Defining empathy: Interconnected discourses of virtual reality’s prosocial impact

Data collection: popular corpus

Consistent with prior work (Foxman, 2018), we collected popular press articles from the top five US publications listed for most-read VR news sites (Fields, 2017), Alexa data insight’s business news sites, computers/tech websites, newspapers (e.g. The New York Times), games news and reviews (e.g. IGN), and the blogging platform Medium. Early adopters of commercial VR turned to these outlets to promote and make sense of the technology and associated concepts, like empathy.

For each website, a manual Google search was performed for articles with “Virtual Reality” and “empathy” as key terms. When an outlet exclusively focused on VR, only “empathy” was searched. Queries and data collection occurred in December 2018.

Inclusion criteria

Articles were eliminated that did not directly address both VR and empathy. For instance, a Wired election blog was removed because, while it mentioned a VR experience about the Republican debates and lack of empathy by Donald Trump, it neither connected the topics nor addressed their relationship (Lapowsky, 2015). Items including the terms only in advertisements were also excluded. Finally, duplicate articles within the same publication were deleted. From an original set of 1689 links, the final corpus consisted of 640 articles from 22 outlets. Table 1 summarizes the corpus by publication type; Figure 1 traces publication frequencies over time.

OPEN IN VIEWER

Table 1.

Popular corpus composition.

Article types	Number of articles
Medium	230
Enthusiast press (e.g. RoadtoVR)	169
Business outlets (e.g. Forbes)	113
Tech outlets (e.g. Tech Radar)	77
News outlets (e.g. The New York Times)	43
Games journalism (e.g. IGN)	8

OPEN IN VIEWER

Figure 1.

Histograms of articles over time. VHIL is Stanford’s Virtual Human Interaction Lab.

Data collection: academic corpus

The authors’ university library system identified papers that contained the terms “virtual reality” and “empathy” as subjects. This initial search produced 50 articles, and several criteria created the focal dataset.

First, we excluded six irretrievable articles, one piece that reviewed work from the existing dataset, and one non-English article. We then supplemented our initial search with proceedings that commonly index VR articles (Association for Computing Machinery). This added 11 articles to our database for a final set of 53 articles. We used all sections of the academic research article in our analyses except for the references. Articles used in this corpus are located in Supplemental Table S1.

Thematic reviewing procedure

Two reviewers performed an inductive qualitative critical analysis of the popular corpus to identify common themes and discourses (Thomas, 2006). They read each article to identify themes surrounding VR and its relationship to empathy: how outlets defined the concept; how they represented its impact on industries and consumers; how they related empathy to specific institutions and practitioners; and finally, the relationship between VR, empathy, and prosocial change. Following the critical tradition (Van Dijk, 1993), reviewers also scrutinized institutions and figures wielding power within the media landscape to specify the uses and effects of the emerging medium. Rather than setting forth an initial set of categories, reviewers allowed them to emerge as they read.

We used the same procedure to identify how researchers defined empathy. Recognizing that the popular press may cover the results of these studies without explaining the research process, comparisons were not made between popular and academic press concerning coverage of measurement. The reviewers did, however, code the academic press for measures of empathy to determine how researchers report claims about VR as an instrument to boost empathy.

Because such qualitative analyses are premised on themes emerging naturally from the text, we sought “trustworthiness” in our findings (Lincoln and Guba, 1985). Following Lincoln and Guba’s (1985) suggestions of prolonged engagement, persistent observation, and researcher triangulation (see Nowell et al., 2017), we ensured credibility with our coding by having two independent raters meet periodically for peer debriefing during reading and analysis phases. Coders also read the entirety of each corpus, strengthening the credibility and validity of their interpretations. They compared, discussed, and checked codes to see whether they could cogently reconstruct similar findings. Such methods and steps to ensure coding trustworthiness have been used in a variety of inductive thematic analyses (e.g. Fjællingsdal and Klöckner, 2017; Kors et al., 2016), including multimethod projects (e.g. Fox et al., 2018). The qualitative analysis thus represented a deep reading of the nuances of the popular and academic press’ framing of VR and empathy, as well as their perceived impact on prosocial change.

