Scoring Accuracy?

[StaceyKoprince]

Hi, yes, answering those questions would take hours. I agree with Anish's advice that you figure out your personal strengths and weaknesses to a very great level of detail. If you're having trouble doing that on your own, it may be worth a few hours of private tutoring - an instructor can help you with something like that.

[victim masterchief]

Andrew (or staff)

I'm writing with your post in mind

"Our algorithm mimics the algorithm of the real GMAT but is not exactly the same, true. GMATPrep does use the real algorithm, but the pool of questions does not seem to be anywhere as deep as that of the real GMAT. The practical effect of this "shallowness" of the pool is that high performance deviates more from the theoretical 50-60% right that pure IRT (item response theory, the theory behind the algorithm of the GMAT which we understand quite well) would predict. In simpler terms, if the pool is shallow, the algorithm cannot give you "tough enough" questions while still meeting the constraints of content breadth and format breadth. Thus you would expect to get more right on a shallow-pool GMATPrep than you would on the deep-pool real GMAT. Our pool is deeper, which leads to a closer fit to the 50-60% right that pure IRT predicts."

So i just took the GMATPrep test and got a Q: 42 V:40 for a 660 :(

The last 3 MGMAT CATs were
Q43 V43 - 720
Q44 V44 - 730
Q44 V40 - 700

My GMAT Focus ranges were
44-49 and
41-48

So my performance on the GMATPrep seems to be low... almost abnormally so. In all my MGMAT CAT's I missed ~16-20 questions per section, but managed to get away with +700 scores, so I'm thinking missing 16-20 on the GMAT PREP is okay. I ended up missing 17 on the Math and 11 on the Verbal - sort of what I would aim for in a good scenerio. I don't think i've remotely approached only missing 11 on the verbal on a MGMAT test. You should also know that my ability to get 600-700 questions correct is probably statistically about the same as my ability to get 700-800 level questions correct [this seems backwards i know].

*I'm HOPING that my poor performance on this is because of the "shallow pool" of questions on the GMAT Prep. In other words because there weren't enough 'tough' questions on the GMAT Prep to offset the volume of questions i got wrong, therefore i had a lower score. Please reasure me that this is the case!?

Also on my GMAT focus i missed roughly half of the questions and managed to get the ranges listed above.

My goal is 700.

Please let me know what to do. Come test day should i mimic my strategy for MGMAT tests i.e. it's okay to miss 16-20 questions knowing I got a decent # of 700+ questions correct? OR should should i allow myself to only miss 8-13 questions knowing I'm not going to have a chance to hit hard enough questions to make up the volume of missed questions. i.e. if i'm only gettin 55~60% of the questions correct, I'm not even going to see enough tough questions to make up for my score?

THANKS !

[StaceyKoprince]

Not abnormally so - that's all well within the standard range you'd expect. Adaptive tests are not anywhere near as precise as people generally assume they are. There are so many reasons for scores to fluctuate up or down on any given day. The general standard deviation. Energy levels. Happening to get a few more questions in your areas of strength vs. your areas of weakness. Making more careless errors than usual for whatever reason. Etc.

The "shallow pool" thing should not automatically cause a deflation in scores - people can score very well on GMATPrep as well. The algorithm adjusts for the fact that the pool is more shallow. More likely, there was a confluence of factors - you weren't quite as "on" as you normally are, you got some "unlucky" questions (more things in your areas of weakness), and you had some bad strings (multiple wrong answers in a row). When that last one happens - especially towards the middle or end - it can really hurt you.

The question about how many to "let yourself" miss - you can't actually control that during the test. Do the best you can within 2 minutes and that's about all you can do. Generally speaking, at a 700 level, it is the case on the official test that you only need to get about 60% of the questions right - but they have to be the right questions, if you know what I mean. If you make even a few more careless errors on sub-700 level questions than you normally make, it'll be tough to hit your goal. So make sure that you give full attention to the problems you think are either easy or right in your range. Don't get so sucked into problems that are harder than your range that you end up having to move too quickly on questions you should be getting right. That's how you can really tank your score quickly.

