CAT and the NCLEX: How Computer Adaptive Testing Decides Pass or Fail

Computer adaptive testing is the mechanism that chooses your next question, sets how long your exam runs, and produces the pass or fail decision. It is also the source of nearly every myth about the NCLEX, because the behavior it produces is genuinely counterintuitive.

This guide covers the mechanics only: what CAT is, the three official rules that end an exam, why you cannot return to a submitted answer, and the unscored pretest items mixed into every test. For what to actually do with that knowledge while sitting the exam, see our companion guide to NCLEX test-taking strategies for the CAT.

What is computer adaptive testing?

Computer adaptive testing, administered by NCSBN, adjusts item difficulty in real time based on your answers. Instead of serving every candidate an identical fixed list, the NCLEX selects each new item from a running estimate of your ability, refined after every response.

The exam opens at moderate difficulty. Answer correctly and the system raises its estimate of your ability and serves something harder. Answer incorrectly and it lowers the estimate and serves something easier. That loop repeats item by item, with your estimated ability moving up or down the scale each time.

This is why two candidates sitting the same exam at the same hour can share almost no questions. Someone performing strongly is routed toward harder content; someone struggling sees an easier sequence. The system is trying to locate your true ability as efficiently as possible, and a run of hard questions usually means it is succeeding.

The passing standard is a fixed criterion rather than a curve, which is a separate subject covered in the NCLEX passing standard. For the item counts themselves, see how many questions are on the NCLEX.

The three ways the NCLEX ends

Exactly three rules can end an NCLEX exam. Knowing them removes most of the guesswork people bring into the testing center, because each one produces a different pattern of exam length, and none of those patterns means what candidates assume.

Both the NCLEX-RN and NCLEX-PN run a minimum of 85 items and a maximum of 150, inside a five-hour limit.

The 95% confidence interval rule. The most common ending. After each item, the system checks whether it can conclude with 95% statistical confidence that your ability estimate sits clearly above or clearly below the passing standard. The moment it reaches that confidence in either direction, the exam stops. That can happen at the minimum item count or far later, depending entirely on how cleanly your answers separate from the standard.

The maximum-length exam rule. If your ability estimate stays too close to the passing standard for the confidence rule to resolve, the system keeps administering items until it hits the maximum of 150. The decision then reduces to one comparison: if your final ability estimate is above the passing standard you pass; at or below it, you fail.

The run-out-of-time rule. If the five-hour limit expires before you reach the minimum item count, the result is an automatic fail, because there are not enough scored responses to support a determination. If time expires after you have passed the minimum, the result is based on the ability estimate from the items you completed.

Your item count tells you almost nothing on its own. A short exam can mean a clear pass or a clear fail. A long exam usually means your performance hovered near the standard, which is not the same as failing.

Pretest items and the no-going-back rule

Two design features surprise candidates more than any others, and both follow directly from the adaptive format rather than from any attempt to make the exam harder. Fifteen items on every exam do not count toward your score and cannot be told apart from the ones that do. Separately, once you submit an answer it is final, with no review screen and no way back. Neither rule exists to catch you out; both fall out of how the algorithm builds your exam item by item, and understanding why makes them much easier to work with.

Pretest items

Every NCLEX includes 15 unscored pretest items distributed among the scored ones. NCSBN uses them to trial content for future exams, and they are written to be indistinguishable from scored items. There is no marker, no section, and no way to identify them.

The only workable response is to treat every question as scored, because the overwhelming majority are and you cannot tell which fifteen are not.

You cannot go back

Once you submit an answer, that item is locked. There is no review screen, no flagging for later, and no changing your mind two questions on.

This is a consequence of the format rather than a policy choice. Your next item was selected using your answer to the previous one, so revising an earlier response would invalidate the sequence that followed it. The practical implication is to make the most defensible decision you can the first time, using elimination and clinical reasoning, then move on without expecting a second look.

What CAT means for your test-day strategy

The mechanics above lead to four conclusions, and all four run against instinct. Difficulty is not feedback, item count is not a verdict, submitted answers are permanent, and every question has to be treated as scored. Candidates who understand the algorithm still lose points by reacting to it emotionally mid-exam, which is the specific failure mode this section is meant to prevent.

1. Stop reading difficulty as a scoreboard. Because the system routes harder content toward stronger performance, a run of difficult questions more often means you are doing well. Reading every hard item as evidence of failure creates a spiral that degrades how carefully you read the next several.

2. Stop counting. Reaching 150 items is not a verdict, and stopping early is not a guarantee. Mental arithmetic about your item count consumes attention you need for the question in front of you.

3. Commit and move on. With no revision possible, build the habit during practice of making a final decision and advancing. Practicing in a format that allows review trains a reflex the real exam will not let you use.

4. Treat everything as scored. The 15 pretest items are unidentifiable, so pacing and attention have to hold from the first item to whichever one turns out to be your last.

None of this substitutes for content mastery. Knowing how to answer NCLEX questions across item types is the part that moves your ability estimate; understanding CAT only stops you misreading the exam while you sit it.

Frequently Asked Questions

What does CAT stand for on the NCLEX?

Computer adaptive testing. It describes a format where each item's difficulty is selected in real time from your performance on previous items, rather than every candidate receiving an identical fixed exam.

How does the NCLEX decide when to stop?

One of three rules. The 95% confidence interval rule stops the exam once the system is statistically confident of a pass or fail. The maximum-length rule applies when your ability estimate stays too close to the standard to resolve, ending at 150 items. The run-out-of-time rule applies if the five-hour limit arrives first.

Can I identify the pretest questions and coast on them?

No. There are 15 unscored pretest items on every exam, scattered throughout and written to look identical to scored items. They cannot be identified, so every question should be treated as if it counts.

Can I skip a question or change an answer?

No. The NCLEX does not allow skipping, and a submitted answer cannot be revisited. Each item is selected from your previous response, which is precisely why review is not built into the format.

Does getting harder questions mean I'm passing?

It is a reasonable signal, not a guarantee. The system tends to raise difficulty when you are answering correctly, so a run of harder items often reflects strong performance. Treat it as a pattern worth understanding rather than a result worth reacting to.

Does reaching the maximum number of questions mean I failed?

No. Hitting 150 items usually means your ability estimate stayed near the passing standard for a long stretch. The outcome still depends on whether that final estimate lands above the standard or at or below it, so a maximum-length exam can end either way.

The bottom line

CAT is not designed to trick you. It is a statistical method for locating your true ability as efficiently as the data allows, and most of what feels mysterious about the NCLEX reduces to three stopping rules, fifteen unidentifiable pretest items, and a no-going-back format that resembles clinical decision-making more than it resembles a paper exam.

None of these mechanics should change how you prepare on content. What they should change is how you interpret the exam while you are inside it: difficulty is not a scoreboard, item count is not a verdict, and every question deserves full attention because you genuinely cannot tell which ones are scored.

Understanding the system removes a layer of anxiety that has nothing to do with your nursing knowledge and everything to do with not knowing how the test behaves. The knowledge itself still comes from NCLEX preparation.