all participants, but they did find a slight effect of doppelgangers for males with low self efficacy.

Response: Thank you for your feedback. Please find below our responses to your comments.

2. While the current work does reference relevant published research investigations (Balienson’s work examining the use of VR for public speaking, in particular), no theoretical justification is provided as to why a virtual model looking like yourself (i.e., doppelganger) should provide a more effective model in order to learn. It seems to be based on intuition and one other result (that should be replicated before taken as a solid finding). If this were designed as a replication of that work, I still think it’s important to. Intuitively, I believe there is reason for participants to accept their doppelganger less than an unknown avatar since the dissonance between doppelganger and themselves may be more apparent. In any case, I think the author’s need to resolve the theoretical justification for their work instead of relying on just one study. Without this I don’t see the merit in publishing this work. Perhaps there is work from the research areas focusing on training/modeling for improving public speaking.

Response: Thank you for your comment. We agree that providing a theoretical framework to our research question related to the use of doppelgangers is needed to add value to our work.

We now provide such a theoretical framework by drawing on research stemming from marketing. More specifically, we build on the self-referencing effect stating that individuals’ learning process is strengthened (meaning that individuals learn faster and remember longer) when the new information is delivered in association with the self. We expect that using trainee’s doppelganger as a role model would thus lead to greater learning because the similarity between the trainee and the role model is maximised. We added this theoretical background on pages 4-5. It now reads:

“Drawing on past findings [7, 14, 15] and research in marketing on the role played by identification in shaping individuals’ attitudes and behaviours [17, 18], we expect that using the doppelganger as the role model will increase training effectiveness. For instance, Ahn and Bailenson [17] have shown across three experiments that participants had more positive attitudes towards a brand and higher purchase intentions when the brand they were exposed to was associated to themselves (either through a photograph or through their avatar) as compared to when it was associated to another person or with only a text-based advertisement. The theoretical background of these findings lies in the self-referencing effect stating that individuals learn better (i.e., learn faster and remember longer) when the new information is delivered in association with the self [19, 20]. In our study, we expect this self-referencing effect to occur through the use of a role model that is maximally similar to the participant: their doppelganger. Using an exploratory lens, we also assess whether participant gender matters. Drawing on past research [7], it is plausible to expect that male participants benefit the most from the use of a doppelganger as the role model.”

3. Is the idea that public speaking can be taught in the one VR session? Why? Is there a methodology regarding public speaking effectiveness that supports this methodology? If yes then this needs to be reconciled with the current methodology. If not then it’s worth stating as much.

Response: Thank you for this comment. We agree that this point needs to be clarified. In our work, we were interested in testing whether the use of doppelganger can help developing public speaking skills. We conducted an experiment composed of two waves designed to test whether the demonstration phase of the IDPF model of training (Bedwell, Fiore, & Salas, 2014; Kraiger, 2003; Salas & Cannon-Bowers, 2001) leads to body language persuasiveness differences depending on whether the role model in the demonstration is similar vs. dissimilar to the trainees.

Literature has highlighted the difficulty to develop interpersonal skills and to teach these skills due to their complexity (Bedwell et al., 2014; Klein, DeRouin, & Salas, 2006). Research has also demonstrated that the practical phase and the feedback phase are crucial in the learning process (e.g., Smith-Jentsch et al., 1996), suggesting that once trainees receive information on the desired behaviour to be learnt and watch a demonstration of it, practice and feedback are required to develop and master these skills. Accordingly, we argue that the development of public speaking skills and more generally interpersonal skills needs to be stepwise. Specifically, we expect that it requires several VR sessions followed by feedback to be able to master these new skills, such that step-by-step trainees can improve their performance by addressing the points raised in the previous feedback.

Hence, we perceive our study as a first step to investigate whether the use of doppelgangers can foster trainees’ interpersonal skills development, and we call for future research before stating clear recommendations related to training development. Nonetheless, our findings are promising because they show that after one single VR session, male participants low in self-efficacy in the DG condition were more persuasive than those in the UA condition. However, it seems plausible to expect an effect of time such that the behavioural improvement based on a single session might not last for a long time. We call for future research to test the effect of such training on a longer period of time, by for instance testing whether the difference between the two groups is still present 1-2 week(s) later. Research should also investigate whether the benefit of this type of training would be strengthened as the number of VR sessions increases, and if yes, to what extent. We now address this point on page 15 and on page 16, respectively:

