Adaptive Testing

Certification and Skills Assessment Group

Microsoft, Inc.

History of Adaptive Testing in High Stakes Testing

Adaptive testing is not new. In fact it has probably been around for centuries as a better way to test. And even since the beginning of the 20^th century, when large-scale testing began, adaptive tests were some of the first tests ever constructed. Alfred Binet created the first adaptive test at the end of the last century, designing an IQ test that began with questions which matched the child’s age and ended when the child could not answered a few questions in a row incorrectly. Binet’s IQ test, which is still in use today in a more modern version, was not computerized, but used individual examiners to administer the tests. Today, both the Stanford-Binet test, and the more popular Wechsler intelligence tests for children and adults, use an adaptive approach for the selection of questions.

Today, with the common use of computers in test delivery, adaptive testing has become more popular and is called computerized adaptive testing or CAT. The computer can make the necessary calculations needed to estimate a person’s proficiency and to choose the questions to present. Several well-known high-stakes testing programs have adopted adaptive testing as their current and future method to test. The Educational Testing Service, the world’s largest testing organization, published the Graduate Record Exam (GRE) as an adaptive test in 1993, and has been slowly reducing its use of the paper version of the test. The Nursing Boards converted completely from paper-based testing to a computerized adaptive test in 1994. Over 400,000 exams for Registered Nurses are given each year. In the information technology industry, Novell successfully introduced CATs into its certification program in 1991. Over 1,000,000 Novell adaptive tests have been given.