The political studies using mimicry and facial identity capture made me doubt everything. I asked my wife if I looked like the candidate I voted for in the last election; she assured me that I looked more like the other guy (although I'm not sure how comforting that is -- I want it to mean that my decisions were based on firm political principles rather than my desire to see those who look like me.) I remember reading about how they kept changing what Betty Crocker looked like, because they wanted the largest number of American women to see themselves in her. As the years went by, Betty Crocker -- an early avatar perhaps? -- changed her look based on what the advertisers could put together as being the quintessential homemaker for each generation. Virtual worlds might be great for students just to move away from a world where everyone in every commercial and classroom "looks like them." Even in the non-human avatars that Dr. Bailenson discussed at the end, the transformed social interaction can broaden their experience, confidence and awareness of others who may not match them visually.
The positive possibilities for students to practice skills (like the example of learning a martial art by "doing" with an instructor in a virtual world) and to gain confidence from virtual selves seems incredibly positive for the classroom. I've taught so many autistic students who struggle with the social cues of face to face interaction. The possibilities suggested by the virtual clues they are given to help them learn appropriate eye contact in social situations would help them not just make others more comfortable in social situations, but as the "virtual mirror" studies suggest, help them succeed. The dangers however of false memories (like swimming with Fudgy the whale) or skewed realities do make me nervous, as does Betty Crocker (below!)
