Tests given by computers measure a person's ability as accurately as paper and pencil tests, do it in less time, and create less test anxiety, according to a four-year study conducted by the University of Missouri-Columbia.
In tailored testing a computer program statistically selects, from the computer's storage, the items a test taker receives. Beginning with the first question, which usually is of average difficulty, the computer goes through a cycle of selecting, administering, and scoring as each item is answered.
When an answer is correct, the next item selected is more difficult. When incorrect, the next item selected is easier.