Multiple-choice items analysis using classical test theory and rasch measurement model

• Main Authors: Abd. Latif, Azizah, Yusof, Ibnatul Jalilah, Mohd. Amin, Nor Fadila, Libunao, Wilfredo Herrera, Yusri, Siti Sarah
• Format: Article
• Published: Serials Publications 2016
• Subjects: LB Theory and practice of education
• Online Access: http://eprints.utm.my/id/eprint/71308/; https://www.scopus.com/inward/record.uri?eid=2-s2.0-84958968806&partnerID=40&md5=03dd0b621f134eab83edb1ad7c3eea1b

The purpose of this study is to analyze the level of difficulty item and person ability using two measurement frameworks, Classical Test Theory (CTT) and Rasch Measurement Model (RMM). A total of 100 undergraduate students from Faculty of Education were responded to final examination paper, Research Methodology. This paper consists of 60 multiple choice questions (MCQ). The value of Cronbach-α(CTT) obtained is 0.62 while Person Reliability (RMM) is 0.59 whereas Item Reliability is 0.96. This study found that there is a slight difference in level of item difficulty index and person ability obtained from CTT and RMM. Additionally, there is no significant difference (p>.05) between item difficulty index (CTT) and item measure (RMM). This study also found that there is no significant difference of person ability (p>.05) estimates by CTT and RMM. RMM is theoretically considered as the superior measurement framework over CTT, this study however found that item and person statistics appeared to be similar for this two measurement frameworks. Thus, the interpretations beyond the philosophy were discussed.