GMAT Strategy

Hi Experts
Please help
I took 2 mgmqt tests and 2 veritas tests and 1 gmat prep test .

My scores are varying from 750 to 670 .I have taken 4 CATs so far .
I am unable to put my finger on the reason for such a huge variation . I am like 2 weeks away from my Test .
At this point of time it seems as if the score depends on a matter of chance as in the combination of quant and verbal scores .
Could you please help me find out the reason for such a huge variation .
Has someone else also experienced such fickleness in scores . Its getting unpredictable . there is difference of +80 between my gmatprep score and veritas score. There is a difference of 50 between my 2 mgmat cats . There is a difference of 70 between my 2 veritas CATS .Dont know how to analyse this .This is getting ridiculous .There is a difference of 70 between my highest veritas score and my lowest MGMAT Score . : :twisted: :

Hello Ramneet,
What score are you aiming for?
My practice test scores were all over the place -- from 480 to 730. My first actual GMAT attempt landed a 690 (49,35) -- messed up on the verbal.
Read these for some more info:
https://www.gmathacks.com/resources/inte ... score.html.
https://www.beatthegmat.com/mba/2010/02/ ... est-scores
https://www.beatthegmat.com/mba/2009/11/ ... ength-test
https://www.beatthegmat.com/mba/2009/09/ ... tice-tests
https://www.beatthegmat.com/mba/2009/10/ ... art-2-of-2

There are a few things to consider:
1. Some companies' tests scores are lower than what you can actually get because they want you to then sign up for their courses.
2. There is a natural variation in the test for a single test taker.
3. Companies' scores are calculated differently, so 2 different companies' scores might not actually comparable.

What I'd recommend:
1. Review the exams -- using the information in the articles above.
2. If you're like me and are doing better on Quant than Verbal, focus on maximizing your Verbal for the highest possible total score. It has been shown that the higher the verbal, the better your overall score: https://www.manhattangmat.com/articles/break-700.cfm.

Use these two weeks to focus on your BIGGEST WEAKNESSES in Quant, and any WEAKNESSES in Verbal. Be sure to pace yourself as you practice. 2 days before your exam, take the last GMAT Prep test -- fully, including the AWA section. That will be the most accurate indicator of your result on test day.

All the best!

--Rishi

Hi Rish8i could you explain
what u meant by

There is a natural variation in the test for a single test taker.

Hi Ramneet,

I took that information from the GMAT Hacks website (https://www.gmathacks.com/resources/inte ... score.html):

There's a natural variation in test scores, due in part to luck, and in part to the questions that pop up on your test. This variation might be as much as 60-80 points. If you took the test three days in row with no additional studying, you might get a 600, a 570, and a 630, simply because of good or bad luck.

Also, just as in anything we do, there is always a natural variation -- I learnt about this in Operations Management courses during Undergrad in the context of a manufacturing environment. One example I remember is about a pencil sharpener -- it can sharpen a pencil with the tip having a width of 5mm on average. This means that some pencils will be 5mm, some will be less, and some will be more.
Let's pretend the requirement is to have all of them sharpened to exactly 5mm, but 4mm and 6mm are also acceptable. Then we would be interested in the standard deviation, how many fall within this range, and how many fall outside of this range. Companies with high quality control are able to keep 99.99% of the products within the acceptable range, which is six standard deviations from the mean (both above and below).
(On a side note, the standard deviation is represented by the Greek letter Sigma "σ". Having an accuracy of 99.99% means that the variation is so small that 3.4 defects per million fall outside of six standard deviations from the mean. This is where the term Six Sigma comes from).
Sorry for the tangent!

So taking the test multiple times will lead to variations in the score, which can be caused by careless mistakes, calculation errors, weakening an argument instead of strengthening it, or as the GMAT Hacks author puts it, bad luck. All we need to do as test takers is to try to maximize that score by using the skills we learn while studying and practicing.

Hope this helps!
--Rishi

rishimaharaj wrote:Hi Ramneet,

I took that information from the GMAT Hacks website (https://www.gmathacks.com/resources/inte ... score.html):

There's a natural variation in test scores, due in part to luck, and in part to the questions that pop up on your test. This variation might be as much as 60-80 points. If you took the test three days in row with no additional studying, you might get a 600, a 570, and a 630, simply because of good or bad luck.

