Deterministic Adaptive Testing With Verification: Score Alignment, Convergence, and Efficiency
Author:
- 21 May 2026
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-7
Abstract
Adaptive testing systems often prioritize statistical optimality at the cost of transparency, predictable test length, and operational simplicity. This study presents and evaluates a deterministic, verificationbased adaptive testing model designed to produce interpretable scores while substantially reducing testing time and item exposure. The model employs a fixed number of iterations, strictly ordered difficulty bands, and a multi-step verification procedure to ensure that routing decisions reflect confirmed performance rather than isolated responses. Adaptive scores are computed using a bounded geometric weighting scheme that guarantees rapid convergence and known score resolution. The adaptive framework is content-agnostic and does not rely on domain-specific calibration, making it applicable across subjects, disciplines, and educational levels wherever ordered item difficulty can be established. Theoretical analysis exhaustively enumerated all possible decision sequences and established explicit bounds on adaptive score behavior. Results demonstrated near-perfect correspondence between adaptive and fixed-form scores under idealized conditions, defining an upper limit for empirical agreement. Empirical evaluation with 235 undergraduate examinees showed strong alignment between adaptive and fixed-form scores, preservation of rank ordering, and small, structured score differences consistent with the verification-based design. The adaptive test reduced item exposure by more than sixty percent and testing time to approximately one third of the fixed-form administration. These findings demonstrate that deterministic adaptive testing provides a transparent, efficient, and broadly applicable alternative to probabilistic computer-adaptive testing in applied educational assessment contexts.Keywords
This publication has 19 references indexed in Scilit:
- Practical Considerations in Item Calibration With Small Samples Under Multistage Test Design: A Case StudyETS Research Report Series, 2024
- Research on FPGA Functional Reliability based on Petri NetJournal of Physics: Conference Series, 2023
- The Development of an Automatic Test Assembly System for a Formative Assessment in Mastery Learning Instruction: Case of the SQL Mastery CourseIEEE Access, 2023
- Statistical harmonization of everyday functioning and dementia‐related behavioral measures across nine surveys and trialsAlzheimer's & Dementia: Diagnosis, Assessment & Disease Monitoring, 2023
- Comprehensive comparison of stroke risk score performance: a systematic review and meta-analysis among 6 267 728 patients with atrial fibrillationEP Europace, 2022
- Seated postural control measure: Italian translation and validation in children with cerebral palsyProsthetics and Orthotics International, 2021
- Measurement precision at the cut score in medical multiple choice exams: Theory mattersPerspectives on Medical Education, 2020
- A concise shoulder outcome measure: application of computerized adaptive testing to the American Shoulder and Elbow Surgeons Shoulder AssessmentJournal of Shoulder and Elbow Surgery, 2019
- Inference in semiparametric binary response models with interval dataJournal of Econometrics, 2015
- Applying fault tree analysis to the prevention of wrong-site surgeryJournal of Surgical Research, 2015