Shocked beyond words...

[mj12g]

Right now I am, shocked, horrified, and ashamed. I just finished taking the GMAT about a half hour ago, and I scored a 700 (44Q/41V). While this may appear to be a decent score, the last time I took the test in September 2009 I scored a 720 (40Q/50V). This time I felt fully prepared: I wasn't hungry, nervous, panicked, tired, cranky, sick, or any other adjective you can think of. I had been scoring well on my practice tests, and I never went below 45 in verbal and also never went below 44 in quant, although that 44 in quant was a single instance; in every other practice test I was scoring 48-50. My actual score in verbal was particularly shocking, as I have never, and I mean NEVER, scored so low on verbal, right back to my very first diagnostic test in 2009. My verbal score is so low that I'm considering calling Pearson and asking if they got my score screwed up somehow. It's that low.

At this point, I don't know what to do. I applied to business schools last year and didn't have much success. This year, my plan was to retake, get a 740 or above, and shoot for some better schools in the first round. Now I'm just lost. I KNOW I can beat this test. In my studying I took the GMATPrep 3 times, scoring 740, 770, and 780. I also took 4 Manhattan GMAT CATs and scored 720, 750, 770, and 740. I'm not really looking for advice, I'm just venting, but now I need to take this test again to prove my mettle.

[prodizy]

I can understand how you feel, but there's nothing to be ashamed of when you gave 100 percent. obviously something's wrong. find out what is it and rock the gmat this time. all the best.

[prateek_guy2004]

I agree with prodizy, never be ashamed...Whatever you have scored its your hard work..

Gmat is a just a part of application...Dont take it That hard mate!

[mj12g]

While I realize that people can be shocked by low scores, I'm still confused by what seems to be a highly improbable statistical anomaly in my verbal score. If I take the mean of all my verbal scores, it cannot be lower than a 46 (and if I assign extra weight to my first actual GMAT verbal score of 50, it's probably closer to 47 or 48). Given the fairly large sample size of about 15 tests, I can be fairly confident that the variance in that score is low. To receive a score at least 5 points below my "true" mean (again, that is probably more like 6 or 7 points) seems almost impossible and appears to be a 2- or 3-sigma event. Has anyone else experienced what seems to be a significant statistical bias downward for individual GMAT sessions? Bear in mind that I have been a historically good test-taker; I scored 1500+ on the SAT and have never been particularly nervous about a test, so I doubt it is a bias due to the conditions inherent in the "real" GMAT.

[baladon99]

Hi mj12g ,
I have the same feeling. I took the test yesterday and scored 640 ( Q 46 V 31 ) my previous score was 650 ( Q 48 V 29) (Last dec) .I thought I did the verbal section decently barring the last RC . I felt that more than 34 questions I attended in the math section were right. I got some really tough questions in math and thought I solved most of them correctly. I even remembered some of the tough problems and solved them again after coming back home.My scores oscillated between 710 and 760 in practice tests . I am really stunned . Don't know what to do

[mj12g]

Hi mj12g ,
I have the same feeling. I took the test yesterday and scored 640 ( Q 46 V 31 ) my previous score was 650 ( Q 48 V 29) (Last dec) .I thought I did the verbal section decently barring the last RC . I felt that more than 34 questions I attended in the math section were right. I got some really tough questions in math and thought I solved most of them correctly. I even remembered some of the tough problems and solved them again after coming back home.My scores oscillated between 710 and 760 in practice tests . I am really stunned . Don't know what to do

That sucks; did you feel prepared this time for verbal and was consistently scoring higher than that? After seeing more and more of these experiences, I'm having trouble believing the GMAT's "30 point standard deviation" claim

[baladon99]

Hi mj12g ,
I have the same feeling. I took the test yesterday and scored 640 ( Q 46 V 31 ) my previous score was 650 ( Q 48 V 29) (Last dec) .I thought I did the verbal section decently barring the last RC . I felt that more than 34 questions I attended in the math section were right. I got some really tough questions in math and thought I solved most of them correctly. I even remembered some of the tough problems and solved them again after coming back home.My scores oscillated between 710 and 760 in practice tests . I am really stunned . Don't know what to do

That sucks; did you feel prepared this time for verbal and was consistently scoring higher than that? After seeing more and more of these experiences, I'm having trouble believing the GMAT's "30 point standard deviation" claim

Yeah.I strived my best to improve my verbal score.In the last few months ,my verbal scores were between 36 and 40 .My V score in the real one is much lower Btw what is that 30 point standard deviation ?

[David@VeritasPrep]

baladon99 -

The 30 point standard deviation is what the GMAC says - if you take the test on consecutive days you should score within 30 points above or below their actual score. It is called the score range.

The LSAT now has a score range that you fall within. They give you the number say (176) but they also say that you would fall within the band of 173 to 179 as an example.

