Dissertation
A biclustering approach for detecting cheating in educational assessments
University of Iowa
Doctor of Philosophy (PhD), University of Iowa
Spring 2025
DOI: 10.25820/etd.007813
Abstract
Cheating detection in educational assessments remains a challenging issue, complicated by increasingly sophisticated cheating strategies and the rise of remote testing environments. In this dissertation, I introduce a novel biclustering approach for detecting cheating by simultaneously identifying groups of examinees and test items exhibiting suspicious response patterns. This biclustering method enables comprehensive detection of cheating behaviors by analyzing response accuracy, response time, and answer choices.
Two comprehensive studies were conducted to evaluate this approach. In the first study, I applied biclustering to real-world assessment data consisting of dichotomously scored multiple-choice items and demonstrated the method’s superior performance over traditional approaches. Analyses further validated its robustness through simulations modeling a variety of realistic cheating scenarios. To enhance realism, the simulations also incorporated additional aberrant responses, such as rapid guessing due to time constraints and low motivation. In the second study, I extended the biclustering method to real-time cheating detection using mixed-format assessments, including dichotomous, polytomous, and multi-part items. This real-time application incorporated enhanced statistical significance testing to maintain low false positive rates and revealed strong detection performance across varying conditions and item types, including timestamp-based detection across different cheating scenarios.
Results across both studies underscore the biclustering approach’s adaptability, computational efficiency, and precision, making it a valuable tool for maintaining test integrity in both post-exam analysis and real-time applications. I conclude by discussing practical implications, limitations, and directions for future research.
Details
- Title: Subtitle
- A biclustering approach for detecting cheating in educational assessments
- Creators
- Hyeryung Lee
- Contributors
- Walter Vispoel (Advisor)Won-Chan Lee (Advisor)Stephen Dunbar (Committee Member)Sanvesh Srivastava (Committee Member)
- Resource Type
- Dissertation
- Degree Awarded
- Doctor of Philosophy (PhD), University of Iowa
- Degree in
- Psychological and Quantitative Foundations
- Date degree season
- Spring 2025
- DOI
- 10.25820/etd.007813
- Publisher
- University of Iowa
- Number of pages
- ix, 110 pages
- Copyright
- Copyright 2025 Hyeryung Lee
- Language
- English
- Date submitted
- 04/28/2025
- Description illustrations
- illustrations, tables, graphs
- Description bibliographic
- Includes bibliographical references.
- Public Abstract (ETD)
- Cheating in educational assessments poses significant challenges, especially as remote testing and complex cheating strategies become more common. In this dissertation, I introduce a novel biclustering approach for detecting cheating by simultaneously identifying groups of examinees and test items exhibiting suspicious response patterns. This biclustering method enables comprehensive detection of cheating behaviors by analyzing response accuracy, response time, and answer choices. Two studies were conducted to evaluate this method. In the first study, biclustering was applied to real-world test data consisting of dichotomously scored multiple-choice items to demonstrate its effectiveness. Analyses highlighted the robustness of this approach through simulations that modeled a variety of realistic cheating scenarios. To make the simulated data more realistic, the model also included other unusual response patterns, such as rapid guessing due to time limits and low motivation. The results showed that biclustering accurately identified cheaters and compromised items across a variety of testing conditions. In the second study, I expanded the method to real-time detection during assessments that included different types of test items, such as multiple-choice and open-ended questions. With its enhanced statistical controls, the biclustering approach was effective in accurately identifying cheating cases, while minimizing the number of times it incorrectly flagged honest test-takers as cheaters. This careful balance shows strong potential for use in live testing environments. Overall, results from this dissertation underscore the adaptability, accuracy, and efficiency of biclustering as a valuable tool for enhancing the integrity, validity, and fairness of educational assessments.
- Academic Unit
- Psychological and Quantitative Foundations
- Record Identifier
- 9984830824802771
Metrics
16 Record Views