Go back and look through that GMATPrep in more detail. I know that they don't give us difficulty level data, so it's hard to interpret fully, but see whether you were making some mistakes you didn't normally make, examine the pattern of right and wrong answers for strings of wrong answers, etc.

[masterchief]

Stacy,

Thanks- you were right, the questions were a lot shorter than the MGAT questions so i probably just mentally dismissed the legitimacy of the test as a whole ; looking back made a couple of really silly mistakes that were probably just a function of that + actually waking up at 8:00AM.

I just thought the verbal was bizarre bc I didn't miss a single reading question which never happens in MGMAT land + got 10 out of the last 11 questions right which also never happens in MGMAT land - so I'm thinking why didn't it just give me tougher questions?

and by 'letting myself miss' i really mean 'give up on after 60 seconds (knowing it'll take 120 to solve if it took me that long to figure out where to begin), eliminate as many choices and guess without worrying too much' I just pick my battles with the toughies, which works out pretty well on MGMAT CATs.

My question is couldn't they just make the test longer to reduce variability? 24 questions (13 are fake) is a pretty small sample size. I'm sure people wouldn't mind a 20 minute extension on each section if that allowed for greater visibility on their CAT

[Masterchief Victim]

Just took a MGMAT - Q47 V41 - 720

I tested the 'energy level' theory this week by eating right- I also started going to bed waay earlier and sleeping longer so that by 8:00, the time I take the test , I'm actually awake. Obviously this experience shouldn't be representative of anything, but here are 2 facts : (1) i felt better able to battle through tough questions and (2) got a quant score 2 pts higher than ever before - and this is only a week after my 42 so my change in 'abilities' can't be responsbile for the +5 in raw

Let's hope the difference is in energy levels at time of test! Wish me luck!

[StaceyKoprince]

Nice!

[lastavalon]

Hi,

I read this thread a week before taking the real gmat today, so now I've taken it, I think I should share my scores too. I'm a little disappointed but it's fairly in line with the simulations I took (below though, grr). I think my score should make it for the INSEAD MBA - I will now focus on the other parts of the application.

So I took all the following in 7 days (I know, it's no good but did not have the choice), and real GMAT on the 9th day.

GMAT Prep 1 710 (forgot to write down Q/V)
MGMAT Free 670 Q45 V36
MGMAT 2 700 Q43 V41
MGMAT 3 700 Q44 V41
GMAT Prep 2 730 Q49 V40 (then went out and got seriously drunk)
MGMAT 4 760 Q49 V45 (seems like alcohol seriously inflated my V score)
MGMAT 5 700 Q46 V39 (should I re-drink?)
I skipped MGMAT 6, decided to rest.

Real GMAT 690 (88%), Q47 (77%) and V38 (83%)

Also, I took the GMAT 6 months ago and got a 610 (68%/58%). On the basis of my 2-tests experience, I'd really say the real test is harder than both GMATPrep and MGMAT. But when I look at the score though, this all seem fairly aligned - so I guess it is all about the pressure of the exam...

[bThe key is really: Practice the test [/b]itself as much as you can. Of course you've got to use exercises books, but to top it well, I would advise anyone to take about 10 simulations minimum before going to the real one. That's the best way to avoid surprises.

In my case, I really don't think I have progressed so much between my first shot and the second. The difference is 6 months doing nothing, 2 weeks reopening exercises books... And a few simulations...that's it...

Thanks to the Manhanttan community for creating these tests, they are great and really helped me. I only wish I found them 6 months ago, I'm sure I would have got a better score at my first shot.

Cheers guys

[StaceyKoprince]

lol at your drinking theory. :)

Congrats on your score and your score improvement! I agree that it was in line with your practice CATS. Yes, I know it would have been nice to see 700 instead of 690, but it really is only a 10-point difference (that is, not statistically significant). It's no different than a 680 vs. 690. I know the psychological impact of seeing a 7 at the beginning instead of a 6, but really - 10 or 20 points won't be the big difference. The average score of INSEAD students (according to their admissions director ) is a 700, so you're right in the mix.

What will make the difference is what you do with your application! Good luck with it - let us know how it goes!