Practice Scores vs. Official Scores

[moodie316]

Hi guys,
During your years of tutoring, what’s the typical range that you’ve seen when students report differences between their scores for practice tests and official exams? I ask this because I sat for the GMAT on June 5th, and I scored an extremely disappointing 600 (Quant - 43, Verbal 30). I scored considerably higher on the practice exams, so I certainly didn’t expect that kind of drop. After composing myself, I took a GMATPrep practice test (I only took MGMAT tests before) and scored a 660 (Quant - 44, Verbal 36). Overall, my history is as follows:

2/11/2009 - MGMAT CAT 1: 640 (Quant - 44, Verbal 34)
4/18/2009 - MGMAT CAT 2: 650 (Quant - 43, Verbal 36)
5/11/2009 - MGMAT CAT 3: 660 (Quant - 42, Verbal 38)
5/17/2009 - MGMAT CAT 4: 670 (Quant - 43, Verbal 38)
6/05/2009 - Official Exam: 600 (Quant - 43, Verbal 30)

6/06/2009 to 6/13/2009: a full week of sobbing

6/14/2009 - GMATPrep 1: 660 (Quant - 44, Verbal 36)

Evidently, I bombed the verbal section royally. My target score was a measly 650, so the 600 that popped up was a mystery to me. I didn’t feel particularly good about the exam after answering the very last question, but I didn’t feel good about any of my practice exams either. My pacing was decent and my guessing was strategic (at least I thought it was). Is this sort of divergence typical, and should I view the GMATPrep test as an indicator that I can do considerably better?

- Oh, and I guess I should provide some background info as well. Two weeks before the exam, I thought to myself "OK buddy-boy, your verbal is somewhat respectable...just get that quant score up and you should be in good shape). So I bought the OG 12th Edition and hammered the newer number properties and word translation problems like it was a part-time job. Could that shift in focus have been my downfall?

In addition, I took a MGMAT online course (Live Online Winter H 2009...J.R. and Kara are wonderful!), but given my work schedule, I had trouble keeping up with the classes and homework. I didn’t want to miss any important points, so I used the excel-based study organizer to complete the assignments over an extended period of time. As a result, I watched the online recordings and finished the course 2-3 weeks after the last live class. Was this a mistake? Did it take too long for me to study? I only have access to the online materials for a few more weeks...should I consider retaking the course or go it alone?

Sorry for the long post...I guess I had more questions than I originally thought =/

[jeffboaz]

Moodie - I believe Stacey said the standard deviation between their tests and the official is about 50.

And I believe she also said the standard deviation between the GMATPrep and the official is about 30.

[StaceyKoprince]

First, call / email the office right away and ask for a Post-Exam Assessment (PEA). This is a free service for our class students - you get a free phone call with an instructor to debrief from your test experience and figure out what to do from here. Your online material access also gets extended for 3 months for free. Call 800.576.GMAT or email studentservices@manhattangmat.com for details.

Yes re: the standard deviation, although that's not quite the right way to explain the GMAC figure (between different kinds of tests).

The number published by GMAC (30 points) is what's called a "within-student" standard deviation. This means that, if a student could somehow take a bunch of official tests right in a row, with no prep in between (and no fatigue!), the collective scores would have a standard deviation of about 30 points. The equivalent figure ("within-student" SD) for our tests is about 50 points.

The SD between our students' last test and the official test is also about 50 points (this is a different figure than the 50 points in the previous paragraph - it just happens that both values are about the same).

Anyway, so, yes, your quant score stayed very steady but your verbal score really dropped. It's important to try to figure out why - you definitely want to discuss that during your PEA. Here are some things to think about / discuss (either here or with the teacher assigned to you for your PEA):

When you took the practice tests, did you also take the essays? Did you spend as much time and mental energy on the practice essays as you did on the real essays?

How was your timing in the verbal section? Did you generally move steadily through the test, giving appropriate time and attention to each question? (1 to 1.5m for SC, 2m for CR, 2-4m to read a passage, 1m for general RC questions, 1.5 to 2m for specific RC questions) Or did you have to rush at times and possibly make random guesses? If you did have to rush and/or make random guesses, on how many questions would you say you did that? Did you do it on a lot of questions in a row or were the guesses scattered? Alternatively, did you move too quickly and finish with a lot of time (>3min) left over?

How was your stamina? How did you feel toward the middle and end of the verbal section? Did you have something to eat and drink on the breaks? Did you get up, walk around, and stretch?

What other differences can you think of between your practice test experiences and your real test experiences? Anything, no matter how small, and no matter whether you think it wouldn't have made a difference to your score. Any differences at all?

[thomaspham259]

I've been reading a few threads regarding possible discrepancies from their MGMAT and their actual scores, and most of these discrepancies arise from their VERBAL, not their QUANT scores.

Is that the case most of you guys have seen with this deviation?

Is the quant sections of the practice test a better predictor of the actual GMAT than the verbal?

[StaceyKoprince]

They are equally predictive (quant and verbal). We calibrate both every two months to make it as close as we can get it, and there is not more of a skew on quant or verbal.

We base our stats on our entire population of test-takers. If you look at only a few data points (as people are doing when they post a thread and a few people discuss their experiences), then you will likely see some kind of skew. But that's not because the skew exists overall - it's because the data set is too narrow. :)