From Person to Personification
How embodying an avatar influences how we communicate in virtual spaces, and the implications for conducting research in the metaverse.
If you could create a virtual version of yourself, what would it (they?) be like? Would you choose to create an avatar who is as close to your ‘real world’ self as possible? Create a more idealised version of yourself? Create a self that represents who you believe that you are rather than who you physically are? Or create a fantastical character which personifies who you aspire to be? Gamers and virtual reality proponents have been exploring these concepts for decades, and now with the rise in metaverse platforms even more people are experimenting with their virtual selves.
The metaverse has also opened up opportunities to conduct research in shared virtual spaces — where participants will join sessions as their avatar. Participants can be invited into a curated space which has been adapted to the needs of the research methodology and area of discussion. For example, a serene, comforting space can be designed where participants are able to openly discuss opinions on financial literacy, or a store can be digitally replicated in which product interactions can be observed. The ability to conduct research virtually, over video or in a virtual world, can be beneficial because it enables the inclusion of people who may not be able to attend sessions in person due to distance, time constraints, mobility or other limitations. What makes virtual spaces unique is the ability to attend as an avatar, which allows participants to show up in a ‘body’ that they feel comfortable with. Whether this be expressing their cultural or gender identity, or being able to create an avatar that makes them feel more empowered, or even protected, when answering personal questions. This individually created embodiment may then allow participants to speak more freely and honestly, removing some of the feelings of vulnerability that may come with being there in person.
In this article I am going to explore how the use of an avatar could impact qualitative research conducted in the metaverse. Considering how it might affect the types of answers participants give in interviews. Or how avatars interact with each other in focus groups and co-creation sessions. Finally, I will then reflect upon how design researchers could augment our practice in response to this.
HYBRID EXPERIENCES
The impact of what you look like in virtual spaces goes far beyond pure appearance and has been proven to affect how you behave and interact with others. For example, a 2007 study has shown that making your avatar taller makes you more aggressive in negotiations, or that ‘just five minutes of role-play in virtual environments as either a hero or villain’ influences people to either reward or punish anonymous strangers. This phenomenon is called the Proteus Effect — where the characteristics of the avatar affects the behaviours of the individual within that virtual space. It has also been shown that this can extend back into the behaviours of the individual in the ‘real world’ e.g., if someone has spent time embodying a more confident and outspoken virtual character, it becomes a learnt behaviour and influences the individual to be more confident in their offline interactions.
As shown above, when embodying an avatar the behaviours of the individual can change as they merge their personalities with the perceived attributes of the avatar — creating a hybrid personality. Consider a scenario where individuals are being interviewed in a virtual space, who would be responding to the questions — the individual or the avatar? Would their responses differ? For example, if asked about their transport and commute preferences would the daily-grind-bus-user or the need-for-speed-racer-avatar respond, or a hybrid of both? This is a slightly extreme example, but when conducting qualitative research we are exploring the experiences and opinions of the individual, and therefore it is important to know whose experiences we are getting. Should we potentially be documenting the character of the avatar which is being used so that we can contextualise their responses? Or if the individual has already spent time inhabiting the racer avatar would the ‘real world’ person have taken on some of their attributes already — and therefore give the same responses anyway? As our experiences continually shape who we are and what we think, whether those experiences happened in ‘real’ or virtual worlds, should we worry about it all?
INTERACTING WITH NON-HUMANS
How we behave is highly influenced by social norms and what is perceived to be ‘acceptable’ by those around us. In interviews, as in everyday life, it has been proven that people alter their responses to be more socially acceptable, or give responses that they think the interviewer wants to hear. People do this because they want to avoid a negative reaction and fear the disapproval of the other. However, when people are speaking to robots, AI or other non-human entities — or believe that they are — the social pressure is dropped and therefore people are more open and honest. For example, a study showed that when soldiers spoke to a therapist after returning home they reported very few ‘mental health’ issues as they did not want to show any weakness. However, when the therapist was replaced with a virtual interviewer, the soldiers opened up about health problems they were experiencing as the fear of judgement had been removed.
Would it be beneficial to the research if we had our participants interact with a trained AI rather than having a human researcher behind the avatar? This could enable the participants to be more open, but would the AI system be able to understand nuances in the participants’ responses and adjust the follow up questions to get the most out of the session, like their human counterpart would?
NON-VERBAL COMMUNICATION
So far I have predominantly been talking about the appearance of avatars, however how realistically the avatar moves also greatly impacts interactions in virtual spaces. Approximately only 7% of how we communicate is verbal (words), and the other 93% is non-verbal. Non-verbal communication is then split into 38% tone of voice (how you say something) and 55% body language, including gestures, gaze, and facial expressions. Currently most virtual worlds are interacted with through laptops, game controllers and VR headsets. Through these devices individuals can use audio to speak to one another but there is very little control over the body movements of the avatars, beyond navigating the space and predefined animations (avatars in Meta Horizons don’t even have the lower half of their body). This means that people lose about 55% of their ability to communicate with each other — that’s huge!
Studies have shown that when participants interact with ‘highly expressive avatars’ there is an increased feeling of social presence and behavioural realism, and consequently participants have better, more enjoyable interactions. Interestingly it was less important how photorealistic the avatar looked and more important how they moved, e.g., keeping eye-contact, being spatially aware of each other and making contextual hand movements. However, to achieve such lifelike body movements takes a lot of tech — cameras, sensors, eye-tracking equipment etc — but it is unlikely that the majority of people will interact with virtual spaces using such high-tech equipment, even when it becomes more commercially available. The majority of people will interact via screens, and more than likely there will be a drive in development towards mobile interfaces and AR (augmented reality), rather than towards fully immersive VR suits which would enable more refined gesture control. How will this lack of non-verbal communication impact how participants interact with each other in collaborative sessions?
IMPACTS FOR RESEARCH
At this stage the honest answer is that we don’t know how embodying an avatar will impact qualitative research in the metaverse because there isn’t much evidence of it yet. We can however find insights and inspiration in parallel studies such as those conducting therapy sessions using VR or projects which construct virtual worlds for research with children. But at this stage it is about us getting our digital hands dirty by building and testing new virtual research strategies ourselves. And importantly it is also about sharing the results so that the wider research communities can benefit from it and build upon the insights. Decentralised spaces, decentralised knowledge.
Considering the above research here are some questions we can ask ourselves when designing virtual research:
- Should we curate the avatars? Would choosing a character for the participant to embody benefit the research in some way? Or maybe the researcher could embody a specific character? e.g., if we wanted to explore retirement planning with participants, would placing them in an older version of themselves help them to consider their future needs?
- Should we train AI avatars as interviewers? Do we as human researchers hinder or help the interview process?
- Should we develop avatars with simulated gestures and eye tracking? Would simulating the non-verbal aspects of communication improve the participant interaction experience?
- Can observational research methods, such as ethnographic studies, be effectively run in virtual spaces? Are observing virtual interactions relevant for designing ‘real world’ spaces — or only virtual spaces?
- Something not explored in this article (but with many interesting studies) is the effects of implicit bias. How do gender, race and ability biases translate into avatar embodiment?
A key thought to finish on is that just because virtual spaces may not be able to effectively mimic how people behave offline, that isn’t necessarily a bad thing. Different does not implicitly mean worse. Instead we need to consider the opportunities that avatars and virtual spaces open up for research e.g., how might virtual capabilities enhance the kinds of research that you are able to conduct? What research scenarios do virtual spaces make possible that are not possible in the ‘real world’? It is exciting to think about what entirely new forms of research the metaverse will open up!
// Check out more insightful design articles on the Design Globant page //
