EJMT Abstract

A method based on mathematical results from non-adaptive group testing and combinatorial design theory for designing diagnostic mathematics tests containing multi-skill questions is described. The goal is to minimize the number of questions asked while covering a large list of skills without sacrificing the ability to pinpoint deficiencies in individual skills provided that the number of deficiencies to identify is low. A set of tools developed in Python 2.7 demonstrating the method is available as a free download for real-life testing needs or research in mathematical knowledge assessment.