DYNAMIC BEHAVIOR OF TEST ITEMS ACROSS LATENT ABILITY CONTINUUM: AN ANALYTICAL STUDY OF THE ITEM RESPONSE THEORY PERSPECTIVE

Abstract

This study examined the dynamic behavior of test items across latent ability continuum using the two-parameter Item Response Theory (2PL IRT) model as a methodological alternative to the limitations of classical test theory (CTT). a sample of 500 university students Completed a 15-item Complex Pattern Analysis test specifically designed to capture variation in item functioning across different cognitive ability levels. results revealed systematic variation in item behavior, with discrimination parameters (a) ranging from 0.76 to 2.90 and difficulty parameters (b) ranging from +2.29 to -2.29, indicating broad coverage of the ability spectrum. Item characteristic curves (ICCs) and information functions (IIFs) showed that each item is a "specialized" measurement tool that achieves its peak accuracy within a specific range of latent ability. The Total Test Information Function (TIF) indicated that the test reaches its peak accuracy in measuring the medium to high ability range. Furthermore, the Differential item functioning (DIF) results demonstrated a complete absence of bias in item behavior across the test. The study concludes that the dynamic nature of item behavior revealed by IRT analysis represents a fundamental shift from a simplified aggregation model to a precise analytical model, with fundamental implications for the design of fairer and more efficient diagnostic tests capable of measuring complexity in advanced mental performance.