Detection of Unscored Questions

[Samy]

Hi,
I have heard that there are some questions that the GMAT asks in the Verbal and Math sections that are not scored and are only for their research purpose. These questions do not contribute toward one's ultimate score and one may end up spending a lot of time on them.

Is there a way we can know which question types are these?
Is there a pattern, like they generally come up between 10-20 question numbers or they are the SC or CR category questions.

Your opinion is appreciated.

Thanks
Samy

[esledge]

Hi Samy,

Those are experimental questions to which you are referring; the GMAC includes them to accurately gauge their difficulty level for use on future tests. There are about 10 such questions in both the Quant and Verbal sections.

There is no way to know whether a particular question is experimental or one that counts toward your score. They are of all question types, and can appear anywhere in the order of problems.

You have to treat all the questions as if they are "real." Certainly don't waste any time on test day trying to figure out which ones are experimental. The flip side of that is advice is that when you find yourself stuck on any question, it is generally not worth spending tons of extra time on it--there is a 25% chance that it wouldn't count towards your score anyway!

[Samy]

Thanks a lot.

Sometimes I do tend to take 4-5 min on 2-3 Questions in every test I take. I find it difficult to leave the problem without doing justice to it. I feel that by revisiting the problem several times I will get it.
Is this worth the time, if I get them right ?
I also run the risk of not completing the last 3 Questions.
But, these lengthy questions could ultimately contribute toward a better score, provided I get them right.
Atleast I wont feel as bad leaving a question since it has 25% chance of not getting scored.
:D

Thanks

[StaceyKoprince]

It's not worth the time, primarily because we are most likely to perform poorly overall on questions on which we spend more than about 3 minutes.

Remember that every single question you see is designed to be answered by someone in 2 minutes. That someone may not be you, however. Generally speaking (and this has been proven in scientific studies), the farther over 3 minutes we go on a particular question, the more likely we are to get it wrong. Having to spend that much time is an indication that you don't really know how to do the problem - because it is designed to be done in 2 minutes and you aren't figuring it out in that timeframe.

You may of course get some of those questions right - but, overall, your performance will be poorer on those questions. On top of that, you get a double-negative, because that extra time you use has to be taken from other questions that perhaps you could actually get right, if you just had the time to do them.

[Samy]

Thanks and I agree. The accuracy really goes down beyond 2.5 min.
However if a Q requires that time in the first 10, should I take that plunge ?
Won't have the courage to leave it....

[StaceyKoprince]

No, definitely not. If you are tempted to go beyond that amount of time, all it's telling you is that you don't know how to do the question. And you're taking that time from another, unknown question towards the end of the test - and that one has a better shot of being an easier one than the one you're currently sitting on that you can't do in 2.5 minutes. So you take a double hit.

[Samy]

Thanks.
I agree.
Is it true that the beginning few Q's carry more weight age ?

I read in the GMAT Bulletin that : "Adaptive Score calculations do not assign differential credit to questions depending on where they appear in the test. The questions in the adaptive test are weighted according to their difficulty and other statistical properties, not according to their position in the test"

Do you follow the same in your Online Simulated CAT's and the above stated correct according to the input you get from students who have appeared ?

This is really important because most of us plan on maximizing accuracy in the beginning.

[StaceyKoprince]

It is a myth that you should spend more time on the earlier questions. We talk about this extensively in our free sessions and in our classes and we build out materials and CATs based upon the way the test actually works, rather than the myth that has been perpetuated for many years.

IRT, the testing theory behind the CAT, does not assign greater weight to earlier questions (or later questions or any questions based solely on position).

[Samy]

But, If you do well in the beginning you tend to get more Difficult Questions.
That means you have the opportunity to score better, because the differential credit system is based on the difficulty level of the questions. Later on you may not see as many difficult questions (i.e if you have not done well enough in the beginning), and even though the credit remains same throughout, you tend to lose out.

Thanks. :)

[dbernst]

Samy, you can significantly improve your score by answering a string of questions correctly no matter where they fall during the section. However, you are in some serious trouble if you spend so much time on the early questions that you do not have an opportunity to complete the section and are forced to guess randomly for the final several problem. As Stacey and Emily stated, it is a complete myth that you should spend an inordinate amount of time on the first 10 quant and verbal questions.

-dan