Answers are grounded in this paper and cite their sources.
Methodological study demonstrates reliable multiple change-point detection in standardized test responses, highlighting practical utility for identifying cheating and test fatigue.
In educational measurement, examinees' response patterns often undergo abrupt transitions due to factors such as cheating, random guessing, or test fatigue. These transition points are defined as change points. Accurate identification of aberrant responses is crucial for ensuring test reliability and validity. However, most existing research on aberrant response detection focuses solely on single change‐point scenarios. Such methods may be inadequate when examinees exhibit multiple behavioral transitions during a test. This study developed a multiple change‐point detection method for response behavior based on the Pruned Exact Linear Time (PELT) algorithm. To address the limitations of fixed penalty criteria in balancing detection power and Type I error control, this study employed an adaptive penalty calibration method based on the false‐positive rate. Simulation results showed that the proposed method controlled the Type I error rate at approximately 5% while achieving overall, single‐change‐point, and multiple‐change‐point detection rates of 75%, 78%, and 70%, respectively. Empirical analysis of the PISA 2015 financial literacy data revealed that approximately 21% of test‐takers exhibited detectable response pattern shifts, with 29% of these test‐takers exhibiting multiple change points, demonstrating the method's applicability in real‐world low‐stakes testing contexts.
No takes yet. Share an insight, caveat, or question.
Lu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: