time allocated for first 12-15 questions?

Expert replies

puniaball: Junior | Next Rank: 30 Posts; Posts: 13; Joined: Fri Mar 07, 2008 9:39 am

by puniaball » Mon Apr 21, 2008 2:14 pm

Can someone please provide thoughts on this?

37 questions in 75 mins translate to about 2 mins per question.. however, due to the computer adaptive nature of the exam, i was told to spend more time on the first 12 or so questions (maybe even 3 mins each) to ensure that they're correct.. i do believe that the first questions are worth more than the last questions of the exam, but does anyone know if I can, say, get the first 15 right, do ok the next 15, and randomly guess (wrong) the last 7?

I'm tempted to test this with a practice exam but then can't guarantee the questions presented will be the exact same.. so kind of defeats the whole purpose of trying..

Like this post

Source: Beat The GMAT — GMAT Strategy | Mon Apr 21, 2008 2:14 pm

Prasanna: Master | Next Rank: 500 Posts; Posts: 418; Joined: Mon Feb 26, 2007 6:41 pm; Thanked: 24 times

by Prasanna » Mon Apr 21, 2008 5:07 pm

puniaball

All questions carry equal weightage on the test. The notion that the initial questions carry more weightage is false. In fact GMAC has clarified on this advising students to treat all questions with equal respect. The test is adaptive and your score can drop drastically if you dont do well in the last 10 questions. So there is no point in wasting practice tests.

Prasanna

Like this post

jamess: Senior | Next Rank: 100 Posts; Posts: 51; Joined: Sat Apr 12, 2008 8:44 am

by jamess » Tue Apr 22, 2008 2:06 am

Prasanna wrote:puniaball

All questions carry equal weightage on the test. The notion that the initial questions carry more weightage is false. In fact GMAC has clarified on this advising students to treat all questions with equal respect. The test is adaptive and your score can drop drastically if you dont do well in the last 10 questions. So there is no point in wasting practice tests.

Prasanna

I am afraid I disagree with your opinion. Please provide the link that shares the GMAC's comment.

I have run simulations on this part and from my knowledge of how the CAT algorithm behaves, in certain cases getting the first few questions right does make a difference to one's overall score. Note that I said in certain cases. This is so because the score moves more in due to the first few questions (when it has less information) than later (when it has more information).

So getting some correct guesses for questions that are harder than your ability put you in a position to receive a higher score.

Check out <a href="https://gmatadvice.blocked">gmatadvice.blocked</a> for useful tips to ace the GMAT.
<a href="https://gmatadvice.blocked/2008/04/ ... tml">ebook on critical reasoning</a> launched.

Like this post

GMAT/MBA Expert

Stacey Koprince: GMAT Instructor; Posts: 2228; Joined: Wed Dec 27, 2006 3:28 pmFollowed by: 694 members; Location: Montreal, Canada; Thanked: 639 times; GMAT Score:780

by Stacey Koprince » Thu Apr 24, 2008 6:15 pm

If you read any OG guides, you will find the info that all questions are worth the same amount. You cannot have run simulations on the GMAT because they will not let you do this on their own actual test and they do not release their algorithm to the public. If you ran simulations on some other software - that's not the GMAT. Most practice GMATs do not actually simulate the real software very well.

The simulation you describe only takes into account the fact that the score can move more earlier on than later (which is true) but does not take into account the severe penalties at the end of the test if you run out of time. And if you spend extra time on earlier questions, you will run out of time - that time has to come from somewhere and it comes from the end of the test.

If you run out of time and leave any questions blank, it results in a 3-percentile-point reduction PER question, including any questions that would otherwise have been experimental questions.

If you run out of time and just guess on any remaining questions but get them all wrong anyway, it results in a 2 to 2.5 percentile point reduction PER question - though not including any experimental questions.

Further, don't think of your score as an average of your performance over the length of the test - as we're used to on regular paper tests. It's more accurate to think of your score as the level you're at WHEN THE TEST ENDS - right at that moment.

Also, if you go beyond about 2m45s on a problem, the longer you go, the more likely you are to get it wrong. All problems are designed such that SOMEONE can do the problem in 2 minutes. If we can't, that just means we don't really know how to do the problem...

Please note: I do not use the Private Messaging system! I will not see any PMs that you send to me!!

Stacey Koprince
GMAT Instructor
Director of Online Community
Manhattan GMAT

Contributor to Beat The GMAT!

Learn more about me

Like this post

jamess: Senior | Next Rank: 100 Posts; Posts: 51; Joined: Sat Apr 12, 2008 8:44 am

by jamess » Sat Apr 26, 2008 4:26 am

I agree with you that there is severe penalty on not answering all questions.

You do agree that the first few questions do matter more (in moving your score). So, if a test is timed well such that the test taker doesn't miss the last questions, he has a higher chance of getting a better score if gets the first few questions right.

