The potential of virtual reality to combat teacher bias and increase empathy

Abstract

Teachers’ biases surface in interactions with students and influence students’ learning and attitudes toward school. Policymakers have increasingly called for anti-bias training for teachers; however, most research has shown little sustained effects. Virtual Reality (VR) interventions to combat bias have shown promising results; however, only a few intervention studies exist. This preliminary study included nine educators from the Northwest of the United States who engaged in a VR experience in which they embodied an avatar experiencing bias from others. They were then interviewed about VR’s potential to increase empathy in education contexts. Our findings highlight important aspects for consideration in designing VR scenarios and their use in anti-bias training. Overall, participants saw great potential in using VR due to its immersive nature. Participants noted that feelings of embodiment related to the graphics’ quality, level of interactivity, and familiarisation with the technology. Participants felt that the experience should be embedded in training that included awareness raising before and debriefing after the VR experience and included strategies to change behaviour. The study discusses specific consideration for the design and implementation of VR experiences to increase empathy and how to avoid participants’ defensiveness and raise awareness of their potentially biased deep-seated beliefs.

Introduction

Just as biases towards minority groups exist in society, they likewise exist in educational institutions (e.g. Blank et al., 2016). It is well-documented that teachers’ biases surface in interactions with students and, as a result, influence students’ learning, academic beliefs, and attitudes toward school (e.g. Brophy, 1983; Cheng and Starks, 2002; De Kraker-Pauw et al., 2016; DeCuir-Gunby and Bindra, 2022; Denessen et al., 2022; Lautenbach and Antoniewicz, 2018; Meissel et al., 2017; Rubie-Davies, 2006).

Policymakers have increasingly called for anti-bias training for teachers. However, there is a dearth of research on such training programmes as they often do not seem to undergo a rigorous evaluation (Carter et al., 2020; Schmader et al., 2022). The research that does exist tends to show that interventions often have little sustained effect (Bezrukova et al., 2016; Carter et al., 2020; Chen et al., 2021; Paluck et al., 2021). Schmader and Colleagues (2022: 1397) note that ‘there is a high demand for theoretically derived and evidence-based interventions aimed at fostering equity, diversity, and inclusion’. While recent reviews on anti-bias training note that the research base is still inconclusive on best practices and effectiveness (Devine and Ash, 2022; Schmader et al., 2022), interventions need to use different cognitive strategies to be effective, including raising awareness of one’s own biases, the nature of bias and its impact, reducing one’s own biases and increasing empathy, and providing strategies to change behaviour (Paluck et al., 2021; Schmader et al., 2022).

Perspective-taking is one of the strategies recommended for building awareness of biases, biased behaviours, and their impact; it may also be effective in increasing empathy (Todd and Galinsky, 2014). The use of Virtual Reality (VR) provides an opportunity to enable an immersive perspective-taking experience in which participants embody another person (Chen and White, 2024; Stavroulia and Lanitis, 2023). However, most research to date on the use of VR has consisted of experimental studies documenting bias in psychology or the medical field, and few intervention studies exist that aim to increase empathy and address bias (Hatfield et al., 2022; Paluck et al., 2021; Tassinari et al., 2022).

We present a preliminary study exploring the potential of VR to combat teacher bias and increase empathy. In our study, educators engaged in a VR experience aimed at increasing empathy for the embodied avatars. The experience included two embodiment scenarios. In the first scenario, the participant embodied a homeless person in his work context and at a shelter. In the second scenario, they embodied a black woman receiving care in a hospital birthing unit. Participants were subsequently interviewed to answer the following research questions: (1) What are educators’ perceptions of the potential of VR to increase feelings of empathy? (2) What are educators’ perceptions of VR scenarios as a professional learning tool to increase empathy in education settings?

