ABSTRACT
A common distinction in educational and psychological testing is the one between speed and power tests. The classical definitions of both types of tests are due to Gulliksen (1950, Chapter 17), who suggested using both the difficulty of the items and the presence of a time limit as criteria. According to his definitions, a speed test would be a test with items easy enough to answer each of them correctly but a time limit that guaranteed that none of the test takers could answer all of them. On the other hand, a power test would have a fixed number of items of increasing difficulty but no time limit. The two types of tests were supposed to be scored differently. For speed tests, the number of items completed within the time limit was assumed to be a proper score, while power tests were assumed to be scored by the number of correct answers. Gulliksen also presented an alternative score for speed tests: the time used to complete a predetermined number of items. Although less practical in the context of paper-and-pencil testing, the alternative was assumed to be entirely equivalent to the number of items completed for the given time limit; that is, except for random error, both were supposed to rank the same set of test takers identically.