Right?

Check out <a href="https://gmatadvice.blocked">gmatadvice.blocked</a> for useful tips to ace the GMAT.
<a href="https://gmatadvice.blocked/2008/04/ ... tml">ebook on critical reasoning</a> launched.

Like this post

mayonnai5e: Legendary Member; Posts: 1018; Joined: Tue Dec 12, 2006 7:19 pmFollowed by: 6 members; Thanked: 86 times

by mayonnai5e » Sat Apr 26, 2008 8:18 am

jamess wrote: This is so because the score moves more in due to the first few questions (when it has less information) than later (when it has more information).

I think you're assuming that the algorithm freezes your score in place once it serves you a question and never recalibrates again. Since no one is really sure how the algorithm works, we cannot make this assumption.

If the algorithm constantly recalculates your current score based on ALL previous questions (you have answered 10 questions that when combined you are considered a 550 person and, after getting this question correct, the algorithm gives you a 575 rating based on the last 11 questions), then your assumption would be wrong. This is because the algorithm would always calculate the score based on all questions irrespective of whether they are at the beginning or end.

https://www.beatthegmat.com/my-blog-erro ... t4899.html
550 =\ ...560 =\... 650 =) ...570 =( ...540 =*( ...680 =P ... 670 =T ...=T... 650 =T ...700 =) ..690 =) ...710 =D ...GMAT 720 DING!! ;D

Learn more about me

Like this post

GMAT/MBA Expert

Stacey Koprince: GMAT Instructor; Posts: 2228; Joined: Wed Dec 27, 2006 3:28 pmFollowed by: 694 members; Location: Montreal, Canada; Thanked: 639 times; GMAT Score:780

by Stacey Koprince » Sat Apr 26, 2008 9:11 am

Yes, mayonnai5e is right. The "swing" from question to question doesn't actually factor in in the way you may be assuming.

Also: "So, if a test is timed well such that the test taker doesn't miss the last questions, he has a higher chance of getting a better score if gets the first few questions right. "

This is true IF you don't have to spend any extra time on those first few questions - because that time has to come from other questions. It's also true, by the way, that IF you do exactly as you would have done on the test anyway but somehow manage to get a couple more questions right ANYWHERE on the test, you'll get a higher score. But that's not a strategy. That's a hope.

Some researchers once showed that if you could get the first 7 questions in a row right and do the rest of the section exactly the same, the score increase would be between about 20 and 60 points. (This was the source of that "spend more time on the early questions" myth.) But the test prep folks didn't read the entire study (I know because I was working in the industry at the time this myth was propagated - I admit, with chagrin, that I didn't read the full study at the time either.) It also said that if you had to guess on just 4 questions at the end as a result of spending more time on 7 questions at the beginning, your score would be the same as if you'd worked through it normally, and if you had to guess on more than 4 questions, your score would actually go down.

The problem? Good luck getting 7 questions in a row right without spending so much extra time that you have to guess on fewer than 4 questions at the end.

The proper way to interpret this data: get the first 7 questions in a row right without spending extra time. And, of course, if you can do that, you don't need anyone's help on the test! I don't have to tell you what kinds of questions you'd be answering if you got even 3 questions in a row right, let alone more than that.

In other words, it's a nice theory, but it does nothing for you in practice. Any time you spend extra time, that time has to come from somewhere. Lack of time leads to wrong answers. And it hurts your score FAR more to get an easier question wrong than it does to get a harder question wrong. In fact, getting a very hard question wrong essentially doesn't hurt your score at all - it just doesn't help your score. But if you get a question wrong that the algorithm had considered "beneath" you, your score will plunge.

Really - don't play with the time (much, anyway - obviously you'll do some in 1.5 and some in 2.5). If you go in with the mindset that you "must" get the first 3 questions right, then you will get sucked into spending way more than that to "make sure" that you get them right - and now you've already blown the entire section.

Like this post

jamess: Senior | Next Rank: 100 Posts; Posts: 51; Joined: Sat Apr 12, 2008 8:44 am

by jamess » Sat Apr 26, 2008 5:02 pm

mayonnai5e wrote:
jamess wrote: This is so because the score moves more in due to the first few questions (when it has less information) than later (when it has more information).
I think you're assuming that the algorithm freezes your score in place once it serves you a question and never recalibrates again. Since no one is really sure how the algorithm works, we cannot make this assumption.

If the algorithm constantly recalculates your current score based on ALL previous questions (you have answered 10 questions that when combined you are considered a 550 person and, after getting this question correct, the algorithm gives you a 575 rating based on the last 11 questions), then your assumption would be wrong.

No. It doesn't negate my assumption. Infact, it is not an assumption.