Combatting teacher bias

Different constructs in the literature are often used synonymously with bias, such as attitudes, stereotypes, and prejudice. However, we agree with Schmader and colleagues (2022) that there are clear delineations between these constructs. In this study, we focused on ‘social biases as the outcome of a set of processes by which the activation of group-relevant cognitions (e.g. stereotypes or attitudes) lead to, or influence, one’s behaviour toward a member(s) of that group’ (Schmader et al., 2022: 1385).

It is noted that bias has cognitive, affective, and behavioural components. While we now have a greater understanding of ‘how and when stereotypes and prejudiced attitudes can shape judgment and behaviour’, there is still ‘a profound gap between this research and its practical application’ (Schmader et al., 2022: 1382). Two types of biases are generally discussed: implicit and explicit bias. Implicit bias results from either a lack of awareness of our stereotypes or negative associations or a lack of effective strategies to regulate our behaviour. Explicit bias occurs when we are aware of our stereotypes but unmotivated to control our behaviour. Ultimately, the goal of anti-bias training is to change biased behaviours, and one way to do so is to increase empathy. Empathy can be divided into affective empathy (feeling what the other person might feel) and cognitive empathy (understanding the other person’s responses and perspectives) (Brydon et al., 2021; Ventura et al., 2020). Empathy is an important base for positive relationships and is seen as an essential skill in teaching (McAllister and Irvine, 2002; Stojiljković et al., 2012).

Perspective-taking to increase empathy

One of the strategies recommended to build awareness of biases and the impact of biased behaviours is perspective-taking (Todd and Galinsky, 2014). Perspective-taking is a self-intervention in which one purposely thinks about what the other person experiencing bias is thinking and feeling. Some studies have used audio, video, or other media to engage participants in perspective-taking and have shown some positive impact on empathy and attitudes towards minority groups (Herrera et al., 2018). The use of VR is now providing an opportunity to enable an immersive perspective-taking experience for participants in which they embody the other person (Brydon et al., 2021; Stavroulia and Lanitis, 2023; Ventura et al., 2020). Some note that perspective-taking and embodiment should be seen as related but distinctive processes, with the latter supporting perspective-taking as participants do not have to rely on their imagination (see Chen and White, 2024).

Virtual reality as a potential empathy machine

Rapid technological advances have seen VR offer powerful, immersive, and interactive learning experiences in highly controlled environments (Ahn, 2021; Blascovich and Bailenson, 2011; Blascovich et al., 2002). However, most VR research has mainly consisted of experimental studies documenting bias in psychology or medicine. Only a few intervention studies exist, and these have mostly been light touch (i.e., under 10 min) and conducted in labs with graduate students. Thus, their applicability to real-life contexts is uncertain (Paluck et al., 2021). Hatfield and colleagues (2022) conducted a systematic review of VR intervention research regarding racial bias. They noted that 61 of the 68 studies reviewed only examined whether bias existed but did not explore how biases could be mitigated. These studies often used VR to enable participants to have contact with avatars presenting an out-group (e.g. avatars of a different race). Some of these studies used VR shooter games to determine whether participants aimed more quickly or more often at targets of a different race. Such studies have, unfortunately, often been shown to increase bias rather than reduce it (Hatfield et al., 2022).

Some studies have experimented with participants embodying an avatar with, for example, a different skin colour, to enable perspective-taking (Groom et al., 2009). Under VR headsets, participants can see and move ‘their body’ (e.g., see their arms or legs move and come across their ‘own’ reflection in a mirror). They can turn in any direction and see, hear, and interact with surrounding events, creating a feeling that they are present or even participating in the event (Bailenson, 2018; Stavroulia et al., 2018, 2019). In one study, participants also heard the character’s inner monologue, gaining insights into how the events affected the character emotionally (Roswell et al., 2020). These studies have shown that embodiment experiences promote mental merging, where, for a time, participants take on behaviour or characteristics associated with their avatar, and these can carry over into the real world and promote empathy and altruism (Herrera and Bailenson, 2021; Schutte and Stilinović, 2017; Van Loon et al., 2018). However, if the avatar lacked a real context or persona, biases again seemed perpetuated. For example, in Banakou et al.’s (2020) study, participants who embodied a black avatar and witnessed negative behaviour from people walking past (e.g. turning away or staring) experienced an increase in bias. Hatfield and Colleagues (2022: 6) described this form of embodiment used to investigate racial bias as ‘virtual blackface’. Virtual reality experiences in which the avatar has a name and history, and which offer views into the avatar’s experiences of bias have shown more positive results in increasing empathy and reducing bias (for a positive example, see Cogburn et al., 2018).

