A web-based architecture and incremental authoring model for interactive learning environments for diagnostic reasoning

Computer-based learning environments have been used to provide simulated settings for problem-based learning. In prior research, a standalone software tool called INDIE was built for authoring and delivering learning environments where students can run simulated tests, collect data, generate hypotheses, and construct arguments. INDIE includes a run-time engine and an authoring tool. Approximately a dozen applications were developed using that system. This dissertation describes a new version of INDIE and a new approach to authoring and deploying learning environments. The new INDIE enhances the authenticity of the learning environment by simulating important real-life aspects such as time and cost, and supporting complex and random test result generation. More importantly, the new INDIE improves the accessibility of learning environments by using a web-based client-server architecture that allows students to interact with learning environments through web browsers. This architecture reduces the difficulty of authoring at run-time by allowing authoring to be done on a centralized server without any change on the client side. In addition, the web-based INDIE authoring tool facilitates authoring with support such as completeness check, preview of authoring results, and automatic retrieval of previously used values. The new INDIE architecture includes the instructor in the feedback loop to complement system feedback and collect materials for authoring. This approach extends system ability to support actions not available to students and allows the instructor to assess student work in the learning environment. Significant up-front development effort can be avoided by authoring the learning environment incrementally during real use. The current INDIE has been used to build Corrosion Investigator, a learning environment where environmental engineering students take the role of consultants helping a paper processing company determine the cause of recurring pipe corrosion. Results from the use of Corrosion Investigator in class show that Corrosion Investigator successfully facilitated the delivery of a problem-based learning module and at the same time reduced the workload for the instructor. Formative evaluation of the authoring tool and the interface for the instructor to provide feedback to students suggest positive results.