Evaluating quality and reliability of final exam questions for probability and statistics course using rasch model

• Published in: International Journal of Engineering and Technology(UAE)
• Main Author: Zulkifli F.; Abidin R.Z.; Razi N.F.M.; Mohammad N.H.; Ahmad R.; Azmi A.Z.
• Format: Article
• Language: English
• Published: Science Publishing Corporation Inc 2018
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85059234320&doi=10.14419%2fijet.v7i4.33.23479&partnerID=40&md5=35f301c767ada27a4a4c4ca4d8260358

Evaluation of the questions' level of complexity for the statistical course was proposed using the revised version of Bloom's taxonomy. The use of Bloom's taxonomy in statistical examination papers allows the degree of difficulty to be pseudo-objectively assessed. Well-constructed questions in the final examination will help in measuring students' abilities based on comprehensive cognitive skills. Therefore, this study used Rasch Model to evaluate the quality and reliability of final exam questions for probability and statistics course. According to research findings, five out of 30 questions are considered as misfit items. It is therefore recommended that these items be removed or rephrased to better suit the students' ability level in a course. Whereas, nine questions have significant differences between taxonomy level and Rasch level that require further analysis. Overall, students view the set of exam questions as simple due to the unavailability of difficult items. Based on this result, it is suggested that the exam questions should undergo verification process from the expert and students should be exposed early to various types of questions with different level of difficulty. © 2018 Authors.