Beat The GMAT a GMAT & MBA community

Brilliant deductions Ian!
I had my doubts about this experiment and was about to comment about it. Your points cover pretty much all aspects!

But there was one point in the thread that appealed to me. There was a discussion about how SC and CR concepts can be 'learned' through preparation but RC skills are pretty much innate. Therefore, it does make sense to place a slightly (not much mind you) higher weight on RC questions.

What are your thoughts?

Thanks David!
My question wasn't aimed at Ian alone...

So you say preparation is the key, anywhere in the GMAT. That's still under my control, so there's hope!

interesting outcomes .

@David@VeritasPrep: I have been following VeritasPrep RC guide and ,undisputedly, it has taught me quite a few things. I could also see improvement in performance especially the way questions expect you to be careful. A thing I wanted to bring to your notice -- that i wrote to Ron also in one of his RC post -- is what is the best way to deal with the passages that use very complicated language (a few GMATPrep RCs) sometimes a sentence makes a paragraph, and that are highly normalized making them densest of densest . Of course, some of SC skills do help but still they are monsters . I can put some examples or screen shots of GMATPRep RCs if I have not effectively communicated my idea.

Looking forward to your ideas. IF you want to consider an article, I am sure many students will get benefited.

Thank you Ian for great insights.
In one of my GMAT training classes, I was told by the instructor that my score would also be influenced (percentile) by how others examinees' performance during that day in that test hall. How much should one worry about that?

_________________
Thank you,
Vj

Huh! People never fail to amaze me! And these guys are training others for GMAT. I know this sounds rude but your coaches are either charlatans or nincompoops. You should ask for your money back first thing tomorrow morning.

_________________
scio me nihil scire

@ David & @ Ian,

Thanks for the inputs. This study has given us a lot of inputs. But based on the inputs making "wild" deductions is NOT LOGICAL. i agree with you.

Still, it depends upon the reader to take the right call!!

_________________
R&D on GMAT algorithm:

http://www.beatthegmat.com/r-d-on-gmat-prep-t67692-15.html#305739

Analyzing scores you guys posted here I wanna share some results with you.

I gathered all your scores in a single table and calculated following statistics:

Average = 1.186394156 points per each correct answer
Stdev = 0.105248529

Conf.level
-99% 0.941165082 points per each correct answer
99% 1.43162323 points per each correct answer

It means that in the worst case scenario you will get 0.941 points for each correct answer, in realistic scenario 1.186 points per correct answer and in the best case scenario 1.4316 points per correct answer.


I do not now how it can be useful for you... just wanted to share

although central tendency measures are good indicators of the population and data distributed normally, they are usually bad indicators of abnormally distributed data sets
since we can't prove 'H(hypothesis): normal distribution true' for GMAT CAT data, our descriptive stats is of little use, correct me if i'm wrong

_________________
Success doesn't come overnight!

I attach some testing results for normal distribution

here's my Prep1 verbal.. I got 37
No mistakes in RC, 3 mistakes in CR and 7 mistakes in SC

I just stumbled on this post, and just have a few things that I also looked in to on my own. The tests I ran had to do with the effect of missing questions early in the test. There is this myth floating around(still not dead) that the first few questions are more important than the rest. This is simply not true. As the GMAT writers say in the Official Guide, the first few questions are used to obtain an initial estimate of your ability, but that is pretty much what those questions do. From that point onward the student has to perform consistently to score high.

I ran these experiments on the Quant section:

Answer first 5 questions incorrectly and then answer the remaining 32 correctly.
Result: A final score of either Q49 or Q50(90 or 95%). I did 6 repeats. There is no way anyone can get a 99% with this performance, because this means you did miss some of the easier questions in the beginning, and those take a greater toll on your score than hard questions.

After missing the first 5 questions, the questions from 6 onwards were easy but the difficulty ramped up to the highest level by Question#12 or so. This means not all is lost in case you bomb the beginning of the test. This is important from a psychological point of view for the student.

At the end of the test the last five questions were easy, I call the last 7 questions of the GMAT the confirmation stage, this is where the algorithm is trying to confirm your score. Meaning if your last 7 questions are hard, then you did pretty well on the test. On actual GMAT my last 7 questions are almost always the hardest on the GMAT, and some of those I struggle to complete in 2 minutes, I tend to leave about 3 min per question for the last 7, if I can. Now in the tests I ran, the last 5 or 6 questions were easy, this is because the software needs to confirm if the first 5 questions I missed was by mistake or I really didn't know the concept.

The point of all this is that you just have to do your best in the beginning and even if you see a very difficult question in the beginning, it is okay if you let it go, in fact for many students this may be extremely important.

And finally a speculation on my part, the 9 experimental questions in the Quant are likely dispersed from Question#10 to 30, because I rarely see a random fluctuation on difficulty level of the questions on the GMAT in the beginning or the end of the test. However, I have seen extremely easy questions on the GMAT in the middle of the test, where they don't follow the difficulty pattern of the test. Let me repeat, this is purely a speculation on my part. Of course as a test writer I would disperse the questions randomly and throughout the test.

Enjoy,
Dabral

_________________
Free Video Explanations: OFFICIAL GUIDE GMAT 13, 12, 11, 10; QUANT REVIEW 2nd, 1st.

Most conservative estimate of GMAT score based on accuracy of GMAT prep
-----------------------------------------------------------------------------------------

Verbal( As it is )
#mistakes score
0-1 = 51
2-4 = 50-45
5-7 = 40-44
8-13= 39-35

Quant(Conservative)
#mistakes score
0-1 = 51
2-4 = 50
5-7 = 49
8-13= 48-45

Conditions:-
1. Never get more than 2 consecutive wrongs.Concecutive wrongs will get your score more lower than mentioned above.
2. First 3 questions on each section are correct.
3. Verbal score is completely skewed especially at 40+.At 40+ range every single question counts and can make or break your score at big level.Suppose 2 persons both got 50 in quants, but one got 770 and other guy got say 720 (overall score),then mere difference between them is 3 or 4 verbal questions correct.
4. If you get first few questions wrong it is very difficult to cross 770+ even if you get all remaining questions correct. However you can cross 740+ even with first few questions( 1 or 2) wrong with ease . Result is insignificant whether you made mistake first or last when compared to number of mistakes provided they are not concecutive.
5. You attempted all questions.


These are my few cents. Experts has to add.