ABSTRACT
There are a number of situations in which we would like to measure a given attribute using two different tests, that is, two different sets of items. It is conventional to refer to two tests measuring “the same” attribute as alternate forms of a test. The need for alternate forms arises most commonly and with greatest urgency in cognitive testing. Tests of general scholastic aptitude or of specific scholastic aptitudes are commonly used to select students into institutions for further learning. The development of alternate forms is motivated by the need for test security, and the contradictory need for freedom of information. The latter need includes the right of an examinee to verify some aspect of the testing process, either as a legal right or as an ethically acknowledged one. It may be necessary to produce a large set of items—an item bank—from which subsets of items can be drawn and used for the measurement of a given aptitude. If the item bank contains enough items, there is no possibility that an examinee having access to it could adopt a strategy of memorizing correct answers. Another situation motivating the development of alternate test forms is one where an educational achievement test, defined by an educational curriculum, is administered at the beginning and end (and possibly at intermediate points) of a course of instruction, to track the learning process. We do not then want to contaminate the measurements by the reuse of items already seen. We note immediately that curricular domains are typically rather complex, and the notion that “the same” knowledge is measured by distinct sets of items can be more problematic, in terms of substantive considerations of content validity, than in the case of some, at least, of the aptitudes.
