Development of an item bank for statistics course using Rasch Analysis

Item bank is a large collection of good test items which their quality is analysed and systematically stored in a computer so that they are accessible to exam setters. The purpose of this study is to develop a new item bank and to measure the validity and reliability of final examination paper for Statistics course. There were 344 students taking this course which comprises 7 structured questions and subdivided into 21 items. The data were analysed using Rasch analysis so that all the item difficulties were linked on the same linear scale along with the students’ measured ability. The results showed that the Statistics course is unidimensional, Cronbach’s alpha value of 0.90, item and person separation more than 2, and person and item reliability of 0.87 and 0.99 respectively. Nineteen items were found to be fitted the measurement model and stored in the item bank while there are two items suggested to be misfit and need to be revised. To summarise, a well-developed item bank can be extremely beneficial to exam setters because it can assist test construction easier, faster, and more efficient. As a result, the test's quality should be higher than it would be without an item bank