Over at Motherboard is a short piece about a virtual gender stereotype experiment conducted at Stanford University. Basic concept: users competed against bots to complete math problems. Regardless of physical world gender, when users were assigned male avatars to compete against female bots, they were most successful. The avatar representations were simply cartoon faces (dark eyes and hair, pale faces). Here’s the paper. I like that they were diligent about standardizing the perceived attractiveness of the avatars and tested whether existing mathematical ability was a factor, based on SAT math scores of participants. Still, could there be something else at work beyond the expected stereotype that males are better at math than females? I’d love to think so, but I’m not sure.
Shifting gears, after I got bored with the lack of endgame for a solo player in TERA (or any MMORPG, really), I thought I’d try a small experiment of my own. How would the experience be different if I played the same class but with a radically different appearance? My level 60 lancer — tank class — Karbyne is rhino-like and scarred, with a name that references the hardest element now known. My second lancer, now level 28, is about as twee as I could manage. She’s short with purple pigtails, stars on her cheeks, and squirrel ears and tail. I dubbed her Katteker, which means squirrel in Low German. I set new rules for myself, too: if someone specifically invited me to join a guild, I would say yes. If someone addressed me directly in conversation, I would answer. However, I wouldn’t seek to socialize any more than I did with my previous character. The first result I found was that Katteker was recruited into a guild at level 17, compared with Karbyne, who reached the level cap guild-less. Secondly, random strangers give Katteker buffs almost every day (for non-gamers, that means they cast spells that give her short-term advantages, like greater endurance or faster healing). The strangers don’t stop to talk, though I usually say a quick “thank you!” in local chat. That never happened to Karbyne.
The most dramatic differences are within me, though. It’s more fun to play as Katteker. Her movements are light-hearted and upbeat rather than the angry, aggressive animations that are standard for Karbyne’s race, and my emotional state is influenced by that. For example, when both characters stand still, Katteker will eventually start to dance, then later to clap her hands on an invisible bug. Karbyne sneers and lunges menacingly. I’m excessively amused by playing an adorable purple squirrel-girl who is a ruthless tank. It might be that the gaming part of my personality is inclined toward cuteness and fun, Ratchet & Clank rather than Call of Duty, so I’m more comfortable playing as Katteker.
So, here’s what I’m pondering today: if I was designing a story-based video game, I might want to manipulate the emotional states and preconceptions of players by forcing them to play with different avatars during specific chapters of the game. Personally, I would want to challenge stereotypes as I did that: featuring female avatars in some logic-based segments and male avatars in emotional ones, and including avatars of assorted races and belief systems as primary characters, not just sidekicks and enemies. But what if I was developing a business or educational application that used avatars? I always prefer a system that allows users to choose and modify their own avatars, but would performance be better if the system forced avatar choice consistent with stereotypes about a particular task? I find it offensive to think about limiting mathematical tasks to male avatars and nurturing tasks to female avatars, for example, but if my goal was peak performance…. Hmm. I think a compromise might be to allow users to pull from a varied set of default avatars for each task or to customize their own. If I felt my best “accounting self” was different than my “presentation self”, I could change avatars or make modifications to my base avatar, rather than having the system choose for me.