Avg Score Increase

[Guest]

What is the average GMAT score increase for someone taking the 9-session Manhattan GMAT prep course?

[RonPurewal]

That question is difficult, if not impossible, to answer, for a couple of reasons.

First off, a majority of our students take the 9-session course without having taken an official GMAT beforehand, thus rendering the idea of "score increase" meaningless. In fact, it's not unusual for a decent fraction of the class to raise their hands during session 1 when the instructor asks "Who here has never seen a single GMAT question before?"**

Second, and more importantly, it won't help you much if we answer this question with a statistic. Your particular score increase will depend to a large extent on your own effort and investment in the course: we expect a lot of you, just as you presumably expect a lot of us. The course involves a substantial amount of homework, most of which is self-graded, and without which you will derive much less utility from the course. Moreover, the same statistic would mean vastly different things to students with different profiles: a score increase of 100 points would quite ordinary for a student beginning the course at the 400-500 level, but remarkable for a student beginning in the mid-600's.

That said, our nine-session course has graduated many students who have gone on to phenomenal GMAT score increases (200+ points in some cases), and the vast majority improve their scores significantly. Our course is also, to put it simply, the best out there, and is taught by the best instructors in the industry.

**This shouldn't be taken to mean that we waste a lot of time on introductory things that would be redundant to students who've already taken the test. We DO walk students through the basics from scratch, for the benefit of those students who indeed have zero experience, but we do so very quickly so that we can dig into the meat of the course ASAP.

[Guest]

While I understand that not all students have taken the GMAT before they take the 9-session course and that results vary from person to person, I still would like to know the average test score increase from someone who has taken the 9-session course.

Manhattan GMAT must keep these statistics, but I haven’t been able to find them anywhere (note: a competitor boasts an average increase of 80-90 points). If you have them by score range from the first test (e.g., 550 - 650, 650+) that would be even better.

Thank you in advance for answering my question.

[StaceyKoprince]

We literally do not keep those statistics ourselves, for the simple reason that the only statistically valid way to make such a comparison is to calculate from an official test to an official test, with our course directly in between. Not enough people follow this formula to calculate a statistically significant number, and it's a waste of time to calculate with the few people who do follow this formula - a non-statistically-significant number does not have any valid meaning.

Most companies that do publish such numbers do so from that company's own test to an official test. Again, this number does not have any real meaning; they are not comparing apples to apples. In the past, some companies have even accused some other companies of providing very difficult first tests to depress initial scores and thereby drive up the "score improvement." We also don't want people to have to wonder whether we're playing those kinds of number games. (In any event, even if a company tries to mimic the real test as best it can, it still can't make it close enough for a statistically valid comparison - plus, students know it's not the real thing, so they are not affected by nerves or other issues that can impact your score.)

The statistically valid and statistically significant number we do publish is simply how our students do when they take the real test after our course. Our students' median score is about a 690, or about the 91st percentile; median (as I'm sure you know from your studies!) means that half our students score higher and half score lower on the official test.

You should also be aware that any company gathers its scores via self-reporting; that is, the students report their scores themselves. Higher-scoring students are more likely to report scores than lower-scoring students, so any published scores or score improvements you see (including ours) are likely a bit inflated from the true number due to this reporting bias. (And if any of our students are reading this and you haven't reported your scores yet, please ignore whether you liked your scores and email your teacher or the office at studentservices@manhattangmat.com - we really want this data from absolutely everyone!)

Moral of the story: beware someone who tries to sell you statistically suspect data and / or doesn't tell you of any inherent weaknesses (ie, reporting bias) and subsequent consequences (ie, inflated scores) in the collection or manipulation of that data.

[Guest]

what stacy said about the statistic numbers reminds me of a critical reasoning question with stats. ---skewed surveys and the question would be ---how you can weaken the passage. gosh...i've been doing too many of these problems....

[StaceyKoprince]

Lol! So right. I sometimes catch myself reading a newspaper article as though it's a CR question and reacting accordingly. The only time it's really bad is when you start doing it at work during someone's presentation or something! :)