Also, just as in anything we do, there is always a natural variation -- I learnt about this in Operations Management courses during Undergrad in the context of a manufacturing environment. One example I remember is about a pencil sharpener -- it can sharpen a pencil with the tip having a width of 5mm on average. This means that some pencils will be 5mm, some will be less, and some will be more.
Let's pretend the requirement is to have all of them sharpened to exactly 5mm, but 4mm and 6mm are also acceptable. Then we would be interested in the standard deviation, how many fall within this range, and how many fall outside of this range. Companies with high quality control are able to keep 99.99% of the products within the acceptable range, which is six standard deviations from the mean (both above and below).
(On a side note, the standard deviation is represented by the Greek letter Sigma "σ". Having an accuracy of 99.99% means that the variation is so small that 3.4 defects per million fall outside of six standard deviations from the mean. This is where the term Six Sigma comes from).
Sorry for the tangent!

So taking the test multiple times will lead to variations in the score, which can be caused by careless mistakes, calculation errors, weakening an argument instead of strengthening it, or as the GMAT Hacks author puts it, bad luck. All we need to do as test takers is to try to maximize that score by using the skills we learn while studying and practicing.

Hope this helps!
--Rishi

so the test might as well depend on the day and your chances of seeing the right questions that you would like to see . Its just a matter of chance whether you come across an Rc that is followed by questions that you might be comfortable or uncomfortable with

Hi I dont know how can scores vary by 80 and still be within the standard deviation

According to GmatHacks , You cant trust anything . LOL

@rishi thanks for the great comment.

@ramneet I am in the same boat as you. Read my post. https://www.beatthegmat.com/understandin ... 91697.html I am also looking for answers.

Hello Ramneet & Prodizy,
Don't be to hard on yourselves. To make you feel better, take a look at my crappy mock exam history:
Test -- Quant -- Verbal -- Overall Score
Kaplan GMAT Exam 1 -- 19 -- 31 -- 480
GMATPrep Exam 1 -- 25 -- 33 -- 490
GMATPrep Exam 1 -- 47 -- 33 -- 650
GMATPrep Exam 1 -- 43 -- 40 -- 680
Veritas Exam 1 -- 45 -- 32 -- 620
Manhattan GMAT Exam 1 -- 47 -- 41 -- 720
Manhattan GMAT Exam 1 -- 46 -- 41 -- 710
Princeton Review Exam 1 -- 33 -- 32 -- 550
Kaplan GMAT Exam 2 -- 37 -- 29 -- 560
Kaplan GMAT Exam 3 -- 34 -- 36 -- 590
Kaplan GMAT Paper -- 47 -- 38 -- 650
Manhattan GMAT Exam 2 40 -- 37 -- 640
Manhattan GMAT Exam 3 -- 48 -- 42 -- 730
Manhattan GMAT Exam 4 -- 47 -- 42 -- 720
GMATPrep Exam 2 -- 48 -- 38 -- 700
Actual GMAT Exam 1 -- 49 -- 35 -- 690

It's all over the place. The Kaplan, Princeton, and Veritas Prep ones are all really low, when compared with some of the MGMAT and GMAT Prep ones. The way I took the first few exams were to do the full thing, including the essays. I got the hang of that and found that I had the stamina for it, so towards the end of the first round of prep, I would only jot down my ideas for the AWA essays (about 6-7 minutes per esssay), and quickly continue on with the rest of the test (to create more time for review while not forgoing any essay practice).

So in my example above:
The Quant mean is 40.9375
The Verbal mean is 36.25
The Overall mean is 636.25

The Quant standard deviation is 8.73547902235476
The Verbal standard deviation is 4.11551940828858
The Overall standard deviation is 78.6507310836969


65% of my test scores are between 557.59 and 714.90073 -- a range of 157 points.

I'm sure that if I take more tests, the mean will rise and the standard deviation will shrink.

So to recap, in the next two weeks, you should do a lot of review.
Take a look at some great advice from Ron in this post:
https://www.beatthegmat.com/550-to-720-i ... tml#330724

Hope all this helps somehow, at least with boosting your confidence!
All the best!

--Rishi

great stuff from rishimaharaj!

It is interesting to see on the same day some people who think that Veritas tests score really low and another who got the highest score from these same tests.

DO take note that rishimaraj's GMATPrep score right before the test was within 10 points of the actual score. So this is the test with the algorithm used on test day.

Luck certainly does factor in a bit - the GMAT indicates that this allows a variation of 30 points above or below.

It is true that 80 points is not within a standard deviation! But 30 points above or below is 60 points of variance...

Focus on the types of questions you missed rather than on the scores. When you get your score on test day there will be no explanation at all. Just a score and hopefully the one that you wanted.

