Tell me if this sounds familiar.

You take a GMAT practice test and get, say, 17 out of 21 questions right on the Quantitative Reasoning section. After finishing the test, you see your Quant score: 81

Later, you take another practice test and get 17 out of 21 questions correct again. Yet, when you see your score, it's a 77.

Wait... what!?

What changed? How is it possible to get the same number of questions right but receive a different score? 

And if your score is a moving target, how can you reliably figure out how many questions you need to get right on the GMAT Exam to score a 705 (or whatever your target score is)?

Understanding GMAT Adaptive Scoring

The reason two test takers can receive different GMAT scores even if they get the same number of questions right is that the GMAT Exam uses an adaptive scoring algorithm. The question difficulty of each question you see depends on whether you got the previous question right or wrong. When you get a question correct on the GMAT, the next question is more difficult; when you get a question wrong, the next question is easier.

With this in mind, let's revisit our earlier hypothetical.

The reason two different test takers who both get 17 out of 21 Quantitative Reasoning questions correct can receive different Quant scores is that they may not get the exact same questions right and wrong. If one candidate gets easier questions wrong while the other candidate gets harder questions wrong, their scores will necessarily differ based on how the GMAT's scoring algorithm interprets those results and predicts each test taker's true abilities.

Indeed, your final GMAT score can vary widely depending on which questions you get right and wrong. The chart below illustrates this reality by presenting a couple of different scenarios and the corresponding expected score ranges.

Source: Graduate Management Admission Council updated percentile rankings data

Let's look again at the scenario where a test taker gets 17 out of 21 questions right on the Quantitative Reasoning section. What stands out from the GMAC's data (row 1) is the wide range of potential scores based on that performance. The median score is 81, which would land you in the 71st percentile according to the latest GMAT percentile rankings. But notice that your Quant score could be as low as the 14th percentile (71), even with just missing four questions. On the flip side, half of test takers (by definition of a median) will end up scoring 81 or higher when they miss four Quant questions.

The key takeaway, then, is that your final GMAT score isn't so much a function of how many questions you get right, but rather which questions you get right.

The Relative Importance of GMAT Questions

While all of this is interesting, it raises perhaps the most important question when trying to determine how many questions you need to get right on the GMAT to hit your target score: If which questions you get right/wrong is more important than how many you get right/wrong... well, which GMAT questions carry the most weight?

At a recent conference we attended at the headquarters of the GMAC, their lead psychometrician, Chris Han, pulled back the curtain on the intricacies of the GMAT scoring algorithm and helped to explain the answer to that question.

The current iteration of the GMAT Exam uses Item Response Theory (IRT) as the cornerstone of its scoring algorithm. IRT assesses a student's ability/proficiency at the individual question level based on a number of factors, the main one being a question's difficulty level. There's a long, complex mathematical equation with lots of Greek letters that is used to translate a right/wrong answer response into a predicted ability/proficiency. We don't need to get lost in the weeds of the formula itself, but a big-picture understanding is helpful for understanding how the computer arrives at your final GMAT score — and it boils down to this:

You GMAT ability (𝛳) is a function of each question's difficulty parameter (b).

Put even more simply, the computer makes a determination about your predicted GMAT score based on whether you get hard/easy questions right/wrong.

Mr. Han shared the following example to illustrate on a small scale how the GMAT's scoring algorithm works across an entire section:

Consider five test takers: A, B, C, D, and E. Now imagine they answer five GMAT questions of varying difficulty, ranging from a b-value difficulty score of 2.23 (hardest) to -1.42 (easiest). In the chart above, let's focus on Test Takers A and D. Each correctly answers 4 out of 5 questions. The difference, however, is that Test Taker A misses the easiest question while Test Taker D misses the hardest question.

Who do you think is the better GMAT candidate? Who should receive the higher score?

As seen in the last row of the table, Test Taker D, who only misses the hardest question, ends up with the highest 𝛳 (predicted ability) score, translating to the highest final GMAT score. Test Taker A still ends up with the second-highest 𝛳 value (it's still better to get four right answers than two or one) but is penalized for missing the easiest question.

This makes sense if you think about it. Someone who truly has the skills and ability to score in the 90th percentile or higher on the GMAT has no business missing the easiest questions on the exam. That's a red flag. By contrast, a below-average test taker might get lucky and guess correctly on some of the harder questions, but missing the easiest question is a tip-off as to his/her true abilities.

So, How Many Questions for a GMAT 705?

Let's get back to our original question. As you now understand, unfortunately there isn't a simple, clear-cut answer to the question of how many questions you need to get right on the GMAT to get a specific score.

But we do know a couple of things for sure.

First, it's always best to get as many right answers as possible. Looking back at the first table, the median Quant score is higher for students who get 19 questions right (85) than for students who get 17 questions right (81). Tip the balance in your favor by minimizing wrong answers.

Second, if you're shooting for the highest possible GMAT scores, do everything you can to avoid getting some of the easiest questions wrong because there is a penalty for doing so. Here are a few tips to help you along these lines:

1. Remember that the first few questions you see on each section are likely to be the easiest questions you see, especially if you're an above-average test taker. As such, spend a little extra time on the first few questions to avoid careless errors.
2. Spend time controlling your text anxiety before starting the exam. Students often make mistakes on the first few questions not because they don't know how to solve them, but because they're nervous and their mind hasn't calmed down yet.
3. Choose a section order (SSO) where your first section is your best section. You're likely to be less anxious — and therefore ready to perform your best right out of the gate — if you start with the content that you're most comfortable with.
4. Avoid careless errors (Listen: "9 Ways to Avoid Making Careless Errors")
5. Don't overthink it. It's human nature to try to figure out how you're doing as you move through the exam. But just because you think you're looking at an easy question doesn't necessarily mean that it is — and even if it is, it may be one of the experimental questions that doesn't count toward your score anyway. Just do the best you can on each question and let the GMAT's computers worry about calculating your final score.

All of that said, if your target GMAT score truly is 705 or higher, that puts you in the 99th percentile so you'll need a 99th percentile performance. To do that, you can't afford to miss more than one or two questions per section at the most. Not only will you need to avoid making careless errors on easy questions, but also you will need to get most of the hard questions right, too. Our complete GMAT Prep Course can help you with that.