“Furthermore, in our work, we tested the effectiveness of the use of doppelgangers in one single VR session, whereas literature on training recommends a stepwise procedure based on four phases to maximize trainees’ learning, namely information, demonstration, practice, and feedback [2, 3]. Specifically, research has shown that the practice and feedback phases are of particular importance [49]. This literature hence suggests several rounds of trials through which, after receiving information and watching demonstration related to the desired behaviour, trainees participate in several practice sessions, followed by feedback, such that, step-by-step, they can improve their performance by addressing in each session the points raised in the previous feedback session. We consider our work as a first step to investigate the effectiveness of doppelganger-based training. Nonetheless, future research is needed to assess the extent to which the effect of one unique VR doppelganger-based session lasts in time as well as whether the effectiveness of this type of training would be strengthened as the number of VR sessions increases.”

“As a research agenda, we suggest to first conduct additional (preferably longitudinal) studies investigating the role played (a) by doppelganger-based training as compared to unknown avatar-based training and (b) by participant gender in the development of interpersonal skills.”

4. It seems that splitting groups up into terciles in order to examine differences at a more ‘meaningful’ group level artificially induces differences that may not be meaningful. In other words, If the sample were much larger and incorporated a wide variety of individuals I would believe that the current sample distribution is representative of the population. However, with a small sample such as this and with such a homogenous sample it's likely we are only viewing one relatively small part of the distribution of scores. By splitting into terciles, there is an assumption that the lowest score from this sample is representative of the population, which is unlikely to be true given the reasons stated above. Therefore, I advise declaring 'good' or 'poor' performers, as all of the scores captured here may be of 'good' performers. If there is reason to believe that this is not the case, then this needs to be stated.

Response: We agree that, due to our low sample size and the homogeneity of our participants, the sample is most likely not representative of the population. Just to be on the same page, we split the self-efficacy measure into terciles, not the performance measure. The latter was the DV and was continuous. With respect to self-efficacy, indeed, we do not know whether our three groups are all situated on relatively high or relatively low levels of absolute self-efficacy. If we were to pretest participants on self-efficacy and then decide which training they should undergo, we indeed would need to test a more representative sample to find the benchmark values on self-efficacy that indicate whether doppelganger training would be particularly beneficial. Such was not the goal of our study. The goal was to test how variables (i.e., self-efficacy) affect nonverbal public speaking performance. So, the terciles simply represent relatively (with respect to our sample but not with respect to the population) higher or lower levels of self-efficacy. The (potential) non-representativeness of the sample is common in experimental research in which the goal is not to provide benchmark numbers but to show how variables affect each other. Moreover, the homogeneity of our sample most likely leads to reduced variance in self-efficacy which then makes it unlikely to find an effect of self-efficacy. So, our study is a very conservative test and most likely underestimates the true effect (that we would find in a more heterogeneous sample). To acknowledge the fact that the terciles are relative to each other in the sample and not absolute measures of the population, we now rephrase at several places in the text and talk about relative high and low self-efficacy, relative meaning when comparing the different terciles within our sample (page 14, lines 323-327). We now also address this point in the Discussion section on page 16 (lines 373-381):

“It has to be noted also that our sample is not representative of the entire population; it is relatively homogenous. Therefore, we do not know whether the people classified as high or low in self-efficacy are really the extremes in the population or whether they represent all rather high, low, or medium levels of self-efficacy. The low self-efficacy group in our study is thus low relative to the other participants in our sample. This is why we talk about relatively low (or high) self-efficacy, relatively meaning with respect to the participants in our sample. If we wanted to preselect trainees based on their level of self-efficacy in order to decide which training is best for them, we would first have to benchmark the self-efficacy score on a representative sample.”

5. Minor Comments:

5.1 IRR = 0.6 seems pretty weak, particularly with a seemingly small effect size. That is the lack of IRR may be substantially contributing to the relatively small differences in the sample.

Response: We agree that a higher interrater reliability coefficient would be ideal. Nonetheless, an IRR of .60 is considered as “good” (Cicchetti, 1994). Furthermore, as mentioned, a low ICC increases the probability of type-II errors because the amount of measurement error increases as the IRR decreases, thus adding noise to the data (Hallgren, 2012). Hence, having a higher ICC could contribute to strengthen the differences in the sample. We now mention this limitation on page 16 (lines 381-385):

“Another discussion point is the reliability of our dependent variable (body language persuasiveness). Although it is considered as good [40], having a higher reliability would be preferable to avoid type-II errors: the lower the interrater reliability, the larger the amount of measurement error and thus the amount of noise in the data [50]. Ideally, future research should reach higher levels of interrater reliability (i.e., ICC > .75).”