Hi David or any other expert , I am considering postponing the exam by 18 more days .This will give me 4 weeks to study from today . The next date available is only after further 10 days ,that is 28 days .
My scores are very erratic and i cant safely predict about them .Also i dont wish to go into the exam having the wrong idea about my abilities and end up badly .
Could you please advise would 4 weeks be sufficient enough time to get a handle on prospective score .
I am a non native speaker . My RC timings aren't really good . I take 10 minutes for an RC .
I manage a 2.5 mints for a CR and 2 minutes for SC
Thanks a Lot

David@VeritasPrep wrote:great stuff from rishimaharaj!

It is interesting to see on the same day some people who think that Veritas tests score really low and another who got the highest score from these same tests.

DO take note that rishimaraj's GMATPrep score right before the test was within 10 points of the actual score. So this is the test with the algorithm used on test day.

Luck certainly does factor in a bit - the GMAT indicates that this allows a variation of 30 points above or below.

It is true that 80 points is not within a standard deviation! But 30 points above or below is 60 points of variance...

Focus on the types of questions you missed rather than on the scores. When you get your score on test day there will be no explanation at all. Just a score and hopefully the one that you wanted.

I think one should take the final gmatprep 8 days before the exam so that he is able to postpone the deal if its not good

4 weeks is actually enough time to do quite a bit of preparation. I would say that you CR time is not off by too much although the sentence correction will need to get a little quicker.

When you say 10 minutes for a reading comp - if that is one with 5 questions you are right on. If it is 4 questions that needs to pick up a little bit as well.

I don't know everything about your circumstances so I cannot help you with your decision, but I would say that there is a good chance that you can get over 700 even if your took the test now given that your scores have varied from 670 on up.

There is a danger in waiting too long and putting too much weight on a single test administration as well. Brian Galvin here at Veritas wrote about when students think that the test that they are about to take is "THE GMAT" when in reality it can be retaken and so less pressure will be there if you think of it is "A GMAT" with others to follow if necessary.

My advice: Pick a date, do your best, study hard and get in there and see what you've got!

Hey guys,

I got a PM about this and wanted to weigh in. Unsurprisingly, I agree with everything David says (especially those parts when he cites things I've already said...).

Another pretty major thought on this:

Don't rely on practice test scores to predict your final score or let you know when you're ready. Those scores are only one factor in your development, and consider:

-The official GMAT maintains that it's accurate within 20 points every time, but that still allows for a 40-point range if you were to perform about the same on consecutive days. (e.g., your true score is 680 but you could go 660 and 700).

-So even schools are more concerned with your ability "range" than your actual score. They know that a 720 is "in the top 10%" but they won't see that 720 as significantly "better" than a 710.

-And those GMAT scores are based on GMAC's extensive budget for R&D. The primary aim of GMAC is to create validity of its scores. It researches each item on the test significantly; it conducts and reports ongoing "validity studies" of how well its scores correlate to success in b-school. It spends around $3,000 on the development and continued monitoring of each question. The Veritas/MGMAT/Knewton/Kaplan tests simply can't be as good. Our primary aims are (or at least should be) to teach, so it's not at all surprising that our best efforts on the practice tests aren't going to be as reliable or precise as GMAC's.


-So... Practice test scores are okay at telling you your general direction (a high-600s practice test score should mean that you're into the 600s or 700s) but what's much more important is your own qualitative assessment of how you've done:

--Are you pacing yourself well enough to feel comfortable with the official test
--Are you aware of the types of mistakes you're making under pressure and are you ready to look out for those?
--Do you feel like you have any glaring weaknesses with a particular question type that you need to address?
--Are you comfortable with your ability to recognize question types and get to work on them?
--Ultimately, do you feel like you're ready to go? Even if the number on the printout is 30-40 points less than what you'd like or your scores are ranging by 60-70 points, do you believe that you're ready?

If so, take the test. Practice test scores are approximations; practice test experiences are more reliable or at least help to round out what you learn about yourself.

I've used this analogy before: if Spain's soccer team plays Italy in a friendly, Spain isn't looking at the final score as even a primary indicator of how it performed. It's looking at how the younger players meshed with the veterans; how the team reacted to different strategic tactics; how the midfielders held up physically playing a more aggressive (and tiring) style; how the defenders reacted when Italy sent more attackers downfield. The score may be 2-1, but the coach had better learn a whole lot more than "2-1" from that experience. The same is true for you - if all you take away from your practice test is "660", you just wasted 3 hours. If you instead take away notes on how you're pacing yourself, which question types tend to bait you into trap answers, which subjects you need to shore up, etc., you have a good idea of where to go and whether you're ready for test day.

Thank You Brian