Authors:
Nils Ulltveit-Moe
;
Sigurd Assev
;
Terje Gjøsæter
and
Halvard Øysæd
Affiliation:
University of Agder, Norway
Keyword(s):
Entropy, Cross Assignment Marking, Learning Management Systems, Efficient Teaching Methods.
Related
Ontology
Subjects/Areas/Topics:
Computer-Supported Education
;
e-Learning
;
e-Learning Platforms
;
Information Technologies Supporting Learning
;
Learning Analytics
;
Learning/Teaching Methodologies and Assessment
;
Metrics and Performance Measurement
;
Simulation and Modeling
;
Simulation Tools and Platforms
Abstract:
This paper proposes an efficient tool-supported methodology for marking student assignment answers according to a knowledge metric. This metric gives a coarse hint of student answer quality based on Shannon entropy. The methodology supports marking student assignments across each sub-assignment answer, and the metric sorts the answers, so that the most comprehensive textual answers typically get the highest ranking, and can be marked first. This ensures that the teacher quickly gets an overview over the range of answers, which allows for determining a consistent marking scale in order to reduce the risk of scale sliding or hitting the wrong scale level during marking. This approach is significantly faster and more consistent than using the traditional approach, marking each assignment individually.