5.2 “Results did not reveal any difference in results between these analyses and those performed using the non-transformed variable” - is it appropriate to run ANCOVA on skewed distribution? I believe one of the assumptions of the ANCOVA model is symmetry, which is violated with skewed distributions.

Response: Results of the Kolmogorov-Smirnov (K-S) test revealed a non-normal distribution, D(76) = 0.17, p < .001. Based on your comments, we further looked at the literature related to data normality. We found that our sample size is considered as large when testing for the normality of a distribution because it exceeds 50 participants (Elliott & Woodward, 2007). In this case, the K-S test might be significant even if the scores slightly differ from a normal distribution. Hence, we further explored our data by looking at the value of skew. This value should be 0 in a normal distribution. In our data, the value is -.183 (SE = .276), which is close to 0 but still different. As recommended (Elliott & Woodward, 2007; Field, 2009), we further assessed the normality of our distribution using two other methods next to the K-S test. First, we visually checked the distribution of the data through a histogram and a P-P plot (probability–probability plot), revealing the data to look normal. Second, we converted the skew value to z-score, = 0.66. This value is lower than 1.96 and is thus non-significant (p < .05), showing that the distribution is normal. Results are similar for the value of kurtosis (-.348, SE = .545, z = .64). Accordingly, we now report on page 12 (lines 282-289) that despite the K-S test is significant, the visual check and the z-scores indicate that the data is normally distributed:

“Results revealed that the distribution of our dependent variable differs from a normal distribution, D(76) = 0.17, p < .001. However, the Kolmogorov-Smirnov test can be significant even if the scores slightly differ from a normal distribution if the sample exceeds 50 participants [42]. Hence, we followed recommended procedures [42, 43] by further exploring the data through a visual inspection (histogram and Probability-Probability plot) and by converting the skew and kurtosis values to z-scores. Both the visual check and the z-scores revealed a normal distribution. Specifically, both z-scores are lower than 1.96 and thus are non-significant (p < .05).”

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(49.1KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0245960.r003

Decision Letter 1

Doron Friedman

10 Nov 2020

PONE-D-20-19196R1

Doppelganger-based training: Imitating our virtual self to accelerate interpersonal skills learning

PLOS ONE

Dear Dr. Kleinlogel,

Please submit your revised manuscript by Dec 25 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

We look forward to receiving your revised manuscript.

Kind regards,

Doron Friedman

Academic Editor

PLOS ONE

Additional Editor Comments (if provided):

Since one of the original reviewers could not follow up the review with the revised version we have allocated another reviewer. Please refer to their comments (reviewer #2).

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #3: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #1: Yes

Reviewer #3: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #3: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

Reviewer #1: Yes

Reviewer #3: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

Reviewer #1: Yes

Reviewer #3: Yes

**********

6. Review Comments to the Author

Reviewer #1: The authors effectively addressed the concerns of the reviewers. In particular, my minor concerns were addressed and those of the other reviewer were, in my opinion, addressed with appropriate references to teh literature and caution to the readers to not jump to potentially false conclusions.

Reviewer #3: 1. I found the manuscript to be clear and concise. The use of doppelgangers to aid in training public speaking skills is a novel application that I find very interesting. Overall, I found no issue with the descriptions of the background or methodologies, and found them easily understood.

2. Why was there not a control that did not include virtual human training? I feel as though there should be more discussion on if just repeatedly speaking in this way tends to lead to equivalent or better. You may be able to say that doppelgangers are better to use than Unknown Agents but without a control how can you tell this is an effective means of training?

3. When it comes to enhancing skills in public speaking there was a lack of prior work provided. The authors heavily relied on the work by Aymerich-Franch and Bailenson but this work focuses on social-anxiety and not improving interpersonal communication skills (except one could argue stress management but that does not seem to be the focus for this manuscript). I would suggest some background on non-verbal gesture or other public speaking skills training be discussed more in depth.

4. 321 -324 "As expected, findings revealed that the use of a doppelganger led to better performance than the use of a same gender avatar role model. However, we observed this effect only among male participants that were relatively low in self-efficacy, as measured following the VR training." The first sentence here is a very strong statement that I don't believe can be stated here considering the small size of the group that once divided up showed the result. I would suggest weakening the statement or leading with the second statement and stating that this provides some evidence towards your hypothesis.

