Item banking with Rasch Measurement: An example for primary mathematics in Thailand
Edith Cowan University
Faculty of Education and Arts
School of Education
This study was conducted in Thailand to create a Mathematics item bank and a Computerized Adaptive Test (CAT) for the students to ‗interrogate‘ the bank. First, 290 multiple-choice test items on mathematical equations were created for an item bank. They consisted of nine aspects: (1) identifying an equation; (2) identifying the true equation; (3) identifying equations with an unknown; (4) finding the value of an unknown that satisfies the equation; (5) identifying a method to solve an equation; (6) finding the solutions to equations; (7) finding a solution to an equation related to a given condition; (8) selecting an equation converted from a verbal problem or a verbal problem related to an equation; and (9) solving an equation problem. Seven papers with 50 items each, containing 40 different items and 10 common items, were administered to 3,062 students of Year 6 (Prathom Suksa 6). There were 409, 413, 412, 400, 410, 408, and 610 students taking part in the 1st to the 7th tests respectively. The data were analysed with the Rasch Unidimensional Measurement Model (RUMM 2010) computer program so that all the item difficulties were linked on the same linear scale along with the student measures of mathematical ability. Ninety-eight test items fitted the measurement model and were installed in the item bank. A computer program for CAT was created, tested, and modified after trialling. A controlled experiment involving the use of CAT with 400 Prathom Suksa 6 students from two primary schools in Ubon Ratchathani province, Thailand, was implemented. Thai students were very supportive of the use of CAT with the mathematical item bank. They showed an interest in CAT and in extending the use of CAT to other subject areas with appropriately developed item banks.