QBank: A Web-Based Dynamic Problem Authoring Tool

Abstract

Widespread accessibility to the Internet and the proliferation of Web 2.0 technologies has led to the growth of online tools for educational content creation, delivery, and assessment. Maintaining high quality of assessment using this medium is made more practical by using tools to author and represent a broad range of assessment problems. A survey of existing problem-authoring tools uncovered two main deficiencies: (a) lack of support for authoring "dynamic" (parameterized) problems, and (b) lack of tools that are independent of a specific publishing format, persistence format, and/or authoring platform.

Dynamic problems are assessment problem templates that support parameterization of the problems by the use of variables. Variables dynamically take values assigned at random to generate different instances of a problem from a template. This provides for greater diversity of authored problems, and permits students to practice with different variations of a problem. In existing problem authoring tools, the problem types supported are often limited to static problems.

A formal definition of an assessment problem structure is presented. This formal definition served as a design aide for a new problem authoring system named QBank, a web-based tool that supports authoring dynamic problems. The proof-of-concept implementation of QBank supports export of questions in CSV format and the Khan Academy Exercise format. The extensible nature of the framework allows future development of features supporting export of authored problems into other publishing and/or persistence formats.