5. 326-329 "We argue that, by watching their virtual self delivering the charismatic speech, participants relatively low in self-efficacy had a greater motivation to put effort into the task and to persevere than those who watched an unknown avatar performing the same charismatic speech." This effect could also be seen as a result of participants with lower self-esteem feeling better about the task after practicing the first time. It would have been interesting to see self-efficacy measured before and after the first public speech as it is unknown if self-efficacy ratings were affected by performing. (Potentially increasing for those that had previous anxiety or the other way around)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #3: No

PLoS One. 2021 Feb 10;16(2):e0245960. doi: 10.1371/journal.pone.0245960.r004

Author response to Decision Letter 1

7 Dec 2020

Reviewer #1

Comment: The authors effectively addressed the concerns of the reviewers. In particular, my minor concerns were addressed and those of the other reviewer were, in my opinion, addressed with appropriate references to the literature and caution to the readers to not jump to potentially false conclusions.

Response: Thank you for your positive feedback.

Reviewer #2

Comment: Why was there not a control that did not include virtual human training? I feel as though there should be more discussion on if just repeatedly speaking in this way tends to lead to equivalent or better. You may be able to say that doppelgangers are better to use than Unknown Agents but without a control how can you tell this is an effective means of training?

Response: We thank you for this comment. We agree that this point needs to be clarified. We investigate the extent to which new technologies – VR – can contribute to improve trainees’ learning process. Effective training is composed of four phases (IDPF model of training), namely information, demonstration, practice, and feedback. Each phase is crucial during the learning process of interpersonal skills (see Bedwell, Fiore, & Salas, 2014) and it is important to investigate the best way of delivering to trainees each phase. In our research, we focused on the demonstration phase of the IDPF model of training. Drawing on past research on the use of doppelgangers as role models, we further documented whether this new way of learning through VR-based demonstration can help trainees improving their interpersonal skills in the specific context of public speaking. To assess the effect of doppelgangers as role models on trainees’ learning, in our study we compare two different demonstration conditions within VR, i.e. an experimental condition in which participants watched their doppelganger as a role model and a comparison condition in which participants watched a virtual person that do not resemble them.

We did not add a third condition that does not include a role model because our work focuses on the demonstration phase of training, hence involving a virtual person as a trainer. Moreover, adding another condition without virtual agents or without VR would have made the comparison difficult because then more than just one feature of the setting would have changed and thus introduced potential confounding variables. In the abstract, we presented the main goal of our work as followed (lines 27-29):

“In the present study, using the technology of immersive virtual reality we investigate the extent to which the use of doppelgangers as role models can boost trainees’ interpersonal skills development.”

We agree that this sentence can bring confusion to the readers regarding our research question. Hence, based on your comment, we decided to clarify this point as followed (lines 27-32):

“In the present study, given that virtual reality is emerging as a valuable training setting, we compare two different demonstration conditions within virtual reality by investigating the extent to which the use of doppelgangers as role models can boost trainees’ interpersonal skills development as compared to a role model that does not resemble the trainees.”

Furthermore, on pages 3-4, we inform the readers that we focus on the demonstration phase, hence involving trainees who observe a trainer – a role model – performing the desired behaviours. Nonetheless, similar to the abstract and based on your comment, we decided to be more precise to avoid confusion about the goal of our research by explaining on lines 73-77 that:

“In the present research, given that VR is emerging as a valuable training setting, our goal is to compare two different demonstration conditions within VR. Specifically, we investigate the effectiveness of using a doppelganger as a role model in the development of interpersonal skills as compared to a role model that does not resemble the trainees.”

Comment: When it comes to enhancing skills in public speaking there was a lack of prior work provided. The authors heavily relied on the work by Aymerich-Franch and Bailenson but this work focuses on social-anxiety and not improving interpersonal communication skills (except one could argue stress management but that does not seem to be the focus for this manuscript). I would suggest some background on non-verbal gesture or other public speaking skills training be discussed more in depth.

Response: Thank you for this comment. Indeed, in our paper we heavily rely on the work by Aymerich-Franch and Bailenson. Their work is central in the field of public speaking VR-based training because, to our knowledge, it is the first study that investigated the effect of doppelgangers as role models on trainees’ learning in the demonstration phase.

Previous research on public speaking has mainly investigated the effectiveness of VR-based training during the practice phase (e.g., Anderson et al., 2005; Gorini & Riva, 2008; Harris et al., 2002; Lindner et al., 2019; Stupar-Rutenfrans et al., 2019). Furthermore, these studies mainly focused on the fear of public speaking as an outcome. Hence, there is a lack of research on public speaking skills development per se involving the technology of VR.

