Hey Gurpinder,

Good question...and unfortunately it's one that doesn't have an easy answer. Your score is calculated using the following factors:

-Number of questions answered correctly
-The difficulty level of those questions
-Number of questions answered overall (with severe penalties for leaving several questions unanswered)

Because of that difficulty level factor - which is a big one on a Computer Adaptive Test - your number correct doesn't give a terrific correlation to your scaled score. For simplicity's sake, let's just look at ten questions and assume that you start at a 50th percentile level. You could get 5 out of 10 correct in multiple ways, including:

Wrong, Wrong, Right, Wrong, Right, Wrong, Right, Wrong, Right, Right

Here, you'd have to think that by getting the first two wrong and then never getting consecutive right answers until the very end you probably stayed lower on the difficulty scale, so your 5/10 would still correspond to a slightly less-than-average score.

You could also go:

Right, Right, Wrong, Right, Wrong, Right, Wrong, Right, Wrong, Wrong

Here, it would seem that your first few right answers would have you at a higher difficulty level, so your correct answers would build you more points and the wrong answers wouldn't hurt you as much, so that 5/10 would probably yield a higher score.


NOTE: This is an intentionally crude demonstration - the GMAT has 37 and 41 questions for its adaptivity to work, so don't think that you HAVE TO get the first two right to get a good score or that getting the first two wrong dooms you. This is just a quick demo!


You may want to check out this blog post on the scoring algorithm, which links to a study conducted by Akil over at BellCurves in which he breaks down test results to show that you can get the same scaled score with wildly different numbers of right/wrong answers:

https://blog.veritasprep.com/2010/08/und ... rithm.html

The best advice I can give you - spend the vast, vast majority of your time working on getting questions right and you won't have to worry about the way it's scored. It's much easier to score well by doing well than by trying to game the system!

Hey Brian!!

Thank you very much for the post. Your advice is absolutely correct!


The reason why I asked this question is because I read alot of posts by people who have beaten the GMAT and said that they got something like 10 q's wrong (lets just say in quant) and got a scale score of 40+. On the other hand, I took a princeton cat and got 9 wrong and got a scale score of 35. I am confused. None of my wrong answers were consecutive. The pattern was like 5 correct 1 wrong, 6 correct 1 wrong, 8 correct 2 wrong.

So not sure! What do you think?

Thanks again!


p.s the link you posted is not active.

Nice catch on that link, Gurpinder - I saw that and thought I fixed it immediately...the HTML formatting just threw a "url" at the end of the actual URL so the hyperlink was off.

Regarding practice tests, they're much, much more useful as practice than assessment - GMAC's scoring algorithm is proprietary and pretty nuanced, and it's therefore tough to replicate. Practice tests typically do an okay job of letting you know the range in which you're scoring, but they do a much better job of helping you to recognize:

-How you're pacing yourself on each section
-Which mistakes you tend to make when working under timed test conditions
-How you hold up over a 3.5 hour test and stay focused on the later verbal passages and questions
-How your mind reacts to questions coming in varying orders (as opposed to doing, say, 30 data sufficiency questions in a row and getting into a DS rhythm)

My guess is that most practice tests are designed to mimic the actual GMAT by checking correlation between practice test and "official" GMAC practice test scores, but that the scoring engines aren't nearly as sophisticated as the official test's. I wouldn't try to read too much into scoring patterns in the practice tests since at best they're similar-to-but-different-from the official test. Learn from the mistakes you made so that you can improve upon them, assess your test-condition strategies, and use the practice tests as an opportunity to improve and not a true assessment of where you stand.