Perspective-taking and embodiment

Popular press writers viewed the idea of occupying someone else’s body in VR as an instrumental affordance of the medium. This trait was fairly prevalent, coded in nearly 16% of articles (100/640). Although embodiment has long been studied as it relates to gaming and virtual worlds experienced through computer screens, evangelists touted VR’s technological prowess in facilitating embodiment: “Virtual reality is uniquely effective at creating . . . [an] immersive perspective-taking experience . . . where users can inhabit digital bodies that look nothing like them . . .” (AppliedVR, 2016). Further, writers suggested that this affordance would catapult the medium into mainstream use: “That could be key to VR fulfilling its destiny . . . allowing us to absorb someone else’s perspective by acting out their life in our own skin” (Constine, 2018). Thus, writers signaled that embodiment contributed to VR’s commercial and practical success.

Immersion and presence

Articles suggested immersion was the bridge between the technology and the development of empathy: “Virtual reality is the really special place where technology and empathy collide. It creates an immersive universe for storytelling unlike any that has come before it” (Graham, 2015). The technical features of immersive devices were lauded for enhancing empathy: “[T]he 360° nature of virtual reality creates the illusion that the spectator is within the image . . . It is from this that empathy and suspension of disbelief stem” (Ducros, 2017). Further, presence provided a new sense of space for the user: “Presence, the feeling of ‘being’ in a mediated space, is the wow factor of virtual reality” (Humphrey, 2017).

While empathy and VR were positively connected, there was a clear tension. Presence, embodiment, and immersion were considered empathy drivers, which is substantiated by research (Bailenson, 2018), but were not situated within the processes of how humans could empathize with each other, nor used to detail how empathy develops more broadly.

The empathy machine

Journalists articulated and connected empathy to VR through a set of metaphors and comparisons. The most frequent association was the portrayal of VR as an “empathy machine,” which appeared in 174 articles and characterized empathy more than any other explication. The term stems from Milk’s (2015) talk, where he used his work to suggest that VR is the best means to elicit empathy in a viewer because it provides users an unmediated experience, which facilitates compassion and human connection. Milk’s assertions are mirrored in the popular corpus and underscore the idea that empathy can be mechanized and operationalized through VR. Freeman (2015) wrote, “Virtual reality is the ultimate empathy machine. It enables us to experience things we wouldn’t typically see in the real world and provides the opportunity to gain a new perspective.” In other words, according to popular press writers, VR is empathy or at least the utmost technical manifestation of it.

Not all authors agreed with the metaphor; some debated the extent, efficacy, or even the necessity of the engine. In Medium posts, it was decried as a “cynical tactic designed to exploit other peoples’ pain and struggles” (Kozuch, 2019). How and by whom such a machine would be deployed was also questioned. However, the idea that VR was an important technological evolution to elicit empathy in viewers was pervasive, even if the implications of such an evolution remained generally uninterrogated.

Walking in someone else’s shoes

A second metaphor, walking in another’s shoes, was tied to the concept of connecting humans to each other:

Virtual reality allows the user to walk in someone else’s shoes instantly from the comfort of their own home . . . [It] connects humans to other humans in a profound way I’ve never before seen in any other form of media . . . (Dittmer, 2017)

Empathy in the metaphor, therefore, was about feeling what others feel from an intimate and personal perspective (Bailenson, 2018). This type of embodiment gave insights into experiences and thoughts that were outside of one’s own skin. It implies an ease to the empathetic experience; a user can proxy another’s life and perspective through VR. At the same time, issues of privilege and access (e.g. whether the user had the right to such an experience), plus the difficulty of living the life of another (both within and outside of the headset) remained absent through the metaphor.

Empathy and transport

A transportive experience—where one can go—was also crucial for VR to arouse empathy. A Forbes article stated that “the ability to interject yourself into different kinds of situations instantaneously . . . is why some call VR an empathy machine” (High, 2018). The results were considered transformative as people could envision catastrophes and create a “global consciousness” (High, 2018). This assumed that if VR exposed users to new experiences and places they would or could not normally go, the encounter would have lasting impact on their worldview.

Empathy as skill

Finally, empathy was understood as a skill, which VR could build and strengthen, particularly regarding training and education. One Medium post directly stated, “Empathy can be a learned communication skill, just like public speaking, writing, recognizing nonverbal cues, or conversational turn-taking” (Krohner, 2015). Here, empathy was not inherently drawn out, as was the case with walking in another’s shoes, but could be developed through VR. Further, it was considered part of a regime of 21st century “soft skills” (Bonasio, 2017) used in everything from patient care to customer service.