What I am saying is that when you get your 11th question right/wrong, your score moves by X points. But when you are at the 31st question, the impact that question has on your score is <X. It is because the algorithm has MORE INFORMATION on calculating your score, and your response to any question averages out together with your response to the previous questions.

This is not to say that the last questions do not matter.

Check out <a href="https://gmatadvice.blocked">gmatadvice.blocked</a> for useful tips to ace the GMAT.
<a href="https://gmatadvice.blocked/2008/04/ ... tml">ebook on critical reasoning</a> launched.

Like this post

jamess: Senior | Next Rank: 100 Posts; Posts: 51; Joined: Sat Apr 12, 2008 8:44 am

by jamess » Sat Apr 26, 2008 5:13 pm

Stacey Koprince wrote: This is true IF you don't have to spend any extra time on those first few questions - because that time has to come from other questions. It's also true, by the way, that IF you do exactly as you would have done on the test anyway but somehow manage to get a couple more questions right ANYWHERE on the test, you'll get a higher score. But that's not a strategy. That's a hope.

Do you agree that getting 7 questions right early in the test would affect the score more than getting 7 questions right later in the test?

Obviously, if your guesses turn out right, it would help ANYWHERE in the test. What I am saying is that it would help more during the early part of the test.

Stacey Koprince wrote: In other words, it's a nice theory, but it does nothing for you in practice. Any time you spend extra time, that time has to come from somewhere. Lack of time leads to wrong answers.

Really - don't play with the time (much, anyway - obviously you'll do some in 1.5 and some in 2.5). If you go in with the mindset that you "must" get the first 3 questions right, then you will get sucked into spending way more than that to "make sure" that you get them right - and now you've already blown the entire section.

Sure I agree with you that one must get "stuck" in the mindset that (s)he has to get the first 10/15 questions right. That can lead to severe mismanagement of time.

But at the same time, I believe that more time/attention should be given to the first few questions in view of the information that they matter more. Now how much more time to give to these questions is up to the test taker and his/her plan on how does (s)he manage time on the test.

This technique is easy to implement for those who have few mins to spare at the end of the test (for ex, for math I generally have 15-20 mins at the end of the test). It is better to spend that time during the first 15 questions than on the last question alone or finish the test early which has no known benefit.

Check out <a href="https://gmatadvice.blocked">gmatadvice.blocked</a> for useful tips to ace the GMAT.
<a href="https://gmatadvice.blocked/2008/04/ ... tml">ebook on critical reasoning</a> launched.

Like this post

GMAT/MBA Expert

Stacey Koprince: GMAT Instructor; Posts: 2228; Joined: Wed Dec 27, 2006 3:28 pmFollowed by: 694 members; Location: Montreal, Canada; Thanked: 639 times; GMAT Score:780

by Stacey Koprince » Sat Apr 26, 2008 8:09 pm

Do you agree that getting 7 questions right early in the test would affect the score more than getting 7 questions right later in the test?

Not if those seven questions were at the same exact difficulty levels, no. The factor that matters in this regard is not the placement of the question but the difficulty level of the question.

No question is worth more than any other question solely based on placement of that question. If you take the exact same question and place it first vs. last, that question will be worth the same amount.

I completely understand why people "instinctively" think the placement matters - I thought so too before I fully knew how the algorithm works - but your instinct is fooling you. That's not how the test works.

Like this post

jamess: Senior | Next Rank: 100 Posts; Posts: 51; Joined: Sat Apr 12, 2008 8:44 am

by jamess » Sun Apr 27, 2008 7:27 pm

Stacey Koprince wrote: Not if those seven questions were at the same exact difficulty levels, no. The factor that matters in this regard is not the placement of the question but the difficulty level of the question.

No question is worth more than any other question solely based on placement of that question. If you take the exact same question and place it first vs. last, that question will be worth the same amount.

It is not possible to get 7 questions (when you answer them correct) of the same difficulty. I assumed that we were clear that the algorithm is adaptive.

As such, the scenario you mention is not applicable to GMAT.

My argument still is that it matters more to get 7 questions right in a row at the beginning of the test than later.

Check out <a href="https://gmatadvice.blocked">gmatadvice.blocked</a> for useful tips to ace the GMAT.
<a href="https://gmatadvice.blocked/2008/04/ ... tml">ebook on critical reasoning</a> launched.

Like this post

GMAT/MBA Expert

Stacey Koprince: GMAT Instructor; Posts: 2228; Joined: Wed Dec 27, 2006 3:28 pmFollowed by: 694 members; Location: Montreal, Canada; Thanked: 639 times; GMAT Score:780

by Stacey Koprince » Sun Apr 27, 2008 8:26 pm

It is not possible to get 7 questions (when you answer them correct) of the same difficulty. I assumed that we were clear that the algorithm is adaptive.

I did not mean that the 7 questions would have one difficulty level. I meant that we were talking about the same 7 questions (with 7 different difficult levels) just placed at different points during the test.

