نوع مقاله : مقاله پژوهشی
نویسندگان
1 دانشجوی دوره دکتری سنجش و اندازهگیری دانشگاه علامه طباطبایی
2 دانشیار دانشکده روانشناسی و علوم تربیتی دانشگاه علامه طباطبایی
3 استاد دانشکده روانشناسی و علوم تربیتی دانشگاه علامهطباطبایی
4 استادیار سازمان سنجش آموزش کشور
چکیده
کلیدواژهها
عنوان مقاله [English]
نویسندگان [English]
Computerized adaptive testing requires a well-designed item pool containing an appropriate number of items to build individualized tests. An optimal item pool should also contain well-balanced items that will achieve optimal item usage and low cost of item creation. One of the methods for designing the blueprint for an item pool is Reckase’s method, which is a Monte Carlo method to determine the properties of an optimal item pool. This study has been presented by the Reckase’s method for designing item pools calibrated with the three-parameter logistic model. Also, in the present simulation, the Sympson-Hetter procedure has been used to control the item exposure rate. To design simulation test items, three approaches including R, MRT, and MTI have been applied. The performance of the simulated item pools has been compared with the operational item pool by considering some evaluation criteria. The results suggest that the optimal pool designed with MRP performs the best, based on test security and measurement accuracy. It does require more items and some highly discriminating items. Tests assembled from MRP pools have smaller test-retest overlap rates and significantly lower percentages of over- and under-exposed items. The results show that the MTI design generally leads to smaller pools and contain items with lower a-parameters. Overall, Considering the Sympson-Hetter procedure, the optimally designed item pools perform better than the operational pool obtained from operational CAT
کلیدواژهها [English]