The TED (Training in Experimental Design) Tutor project aims to develop a computer-based intelligent tutoring system to improve science instruction in late elementary through middle school grades focusing on the conceptual understanding and procedural skills of designing and interpreting scientific experiments.
TED's instruction focuses on the conceptual understanding and procedural skills of designing and interpreting scientific experiments. The central goal of the project is to develop an intelligent computer tutor that adapts to individual learners and helps to achieve robust mastery on a variety of internally and externally valid assessments.
Our research and development process iterates through a series of increasingly computerized and adaptive modules. Beginning with human-delivered instruction with physical materials, we have incrementally developed what is now a fully-functioning semi-adaptive computer tutor that consists of a pretest, a video introduction to the lesson, pre- and post near-transfer assessments, training requiring students to evaluate experiments, and finally a far-transfer post-test. Recent design-implement-assess-revise cycles indicated that learning outcomes of both middle- and low-SES students using TED are equal to or greater than outcomes of students given the same one-to-one instruction by a human tutor using physical materials. Furthermore, in a recent study with 8th-graders in a local magnet school, training in TED was significantly more efficient than a hands-on lesson from the district's curricular materials.
We are now further increasing the intelligence of TED by integrating menu-based responses with a Bayesian engine, which will allow TED to assign students to particular initial instruction pathways based on their incoming knowledge levels as well as provide remedial instruction when necessary.