An Intelligent Teaching System for Database Modelling
Pramuditha Suraweera
Department of Computer Science
University of Canterbury
Abstract
Database (DB) modelling, like other analysis and design tasks, can only be learnt through extensive practice. Conventionally, DB modelling is taught in a classroom environment where the instructor demonstrates the task using typical examples and students practise modelling in labs or tutorials. Although one-to-one human tutoring is the most effective mode of teaching, there will never be sufficient resources to provide individualised attention to each and every student. However, Intelligent Teaching Systems (ITS) offer bright prospects to fulfilling the goal of providing individualised pedagogical sessions to all students. Studies have shown that ITSs with problem-solving environments are ideal tools for enhancing learning in domains where extensive practice is essential. This thesis describes the design, implementation and evaluation of an ITS named KERMIT, developed for the popular database modelling technique, Entity Relationship (ER) modelling.
KERMIT, the Knowledge-based Entity Relationship Modelling Intelligent Tutor, is developed as a problem-solving environment in which students can practice their ER modelling skills with the individualised assistance of the system. KERMIT presents a description of a scenario for which the student models a database using ER modelling constructs. The student can ask for guidance from the system during any stage of the problem solving process, and KERMIT evaluates the solution and presents feedback on its errors. The system adapts to each individual student by providing individualised hint messages and selecting new problems that best suit the student.
The effectiveness of KERMIT was tested by three evaluations. The first was a think-aloud study to gain first-hand experience of the student's perception of the system. The second study, conducted as a classroom experiment, yielded some positive results, considering the time limitations and the instabilities of the system.
The third evaluation, a similar classroom experiment, clearly demonstrated the effectiveness of KERMIT as a teaching system. Students were divided into an experimental group that interacted with KERMIT and a control group that used a conventional drawing tool to practice ER modelling. Both group's learning was monitored by pre- and post-tests, and a questionnaire recorded their perception of the system. The results of the study showed that students using KERMIT showed asignificantly higher gain in their post-test. Their responses to the questionnaire reaffirmed their positive perception of KERMIT. The usefulness of feedback from the system and the amount learnt from the system was also on a significantly higher scale. Their free-form comments were also very positive.