Ability Estimation of Senior High School Students in Mathematics using the Classical Test Theory and Three-Parameter Item Response Theory Model Bayah J. Amiruddin, Mercedita Langamin
Mindanao State University - Tawi-Tawi
Abstract
This study aimed to estimate and compare the academic abilities of the senior high school students of the Mindanao State Univeristy-Tawi-Tawi College of Technology and Oceanograpgy - College of Arts and Sciences in Mathematics using the Classical Test Theory (CTT) and the Item-Response Theory (IRT) models. Specifically, the Bayesian modal or maximum a posteriori (MAP) under the 3PL model of IRT was used as ability estimator. The study involved a total of 52 Grade 11 students in a descriptive-quantitative approach using Microsoft Excel, R, and SPSS software in the data calculation. The findings generally revealed that 30 (57.69%) students passed and 22 (42.31%) failed under the CTT approach. However, under the IRT approach, 34 (65.38%) passed while the remaining 18 failed. Moreover, when the CTT and IRT results were compared, it produced significant results indicating that transmuted grades of the students were higher under IRT than CTT. Thus, the null hypothesis stating that, there is no significant difference between the actual performance in CTT approach and the estimate ability in IRT approach was rejected. Furthermore, the results of the two approaches showed significant relationship further providing a better picture of the comparison. Hence, the null hypothesis for the relationship between the two sets of data was rejected. In this study, IRT appeared to provide more details/information about student performances/abilities than the CTT and thus the use of IRT in the ability estimation is recommended. In addition, Bayesian modal or MAP appeared to work well as ability estimator and thus its employment is likewise encouraged.
Keywords: Ability Estimation, Item Response Theory, Classical Test Theory, Bayes Modal Estimation, Three Parameter Logistic Model.