In our paper, we wrote that traditionally, the demonstration phase involved the use of videos, however nowadays new technologies allow new ways of designing training (lines 62-66):

“In the present research, we focus on the effect of role models for learning and training. Role models are part of the demonstration phase in which trainees observe individuals performing the desired behaviours so that they can learn through behavioural mimicry [4, 5]. Traditionally, demonstration involves the use of videos allowing trainees to watch examples of desired and undesired behaviours [2]. With the emergence of new technologies such as immersive Virtual Reality (VR) and the use of virtual humans for training, there are new possibilities available for training [6].”

We agree that it would be informative for the readers to discuss in more detail past research. We now briefly present the current state of research on lines 66-72:

“Past research has documented VR as an effective tool in interpersonal skills development and more specifically in public speaking training [7-11]. However, these studies have focused on the practice phase of training by showing positive outcomes after trainees’ participation to VR-based training sessions. Furthermore, these studies have mainly assessed VR as a tool to reduce public speaking anxiety, hence neglecting other interpersonal skills. To date, there is thus a lack of research on public speaking development per se involving the technology of VR.”

Comment: 321 -324 "As expected, findings revealed that the use of a doppelganger led to better performance than the use of a same gender avatar role model. However, we observed this effect only among male participants that were relatively low in self-efficacy, as measured following the VR training." The first sentence here is a very strong statement that I don't believe can be stated here considering the small size of the group that once divided up showed the result. I would suggest weakening the statement or leading with the second statement and stating that this provides some evidence towards your hypothesis.

Response: We agree that, as mentioned on lines 384-386, “Our findings should be interpreted with caution due to our relatively small sample size […]” and hence these statements are too strong. Accordingly, as suggested, we toned down the statements related to our findings. It now reads on lines 333-340:

“Results went in the expected direction by showing that the use of a doppelganger helped improving performance as compared to the use of a same gender avatar role model. However, findings suggested this effect to be specific to male participants that were relatively low in self-efficacy, as measured following the VR training. These results provide some evidence towards our expectation that individuals relatively low in self-efficacy would benefit the most from doppelganger-based training than individuals with relatively higher levels of self-efficacy.”

Comment: 326-329 "We argue that, by watching their virtual self delivering the charismatic speech, participants relatively low in self-efficacy had a greater motivation to put effort into the task and to persevere than those who watched an unknown avatar performing the same charismatic speech." This effect could also be seen as a result of participants with lower self-esteem feeling better about the task after practicing the first time. It would have been interesting to see self-efficacy measured before and after the first public speech as it is unknown if self-efficacy ratings were affected by performing. (Potentially increasing for those that had previous anxiety or the other way around).

Response: Thank you for this comment. Indeed, it would be interesting to capture variation of participants’ level of self-efficacy throughout the training. Hence, it would allow assessing whether the public speaking task leads to increase or weaken participants’ level of self-efficacy depending on how they perform, as well as assessing how this variation affects their subsequent performance. We added this idea of future research on lines 404-407. It reads:

“Furthermore, it would be interesting to study whether trainees’ level of self-efficacy varies throughout the training and if yes, to assess how and the extent to which this variation is affected by the performance of trainees as well as the extent to which it affects their subsequent performance.”

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(45.3KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0245960.r005

Decision Letter 2

Doron Friedman

12 Jan 2021

Doppelganger-based training: Imitating our virtual self to accelerate interpersonal skills learning

PONE-D-20-19196R2

Dear Dr. Kleinlogel,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements. Congratulatons!

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Doron Friedman

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

Reviewer #3: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #3: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #3: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

Reviewer #3: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

Reviewer #3: Yes

**********

6. Review Comments to the Author

Reviewer #3: I believe with the revisions the authors have made to what they claimed to be exploring has helped to greatly clarify the paper and its claims. With these revisions I believe this paper is now in an acceptable state.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #3: No

PLoS One. doi: 10.1371/journal.pone.0245960.r006

Acceptance letter

Doron Friedman

15 Jan 2021

PONE-D-20-19196R2

Doppelganger-based training: Imitating our virtual self to accelerate interpersonal skills learning

Dear Dr. Kleinlogel:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Prof Doron Friedman

Academic Editor

PLOS ONE

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(49.1KB, docx)}

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(45.3KB, docx)}

Data Availability Statement

The datasets analysed during the current study, the data dictionary, and the syntax are available in the OSF repository, osf.io/8uedt.

[pone.0245960.ref001] 1.Klein C, DeRouin RE, Salas E. Uncovering workplace interpersonal skills: A review, framework, and research agenda In: Hodgkinson GP, Ford JK, editors. International review of industrial and organizational psychology. New York: Wiley and Sons, Ltd; 2006. p. 80–126. [Google Scholar]

[pone.0245960.ref002] 2.Bedwell W. L., Fiore SM, Salas E. Developing the future workforce: An approach for integrating interpersonal skills into the MBA classroom. Academy of Management Learning & Education. 2014;13(2):171–86. [Google Scholar]

[pone.0245960.ref003] 3.Salas E, Cannon-Bowers JA. The science of training: A decade of progress. Annual Review of Psychology. 2001;52:471–99. 10.1146/annurev.psych.52.1.471 [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref004] 4.Bandura A. Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall; 1986. [Google Scholar]

[pone.0245960.ref005] 5.Bandura A. Social cognitive theory: An agentic perspective. Annual Review of Psychology. 2001 2001;52:1–26. 10.1146/annurev.psych.52.1.1 [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref006] 6.Schmid Mast M, Kleinlogel EP, Tur B, Bachmann M. The future of interpersonal skills development: Immersive virtual reality training with virtual humans. Human Resource Development Quarterly. 2018;29:125–41. [Google Scholar]

[pone.0245960.ref007] 7.Anderson PL, Zimand E, Hodges LF, Rothbaum BO. Cognitive behavioral therapy for public-speaking anxiety using virtual reality for exposure. Depression and Anxiety. 2005;22:156–8. 10.1002/da.20090 [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref008] 8.Gorini A, Riva G. Virtual reality in anxiety disorders: The past and the future. Expert Review of Neurotherapeutics. 2008;8(2):215–33. 10.1586/14737175.8.2.215 [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref009] 9.Harris SR, Kemmerling RL, North MM. Brief virtual reality therapy for public speaking anxiety. Cyberpsychology & Behavior. 2002;5(6):543–57. [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref010] 10.Lindner P, Miloff A, Fagernäs S, Andersen J, Sigeman M, Andersson G, et al. Therapist-led and self-led one-session virtual reality exposure therapy for public speaking anxiety with consumer hardware and software: A randomized controlled trial. Journal of Anxiety Disorders. 2019;61:45–54. [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref011] 11.Stupar-Rutenfrans S, Ketelaars LEH, van Gisbergen MS. Beat the fear of public speaking: Mobile 360° video virtual reality exposure training in home environment reduces public speaking anxiety. Cyberpsychology, Behavior, and Social Networking. 2017;20(10):624–33. 10.1089/cyber.2017.0174 [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref012] 12.Aymerich-Franch L, Bailenson J. The use of doppelgangers in virtual reality to treat public speaking anxiety: A gender comparison. Proceedings of the International Society for Presence Research 2014. p. 173–86.

[pone.0245960.ref013] 13.Collins RL. For better or worse: The impact of upward social comparison on self-evaluations. Psychological Bulletin. 1996;119(1):51–69. [Google Scholar]

[pone.0245960.ref014] 14.Lockwood P, Kunda Z. Superstars and me: Predicting the impact of role models on the self. Journal of Personality and Social Psychology. 1997;73(1):91–103. [Google Scholar]

[pone.0245960.ref015] 15.Marx DM, Roman JS. Female role models: Protecting women’s math test performance. Personality and Social Psychology Bulletin. 2002;28(9):1183–93. [Google Scholar]

[pone.0245960.ref016] 16.Marx DM, Ko SJ, Friedman RA. The ‘‘Obama Effect”: How a salient role model reduces race-based performance differences. Journal of Experimental Social Psychology. 2009;45:953–6. [Google Scholar]

[pone.0245960.ref017] 17.Rich GA. The sales manager as a role model: Effects on trust, job satisfaction, and performance of salespeople. Journal of the Academy of Marketing Science. 1997;25(4):319–28. [Google Scholar]

[pone.0245960.ref018] 18.Taylor SE, Lobel M. Social comparison activity under threat: Downward evaluation and upwards contacts. Psychological Review. 1989;96(4):569–75. 10.1037/0033-295x.96.4.569 [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref019] 19.Bailenson J, Patel K, Nielsen A, Bajscy R, Jung S-H, Kurillo G. The effect of interactivity on learning physical actions in virtual reality. Media Psychology. 2008;11:354–76. [Google Scholar]

[pone.0245960.ref020] 20.Fox J, Bailenson JN. Virtual self-modeling: The effects of vicarious reinforcement and identification on exercise behaviors. Media Psychology. 2009;12:1–25. [Google Scholar]

