GMAT Prep vs. MGMAT scoring methodology

[rajneeshjindia]

Hi
I am writing my GMAT in another 3 weeks and there's a little thing which has been puzzling me for a while.
I have written a few GMAT Prep and Manhattan practise exams.
I scored 660 on one of the GMAT Prep exams and also on one of the MGMAT exams.

Code: Select All Code

          Quant                          Verbal
         Correct    Incorrect       Correct    Incorrect
GMAT Prep  26        11                 32           9
MGMAT       20          17                21           20


What's puzzling me is that even though the number of questions I answered correctly in both the exams is very different, the overall score turns out to be the same.
Isn't it so that all questions carry equal weightage?
Any insights are greatly appreciated.
Thanks
Rajneesh

[leisuresoul.avi]

All question don't have equal marks. You start your section with easiest type with have lowest marks waitage then as you start giving right answer, the difficulty level also increase as well as marks/question and if u give wrong answer then a easier than previous question comes to which have less waitage. Thats what CAT is all about !

[rajneeshjindia]

Hi leisuresoul
Thanks for the reply. Even I'd imagine the same thing as you said but if you have a look at the scoring methodology of any of the GMAC's paper based tests, you'll see that each of the question is assigned an equal weightage, irrespective of its difficulty level and position in test.
Thanks
Rajneesh

[leisuresoul.avi]

The final test of Gmat would be a CAT, so you should prepare with those tests only which follows CAT methology. I am unaware about GMAC paper but i will be shocked if they don'thave CAT methology in practice test.

[rajneeshjindia]

The ones that I am talking about are the tests available on mba.com for purchase. Check out this link -->
http://www.mba.com/mba/store/ProductInfo?SKU=1103260X

Anyway, thanks for the reply. I guess you are right and I should just concentrate on getting other things straight rather than working out the scoring methodology.
Thanks
Rajneesh

[StaceyKoprince]

The paper-based tests use a completely different scoring methodology than the CATs. Do NOT base your understanding of the CAT algorithm on the paper tests! :)

As a general rule, CATs are not scored based on percentage correct. The algorithm is much more complex, and the "weighting" applied to a question depends upon many other factors. (It's not really a weighting - that's not what's happening on the test.)

Some things to keep in mind:
- the effective per-question penalty increases if you have a string of wrong answers in a row; the longer the string, the higher the per-question penalty
- the effective penalty for a wrong answer is higher if the question is below the level at which the test currently expects you to perform (based on your performance until that point); the lower the rating of the question, the higher the effective penalty if you get it wrong
- while the test does follow the "give a harder question if you get the previous question right" rule, the "harder question" is NOT just based on the previous question; eg, if the previous question was a 650 level, that doesn't necessarily mean the next one will be higher than 650. Rather, the test will give you something harder than your collective performance to that point. That last question may have been a 650, but your collective performance may be only a 600, so the next question will just be higher than 600, not 650. (And it doesn't even work exactly this way - the questions are not rated based upon the numbers I just gave, but the way they are rated is too complicated to explain in a forum post.)

Anyway, you're right that you shouldn't worry about this too much, but you should understand the basic fact that the test is NOT rated based upon percentage correct, so you should NOT be trying to get every problem right (the way you would on a paper-based test). That strategy will typically cause your score to be lower than it could be (because, when people try to get everything right, they spend too much time, and the penalty for running out of time or getting strings of questions wrong in a row is severe).