A Consideration on Point Allotments to Items and Score Adjustment in Testing: Comparison to the Item Response Theory
item response theory
difference reduction method between two empirical distribution functions
score transform via median value shifting
In many situations, regardless of whether we are aware of it or not, point allotment to items in testing and score adjustment are often performed without deep considerations; for example, in final examination in universities, teachers try to find the appropriate point allotments for scores, or to set the appropriate threshold to discriminate the successful group and the failed group by using their own methods. However, such the methods of point allotment and score adjustment may affect the examinees’ accurate evaluation for abilities. The item response theory, IRT, is one of the famous methods to evaluate examinees’ abilities and items (problems) difficulties simultaneously accurately and efficiently. In this paper, we consider typical situations in testing to compare the results from point allotment methods and score adjustment methods with those from the item response theory. We have found that score adjustment methods have possibilities to disturb the orders of scores arranged by using the raw scores. However, the IRT may consistently leave the orders as they were. In addition, the IRT arranges appropriate scores. To evaluate the examinees’ abilities accurately, it is recommended to use the IRT method rather than to use point allotment or score adjustment methods.
Departmental Bulletin Paper