COMPUTERIZED ADAPTIVE TESTING BERBASIS IRT: PRINSIP, MEKANISME, DAN SIMULASI SEDERHANA

Abstract

The transformation of assessment from linear to adaptive forms has become a crucial necessity to enhance the precision of educational measurement. This study aims to dissect the working mechanisms of Computerized Adaptive Testing (CAT) through a simulation approach using Microsoft Excel. The method employed is a quantitative simulation integrating the Maximum Likelihood Estimation (MLE) algorithm and an item bank calibrated using the 1-PL Item Response Theory (IRT) model. The novelty of this research lies in the application of hybrid stopping rules, combining precision criteria (SEM ≤ 0.30) and stability criteria (∆SEM < 0.03) to optimize test duration. Simulation results demonstrate that the adaptive mechanism effectively achieves participant ability estimation convergence with high accuracy within a range of 12-15 items. These findings prove the effectiveness of CAT in producing efficient and stable measurements, while providing a transparent overview of the functional logic of adaptive assessment for educational practitioners.