In one of the few such studies in education to date, Stavroulia and colleagues (2018, 2019) explored the design of a virtual classroom for teacher training, which had a small but positive impact on teachers’ empathy towards diverse students. In Stravoulia and colleagues’ (2023) latest study, teachers could talk to diverse student avatars or embody a refugee student and a student with a prosthetic leg. Only the teachers who embodied the students had increased empathy scores after the VR experience. While this study shows the potential for VR to combat bias, more research is needed on how teachers experience VR as a tool for increasing empathy and what considerations should be taken regarding VR scenario design and technical implementation.

Methods

This study used a qualitative research design in which participants were asked to take part in a 10-min VR experience and a one-hour interview afterwards about their experience. We invited educators from a state in the Northwest of the United States to participate in the study. Nine educators participated in the study. They held a range of roles at the tertiary, district, and school level. Six participants were female, and three were male; they ranged in age from 30 to 60 and included a range of racial/ethnic identities, including African American, Asian, and Middle Eastern. Only one participant had previous experience with VR; none had experience embodying an avatar in a virtual environment.

Virtual reality experience

The VR experience took about 10 min and included two VR scenarios in a beta development version. The VR scenarios aimed to create empathy by having participants embody a homeless man at work and at a shelter, and in the second scenario, a woman of colour as a patient in a hospital birthing unit. Both characters experienced biased behaviour from other characters in the scenarios, including colleagues and medical personnel, for example, expressing negative views about the roots of homelessness and blaming the mother for complications based on stereotypes. The VR experience was programmed to create an artificial but authentic environment rather than using a filmed environment. Participants could look around the environment and move their avatar’s arm and head; however, they could not interact with the environment (i.e., move around in the room, move objects, speak). The VR experience thus presented what is classified as a limited mobility VR (Chen and White, 2024). Participants could hear the characters’ conversations with others in the scenarios. They further heard their avatars’ inner monologues, providing the participants with information on their avatars’ feelings and thoughts about their experiences and how these negatively impacted them. The VR experience also included short audio clips alongside still photographs of people in similar situations to the embodied avatars (i.e., experiencing homelessness and bias in healthcare settings) after each scenario. These spoken stories provided examples of experiences of bias from real-life contexts, thus supporting the ‘constructed’ scenario in which the participants had embodied an avatar.

Participants were seated for the VR experience and received a brief introduction to VR and the theme of the scenarios to ensure they were aware of the sensitive topics, the risk of VR creating nausea, and how to withdraw from the experience if needed. Participants wore a headset (Oculus Quest 3) and held controllers and had to press a button appearing in front of them to start the experience. Participants could ask questions or take off the headset anytime during the experience.

Data collection and analysis

Interviews of around 1 hour were conducted directly after participants had gone through the VR experience. We used a semi-structured interview protocol and follow-up prompts to collect rich data about participants’ experiences (Ruslin et al., 2022). Three interviews were conducted with individual participants; three were conducted in pairs. The interviews focused on participants’ perceptions of the VR experience, factors that helped and hindered feelings of embodiment and empathy, and how the VR experience could be adapted for an education context (for example, as part of a training course). The interviews were audio-recorded and transcribed by the researchers.