[pone.0245960.ref021] 21.Fox J, Bailenson JN, Binney J. Virtual experiences, physical behaviors: The effect of presence on imitation of an eating avatar. Presence: Teleoperators and Virtual Environments. 2009;18(4):294–303. [Google Scholar]

[pone.0245960.ref022] 22.Ahn SJ, Bailenson JN. Self-endorsing versus other-endorsing in virtual environments. Journal of Advertising. 2011;40(2):93–106. [Google Scholar]

[pone.0245960.ref023] 23.Martin BAS, Kwai-Choi Lee C, Yang F. The influence of ad model ethnicity and self-referencing on attitudes: Evidence from New Zealand. Journal of Advertising. 2004 2004/12/1;33(4):27–37. [Google Scholar]

[pone.0245960.ref024] 24.Ahn SJ, Bailenson J. Self-endorsed advertisements: When the self persuades the self. Journal of Marketing Theory and Practice. 2014;22(2):135–6. [Google Scholar]

[pone.0245960.ref025] 25.Kuiper NA, Rogers TB. Encoding of personal information: Self-other differences. Journal of Personality and Social Psychology. 1979;37(4):499–514. [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref026] 26.Greco LM, Charlier SD, Brown KG. Trading off learning and performance: Exploration and exploitation at work. Human Resource Management Review. 2019;29(2):179–95. [Google Scholar]

[pone.0245960.ref027] 27.Colquitt JA, LePine JA, Noe RA. Toward an integrative theory of training motivation: A meta-analytic path analysis of 20 years of research. Journal of Applied Psychology. 2000;85(5):678–707. 10.1037/0021-9010.85.5.678 [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref028] 28.Thoresen JC, Francelet R, Coltekin A, Richter K-F, Fabrikant SI, Sandi C. Not all anxious individuals get lost: Trait anxiety and mental rotation ability interact to explain performance in map-based route learning in men. Neurobiology of Learning and Memory. 2016;132:1–8. 10.1016/j.nlm.2016.04.008 [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref029] 29.Herrero AI, Sandi C, Venero C. Individual differences in anxiety trait are related to spatial learning abilities and hippocampal expression of mineralocorticoid receptors. Neurobiology of Learning and Memory. 2006;86(2):150–9. 10.1016/j.nlm.2006.02.001 [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref030] 30.Bandura A. Self-efficacy: Toward a unifying theory of behavioral chance. Psychological Review. 1977;84(2):191–215. 10.1037//0033-295x.84.2.191 [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref031] 31.Bandura A. Social cognitive theory: An agentic perspective. Asian Journal of Social Psychology. 1999;2:21–41. [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref032] 32.Bandura A, Wood R. Social cognitive theory of organizational management. Academy of Management Review. 1989;14(3):361–84. [Google Scholar]

[pone.0245960.ref033] 33.Ackerman PL, Kanfer R, Goff M. Cognitive and noncognitive determinants and consequences of complex skill acquisition. Journal of Experimental Psychology: Applied. 1995;1(4):270–304. [Google Scholar]

[pone.0245960.ref034] 34.Chemers MM, Hu L-T, Garcia BF. Academic self-efficacy and first-year college student performance and adjustment. Journal of Educational Psychology. 2001;93(1):55–64. [Google Scholar]

[pone.0245960.ref035] 35.Locke EA, Frederick E, Lee C, Bobko P. Effect of self-efficacy, goals, and task strategies on task performance. Journal of Applied Psychology. 1984;69(2):241–51. [Google Scholar]

[pone.0245960.ref036] 36.Multon KD, Brown SD, Lent RW. Relation of self-efficacy beliefs to academic outcomes: A meta-analytic investigation. Journal of Counselling Psychology. 1991;38(1):30–8. [Google Scholar]

[pone.0245960.ref037] 37.Stajkovic AD, Luthans F. Self-efficacy and work-related performance: A meta-analysis. Psychological Bulletin. 1998;124(2):240–61. [Google Scholar]

[pone.0245960.ref038] 38.Zimmerman BJ. Self-efficacy: An essential motive to learn. Contemporary Educational Psychology. 2000;25:82–91. 10.1006/ceps.1999.1016 [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref039] 39.Gist ME, Stevens CK, Bavetta AG. Effects of self-efficacy and post-training intervention on the acquisition and maintenance of complex interpersonal skills. Personnel Psychology. 1991;44(4):837–61. [Google Scholar]

[pone.0245960.ref040] 40.Stevens CK, Gist ME. Effects of self-efficacy and goal-orientation training on negotiation skill maintenance: What are the mechanisms?. Personnel Psychology. 1997;50:955–78. [Google Scholar]

[pone.0245960.ref041] 41.Mattick RP, Clarke JC. Development and validation of measures of social phobia scrutiny fear and social interaction anxiety. Behaviour Research and Therapy. 1998;36:455–70. 10.1016/s0005-7967(97)10031-6 [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref042] 42.Lacroix F, Zhou B. Échelles traduites en français: Anxiété sociale et Taijin Kyofusho: Concordia University; Université de Montréal2015.

