That's exactly how the test works - you will get a lot of questions wrong no matter what your score is. Most people get about half of the questions wrong - even at a 700 level!
So we really can't say whether we did well or poorly simply based on the number of questions right / wrong. The test is actually scoring us based on the difficulty level of the questions we're answering, not the pure percentage correct.
This is probably the easiest way to think about it:
On static, paper tests, everybody takes the same test. So the paper test includes questions from very easy to very hard and the score is based on the percentage that someone can get right.
On CAT tests, everybody takes a different test. Because the test can customize the mix of questions, it doesn't have to rely on percentage correct to calculate a score - and percentage correct actually does have a key limitation to it.
Let's say you and I both answer a bunch of 600 level questions and we both get them 100% right. Is our level 600? All we really know is that our level is at least 600. Maybe I can also answer all the 610 level questions right, but after that I start to get some wrong. Maybe you can answer all the questions up to 650 right, but then you start to get some wrong. Hmm. What does that mean?
What I really want to do (as the test algorithm) is find the level below which you tend to get questions right and above which you tend to get questions wrong. Then I know I've really found your true scoring level. So I give you questions that are clustered around what you're getting right and wrong as you take the test in an effort to find your "50/50" level - the level at which you get most of the questions below that level right and most of the questions above that level wrong.
Okay, so if that's how the test works, what does that mean for us in terms of being prepared to take the test? It's not based on percentage correct. I will get about half of the questions right and about half wrong. This is true up until the low 700s; above that, things do start to skew to more right, BUT even 99th percentile testers are getting a surprisingly high percentage of the questions wrong. 99th percentile does not mean we're getting them all right!
So I need to know, before I go in there, that whatever I can do, the test can give me something I can't do. And it will. When that happens, I can't let myself get so sucked in that I blow a bunch of time, or get so upset that I'm still thinking about it on the next question (or the one after that...). I just have to pick something and keep going. That means I have to (a) know how much time is too much time, so I don't go over, (b) have some other means of picking an answer, so I can put SOMEthing down and keep going, and (c) understand that this is how the scoring works so that I don't get totally flustered when I have to let a question go.
I guarantee that you'll have to let questions go during the test. HOW you choose to do that is up to you (and could have a significant impact on your score).
Please note: I do not use the Private Messaging system! I will not see any PMs that you send to me!!
Stacey Koprince
GMAT Instructor
Director of Online Community
Manhattan GMAT
Contributor to Beat The GMAT!
Learn more about me