We were aware of the theoretical knowledge we bring to the analysis process and thus engaged in a reflective thematic analysis (Braun and Clarke, 2022; Clarke and Braun 2013) of the interview transcripts. In our iterative coding process, we used a mix of deductive and inductive approaches. We used NVivo as a qualitative data organisation tool for the analysis. Our very broad initial themes aligned with our research questions: (1) the potential to increase empathy and combat bias; (2) the potential application to an education context, and (3) technical aspects of the VR scenarios that helped or hindered participants’ experience of embodiment and relating to the characters. Within these themes, we coded inductively to identify sub-themes that participants raised. One researcher conducted the initial analysis and shared a list of themes with related data excerpts. We reviewed the data as a team to refine the themes and definitions to ensure clarity and delineation of themes. Decisions on the coding structure involved revisiting data sets and looking for disconfirming evidence, checking understanding of themes and data excerpts until a consensus was reached on a final coding structure. A second coding iteration applied the agreed-upon coding structure, and the complete coding was reviewed as a team. The review and discussion of discrepancies involved both researchers to ensure trustworthiness. Further, we documented all coding decisions and created a detailed code book to ensure rigour, credibility, and transparency (Shenton, 2004). Regarding the trustworthiness of the data and its generalisability, the small sample size needs to be acknowledged. Further, the recruitment for the study was an open call for participants, and hence, the sample only included participants who seemingly had an interest in the application of VR for anti-bias training. The sample could, hence, have an influence on the findings and limit their wider generalisability. The findings only present an initial exploration of the potential of VR for increasing empathy in teachers. It would be beneficial for future research to include a larger and more diverse sample to better understand the potential of VR across different education contexts and demographics.

Findings

We present the findings regarding the three main themes below: the potential of VR to increase empathy and combat bias, the potential application to an education context, and technical aspects of the VR scenarios that helped or hindered participants’ experience of embodiment.

The potential of VR to increase empathy and combat bias

Three themes emerged from the data related to the potential of VR to increase empathy and combat bias: feelings of immersion, increased empathy through perspective-taking, and enabling participants to address their own biases.

Feelings of immersion

All the participants perceived VR as a potentially impactful tool for an anti-bias training course. They reported feeling more focused than in typical training sessions in which they would listen to a speaker, review written scenarios, or engage in discussions with colleagues. They attributed the increased focus to their other senses being dulled and the lack of distractions while in the VR experience. Almost all participants noted that the impact was heightened because they were unable to multi-task or let their minds wander while they were wearing the VR headset. As Participant 2 noted, ‘It blocks everything else out. Right? Yeah, literally puts your blinders on. So, you can't really focus on anything else’. Participant 5 echoed this, saying, ‘It certainly increases the intensity of the experience than if I were just watching it on a screen, for sure … It was … very, very intense information. And certainly, looking at it from a first-person perspective, I think, is impactful’.

Participants felt that the novelty of the VR format increased its potential for affecting greater retention of key messages when used for professional learning. As Participant 7 put it,

Teachers sit through hours of training every single year. And the novelty of something different and being immersed in a different experience sticks with you differently than coming to another PD where you can learn about similar information … [I] think there is something there, I think we are always trying to think about, especially for adult learners, how do you have information stick?

Participant 2 reflected that, due to the individual nature of experiencing the VR, they were freer to let themselves respond authentically to the scenarios; in contrast, they felt more cautious and under observation by others when engaging in a training in a room with colleagues:

I felt more free just to have whatever thoughts or feelings I was going to have about it than I feel like I am able to in a class setting or if I am sitting in the library where the whole staff is together. We are all doing something together, and I feel like I need to be a little bit more cautious about how I am interacting with that situation.

Some participants thought that experiencing the VR on their own may reduce the defensiveness that teachers might otherwise have towards their pedagogical practices. However, others noted the tendency to automatically ‘make excuses’ for the biased behaviours witnessed, especially if the bias was too blatant. As Participant 6 said,

When I watched the video of the jerky doctor being super, super mean, I distanced myself from that person. And I distance myself from anybody who would be like that, just like when I watch videos of really punitive discipline-oriented teachers. You think, ‘Well, that that is not us, like, that is a different … they may be the same skin colour as me or same gender as me, but that is not me.

