For two tests, a link between their scores is a transformation from a score on one to a score on the other. Linking methods can then be divided into three basic categories, called prediction, scale aligning and equating, and several subcategories. Testing professionals and researchers need to understand the differences among these methods and the circumstances when one method is more relevant than another. Also, if required, their need to be able to put in plain words these distinctions to test users. In this paper we use the framework developed by Holland and Dorans (2006) to give a brief account of these methods and suggest some practical applications of them.

Linking, equating, scale aligning, prediction, comparability.