MGMAT CAT Quant harder than PowerPrep

BREAKING: Target Test Prep releases Brand New 2026 On Demand GMAT prep course

Redeem

This topic has expert replies

Post new topic Post Reply

Previous Topic Next Topic

erjamit: Master | Next Rank: 500 Posts; Posts: 170; Joined: Mon Oct 22, 2007 4:36 am; Thanked: 7 times

MGMAT CAT Quant harder than PowerPrep

by erjamit » Sun Jul 20, 2008 3:25 am

Hi,

I am taking MGMAT CATs every weekend. I found that MGMAT Quant is quite tough and its definitely harder than PowerPrep's quant. Is it correct or real GMAT quant is of same level as MGMAT's.

On MGMAT Quant my accuracy is close to 60-65% while on OG or old GMAT exams it is 75-80%.

Also, is the same correct for MGMAT Verbal also.

However, I would like to congratulate MGMAT team for making such excellent CATs. These CATs are worth the money spent for them.

Thanks
Amit

Quote

Source: Beat The GMAT — GMAT Strategy | Sun Jul 20, 2008 3:25 am

GMAT/MBA Expert

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

by Stacey Koprince » Mon Jul 21, 2008 9:54 am

A lot of my students do tell me they think our quant is harder than the real thing (though not everybody thinks so). I think there are two main reasons for this:

1) We don't include experimental questions in the mix, but the real test does. Experimental questions come at all difficulty levels, so the higher your score, the more likely your experimental questions will be lower-level compared to what you can do. That means you get a mental "break" in the midst of the test when you get these questions.

2) Some of our hardest quant questions are too computation intensive. As we get the data, we are editing or dropping these questions, but we've still got some in there that take a bit too long to solve.

As for the verbal, I generally have an equal number of people telling me it's easier and harder than the real thing. In fact, on days that I answer questions in the forums, it's not uncommon for me to see one post complaining that the verbal is too easy and another complaining that it's too hard. That usually means we've got it about right.

The "no experimental questions" thing is true for the verbal section too, but that doesn't have as big of an impact at the higher levels as the quant experimentals do for high-level quant testers.

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

Quote

erjamit: Master | Next Rank: 500 Posts; Posts: 170; Joined: Mon Oct 22, 2007 4:36 am; Thanked: 7 times

by erjamit » Mon Jul 21, 2008 11:40 am

Thanks Stacey for your opinion. However, I couldn't understand the following. Can you pls explain in some detail.

The "no experimental questions" thing is true for the verbal section too, but that doesn't have as big of an impact at the higher levels as the quant experimentals do for high-level quant testers.

Amit

Quote

GMAT/MBA Expert

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

by Stacey Koprince » Mon Jul 21, 2008 11:47 am

Our tests don't have experimental questions for quant or verbal, where the real test does for both quant and verbal. The experience that some people have (of thinking the real test is easier due to getting a lot of lower-level experimental questions) really only happens on quant, though - people don't seem to notice as much of a "mental break" on the verbal side (via getting lower-level experimental questions there).

Quote

erjamit: Master | Next Rank: 500 Posts; Posts: 170; Joined: Mon Oct 22, 2007 4:36 am; Thanked: 7 times

by erjamit » Thu Jul 24, 2008 6:54 am

Stacey how do we know that a question is an experimental question. What happens if an experimental question is incorrect. Do we get an easier question. My assumption is experimental questions are not accounted for the GMAT score.

Thanks
Amit

Quote

GMAT/MBA Expert

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

by Stacey Koprince » Thu Jul 24, 2008 8:00 am

You don't know whether something is experimental - they look just like the questions that do count. And, no, the experimental questions do not factor into the difficulty level of the questions that count.

There are three basic ways that the experimental questions will or could affect your test:

1) Together with the questions that count, they "cover the spread" of question types and content areas that we're all supposed to get (so you wouldn't, for example, get 6 geometry questions that count and another 6 that don't count - you'd get around 4-8 geometry questions total, some of which may count and some of which may be experimental).

2) If you get sucked into one, that could cost you big time later in the test on real questions.

3) If you leave any questions blank at the end, the experimental ones now DO count toward your score - any question left blank, whether regular or experimental, results in a 3 percentile point drop (per question) in your score.

Quote

skrutty: Newbie | Next Rank: 10 Posts; Posts: 2; Joined: Wed May 21, 2008 2:40 am

by skrutty » Thu Jul 24, 2008 11:06 am

MGMAT CAT Free Test - This is the first test I took and I got a 48 score on quant (86% percentile) and got about 15 wrong...

How does this all add up. 48 looks to good for about 15 wrong answers...

Quote

GMAT/MBA Expert

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

by Stacey Koprince » Thu Jul 24, 2008 11:08 am

That's how an adaptive test works. Your score is not based on a linear scale of percentage correct, the way a standard paper-based test is. Rather, you will be given more higher-level questions if you are a higher-scoring person - but you will still get a lot of questions wrong, because the mix you are being given is much tougher than that given to a lower-scoring person.