So that is with all standardized tests. They say that the number is close to what you would score, but not necessarily exact. Now lots of people have taken the test multiple times over the years and had the exact same score, but a variation is more common.

Of course the variation does not take into account the effects of things like studying so that you are better prepared, or lack of sleep or a cold or things like this. The 30 points above and below is just the variation in the test so it is -- all other things being equal.

[David@VeritasPrep]

mj12g -

Just a couple of thoughts on a score that falls well outside of the normal distribution of your scores.

The way the GMAT is scored, missing a few questions at an easier level of difficulty can really lower your score. It may be that you made a few small mistakes on questions that are below your level and if this happened enough, the computer scoring algorithm will no longer trust you to rise above this level and even if you get tougher questions right it will place you back down into that lower score. It could be a matter of a few errors on questions that you would be expected to get right. So the test would feel the same to you as you took it and you would still some tougher questions right.

Also, do not forget the experimental questions. These can influence the score as well. On the verbal that is something like 9 questions that do not count. If you get these questions all correct and happen to miss others that do count then this could impact your score. Statistically this is not supposed to happen if you take a large population into account. But for an individual it could be a factor.

One or both of these factors might be at play: Missing a few questions - for whatever reason - that you would be expected to get right causing the computer to calibrate your score lower even though you might have done well on the tougher questions and/or getting all of the experimentals right so that every question you missed counted against you.

Just some thoughts. My bet is that if you could take it again tomorrow your score would be back closer to the levels you have seen in practice.

Going forward just remember that the most important questions to get right on the test on the ones that when you look at the question you say "I can definitely do this one." In that case, take the time and focus and get those right, they are very important.

[prodizy]

@david the experimental questions point you raised sounds logical. But, getting few easy questions wrong and the exam not trusting you because of that makes no sense at all. After all it's a computer ADAPTIVE test. I would be really scared if what you are saying is true.

[David@VeritasPrep]

Yes. It is an adaptive test. Just not nearly as adaptive as you think. Especially not on the verbal side.

Check out this posting by Ian Stewart he has written it all up pretty thoroughly here. It is on the bottom of the page. (Page 5 in case this does not take you to that page directly).

https://www.beatthegmat.com/r-d-on-gmat- ... 92-60.html

Let me know what you think.

Also check out the postings between Ian and I on this thread. https://www.beatthegmat.com/cat-question ... 85955.html

By the way, no reason to be scared by this. Just follow my advice to students from day one. Focus on getting the questions right that you can get right. On the verbal side, that better be the vast majority of them if you want to get a 50!

[baladon99]

baladon99 -

The 30 point standard deviation is what the GMAC says - if you take the test on consecutive days you should score within 30 points above or below their actual score. It is called the score range.

The LSAT now has a score range that you fall within. They give you the number say (176) but they also say that you would fall within the band of 173 to 179 as an example.

So that is with all standardized tests. They say that the number is close to what you would score, but not necessarily exact. Now lots of people have taken the test multiple times over the years and had the exact same score, but a variation is more common.

Of course the variation does not take into account the effects of things like studying so that you are better prepared, or lack of sleep or a cold or things like this. The 30 points above and below is just the variation in the test so it is -- all other things being equal.

Thanks for this insight. So can we say that , in terms of B school admissions , my scores 650 and 640 are almost the same ? ? ( 650 - last Dec and 640 -two days back)

[David@VeritasPrep]

Surely those two scores are very close together. There are some places that have given particular students a solid cutoff that they must exceed so it would be a big deal in such a case but normally not a big difference I imagine.

[pemdas]

@David, what you have explained sounds like linear relationship among the CAT scoring factors. What about people who have been tested on that same day or people who have been asked the same questions, i.e. percentile ranks? Some people drift better towards the end, some feel stronger at the start of exam. From what I know about GMAT CAT, people who consume more time with less thinking score lower than do people who consume less time with more thinking.

[David@VeritasPrep]

Not sure what you mean?

What do you mean people tested on the same day? Are you saying that you are being compared against people who take the test the same day that you do? But you get your score immediately - even if you are the first person in the world to take it that day. And the difficulty of the questions and other factors have nothing to do with people taking it the same day that you do.

As to how people scored on the same questions - that is exactly how the difficulty is determined. Questions are experimental and they are offered as actual questions and the percentage of people that get these right determines difficulty (as well as perhaps a couple of other factors).

Ian has explained in the post that in the end all questions count in terms of scoring (except for experimental questions). When you get questions right helps to determine how difficult subsequent questions will be. For example lots of right answers in the beginning gives you tougher questions and that can help with the scoring so long as you can get the tough questions that you see later correct.

I will put it in terms that can only help and cannot hurt no matter what the peculiarities of scoring. "When you see a question that you can get right, do what it takes to get it right." People spend way too much time on questions that they do not understand and will not likely answer correctly and they try to save time on the ones that they think are "easy" this is a huge mistake and Ian gives us one of the reasons why.