A computer-adaptive test (CAT) is a type of assessment that adapts to the test-taker’s ability level in real-time. The computer algorithm adjusts the difficulty of the questions based on the test-taker’s responses to previous questions. If the test-taker answers a question correctly, the next question will be slightly more challenging, while an incorrect answer will result in an easier question. By continually adjusting the difficulty of the questions, the test can more accurately measure the test-taker’s knowledge and skills in a shorter amount of time than traditional tests.

CATs are commonly used for high-stakes exams, such as college entrance exams, professional certification exams, and licensure exams. They can help to reduce test anxiety since the test-taker is not confronted with questions that are either too easy or too difficult. Additionally, the adaptive nature of the test ensures that each test-taker is presented with a unique set of questions, reducing the likelihood of cheating. Overall, the use of computer-adaptive testing has revolutionized the way in which tests are administered and has proven to be an effective method for measuring knowledge and skills.

The Computer Adaptive Test (CAT) format is used in various standardized tests, including the Graduate Record Examination (GRE), Graduate Management Admission Test (GMAT), Medical College Admission Test (MCAT), and the Armed Services Vocational Aptitude Battery (ASVAB) and Digital SAT.

Algorithm

The algorithm of the Computer Adaptive Test (CAT) involves a sophisticated process of selecting questions based on the test taker’s performance. Here are the basic steps:

1. The test begins with a question of moderate difficulty.
2. Based on the test taker’s response, the algorithm assigns a score that reflects the individual’s estimated proficiency level.
3. Using this score, the algorithm selects the next question. If the test taker answered the previous question correctly, the next question will be slightly more difficult. If the test taker answered the previous question incorrectly, the next question will be slightly easier.
4. This process continues throughout the test, with the algorithm continually adjusting the difficulty of the questions based on the test taker’s performance.

The algorithm uses item response theory (IRT), which is a statistical model for measuring abilities and skills. IRT considers the difficulty of each question, the ability of the test taker, and the probability of the test taker answering the question correctly. The algorithm uses this information to estimate the test taker’s proficiency level and select the next question.

The goal of the CAT algorithm is to provide an accurate and efficient assessment of the test taker’s abilities. By selecting questions that are most informative, the algorithm can provide a precise estimate of the test taker’s proficiency level with fewer questions than a traditional test. This can save time and improve the precision of the test results.

The figure above demonstrates a possible adaptation CAT to DSAT. The first module consists of 22 medium-difficulty questions. If the student responds to 16 correctly out of 22 questions, he/she moves to module 2, a set of questions with corresponding difficulty. After finishing module 2, the system decides his/her score matching to the number of right questions. In this case, the test-taker can receive between 500 to 650 points. If a student does not perform well in the first module then the score he/she can receive at the most 500 points. In each module, the difficulty also could be changed by the response to the previous question.