Your previous quote:

Do you agree that getting 7 questions right early in the test would affect the score more than getting 7 questions right later in the test?

The difficulty level of a question will impact your scoring differently depending upon how you were scoring prior to that question. So if you want to talk about getting 7 questions right at one point vs. another point, you at least have to compare the same 7 questions, or questions of equivalent difficulty levels to those 7 questions. And even that is too simplistic, because your scoring level will be different at different points on the test.

If you and I were both to get the exact same question at #11, but you went into that question "scoring" 700 and I went into that question "scoring" 650, its impact on our scores would be different even if we both got it right (or both got it wrong). It isn't the case that, at #11, your score moves a certain amount up or down solely (or even primarily) due to the fact that you're on #11.

A better way to think about the wider range of possible difficulty levels on the questions offered to you earlier in the test is simply that, because the test has less data on you, it's allowed to be a little less precise in deciding what to give you next from its database. It's problematic to think of it as "the test is choosing the question based upon the current score it is calculating for me" because that leads people to think things like "the early questions are worth more." They're not.

Like this post

GMAT/MBA Expert

Stacey Koprince: GMAT Instructor; Posts: 2228; Joined: Wed Dec 27, 2006 3:28 pmFollowed by: 694 members; Location: Montreal, Canada; Thanked: 639 times; GMAT Score:780

by Stacey Koprince » Sun Apr 27, 2008 8:45 pm

p.s. we may just have to agree to disagree on the algorithm discussion.

The people who make the test say that the early questions are not worth more than questions elsewhere on the test. Even if you believe they would lie about something like this - if they did, they would long since have been sued for spreading misinformation that caused people to perform more poorly on their test. There are organizations that exist just to serve as watchdogs for this purpose. For example, check out www.fairtest.org.

Like this post

jamess: Senior | Next Rank: 100 Posts; Posts: 51; Joined: Sat Apr 12, 2008 8:44 am

by jamess » Thu May 01, 2008 5:54 pm

Stacey Koprince wrote: If you and I were both to get the exact same question at #11, but you went into that question "scoring" 700 and I went into that question "scoring" 650, its impact on our scores would be different even if we both got it right (or both got it wrong). It isn't the case that, at #11, your score moves a certain amount up or down solely (or even primarily) due to the fact that you're on #11.

A better way to think about the wider range of possible difficulty levels on the questions offered to you earlier in the test is simply that, because the test has less data on you, it's allowed to be a little less precise in deciding what to give you next from its database. It's problematic to think of it as "the test is choosing the question based upon the current score it is calculating for me" because that leads people to think things like "the early questions are worth more." They're not.

Sorry for the late reply.

I find the two paragraphs that you wrote as contradicting. In the first para you tell me that the amount our score moves depends upon our scores at that particular question (which is obviously a function of how i/you answered the previous questions). Then in the next para you say that its problematic to think that the test is choosing questions based upon the current score it has calculated for me. I may have been wrong in interpreting it, but I find the two do not match.

My opinion is that the test does deliver questions based on its current assessment of our score. So I agree with your statement in para 1. And a corollary of that is that "getting questions right at the beginning of test is worth more".

Check out <a href="https://gmatadvice.blocked">gmatadvice.blocked</a> for useful tips to ace the GMAT.
<a href="https://gmatadvice.blocked/2008/04/ ... tml">ebook on critical reasoning</a> launched.

Like this post

GMAT/MBA Expert

Stacey Koprince: GMAT Instructor; Posts: 2228; Joined: Wed Dec 27, 2006 3:28 pmFollowed by: 694 members; Location: Montreal, Canada; Thanked: 639 times; GMAT Score:780

by Stacey Koprince » Thu May 01, 2008 6:02 pm

I could've worded that more clearly. I didn't actually say in the 2nd para that this is not how the test works. I said it's problematic for us, the non-researchers who didn't make the algorithm, to think about it that way because we then make certain assumptions which lead us to false conclusions. That's all.

Like this post

Post new topic Post Reply

GMAT Course Reviews

559 reviews

1168 reviews

431 reviews

359 reviews

147 reviews

5/5

5 Star (560 Reviews)

"Target Test Prep is the closest to the official version of the GMAT exam, about 99% accuracy in terms of the quality and quantity of information. The course has excellently created singular sets of focused lessons and tests for every possible topic that one could come across in the official GMAT exam."

"The TTP course maximizes the efficiency of the time you spend studying. It will take time and effort but I could almost guarantee that if you complete the course exactly as it is laid out you will get an amazing score. They also have a very responsive team willing to help with any questions you might have."

"TTP has two things that I think no other test prep company offers: A teaching approach that reinforces understanding and an attitude that will give you the mental preparedness needed to succeed on the test. TTP gives you a deep understanding of the concept you need to know while teaching you how to think."

Write a Review