Modern GRE, GMAT, LSAT, SAT and MCAT are called computer adaptive tests (CATs) because computer adapts the questions to the level of your skill and performance. Sometimes before all psychological tools were offered in pencil and paper to measure different aspects of your personality and intelligence. It is not too far when all psychological tests shall adapt to computer based algorithm.
The computer is given access to a large number of test items classified in question types like graphs, antonyms, math, quantitative etc and arranged in order of difficulty.
When you face Computer Adaptive Tests, the computer program continuously updates your estimation from the very first question to the last item. Initially, you are supposed an average person. When you answer the first item correctly the next question becomes a bit difficult. When you again answer correctly, another harder item is offered. The computer continues increasing the level of difficulty until you start answering wrongly.
Watch Computer Adaptive Test Article on Video
The CATs environment is different from the normal testing session. You sit in a computer lab or testing center while a monitor is placed in front of you. You may be alone in the examination room or other candidates may be sitting in nearby cabins.
You are required to start your test by clicking an icon on the computer screen. First question appears and you are required to answer it by making a choice of the multiple options. Then you are required to confirm your answer by clicking a box in the same screen to move forward. Then a new question appears on the question and you repeat the process until you reach to the final question.
This mode of testing is different from the normal psychological testing session on a number of accounts.
1- In a normal pencil-paper standardized test, you are given specific items in the booklets of questions and answers. Every subject receives the same number and kind of items. Everyone has to face the same level of difficulty or ease. These easy and hard items are like adding constants to someone's score. They provide relatively little information about the subject's ability level. Consequently, large numbers of items and examinees are needed to obtain a modest degree of precision. However, in a CATs you may see the questions entirely different from the candidate sitting next to you. Your ability level relative to a norm group can be iteratively estimated during the testing process. You can be given the items that maximize the information about your ability level from your responses to different items. You will receive items which are very easy or very hard depending upon your responses to your current answers.
2- In a normal standardized psychological test you may skip from one question to another by leaving many questions unanswered and then come back to make a well calculated choice. You can answer the questions in any order as you like. However, in these tests you are not allowed to skip through the questions. When you skip a question, you can’t come back to answer it later on. You are required to answer the questions in the order as they are offered on your monitor. You need to calculate your estimated time that you can allocate for each question. You should not spend too much time on a question for the correct answer. When you are not sure of your answer, you should make a choice before moving forward. Keep in mind that most of the computer adaptive tests don’t penalize you for your wrong choices.
3- In a normal psychological test, you may have second thought about your choice. You can go back, erase your previous choice and select a different option. However, this luxury in not available in the adaptive tests.
4- In normal standardized psychological you are free to make some marks, cross out incorrect options, underline keywords, and highlight questions to consider them later. However, this facility is not available in computer adaptive tests.
The psychologists build computer adaptive tests with the following constructs:
The items pool
The pool of items must be calibrated with a psychological testing model. Typically item response theory (IRT) model is applied in the computer adoptive testing.
Computer adaptive tests assume that you are average and the first question is always an average question.
Item Selection procedure
The IRT places you and the items on the same metric known as theta. So, when you answer your first item, it starts measuring your level of capability and the process continues till completion of the test.
Item scoring procedure
When you answer a question correctly, your ability is scored higher and when your answer is wrong, your capability level falls. Using the procedure each item is score in accordance with the already set value of the item.
The computer adaptive tests (CATs) need a termination criterion otherwise you shall continue answering the questions until the whole item pool exhausts. Often computer adaptive tests are terminated when your standard error of measurement falls below a certain level.
We find value in differences between learning, interpreting and overall opinions. Please share your thoughts freely about this topic, but always remain respectful. Thank you for your contribution.