SPP 1293 Subproject - Rule-based Item Generation of Algebra Word Problems Based upon Linear Logistic Test Models für Item Cloning and Optimal Design

Basic data for this project

Description

The main goal of this project is to combine techniques from cognitive analysis, item response modeling, and optimal design theory to develop procedures for rule-based item generation and automated test assembly. The area of application is testing of statistical knowledge using word problems. During the first two years of this project a hierarchical response model for families of test items generated using cognitive rules was developed and made statistically tractable. The results were published in one manuscript for an international journal article and one book chapter. In addition, 14 cognitive rules for the generation of statistical word problems were developed and used to generate items. The rules and items were tested extensively in five different field studies. The results were published in two manuscripts for international journal articles. Further, a functional design for the automated item generator was drafted and a prototype for a text generator system developed. Finally, work on the application of optimal design theory to the automated assembly of tests from rule-based items pools has begun and its use in real-world environments has been explored. The results have been published in two manuscripts for international journal articles.