Participants felt that embedding the VR in broader anti-bias training might reduce the tendency to become defensive.

Increasing empathy through embodiment and perspective-taking

Participants reported feeling they understood the characters’ perspective more deeply through VR than traditional training. Participant 5 reflected on the embodiment experience, noting its impact:

It is just a different perspective when you're trying to look out of the eyes of someone else. And really … the feeling that you get when you are literally standing in front of that person checking you in, in a shelter, it just, it puts you there in a way that is very different than if you are reading a story. … With the internal monologue, it just feels like a different presentation, kind of a jolt; this is what we're really talking about … this is what it could really feel like when you're approaching the stand to check-in.

The first-person perspective of what the character is hearing, seeing, and thinking was reported as increasing understanding of the avatar’s perspective. Participants noted that teachers can become desensitised to the systemic issues that present themselves in a myriad of ways in their classrooms; they felt that the VR format could break through teachers’ ‘walls’ and increase teachers’ empathy for their students’ realities. As Participant 2 put it,

One thing that I have a lot of questions about, as an educator, is what my students’ experiences are outside of the classroom, and how that would impact their internal monologue in the classroom and how tired they are or why they are not getting their homework done, or whatever it is, right? My reaction to that is sort of sterile because I do not understand their context very well. If I had at least some perspective into what their context might be, that would help to inform my decision at that moment as to … what my homework policies are and that kind of thing in my classroom.

Participants felt that while the perspective-taking element was instrumental, the scenarios needed to resonate with classroom situations. It was hard for them to extrapolate from non-education scenarios to consider their biases and how they impacted their behaviour with students in the classroom.

Combatting bias

When asked about the potential of VR to reduce teacher bias, participants identified the challenge of addressing unconscious bias. Participant 2 noted, ‘It's going to be, I think, very difficult, just because of how … individualised, everybody’s own biases are and where people come from, like the context that they assume about other people’. Rather, our participants felt it might be more impactful to focus on more systemic biases that exist in the culture of education systems; this could be a starting point to enable teachers to gain an overall awareness of the pervasiveness of bias.

Participants noted that it was important for the training experience to be moved away from shame and towards owning and identifying biases so that teachers can move towards dismantling them, both in themselves and within the educational system. Participants commonly expressed the sentiment that there is an understanding that bias exists, but there is also discomfort in acknowledging that teachers behave in ways that perpetuate these biases. As Participant 4 noted,

Part of it is just like, is the moving away from the shame piece of it and owning and identifying it, like recognising that [bias] exists, not having to [say] … this is a reflection of who I am, but let's talk about where this bias comes from and what I have to do to dismantle that.

Participant 7 drew attention to the impetus for changing biased behaviour and suggested that capitalising on teachers’ belief systems of concern and care might catalyse actual behavioural change, rather than the impetus coming from a place of forced allyship:

There seem to be two competing views of equity and inclusion based on the way it is presented. … [Framing it as] white privilege … that your bias, your prejudicial behaviours are negatively impacting people of colour, or people different than you … does that lead to change? Or… can it be something … calling on the humanity of their lived experience and saying … it is in our shared interest to love and respect and treat people as people through an asset, you know, that these people have assets as individuals? Does that lead to more change? And I don't really know which one, to tell you the truth. I feel like in education, we moved into this, like we are going to be actively anti-racist institutions, even though that framing of that doesn't sit well with most of our constituents.

This caution exemplified the care with which the VR scenarios need to be crafted in order to provide an entry point into understanding the effects of biased behaviours and a platform in which to evaluate how these behaviours resonate within teachers’ pedagogical practice.

