An exchange algorithm for optimal calibration of items in computerized achievement tests

The importance of large scale achievement tests, like national tests in school, eligibility for university, or international assessments evaluation students, is increasing. Pretesting questions the above mentioned done to determine characteristic properties by adding them an ordinary test. If computerized are used, it has been shown using optimal experimental design methods that efficient assign pretest examinees based on their abilities. specific distribution abilities available considered and restricted designs applied. A new algorithm developed which builds equivalence theorem. It discretizes space with possibility change grid adaptively during run, makes use exchange idea filters computed designs. illustrated how works through some examples as well convergence can be checked. flexible used even if different models assumed questions.