We started developing Cut with the goal of running economic games on desktop browsers. We were looking to match human subjects with each other and have them interact in real time. In addition, we had a feature for experimenters and researchers to specify agents to be matched with subjects. Agents could be customized in terms of profile information, avatar, and interaction strategies (e.g., selfish, generous, rational, etc).
Another goal was to make Cut user-friendly, not just for subjects but also for experimenters and researchers. To this end, we added an experiment designer, which made it possible to attach various survey items such as multiple choice questions, likest scales, and matrices as well as economic game in a serial fashion to create a complete experiment.
We were offered integration with Amazon Mechanical Turk. Once the experiment was designed and ready to launch, it could be launched as a human intelligence task (HIT) on MTurk, provided the user had money in their Amazon account.
Our subsequent goal was to offer descriptive previews of collected data and give some elementary statistics. This would complete the loop, making Cut on one-stop solution to design studies with economic games, collect data, and see results. These features significantly simplified designing and running studies, and allowed us to use Cut as a teaching tool, and resources for students to develop studies.
The following video shows how this now legacy system worked:
We received funding from New SchoolNew Challenge competition, and the Association for Psychological Science (APS) Fund for Teaching and Public Understanding of Psychological Science.
We have since moved on to an more scalable architecture based on improved technologies to meet new goals. We’ll talk about these new goals and current state of the project in another post.
This has been a long journey, and still ongoing.