Methods of Investigation


The first phase of this study was to create a MUSH environment based on the MUSH 2.0.59pl3 server. This environment subsequently became known as WriteMUSH. Within this MU* environment was created a single virtual "room", described as a writing lab. Software was added, in the form of JED, (Jim's Editor) a simple text editor system using MUSH softcode and macros for the TinyFUGUE MUSH client. Also created was Igor, a character whose purpose is to log all communications in the writing lab, as well as all interactions with JED.

The next phase involved getting the participants. Each was given the experimental notification form as required by the sponsoring university. Participants were asked to create characters on the MUSH using their own first names.

Running the study:

When the scheduled time for the study arrived, participants were informed of the nature of the project, that their interpersonal communications in the "writing lab" would be logged, that their interactions with the editor would be logged, and that they had 3 hours to create a short story. The study administrator remained in the "writing lab" virtual room for purposes of technical support.

At the end of the study, JED was frozen to prevent data being lost, the log files were closed, and an informal debriefing session, not part of the study, was conducted.

Processing the results:

The resulting logfiles was broken down into text events, and each text event was numbered in the main log file. The log files were created in such a way that an individual text event - a person sending a message to another person, a person adding a paragraph to the story with JED, and so on - is treated as an arbitrarily long single line. Using the cat function in Unix, these arbitrarily long lines were numbered.

The main log file was processed into the various categories given in the coding scheme (below). This was done by editing the main log file with the Elvis variant of the Unix vi editor, which wraps the text events into 80 column lines on the display without actually breaking text events up in the file. The yank and copy functions were used to move individual text events from the master file into the subfile or files they belonged in.

The subfiles were in turn analyzed using the Protocol Analysis method first explored in writing research by Flower and Hayes. (Flower and Hayes, 1981) While the data gathered cannot be considered a true protocol analysis since it did not capture self talk during the creation process, this method of analysis nevertheless is most effective at breaking down the process by which the collaboration and creation of the story took place by analyzing the communications between participants - who were geographically separated, and therefore only able to communicate through the MUSH.