Adaptation to an education context

Participants saw possibilities for ways VR could be adopted as a tool for a training course for teachers, but they stressed the need for the scenarios to resonate with an education context. Lack of proximity to the scenario contexts was viewed as a barrier to forming empathy for the characters and situations. Participants suggested the VR experiences could have teachers experience a ‘day in the life’ of their students to build a greater understanding of their context. As Participant 2 noted,

As a high school secondary teacher, I am more vulnerable to that I only see this kid in this very specific context, you know, like, I teach math. So I see you in math class, and I see you never anywhere else. So I have no idea what you are like, and the rest of your life. So it is much easier to assume that people are more unidimensional … Being able to see students kind of throughout … their day or throughout their lives, or, you know, just little snippets of what it is like that has some unifying theme to it, whether it is the impact of homelessness on your school life … and tying a story together about that, or if it is how your race impacts where you are, or your gender, or your gender identity, or whatever it is.

Participants suggested that including multiple ‘snippets’ could help teachers understand how subtle microaggressions accumulate to cause harm.

A common view was that including positive and negative scenarios with internal monologues of students experiencing varied behaviours would be beneficial. As Participant 2 said,

That might be compelling for teachers to want to change what they are doing. If they see, ‘Oh, yeah, I have that behaviour too. And this is a different way that I could do it’ by seeing this positive example. Or if nothing else, even if they are like, ‘for reasons x, y and z, I don't want to do what's happening in that positive example. But at least I can see how that positive example impacted the student from their perspective. So then I can think about ways of okay, well, how can I do that in a way that I am okay with, even though it is not what that positive example was’? So even from that perspective, I think having [positive] example[s] would be beneficial.

The idea of including positive examples was linked to the ultimate goal of not just increasing teachers’ understanding that bias exists but increasing their motivation to change their own behaviours that might be contributing to systemic bias. Our participants were unanimously interested in this idea of VR as a purposeful vehicle for change.

Technical aspects of the VR scenarios

Participants reported a varied sense of their embodiment of the avatars during the VR scenarios. Generally, they felt immersed, as described above, and compassionate for the avatars, but, as Participant 9 put it, ‘in neither scenario did I really feel like I was that person’. Our participants suggested that greater embodiment would occur if the scenarios were more interactive. As Participant 1 said, ‘If I had had control to … make decisions about how this was unfolding, I probably would have felt a lot more embodied’. Participant 7 echoed this, suggesting that, in the scenario with the homeless man,

he is in the kitchen with his co-workers … if I were given the choice of … do I go in this room with these people, or do I go here, where I am alone … you know, those kinds of choices where I would be involved in the story because I was not involved in the story.

Participant 5 suggested adding a time element in, for example, in the homeless scenario,

Like, you know, if you had to walk up and talk to the [shelter administrator] … and then they say, ‘Oh, you can have this bed … [but] you cannot check in for 15 more minutes’ … and then you had to turn away and go try to sit somewhere for 15 minutes, I could see how that [could] be impactful.

Participants also felt that the technical drawbacks of the VR scenarios were disruptive to the immersive experience. As Participant 5 said, ‘They’re so close to humans. I almost wish they were just a little more cartoony because there’s the creepiness factor’. Participant 3 noted, ‘It was a little bit glitchy. It's a little bit distracting [that] … it was kind of like jerky movement [which felt] unnatural’. Participants suggested that either increasing the realism of the characters or designing them to be purposefully cartoonish would increase the ability to engage with them without the obstacle of straddling the slightly absurd line between the two. Some felt 360-degree videos would be more immersive than the three-dimensional avatar environment; however, most had not seen any video material in VR.

Finally, some participants felt it would be beneficial to have time to familiarise themselves with the virtual format prior to engaging with the scenarios. This would aid in focusing solely on the stories rather than simultaneously adjusting to the technology. Participant 5 suggested that ‘a warm-up video that was less emotion-driven so that someone gets a chance to get used to it if they’re new to [VR] before they have, you know, have the sort of emotional pull that those videos have’.

