If you've heard that some questions on the GMAT® don't count toward your score, you might be wondering whether you can identify them and skip them. Or whether a question that seems bizarre might be one you can safely ignore.
That's a reasonable thing to wonder about. The short answer is: some questions on the GMAT® probably don't count, but you can't tell which ones they are, and trying to figure it out can hurt your score.
Here's what experimental questions are, why they exist, and the one way knowing about them can actually help you on test day.
What Experimental Questions Are
Experimental questions, sometimes called pretest or research questions, are questions that GMAC includes on the exam to collect data before the questions become scored items. Test takers answer them as part of the normal exam, but the answers don't affect the test taker's score. GMAC uses the data from those answers to calibrate the question's difficulty, check its fairness across different groups, and decide whether it's ready to become a real scored question on future exams.
This is standard practice in computer-adaptive testing. Any adaptive test needs a steady pipeline of new questions to keep the question pool fresh, and the only way to know how a new question performs under real test conditions is to put it in front of real test takers. GMAC can't simulate that in a lab. They need actual performance data from the actual exam.
So some small number of questions on your GMAT® are likely there for research purposes. They look like regular questions, they're mixed in with the scored questions, and you won't see any label or marker distinguishing them from questions that count.
Does the GMAT® Focus Edition Have Experimental Questions?
GMAC doesn't officially confirm that the current GMAT® includes experimental questions. They don't publish information about how many unscored items appear on each section, and the official materials don't use the term "experimental" at all.
But here's what we know. On the classic GMAT®, which was retired in early 2024, experimental questions were an open secret. GMAC confirmed that some questions were used for research purposes and didn't count toward scores. When we were working with the classic GMAT® years ago, we estimated that roughly 8 out of 31 quant questions didn't count for a test taker's score. The verbal section had a smaller number, probably around 3.
The current GMAT® (formerly called the Focus Edition) is shorter than the classic version. It has 64 total questions across three sections instead of the 80 questions on the old format. A shorter exam with fewer questions means less room for unscored items, so the number of experimental questions has likely decreased. But adaptive testing still requires new questions to be calibrated, which means some form of pretesting is almost certainly still happening.
We can't tell you exactly how many experimental questions are on the current exam because GMAC doesn't share that information. What we can tell you is that the number is probably small, and the practical impact on your score is minimal.
Why You Can't Identify Experimental Questions
This is the part where a lot of test takers go wrong. They hear that experimental questions exist, and they start trying to spot them during the exam. The logic usually goes something like: if a question seems weird, unusually difficult, or unlike anything I've seen in my prep, it must be experimental.
That approach has a serious flaw. The GMAT® is adaptive, which means it gives you harder questions when you're doing well and easier questions when you're struggling. A question that feels strange or unusually hard might be experimental. But it's far more likely to be a real scored question that the algorithm has served you because you're performing at a level where hard questions are the norm.
If you decide a question is experimental and stop trying on it, two things can happen. If you're right and it's experimental, nothing changes. If you're wrong and it counts, you've given away a point on a question you might have answered correctly. The downside of being wrong is much larger than the upside of being right.
There's also no reliable pattern to what experimental questions look like. They're not always the hardest questions, and they're not always the weirdest. A new question being pretested could be easy, hard, or anywhere in between. That's the whole point of pretesting: GMAC needs data across the full difficulty range. So the "weird question = experimental" heuristic doesn't hold up.
The One Way Knowing About Experimental Questions Helps
There is one productive use for this information, and it has nothing to do with identifying which questions are experimental. It's about managing your mental state during the exam.
If you hit a question that throws you off, one that seems impossible or unlike anything you prepared for, it might trigger a spiral. You might start questioning your prep. You might start questioning whether you're doing well. You might spend extra time on the question, which could eat into your time for the rest of the section.
In that moment, telling yourself that the question might be experimental can be a useful mental reset. You don't actually believe it, and you shouldn't act on it by giving up on the question. But using it as a brief pressure release can help you let go of the anxiety, make your best guess, and move on to the next question with your confidence intact.
The goal isn't to convince yourself the question doesn't count. The goal is to keep one hard question from derailing the rest of your section. Give it your best effort, pick an answer, and reset.
What This Means for Your Study Strategy
Knowing about experimental questions shouldn't change how you prepare. Every question might count, so you prepare for every question as if it does. The study habits that raise your score are the same whether 5 percent of questions are experimental or 15 percent are:
Treat every practice question as a real question. Build your timing discipline around answering everything. When you review your practice tests, analyze every missed question, because you won't know which ones you'd be safe skipping on the real thing.
The students who score well on the GMAT® tend to be the ones who answer every question as well as they can, manage their time across the full section, and don't let one hard question shake their focus.
FAQ
Are there experimental questions on the GMAT® Focus Edition?
GMAC doesn't officially confirm experimental questions on the current GMAT®, but the practice of including unscored research questions is standard in computer-adaptive testing. The classic GMAT® confirmed their use, and the current shorter version likely includes a smaller number. You won't be able to identify which questions are experimental, so the practical impact on your strategy is minimal.
How many experimental questions are on the GMAT®?
On the classic GMAT®, test experts estimated roughly 8 experimental questions on the Quant section (out of 31) and around 3 on the Verbal section (out of 36). The current GMAT® is shorter with 64 total questions, so the number of experimental questions is probably lower. GMAC doesn't publish the exact number.
Can you tell which questions are experimental on the GMAT®?
No. Experimental questions look like regular scored questions. There's no label, no pattern, and no reliable way to identify them during the exam. Questions that seem unusual or very hard are usually adaptive difficulty, not experimental items.
Should you skip questions you think are experimental?
No. Since you can't identify experimental questions with any reliability, skipping a question you think is experimental means risking a real scored point. Treat every question as if it counts.
Do experimental questions affect your GMAT® score?
Experimental questions don't directly affect your score, but the time you spend on them can affect your pacing for the rest of the section. Every minute you spend on an unscored question is a minute you can't spend on a scored one. That's another reason to maintain consistent timing discipline across all questions rather than over-investing in any single problem.
Why does the GMAT® include experimental questions?
Computer-adaptive tests need a constant supply of new scored questions to keep the question pool current. To calibrate a new question's difficulty and verify its fairness, GMAC needs performance data from real test conditions. Experimental questions are the mechanism for collecting that data before a question becomes scored.
Want to learn even more?
Understanding how the GMAT® scoring algorithm works is one of the highest-leverage things you can do for your prep. If you haven't yet, read our complete guide to how the GMAT® scoring algorithm works — it covers adaptive difficulty, how the algorithm assigns your score, and why getting more questions right doesn't always mean getting a higher score.
For timing strategy across all three sections, our GMAT® timing strategy guide walks through when to let go of a question, how to pace each section, and how to train yourself to move on without panicking.
And if you're hitting a wall in your prep, our guide on breaking through a GMAT® score plateau covers the most common reasons scores stop improving and what to do about it.
You can also listen to our podcast series on Spotify, Apple Podcasts, or YouTube for full episodes on GMAT® strategy, study methods, and test-day preparation.