Authors:
Sebastian Kucharski
1
;
Florian Stahr
1
;
Iris Braun
1
and
Gregor Damnik
2
Affiliations:
1
Chair of Distributed and Networked Systems, TUD Dresden University of Technology, Dresden, Germany
;
2
Chair of Didactics of Computer Science, TUD Dresden University of Technology, Dresden, Germany
Keyword(s):
Automatic Item Generation, AIG, Assessment, Cognitive Model, Item Model.
Abstract:
Studying at German universities is often associated with a passive mode of learning. Using learning tasks and (self-)test items is an effective way to address this issue. However, due to the high cost of creation, these materials are rarely provided to learners. The approach of Automatic Item Generation (AIG) allows for the resource-efficient generation of learning tasks and (self-)test items. This paper demonstrates, after presenting general ideas of AIG, how tasks or items can be automatically generated using the AIG Model Editor designed at TUD Dresden University of Technology. Subsequently, items generated using the AIG approach are compared with items created in a traditional manner. The results show that automatically generated items have comparable properties to traditionally created items, but their generation requires much less effort than the traditional creation, thus making AIG appear as a promising alternative for supporting active learning at universities.