Using Automated Knowledge Representation Methodologies to Cross-Validate Grading in Third-Level Education: A Cross-Validation Study

• In Event: Cognition and Assessment Paper Session

Abstract

Grading assignments and papers is a huge task throughout all academia.Several new methodologies of language-based assessment are available from research. In this study, we use an automated and computer-based analysis and seven graph theoretical comparison measures to investigate for two different disciplines - educational science and economic science - how much of the grading can be predicted by automated comparisons between an assignment and its main referring literature as chosen by the student writers (N=50).We found considerable matches between the automated measures and classical grading in the field of economic science and none for educational science.

Authors

• Pablo Nicolai Pirnay-Dummer, University of Freiburg

• Kausalai K. Wijekumar, The Pennsylvania State University