Discussion

The study explored the potential of VR to support efforts of combatting teacher bias and increasing empathy. First, participants noted that the VR experience enabled them to understand another’s perspective due to it capturing their attention, immersing them in an environment, and having them embody another person. These sentiments echo other studies and reviews showing the potential of VR for tackling bias through perspective-taking (e.g. Chen and White, 2024; Paluck et al., 2021). However, our study also highlights important aspects for consideration in designing VR scenarios and their use in anti-bias training with teachers. These include the contextualisation of the scenarios, participants’ level of familiarity with VR, and technical implementation.

First, participants felt VR had great potential to increase empathy but noted that a contextualisation of the scenario was essential. Several design aspects were deemed important, including the type of biased behaviours displayed (e.g., overt racism increased defensiveness), the proximity of scenarios to participants’ own experiences (e.g., classroom and student perspective), and ways to display the impact of ongoing and/or systemic bias (e.g., through including different ‘scenes’). Other researchers have called for the contextualisation of scenarios and avatars to avoid ‘virtual blackface’ (Hatfield et al., 2022), but few have focussed on how the contextualisation of scenarios might impact participants’ experiences and application to their contexts.

Hearing the inner monologue of the embodied character was seen as enabling perspective-taking, a design aspect not widely described in other studies. The internal monologues helped participants identify and understand biased behaviours and conveyed to them that hurt occurs even when those being harmed do not vocalise their trauma. Often, people not sensitised to bias do not notice biased behaviour.

Contextualisation, or the lack thereof, might explain some of the mixed results in the very recent studies using VR (Chen and White, 2024). As pointed out in recent reviews by Chen and White (2024) and Tassinari et al. (2022), studies often do not share specific details of the scenario design in publications; hence, it is difficult to examine how design aspects might influence results and to compare study results. Such detail is important, as Stelzman et al. (2021) showed that participants’ bias was reduced towards people with schizophrenia only when they liked the person with schizophrenia they encountered in the scenario.

While content and contextualisation might be important, participant characteristics might also play a role, as some studies noted gender differences in feelings of embodiment with male participants finding it more challenging to embody female avatars (e.g. Kessler and Wang, 2012). Here, content and technical design go hand in hand, and other technical design aspects also seem to influence the experience and the level of embodiment. In our study, participants were unfamiliar with VR and thus seemed initially distracted by the experience of VR itself. Thus, researchers should consider the intended audience for the experience or include a familiarisation period or exercise in the VR experience before the actual ‘intervention’ scenario. Familiarisation with VR is especially important if the experience includes interactivity. While interactivity is considered an enabling factor for embodiment feelings (Chen and White, 2024), it could also act as a distractor if the participant struggles with the activity and controls.

A final technical aspect is the quality of the graphics, which is seldom discussed in studies; low quality graphics can undermine feelings of embodiment, either by being too unnatural as noted by participants in this study, or by very closely resembling humans in many respects but not being quite convincing, leaving participants with an unsettled feeling called the ‘uncanny valley effect’.

Additionally, to adapt the VR experience for use with teachers, participants in this study noted that it should be embedded in a training course that included awareness raising before and debriefing after the VR experience, as well as strategies to change behaviours. This echoes findings from prior research that noted the limited efficacy of VR interventions without follow-up (Hatfield et al., 2022).

The opportunity to take someone else’s perspective can raise awareness of participants’ own bias and might trigger behaviour change. However, to support participants to move beyond an awareness of bias to changing biased behaviours, an intervention that includes a VR component should combine cognitive and emotional strategies (Devine and Ash, 2022; Paluck et al., 2021). As Carter and colleagues (2020) note, training often merely raises awareness of bias and can sometimes trigger defensive reactions when participants are confronted with their biases without being given strategies to move forward. Our participants warned that such defensiveness was likely if teachers felt blame and shame for the experiences depicted in the VR scenarios, rather than being instructed to reflect on the emotional experience. Thus, it is important to include strategies for change in either an accompanying training or in the VR experience.

