Hybrid data analytic technique for grading fairness
| DOI | https://doi.org/10.1108/DTA-01-2022-0047 |
| Published date | 20 April 2022 |
| Date | 20 April 2022 |
| Pages | 18-31 |
| Subject Matter | Library & information science,Librarianship/library management,Library technology,Information behaviour & retrieval,Metadata,Information & knowledge management,Information & communications technology,Internet |
| Author | Thepparit Banditwattanawong,Arnon Marco Polo Jankasem,Masawee Masdisornchote |
Hybrid data analytic technique for
grading fairness
Thepparit Banditwattanawong
Department of Computer Science, Faculty of Science, Kasetsart University,
Krung Thep Maha Nakhon, Thailand
Arnon Marco Polo Jankasem
Research Division, D.D.K.2019 Limited Company, Chonburi, Thailand, and
Masawee Masdisornchote
School of Information Technology, Sripatum University,
Krung Thep Maha Nakhon, Thailand
Abstract
Purpose –Fair grading produces learning ability levels that are understandable and acceptable to both
learners and instructors. Norm-referenced grading can be achieved by several means such as zscore, K-means
and a heuristic. However, these methods typically deliver the varied degrees of grading fairness depending on
input score data.
Design/methodology/approach –To attainthe fairest grading, thispaper proposes a hybrid algorithmthat
integrateszscore, K-means andheuristic methods with anovel fairness objectivefunction as a decision function.
Findings –Depending on an experimented data set, each of the algorithm’s constituent methods could deliver
the fairest grading results with fairness degrees ranging from 0.110 to 0.646. We also pointed out key factors in
the fairness improvement of norm-referenced achievement grading.
Originality/value –The maincontributions of this paper arefour folds: the definitionof fair norm-referenced
gradingrequirements, a hybridalgorithm for fairnorm-referenced grading,a fairness metric fornorm-referenced
grading and thefairness performance results ofthe statistical, heuristic and machinelearning methods.
Keywords Student grading, Norm-referenced achievement, Fair assessment, Fairness measurement,
Algorithm, Clustering, Zscore, K-means, Heuristic, Hybrid technique, Ensemble technique, Decision function
Paper type Research paper
Introduction
Norm-referenced grading, unlike criterion-referenced grading, systematically compares each
individual score to relative criteria that are defined based on the scores of all individuals to
obtain an appropriate grade with or without a condition (Wadhwa, 2008). This paper focuses
on unconditionally norm-referenced grading. The relative criteria can be determined by a
conventionally statistical means (i.e. zscore), a heuristic (Banditwattanaw ong and
Masdisornchote, 2021), or machine learning methods (Banditwattanawong and
Masdisornchote, 2021). Despite these available methods, a common question to norm-
referenced grading is how to clarify the fairness of grade results to all stakeholders (e.g.
learners and co-instructors) instead of leaving them with grade disagreement and concerns
about bias and discrimination. Two common forms of unfair grades found in norm-referenced
grading are as follows.
First, the apparent score intervals of different grades are diverse rather than identical. For
instance, suppose that grading certain course achievement using an A-B-C-D-F symbol
DTA
57,1
18
This work is financially supported by the Department of Computer Science, Faculty of Science,
Kasetsart University, Thailand.
Data Availability: The data used to support the findings of the study are included in the article.
Conflicts of Interest: The authors declare that there is no conflict of interest regarding the publication
of this paper.
The current issue and full text archive of this journal is available on Emerald Insight at:
https://www.emerald.com/insight/2514-9288.htm
Received 31 January 2022
Revised 2 April 2022
Accepted 5 April 2022
Data Technologies and
Applications
Vol. 57 No. 1, 2023
pp. 18-31
© Emerald Publishing Limited
2514-9288
DOI 10.1108/DTA-01-2022-0047
To continue reading
Request your trial