Together, these four metaphors set the caveats and bounds of what empathy could do. While providing insights into potential impact, each metaphor identified empathy as a trait to be drawn out or reinforced through immersive technologies. Consequently, journalists tied the term to VR devices, rather than independently focusing on it.

Prosocial change for people

By allowing users to inhabit the body and life of another, VR’s empathetic experiences could generate greater appreciation of different races, genders, the elderly, the homeless, the disabled, and the prison population, among others.

[P]ut people in each other’s shoes to increase empathy. A man can see himself . . . as a woman, or a person can see herself as a different race. Sighted people can perceive the world the way that someone with cataracts or macular degeneration does. (Downey, 2016)

One article similarly suggested students could better understand and empathize with those struggling with dementia based on enthusiastic, if inconclusive, anecdotes from a pilot study working with seniors (Haptical, 2018). In fact, most articles pointed to how the technology could lead to deeper compassion. The notion that empathy had such broad prosocial impact reflects a kind of homogenization (e.g. marginalized groups may be better understood through an empathetic experience). The assumptions portrayed in the popular press seem difficult to validate and generalize problems of different groups, at least when it comes to empathy.

Prosocial change for place

Journalists reported that transporting someone to a place far from the comforts of home helped them to empathize with the afflictions of war, climate change, and other calamities. Two VR experiences—about Syrian refugees and wildlife conservation in Africa—were described as presenting “a world so complete in its presence you feel a deep, moving empathy for the people and their struggle to live in harmony with nature” (Fink, 2018). After such exposure, there was optimism that donations would increase—articles mentioned that Clouds Over Sidra doubled donations at a conference after it was viewed. More broadly, transporting the observer to battlegrounds and wildlife reserves was significant because it was an experience they might never have otherwise: “. . . [Y]ou really don’t want to take somebody to a war zone and put their life in danger, or to a refugee camp—it’s simply not practical . . .” (Haidamus, 2015).

Thus, places beset by tragedy and conflict comprised most potential virtual trips to induce empathy, such as following poachers in Africa, earthquake sufferers in Nepal, and populations devastated by natural disasters. VR might also give insight into environmental crises such as climate change (Markowitz et al., 2018). One article described a nature production as creating a “visceral understanding of environmental issues through immersive data storytelling, and the physical sense of presence and increased empathy that VR provides” (Business Journal, 2018). Overall, they suggest that transportation might propel social action.

Together, most journalists were optimistic that VR could have broad and vital influence by eliciting empathy. At the same time, even if VR was the answer to societal woes, direct, long-term, and extensive effects beyond specific anecdotes surrounding fundraising remained underreported. This occurred in part because much research on the subject is ongoing and therefore cannot guarantee the monumental change portrayed by some mainstream media.

Defining empathy

The most cited references used to define (and measure) empathy (11/53) were from Davis (1980) who described it as one’s responsiveness to the experiences of another and the ability to relate to another’s point of view, or perspective-taking. Articles that addressed empathy through this perspective-taking lens presented specific behavioral or psychological variables that explained how it could be accomplished.

Empathy was associated with other traits and emotions, which emphasizes the multifaceted nature of the term (Zaki, 2019). For example, psychologists Wallach et al. (2010: 4) wrote, “empathy and presence use the same projective ‘tool set.’” Similarly, in a study of persuasive games, designers Kors et al. (2016: 92) explained, “With this feeling of being ‘somewhere’ else, the feeling to stand in ‘someone’ else’s shoes becomes possible, which in turn presents ample opportunity to also think and feel like this other person (i.e. empathy).” For these authors, it was unclear whether presence is equivalent to or aids empathy. Further, these attributes could be either appearances of an empathetic response or a reaction to the novelty of the experience. Thus, a set of connections between empathy and other related concepts emerged: from presence to facial expressions and compassion. These relationships further reinforced the difficulty of defining the term within any single field or subfield.

Measuring empathy

As expected, measuring empathy and the means to do so resulted in mixed findings. Only half (27/53) of the papers in the academic corpus offered any measures of empathy in their research. When they did, the most common tool (n = 11) was the IRI (Davis, 1983). Due to the robustness of the IRI, researchers have adapted and used it in many fields. In studies where the IRI was not used, authors created their own self-reported benchmarks and employed field-specific measures of empathy (e.g. Ando et al., 2011) or different methodological approaches such as semi-structured interviews and/or biofeedback (Kim et al., 2009). Scholars in fields like history, philosophy, and design approached the subject more theoretically.

