GMAT Strategy

Most high-level students report timing issues on the real test even when they didn't have them on practice tests. The biggest reason seems to be that people don't want to "let go" when they hit a question that's really too hard for them - instead, they spend too much time, get behind, and have to play catch up or guess at the end. (Even if they were able to "let go" on practice tests, during the real thing, they panic and revert to "school math mode" which tells us to try our best on every problem.) Net result: the score is not as high as it should be.

We want to make sure that our high level quant questions really push our students to deal with this issue BEFORE they get to the real test. What do you do when you hit something you can't do? How do you make sure you maintain your pacing throughout the section? What do you do if you get behind? You need to have answered those questions before you take the real thing, and our higher-level questions force you to deal with the issue so that you can make those adjustments successfully.

Thanks - we really appreciate your feedback. We're constantly looking for ways to improve and can use all the help we can get!

I forgot to add one other thing that is tougher to get around and does affect your perception if you are a higher-level math person. The official test inserts experimental questions into the mix, so if you are high-level, you essentially get little "breaks" here and there, where you see easier questions. These questions aren't actually a part of your scoring algorithm though - they don't count at all. But you do get that mental break for a couple of minutes.

We didn't want to artificially adjust the scoring algorithm just to give people these breaks - because then it would actually alter the score you receive and not be as comparable to the GMAT. We're working on building an experimental question set-up into our tests so that we mimic the real thing even more, but that probably won't be live for quite some time. It's pretty complex.

That standard deviation range I listed was specifically for a score of 650.

So, on the official test, the standard deviation is about 30 points. That means - again, on the official test - that if you get any particular score, you have a 2/3 chance of scoring within 30 points of that score if you take it again and a 1/3 chance of scoring outside of that range. So if you score, say, a 650, you have a 2/3 chance of scoring between a 620 and a 680 if you took it again the next day (which you can't do in actuality), and a 1/3 chance of scoring below a 620 or above a 680.

On our test, the standard deviation is about 50, not 30, so the range is a bit wider. It's not that much wider than the official test's SD though. (And, again, any non-official test is going to have an SD that is wider than the official test - it's not possible to be more accurate than the real thing.)