Influencing Conversation Through the Design of 3D Virtual Wolds – Part 1 (Avatar Design)
Conversation plays a significant role in education, regardless of discipline, and it has a particularly esteemed role in humanities education as a core element of the learning experience. Through conversation, learners are given the chance to actively engage with words, ideas, and peers. Discussions are prominent in language arts and social studies classrooms of all ages, as they support open-ended, collaborative interpretation of texts and ideas.
Digital technology has ushered in a new set of tools for facilitating discussions. There are many services that enable group text or voice chatting online, and video chat services like Skype enable video communication between teachers and students, creating a strange mimicry of face-to-face interaction. Rather than relying on physical proximity like face-to-face conversation, these new tools are mediated through the internet and digital screens. Another new tool, which is the subject of this review, the three-dimensional virtual world, is a computer-simulated environment that creates a perception of 3-D space, wherein users are often represented by and control an avatar, a digital representation of the user that may or may not look like the user at all. If support is in place for more than one avatar in the environment at once, these worlds are called multi-user virtual environments (MUVEs).
Within multi-user virtual environments (MUVEs), users are able to move around a pre-designed space, communicate with other avatars through voice, text, or non-verbal movements, and consume or create different kinds of content based on the affordances of the particular platform and the permission settings of those administering the space. The use of MUVEs as learning environments is generating interest in the educational community as rapidly as the underlying technologies of virtual reality improve. At the time of writing this review, Sony has just unveiled a virtual reality headset, and one of the most successfully funded projects on Kickstarter, a site where projects get crowdfunded, is Oculus Rift, a next generation virtual reality device that will be released in 2014. MUVEs are popping up on the internet so quickly that it is difficult to keep track of them, one of which, “Second Life”, has received the lion’s share of research.
The subject of this review is how the design of MUVEs impacts avatar-mediated conversation within them. There is ample research suggesting that the use of avatars can generate a sense of “co-presence” among users, the feeling of being together with others, and also “social presence”, the degree to which learners can present themselves socially and emotionally to others (Traphagan, 2010; Garrison 2007). Both of these are arguably prerequisites for engaged conversation. Less research, however, exists about what other aspects of the MUVE’s design, beyond the use of avatars, affects co-presence, social presence, and other factors that lead to meaningful conversation. It has long been recognized that the design of physical learning spaces has an influence on the learning that takes place within them, as do dress codes, and more attention needs to be given to the design of spaces and avatars in virtual worlds. To what extent does the research about physical learning space design apply to virtual spaces? Also, to what extent should virtual spaces even try to resemble physical reality at all?
After a review of some of the literature that seeks to address this constellation of questions, suggestions are posed for further research that aims to fill some of the gaps. Tentative suggestions are also posed for those designing or using MUVEs with the intent of promoting meaningful and engaging conversation within them.
It is hypothesized that communicating through avatars is fundamentally different from other forms of communication such as face-to-face interaction, video teleconferencing, audio-only conversation, or text-based chatting. Bailenson and Beall (2006) argue that one of the key reasons for this is the ability of those manipulating avatars to “systematically filter their physical appearance and behavioral actions in the eyes of their conversational partners.” Avatars may or may not look like the people manipulating them, and many MUVEs do not require avatars to look like humans at all. Avatars may also be doing non-verbal activities such as waving or nodding that their controllers are not. Avatar communication is also different from video or text chat since it situates a conversation in a space, albeit a graphically simulated space, creating a sense of coming together that is absent from video chats, which instead usually provide visual reminders that the communicating parties are separate in space.
This sense of coming together with others, to frame it in broad terms, is often referred to as “co-presence”. Bente et al. (2008) describe co-presence as one dimension of social presence, a broader construct which also includes mutual comprehension, emotional closeness, social relatedness, and behavioral contingency. These are factors which play a role in the shaping of conversation. While hypothesized that the use of avatars increase social presence, one potential problem is the layer of superficiality between users and their avatars. Under the assumption that social presence is related to and perhaps even necessary for the development of trust, Bente et. al. formulate the problem (295):
Given the capacity of avatars to personalize and situate mediated encounters, positive effects on trust could be expected. On the other hand, anonymity and artiﬁciality could also neutralize or even invert such effects.
Of course, in face to face interaction there is always the possibility that someone will not be saying things that they mean or behaving as they want to, but this possibility is perhaps heightened when a user is interacting through an avatar, and perhaps more so when interacting through an avatar that bears no physical resemblance to the user (Wang, 2011, p. 622). Bente rightfully recommends that future studies on MUVEs consider possible downsides such as “increased uncertainty” and “reduced self-disclosure”(310), and Wang is concerned about more foreboding sounding outcomes like “identity confusion” and “reality confusion”, neither of which seem conducive to authentic interpersonal communication. Laura Nicosia (2008), an enthusiastic Second Life educator and theorist, gushes that “inhabiting an avatar may be an invitation to take on a new personality”. However, for the purpose of open-ended discussion, having a learner accept this invitation may not be a positive outcome. What if this “new personality” is not as open-minded as its previous personality, or as expressive, and what does the design of the avatar have to do with it? There is an interesting study which details how an educator embraced the possibility of “identity confusion” through the use of role-playing activities (Jumalaya, Chee, Ho, 2009). Users were given roles to play with perspectives to adopt and were asked to have conversations with each other, virtually stepping into someone else’s shoes. So, avatar design can be thoughtfully employed to support such experimental conversational scenarios.Further studies should be done on the communicative patterns of users with avatars that physically resemble their real-life counterparts as compared to users with avatars that look nothing like them.
There is growing evidence which highlights the influence of avatar design on user communication and satisfaction, including some fascinating studies about how people interact differently if their avatar is taller, shorter, more or less conventionally attractive, etc (e.g. Yee, Bailenson, Duchenaut, 2009). Yee et al. describe how users with more attractive avatars are friendlier and more extraverted than those with unattractive avatars. Having attractive avatars in the virtual world could have a serious impact on conversational quality and quantity (294). When users occupy taller avatars, they communicate with more confidence. As to why this happens, the argument goes that since tall people are perceived to be more confident, the users with taller avatars conform to this expectation and do indeed communicate with more confidence. This raises the interesting question of who should have tall avatars, since in order for some to be noticeably taller, others have to be smaller. Other design choices which at first thought may seem trivial usually deserve more thought. Take color, for instance. One study found that when users are given avatars with black robes, they express more interest in exhibiting antisocial behavior than their white-robed companions (Merola, Penas, & Hancock, 2006).
There is also interest in the research community about the impact of anthropomorphic avatars compared to fantastical ones, and the impact of photorealistic avatars compared to more abstract ones. What about handing some aspects of the design choice, if not all of them, to the to the user? According to Ward and Sonneborn (2009), the ability for users to influence the design of their avatars is one of the key features of a virtual world that can affect collaboration. As a word of caution though, it should be noted that the way avatars look has a bearing on the tone and atmosphere of the surrounding virtual space as perceived by others. For example, allowing users to represent themselves as robots, animals, or symbols will shift the aesthetic of the virtual space from realistic to fantastic.
More research needs to be done on how changes in avatar design influence user behavior and communication. What sorts of designs make users more willing to communicate? Since most studies at present have been done on anthropomorphic avatars, more attention should be given to alternatives. From an ethical perspective, it is worth wondering to what extent educators should try to influence the communicative behavior within a virtual world by exercising control over avatar design.
Part Two Forthcoming..