http://www.ncsbn.org/testing/candidates_info_cat.asp text: Computerized Adaptive Testing (CAT) Overview Computerized adaptive testing (CAT) is a method for administering tests which uses current computer technology and measurement theory. The NCLEX® examination administered via CAT uses standard NCLEX examination multiple-choice questions. With CAT, each candidate's test is unique: it is assembled interactively as the individual is tested. As the candidate answers each question, the computer calculates an ability estimate based on all earlier answers. The test questions, which are stored in a large item bank and classified by test plan area and level of difficulty, are then scanned and the one determined to measure the candidate most precisely in the appropriate test plan area is selected and presented on the computer screen. This process is repeated for each question, creating an examination tailored to the individual's knowledge and skills while fulfilling all NCLEX test plan requirements. The examination continues in this way until a pass or fail decision is made. CAT provides greater measurement efficiency, as it administers only those questions which will offer the best measurement of the candidate's ability. The way a CAT examination works is very similar to the way an educator might administer an oral examination. After a candidate's response to the first question asked, the educator assesses the candidate's ability level, then asks another question, based on this preliminary assessment. The candidate's answer to this second question provides the educator more information about the candidate's ability, and the educator's assessment is becoming more precise. This process continues, with the educator asking either easier or more difficult questions, based on the candidate's responses. This type of examination, like a CAT examination, is tailored to each individual candidate, and only those questions which will measure the candidate's ability the best are administered. During an oral examination, the educator makes an assessment of the candidate's ability, then chooses the appropriate level of difficulty for each question administered. In a CAT examination, this process is built into the computer system. NCLEX examination decisions are not based solely on how many questions a candidate answers correctly, but also on the difficulty of the questions a candidate answers correctly. CAT administers questions with difficulty levels so that each candidate will answer about half correctly; these questions provide the most information. Thus, all candidates answer about 50 percent correctly: passing candidates answer 50 percent of more difficult questions correctly, and failing candidates answer 50 percent of easier questions correctly. Even though candidates may answer different questions and different numbers of questions, the NCLEX examination administered using CAT is fair to every candidate. All examinations conform to either the NCLEX-RN® or NCLEX-PN® Test Plan which controls inclusion of important nursing content, and all candidates have ample opportunity to demonstrate their ability, as the examination won't end until stability of the pass/fail result is assured or time runs out. How CAT Works: A Candidate Primer The goal of computerized adaptive testing, or CAT, is to determine your ability, based on the difficulty of questions you can answer correctly, not how many questions you can answer correctly. This is a fundamentally different approach than is used on paper-and-pencil tests, where everyone receives the same questions. CAT examinations are individualized. We know the exact difficulty of the approximately 1500 questions in each operational NCLEX examination item pool because each has been taken as a pretest question by hundreds of candidates and then statistically analyzed. Picture the questions all lined up, from easiest to hardest. If we asked you the easiest questions, you'd get most of them right. If we asked you the hardest, you'd probably get most wrong. As we move from easy to hard, there will come a point where you go from getting more right to getting more wrong. This is the point where you are answering 50 percent correctly. Questions harder than that you would probably answer incorrectly (you'd get some right, but more wrong); questions easier than that, you would probably answer correctly. The goal of CAT is to find that point for you. That point is different for everyone. Nursing experts could probably answer at least one-half of the hardest questions we could ask. We'd have to ask beginning nursing students the very easiest ones before they could answer even one-half correctly. You probably fall somewhere between those two points. First, the computer asks a relatively easy question, and if you answer it correctly, it asks a somewhat harder question. As you continue answering correctly, the questions get harder and harder. When you start missing questions, they get easier until you start answering them right again, then they get a little harder. Each time you answer one correctly, the next is harder. Each time you answer one incorrectly, the next is easier. This process continues as it zigzags, narrowing in on the point where you answer 50 percent correctly, e.g., one right, then one wrong. That point represents your ability level. This is why everyone ends up correctly answering about 50 percent of the questions he or she is asked. After you have answered the minimum number of questions, the computer compares your estimated ability level to the passing standard and makes one of three decisions: One, if you are clearly above the passing standard, you pass and the examination ends. Two, if you are clearly below the passing standard, you fail and the examination ends. Three, if your ability estimate is close enough to the passing standard that it's still not clear whether you should pass or not, the computer continues to ask you questions. "Clearly" passing or failing is defined as when the "gray zone" around your ability level falls entirely above or below the passing standard. The gray zone is the region within which your estimate might vary if you answer more questions. The gray zone shrinks a little after each question because your estimate is based on more information. After each question, your ability level and the gray zone are recomputed, adding your new response to all of your previous answers. When the gray zone is entirely on one side or the other of the passing standard, you've clearly passed or failed and the examination ends. Of course, some people's ability level is very close to the passing standard. For some of these people, all the questions in the item pool still might not be enough to make it "clear" whether they should pass or fail. When a candidate's ability level is very close to the passing standard, the computer continues to administer questions to them until the maximum number of items is reached. At this point, the computer disregards the gray zone and simply looks at whether the final ability estimate, based on every question answered, is above or below passing. If you are above it, you pass. If not, you fail.