Acknowledging empirical limitations

A final theme emerged in the papers’ limitations sections. Studies frequently concluded that measures of empathy might tap a host of psychological dimensions and/or that the operationalization of empathy was lacking across the relevant literature. In addition, authors stated other contributing factors (instead of empathy) that might affect findings. For example, self-association (along with issues in self-reporting) appeared in several discussion sections; VR experiences also made subjects more self-aware, rather than empathetic toward others, which reinforced preexisting pro- and anti-social tendencies.

Our qualitative themes reveal a web of convergent concepts when defining empathy and VR. Their relationship is nuanced and built upon decades of research. This pattern also extended to tempered language around findings about prosocial change. While many scholars discovered prosocial benefits to VR, such as feeling more connected and willing to sign petitions about homelessness (Herrera et al., 2018), conclusions were incremental and appropriately contextualized within each study. As might be expected, these accounts differ from the more generalized and aspirational coverage espoused by the popular press.

Popular corpus

Eight components were extracted from the popular corpus, cumulatively explaining 11.79% of total variance, an amount consistent with other evaluations of natural language (Millar and Hunston, 2015). Table S2 in the Supplemental Materials contains extracted components, and italicized words below are items from each component.

Many articles described VR content creation, particularly elements that play a role in developing and experiencing an immersive world (e.g. audience, viewer, story; Component 1). By describing play and what can happen with the technology (e.g. create real and important experiences; Component 2), readers are presumably told about the possibilities of how to sense, create, and feel genuine VR experiences (Component 5). These three components are functional in nature and describe characteristics of audience, technology, and experience.

Two components also related to VR hardware, such as the characteristics of immersive video (Component 3), headsets in the current marketplace, and headset manufacturers (Component 4). Relatedly, writers mentioned the commodification of VR through connections to industry (Component 6). By putting VR in economic terms (e.g. company, market), writers acknowledged that industry is a major contributor to VR’s popularity and plays a vital role in its future. These components collectively underscore the importance of communicating industry as an economic driver and producer of VR.

Finally, the aspirational nature of VR was also mentioned in two components. Words such as impact, powerful, and important clustered in Component 8, which described the perspective-taking potential of virtual experiences. The ability to “put yourself in someone else’s shoes” and understand their point of view was often mentioned by the popular press. Consistent with the qualitative analysis, writers described the aspirational nature of VR for empathy by suggesting that it may have an impact on emotion in the future (Component 7), but they struggled to articulate exactly how the technology may accomplish this.

Together, the PCA data confirm that empathy is defined by the popular press as an aspirational form of communication between people and technology. However, the relationship between VR and empathy is also defined by ties to industry and the desire to convince readers virtual experiences are real, meaningful, and impactful.

Academic corpus

Three components were extracted from the academic corpus, explaining 35.18% of total variance (Table S3 in the Supplemental Materials). The first component described positive empirical effects for the relationship between VR and empathy: An interaction in VR can provide a high level of support to others is a research finding reflected in the empirical data. The second component describes how scholars justify their use of VR in their studies. Work in a virtual environment can provide support for their research question. Finally, Component 3 shared a consistent theme with the popular corpus by describing the potentiality of VR. Words such as future, give, allow, and empathy indicate that academics view the relationship between VR and empathy as aspirational.

All told, there are several commonalities across corpora. First, writers for the popular and academic presses justified the use of VR in prosocial settings. As one might expect, justifications emerged through different means, since the audiences of these presses were dissimilar. For example, the qualitative analysis suggests that the popular press uses key figures (e.g. researchers, practitioners) to describe serious use cases for empathy. Academics, as observed in the automated text analysis, addressed VR through research questions, empiricism, and prior evidence. Second, both presses also converged on the potentiality of the medium and the belief that a prosocial empathetic experience could be developed or enhanced with VR. The qualitative analysis of the popular press revealed that the relationship between VR and empathy is aspirational for users. The automated text analysis furthered such observations by showcasing how VR and its related industry could relate to empathy in users.

Together, the automated text analyses largely echoed the qualitative coding, especially for the academic corpus. In the popular corpus, however, the automated text analysis identified unanticipated themes, including key headset manufacturers and discussions about the medium’s future. This latter theme is different from potentiality because it focuses less on prosocial change and more on the medium. Specifically, it highlights expectations of the medium’s hopeful longevity as expressed by the popular press. These findings suggest that qualitative and quantitative techniques have complementary value in understanding complex phenomena such as VR, empathy, and prosocial change.