The Sample Size Effect on Determining the Question Bank Items

Abstract

The study examined the effect of changing the sample size on the scoring of the item bank using the partial credit model (PCM), by using the test reference test designed by the researcher, and the item bank was estimated through using three different size samples (1000, 800, 400), Each time, two images were re-equated for two tests (easy and difficult) to obtain estimates of the item’s difficulty, and in the end, The results showed that the best scoring for the items corresponded to the sample size (1000) where the proportion of inappropriate items was (5.1%), when the sample decreased to (800), which is less than the ratio when the sample size decreased from (800) to (400) where it increased The ratio of inappropriate items to (8.5%) of the total number of items, indicating an increase in the quality of the scoring by increasing the sample size, and the study recommended that the sample size used should not be less than (1000) when using polytomous models.