Finally, several challenges need to be noted in developing and implementing VR experiences in education settings. The first challenge is the cost of the VR experience development, the VR headsets, and potential technical support. Developing high-quality VR experiences is costly, with estimates of US$10,000 for a couple of minutes for a programmed experience. The option of 360-degree filming can be less costly but is not adaptable. Once filmed, there are only a few changes that can be made. Filmed experiences also offer fewer possibilities of interactivity, for example, the participant cannot move or pick up items in the experience. While the cost of VR headsets has reduced in the last decade, there are still purchasing costs, the cost to hire technical staff, and the time involved in their set-up and upkeep. However, universities and initial teacher education programmes, training providers, and, to some degree, schools are investing in VR for teaching purposes (see Huang et al., 2023). Another challenge could be seen in educator resistance to using new technology; however, with the introduction of VR in initial teacher training and in-service training, and in classroom teaching, the technology will become more familiar. As with any new technology, and as Carpenter and colleagues (2023) note, the adoption of VR typically depends on teachers’ perceived readiness and confidence in their technological skills and a belief that VR has potential benefits for themselves or their students’ learning.

Our study’s findings thus highlight the infancy of research into the use of VR for increasing empathy and combatting bias and point to several areas for further exploration. First, intervention studies are needed that measure the impact of such VR experiences on teachers’ empathy and beliefs. Important questions exist in whether there are significant effects of VR and whether these effects are sustained and result in long-term behaviour changes. Thus, longitudinal research would be valuable in assessing the durability of these effects and determining whether additional interventions are needed to reinforce the initial impact.

Furthermore, as we noted, there are considerable design decisions regarding the VR experience and what design aspects or what kind of experience is most impactful is another area for future research (see also Hatfield et al., 2022). However, what is needed for such exploration are researchers with the technical know-how or access to technical support, funding for development, and time to consult and develop VR scenarios in consultation with teachers and communities they are trying to serve to ensure contextualisation and to avoid perpetuating bias or raising defensiveness.

Policy implications

Addressing persistent inequities in educational outcomes requires governments, communities, and teachers to acknowledge – and disrupt – the systemic bias that hinders opportunities for all students to be successful. Combatting bias requires shifts in beliefs and behaviours, and perspective-taking through virtual reality may spur individual action through increased empathy. Policies targeting diversity, equity, and inclusion can prompt wider efforts to enact change. Indeed, jurisdictions around the globe have embraced policies aimed at reducing bias (Smith et al., 2023), but we believe change requires a coordinated effort to bridge policy aspirations and realities in the classroom.

There is a social aspect to policy implementation, especially in schools and especially when a policy asks teachers to try something new (Coburn and Stein, 2006; Dumas and Anyon, 2006; Honig, 2006; Little, 2003; Spillane et al., 2006). Virtual Reality experiences in which teachers embody a student experiencing bias could raise teachers’ awareness of their own biases, increase their empathy, and lead them to critically engage with their beliefs and behaviours. However, considerable care needs to be taken in the design and use of VR scenarios to reduce or avoid participants’ defensiveness and to enable participants to acknowledge deep-seated beliefs and engage in open discussions about how to embrace diversity in their classrooms.

Furthermore, such learning experiences are less likely to succeed without the requisite ‘social infrastructure to foster learning’ (Wenger, 1998: 225). Hence, a VR experience, in which teachers are immersed in perspective-taking to jolt them into changing behaviours, needs to be embedded within broader professional development that provides opportunities to engage in sense-making (Spillane, 2000; Spillane et al., 2006). Policy that supports such time for learning can turn aspirations into outcomes. Policies that then address systemic issues of bias are also needed to drive long-term efforts to dismantle bias.