ABSTRACT

In a recent paper by Willse, Ackerman, and Luecht (2012, 2), the test structure referred to as computerized adaptive multistage testing (MST) was introduced as “a modern way of designing, assembling, and administering self-contained, auto-adaptive units for a wide variety of potential purposes.” This statement quite elegantly frames the MST design as a separate and distinctive approach to measurement independent of content or construct. MST requires many highly interdependent design decisions to be made throughout the process of test development, but the consequence is a highly flexible and customizable assessment that ultimately can provide very good measurement results according to all of the usual standards for evaluating such things as reliability and precision, validity, cost, and so on.