A comparative analysis of item difficulty and Discrimination parameters estimation in classical test and item response theories.

Abstract

The global concerns for attaining the best practices in assessment prompted the comparison of classical test theory (CTT) and item response theory (IRT) analysis in psychometrics. These frameworks provide item information needed for the development and use of test items that are capable of estimating examinee’s ability and item characteristics. The study compared psychometric properties of item difficulty and discrimination indices in classical test theory and item response theory by the use of BILOG MG3 software. Ex post facto design was adopted for the study. The population for the study consisted of 11,538 candidates’ responses of candidates who took Type L 2020 Unified Tertiary Matriculation Examination (UTME) Mathematics paper in Akwa Ibom State, Nigeria. The sample of 5,192 (45%) responses was randomly selected through stratified random sampling technique. Two research questions were raised to guide the study. BILOG-MG3 computer software was used to calibrate the candidates’ responses in a 2-Parameter logistic model. The results revealed that some items showed similarities and differences at the various item difficulty indices at both CTT and IRT classification range of values. Also, in CTT item discrimination estimation, point-biserial values and IRT classification range values, showed several degrees of similarities and differences in item calibrations at various indices. However, it was concluded that CTT and IRT are comparable in their classification range values, but IRT provided more reliable information about item classification range values than CTT in psychometrics.