[pone.0245960.ref043] 43.Bergeron J. State-trait anxiety in French-English bilinguals: Cross-cultural considerations In: Spielberger CD, Diaz-Guerrero R, editors. Cross-cultural anxiety Washington, DC: Hemisphere/Wiley; 1983. p. 157–76. [Google Scholar]

[pone.0245960.ref044] 44.Spielberger CD, Gorsuch RL, Lushene R, Vagg PR, Jacobs GA. Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press; 1983. [Google Scholar]

[pone.0245960.ref045] 45.McGraw KO, Wong SP. Forming inferences about some intraclass correlation coefficients. Psychological Methods. 1996;1(1):30–46. [Google Scholar]

[pone.0245960.ref046] 46.Schwarzer R, Jerusalem M. Generalized Self-Efficacy scale In: Weinman J, Wright S, Johnston M, editors. Measures in health psychology: A user’s portfolio Causal and control beliefs. Windsor, England: Nfer-Nelson; 1995. p. 35–7. [Google Scholar]

[pone.0245960.ref047] 47.Elliott AC, Woodward WA. Statistical analysis quick reference guidebook: With SPSS examples. Thousand Oaks, CA: Sage; 2007. [Google Scholar]

[pone.0245960.ref048] 48.Field A. Discovering statistics using SPSS. London: Sage; 2009. [Google Scholar]

[pone.0245960.ref049] 49.Koenig AM, Eagly AH, Mitchell AA, Ristikari T. Are leader stereotypes masculine? A meta-analysis of three research paradigms. Psychological Bulletin. 2011;137(4):616–42. 10.1037/a0023557 [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref050] 50.Spencer SJ, Steele CM, Quinn DM. Stereotype threat and women’s math performance. Journal of Experimental Social Psychology. 1999;35:4–28. [Google Scholar]

[pone.0245960.ref051] 51.Steele CM. A threat in the air: How stereotypes shape intellectual identity and performance. American Psychologist. 1997;52:613–29. 10.1037//0003-066x.52.6.613 [DOI] [PubMed] [Google Scholar]

[pone.0245960.ref052] 52.Latu IM, Schmid Mast M, Lammers J, Bombari D. Successful female leaders empower women’s behavior in leadership tasks. Journal of Experimental Social Psychology. 2013;49:444–8. [Google Scholar]

[pone.0245960.ref053] 53.Latu IM, Schmid Mast M, Bombari D, Lammers J, Hoyt CL. Empowering mimicry: Female leader role models empower women in leadership tasks through body posture mimicry. Sex Roles. 2019;80:11–24. 10.1007/s11199-018-0911-y [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0245960.ref054] 54.Smith-Jentsch KA, Salas E, Baker DP. Training team performance-related assertiveness. Personnel Psychology. 1996;49(4):909–36. [Google Scholar]

[pone.0245960.ref055] 55.Hallgren KA. Computing inter-rater reliability for observational data: An overview and tutorial. Tutorials in Quantitative Methods for Psychology. 2012;8(1):23–34. 10.20982/tqmp.08.1.p023 [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Doppelganger-based training: Imitating our virtual self to accelerate interpersonal skills learning

Emmanuelle P Kleinlogel

Marion Curdy

João Rodrigues

Carmen Sandi

Marianne Schmid Mast

Roles

Abstract

Introduction

Materials and methods

Experimental design

Participants and recruitment

Materials

Procedure

Session 1

Session 2

Fig 1. Virtual conference room.

Fig 2. Charismatic speech view.

Fig 3. Meeting of the virtual person.

Measures

Body language persuasiveness

Self-efficacy

Identification with the virtual person

Results

Fig 4. Body language persuasiveness.

Discussion

Acknowledgments

Data Availability

Funding Statement

References

Decision Letter 0

Doron Friedman

Roles

Author response to Decision Letter 0

Decision Letter 1

Doron Friedman

Roles

Author response to Decision Letter 1

Decision Letter 2

Doron Friedman

Roles

Acceptance letter

Doron Friedman

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases