The experiment I am involved in simulates an emergency planning scenario. There are three participants who each take a different role - the Public Works collaborator, the Mass Care collaborator, and the Environmental collaborator. The scenario goes as follows - some people are stranded and need to be rescued and taken to a nearby shelter - the participants' task is to choose the best shelter. The public works expert knows about things like road construction, power outages, and washed out roads. The environmental expert knows about inclement weather - thunderstorms, extremely cold temperatures, etc, and things like where state parks are located and the possiblity for things like mudslides or lightning. The mass care expert knows about the capacity of shelters (for instance, that the capacity of a shelter can be overfilled by a certain percentage), health problems the people being rescued may have, and things like that.
The experiment makes use of a concept called a hidden profile, which means that each participant only has some of the information. In order to pick the most optimal shelter, the members have to collaborate and pool together their information. (For more about hidden profile studies, I would reccomend reading pretty much any of Garold Stasser's papers, but in particular "Hidden Profiles: A Brief History", for a quick introduction.)
The first stage of the study was conducted with a paper prototype - each member had an individual map, which only they could see, and there was a group map which everyone could look at. The second stage of the study was conducted with a software prototype which was developed from requirements gathered during the use of the paper prototype. The software prototype is built off of an already existing program - BRIDGE. BRIDGE stands for Basic Resources for Integrated Distributed Group Environments, and is meant to faciliate synchronous collaboration and learning. The software prototype is mainly what I have been working on. I've been working on adding additional features which further support collaboration between the users.
I also worked with Helena, who is a graduate student working with Gregorio. We spent some time discussing the rationale that participants used to choose a shelter, and how that factored into the concept of common ground. Common ground is what the majority of Gregorio's graduate work has been on - it ties in with the idea of activity awareness. Activity awareness encompasses knowing what your team members know, their attitudes, their goals, and how the view of the shared plan evolves over time. Common ground is defined as the process of maintaining activity awareness. Helena and I worked together to come up with a coding scheme that would enable us to analyse the rationale behind particpants' choices. We coded a few transcripts, but unfortunately did not have enough time this summer to take the next step - analysis of the transcripts using Rhetorical Structure Theory (RTS).
For a more in depth overview of the software prototype and its features, as well as a look at how the concepts of common ground and activity awareness tie into geocollaboration, take a look at my final paper.