Research in Second Life

Research in Second Life

Our group is interested in doing research in Second Life, with a focus on

how gender issues affect communication.  Second Life is most definitely a

form of CMC, but unlike AIM or e-mail, there is a visual element

integrated right into the medium.  One of the things we would like to find

out is whether the appearance of another party’s “avatar,” or virtual

representation, affects how we communicate with them.  Alternatively, how

does the appearance of our own avatar affect how we communicate with

others?

To collect data, we intend to pre-construct Second Life avatars for use by

experimental subjects.  These subjects will use these avatars (either by

choosing them or being assigned them) and sign onto Second Life, where

they will be asked to hold a conversation with another experimental

subject who has also either been assigned or allowed to choose one of our

pre-built avatars.  In order to quantify our research, we will first

record and store chat transcripts from the Second Life conversations of

all of our research participants.  In addition to this, we will have

participants complete a post-experiment survey asking them to give their

impressions of various aspects of the experiment.

For the details of the experiment design, we see two options:

Intersubject comparison: We would randomly assign participants to either

control or experimental group. The control group would have an avatar of

their gender, the experimental group would be the opposite gender. We

could see if people interact different if they are taking on the identity

of the opposite gender by comparing the results of the two groups.

Intrasubject comparison: Each subject would briefly interact being one

gender, then stop, be assigned an avatar of the other gender and interact

as that gender. Thus, we would have data for both same as self and

opposite to self for each subject. This would test within each subject if

each person acts differently (language wise), as compared to the

Intersubject design where we are looking for differences between subjects.

Both designs are valid and depending on the statistician you talk to one's

better than the other.  We are still deciding which to choose.

We have already taken a look at some of the relevant aspects of Second

Life, including options for customizing avatars and what gestures

(nodding, shaking one’s head, etc.) are available to be used within the

world.

Anshu Agarwal

Grace Pusavat

Nicholas Fajt

Rodney Eng