TY  - JOUR
AU  - Mahboob, Maleeha
PY  - 2026
DA  - 2026/05/17
TI  - Misclassification Analysis in Automated Bloom’s Taxonomy Classifiers: A Data-Centric Perspective on Educational Software
JO  - ICCK Journal of Software Engineering
T2  - ICCK Journal of Software Engineering
JF  - ICCK Journal of Software Engineering
VL  - 2
IS  - 2
SP  - 138
EP  - 155
DO  - 10.62762/JSE.2026.118512
UR  - https://www.icck.org/article/abs/JSE.2026.118512
KW  - bloom’s taxonomy
KW  - automated assessment
KW  - cognitive level classification
KW  - educational data analysis
KW  - error analysis
KW  - dataset diagnostics
KW  - instructional verbs
KW  - question stems
AB  - Automated classification of assessment questions according to Bloom’s taxonomy is increasingly used to support curriculum design and educational analytics. Many existing approaches rely heavily on instructional verbs as proxies for cognitive demand, despite longstanding concerns about their interpretive reliability. This paper adopts a data-centric perspective to examine why verb-centric Bloom-level classification remains fragile when applied to authentic multiple-choice question (MCQ) stems. The study is based on a custom, single-domain dataset of MCQ stems annotated according to the revised Bloom’s taxonomy, with intentional class imbalance preserved to reflect realistic assessment practices. Through dataset diagnostics and systematic qualitative inspection, the analysis highlights plausible sources of misclassification risk arising from linguistic ambiguity, task–verb mismatch, conceptual proximity between adjacent cognitive levels, and domain-specific semantic effects. Rather than proposing new classification models, the paper introduces a formal error taxonomy and an analytical framework that explains the interactions producing structured, non-random misclassification patterns, without reporting performance scores or model comparisons. These findings expose recurring failure patterns in automated Bloom-level classification systems. From a software engineering perspective, the analysis informs system evaluation and reliability by clarifying where verb-centric classifiers are prone to systematic misinterpretation during assessment analytics and decision support. The main contributions are threefold: (1) a data-centric examination of Bloom-level misclassification grounded in authentic computer science MCQ stems; (2) a structured taxonomy of recurring error types linked to linguistic ambiguity, task interpretation, and dataset characteristics; and (3) clarification of the implications of these error patterns for system evaluation and reliability in educational software that relies on automated Bloom-level classification.
SN  - 3069-1834
PB  - Institute of Central Computation and Knowledge
LA  - English
ER  - 

@article{Mahboob2026Misclassif,
  author = {Maleeha Mahboob},
  title = {Misclassification Analysis in Automated Bloom’s Taxonomy Classifiers: A Data-Centric Perspective on Educational Software},
  journal = {ICCK Journal of Software Engineering},
  year = {2026},
  volume = {2},
  number = {2},
  pages = {138-155},
  doi = {10.62762/JSE.2026.118512},
  url = {https://www.icck.org/article/abs/JSE.2026.118512},
  abstract = {Automated classification of assessment questions according to Bloom’s taxonomy is increasingly used to support curriculum design and educational analytics. Many existing approaches rely heavily on instructional verbs as proxies for cognitive demand, despite longstanding concerns about their interpretive reliability. This paper adopts a data-centric perspective to examine why verb-centric Bloom-level classification remains fragile when applied to authentic multiple-choice question (MCQ) stems. The study is based on a custom, single-domain dataset of MCQ stems annotated according to the revised Bloom’s taxonomy, with intentional class imbalance preserved to reflect realistic assessment practices. Through dataset diagnostics and systematic qualitative inspection, the analysis highlights plausible sources of misclassification risk arising from linguistic ambiguity, task–verb mismatch, conceptual proximity between adjacent cognitive levels, and domain-specific semantic effects. Rather than proposing new classification models, the paper introduces a formal error taxonomy and an analytical framework that explains the interactions producing structured, non-random misclassification patterns, without reporting performance scores or model comparisons. These findings expose recurring failure patterns in automated Bloom-level classification systems. From a software engineering perspective, the analysis informs system evaluation and reliability by clarifying where verb-centric classifiers are prone to systematic misinterpretation during assessment analytics and decision support. The main contributions are threefold: (1) a data-centric examination of Bloom-level misclassification grounded in authentic computer science MCQ stems; (2) a structured taxonomy of recurring error types linked to linguistic ambiguity, task interpretation, and dataset characteristics; and (3) clarification of the implications of these error patterns for system evaluation and reliability in educational software that relies on automated Bloom-level classification.},
  keywords = {bloom’s taxonomy, automated assessment, cognitive level classification, educational data analysis, error analysis, dataset diagnostics, instructional verbs, question stems},
  issn = {3069-1834},
  publisher = {Institute of Central Computation and Knowledge}
}