How close can Manhattan GMAT predict real GMAT score?

[DucD465]

Hi everyone,

I am planning to take the GMAT in two weeks. I aim to get a 750 because (1) I did sample questions in the Official guide and I felt I can get a 750, and (2) I need a high GMAT score in order to be competitive for some scholarships.

Anyway, I took my 1st computer sample test (Manhattan CAT) last week and got a 650 (Q: 42, V: 37, IR: 6.53). I was disappointed but I thought one of the reasons was that I was not prepared to do the exam on a computer and therefore, I did much slower and skipped many questions at the end. I only practiced on paper in the past.

Today, I took the 2nd CAT (there was no practice since my 1st CAT) and I felt quite well after finishing it. I tried to answer each question as fast as possible. For example, when I thought C was right, I immediately selected it and moved on without even reading D and E options. In my 1st CAT, I would read all options, and reevaluate them a second time. I got 780 (Q:51. V:45, IR: 6.6). Of course, I am happy with the score but I also don't really know how accurate this score is. Since I moved really fast, I felt a higher level of uncertainty about my score. I don't know whether I was just lucky with my 2nd CAT.

My questions:
- How accurate is Manhattan GMAT score as a prediction for the real score?
- Is a 780 fluctuating to below 750 normal? To rephrase the question: If a person's "real" score is below 750, is it very likely that he can earn a 780 by luck? I am not much familiar with GMAT scoring and don't really know how significant can 20 or 30 points mean.

[StaceyKoprince]

Our practice CATs have a standard deviation of about 50 points (the real test has a standard deviation of about 30 points).

What this means: if you were somehow able to take the real test two days in a row under "identical" conditions (meaning, you're not tired out on the second day! and you don't study / learn more between the two tests), then your score on the second day would likely be within + or - 30 points of your score on the first day (though no guarantees, of course). The same would be true for our test, except the likely range would be + or - 50 points.

There is a difference between a 750 and a 780, but not an enormous one - even on the real test, that is not considered a statistically significant difference.

Here's the good news: you earned that 780. It would be like winning the lottery to get THAT lucky. :) Yes, there's some variation, so you're not guaranteed the same score the next time you take a test (whether practice or real), but you are likely well positioned to get a great score on the test.

Solidify the result: try GMATPrep next.

Oh! One caveat. Did you do the essay and IR sections? And did you take them seriously? If not, make sure to do that on GMATPrep. Those sections can tire you out and that can negatively affect your Q and V scores.

I would also recommend checking all 5 choices on verbal. Sometimes you do see something that looks good, but a later answer is still better. Just run through them all. If you're confident, you don't need to double-check.

Let us know how GMATPrep goes!

[DucD465]

It's been a while but I think my data can help other people who take the test in the future. I took the GMAT 2 weeks after my last post and got a 770! I did take another CAT before the real GMAT and got 750. I also took a GMATPrep CAT and got 750 as well.

[StaceyKoprince]

Excellent - thanks for letting us know. Congratulations and good luck with applications!