Detecting personality from how people customize their virtual city
[caption id=”attachment_2870" align=”aligncenter” width=”750"]
Even at the same level of the game, with access to the same items, you can see that people choose to design their virtual spaces very differently. Who do you think has the more neurotic personality?[/caption]
After two years of rejections from very traditional psychology journals, I’m very pleased to report that my paper on spatial customization and personality has been finally accepted for publication at Computers in Human Behavior (download here). I was very much inspired by Sam Gosling’s studies that looked at how people’s offices and bedrooms were reflective of their personality, and all the wonderful research on avatars and personality by Katie Bessiere and Nick Yee to name a few.
The idea for this paper is very simple, very intuitive. Simulation games like Sim City, Farmville, etc., allow people to design their own space. But does that space reflect their true personality?That was the main research question guiding this study. You may think, of course it does, but it’s not that simple. There has been a lot of research about avatar customization and personality, and some say that people tend to create avatars that are closer to their ideal personality rather than their real personality. That makes sense because a game is a fictional environment, so people can be whoever they want to be. At the same time, you could also make the opposite argument, and that some aspect of the player is bound to be reflected in whatever they design.
As is with social science, it’s a simple question, but a hard one to “prove” through statistical means. More details are in the paper, but I had people fill out a survey asking different aspects of themselves, including a number of personality measures. After a few weeks, I had them play a simulation game (in this case, Cityville, because at the time the game was relatively new and these people had never played this game) for one week, telling them that it was a study about how much they enjoyed the game. During this time, I took screenshots of their progress in the game. When you are at a low level, there are very limited options in terms of what you can build in your city, but it was interesting to see how people were creating completely different cities even with those limited options.
I then took screenshots of all the players at the same level (so we could control for the type of buildings they could build, amount of money, etc.). Two people then looked at the screenshots and answered questions about what they thought the personality of the person who created that city was. I then compared what the observers thought with what the players had reported and found that the observed personality was very close to players’ real personality rather than their ideal personality. Since the game only had a limited set of items, this suggests that the arrangement of items in a personalized virtual environment reflects the personality of the creator. This may be a useful rationale for creating unobtrusive behavioral measures for personality. Much of personality is based on self-report, but the wording of the questionnaires can sometimes be an issue, especially when translating into different languages, or administering to younger people or people in different cultures. Designing a virtual city does not require any language and is something that even little children can do.
Of course, further investigation is required to see if there are similarities and differences in different genres of virtual environments. Even if the concept of customizing a virtual space is similar, we don’t know if the results we found in Cityville would also apply to Farmville, Pet Society, or even less restrictive virtual environments such as Second Life. So many directions for the next study!