Quote

skrutty: Newbie | Next Rank: 10 Posts; Posts: 2; Joined: Wed May 21, 2008 2:40 am

by skrutty » Thu Jul 24, 2008 11:15 am

Stacey,

Thanks for the information....

Oh boy ! My focus so far has been to on 100% accuracy. I am trying to gather/solve all kinds of problems and trying to develop a problem solving pattern. Not only is this consuming all my time, but it never seems to end...

And I still seem to keep getting suff wrong..(BTW ..I completed all the match in math)

What would be a good point to give up all this ? I mean is a 48 good enough ?

How do i overcome this "perfectness" creep ?

Quote

GMAT/MBA Expert

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

by Stacey Koprince » Thu Jul 24, 2008 11:19 am

Yes, it will never end if you're trying to hit 100% - not going to happen! (Not to mention, that will tend to mess up your timing, in which case there's a good chance your score will drop below what it should be - if you run out of time or have to rush on a bunch of questions.)

48's a great score. If you can match that with a great verbal score, you're good!

Quote

jsl: Master | Next Rank: 500 Posts; Posts: 203; Joined: Wed Jun 04, 2008 1:01 am; Location: Windsor; Thanked: 5 times; GMAT Score:650

by jsl » Mon Sep 08, 2008 12:55 am

Hi Stacey,

I'd like to provide a suggestion to improve the realism of MGMAT CATs. Why don't MGMAT create an algorithm to simply "ignore" the result of a handful of questions each time a user takes the test? MGMAT could create a simple algorithm which would then more closely mimic the actual test. I accept that it is almost impossible to mimic the actual test; however, if MGMAT add some "true probability" into the practise CATs, it will do a couple of things:

1. help candidates to realise that probability and luck plays a part in the candidate's score
2. develop the ability and reasoning to skip questions

I'm sure there are implementation difficulties in the above - notably that it could completely mess up the MGMAT scoring algorithm. In fact, you could even "score" the result in exactly the same way and in the summary report, why not just "state" that a handful of questions were experimental. This may help candidates realise they should keep timing in check.

Anyhow, this is just a quick idea I had since I come from a computer science/financial background - the markets revolve around probabilities! Besides, if MGMAT do this, it would create a USP which no-one else on the test prep market has.

Following the above, would you happen to know whether... when you get an experimental question, and you get it right, assuming the next question is a scored question, will the GMAT give you an harder "real" question next? I.e. does the correctness/incorrectness of an experimental question affect the level of question difficulty?

Quote

GMAT/MBA Expert

Ian Stewart: GMAT Instructor; Posts: 2623; Joined: Mon Jun 02, 2008 3:17 am; Location: Montreal; Thanked: 1090 times; Followed by:355 members; GMAT Score:780

by Ian Stewart » Mon Sep 08, 2008 3:52 am

jsl wrote: Following the above, would you happen to know whether... when you get an experimental question, and you get it right, assuming the next question is a scored question, will the GMAT give you an harder "real" question next? I.e. does the correctness/incorrectness of an experimental question affect the level of question difficulty?

When you're taking the GMAT, you're really taking two different tests at once, one real test, and one experimental test. The questions just happen to be mixed together. One test does not influence the other; it doesn't matter how you do on the experimental questions, since they won't influence either the real questions you see, or the score you receive. The reverse is also true; no matter how well or badly you perform on the real questions, you will still see the same set of experimental questions that were randomly chosen for you when you sat down at the beginning of the test session.

The real and experimental questions only interact in one way: after your test is over, your performance on the real questions is used to judge how difficult the experimental questions you answered are.

For online GMAT math tutoring, or to buy my higher-level Quant books and problem sets, contact me at ianstewartgmat at gmail.com

ianstewartgmat.com

Quote

jsl: Master | Next Rank: 500 Posts; Posts: 203; Joined: Wed Jun 04, 2008 1:01 am; Location: Windsor; Thanked: 5 times; GMAT Score:650

by jsl » Mon Sep 08, 2008 4:43 am

Ian Stewart wrote:
jsl wrote: Following the above, would you happen to know whether... when you get an experimental question, and you get it right, assuming the next question is a scored question, will the GMAT give you an harder "real" question next? I.e. does the correctness/incorrectness of an experimental question affect the level of question difficulty?
When you're taking the GMAT, you're really taking two different tests at once, one real test, and one experimental test. The questions just happen to be mixed together. One test does not influence the other; it doesn't matter how you do on the experimental questions, since they won't influence either the real questions you see, or the score you receive. The reverse is also true; no matter how well or badly you perform on the real questions, you will still see the same set of experimental questions that were randomly chosen for you when you sat down at the beginning of the test session.

The real and experimental questions only interact in one way: after your test is over, your performance on the real questions is used to judge how difficult the experimental questions you answered are.

Hi Ian,

This is a gem! Thanks for clarifying. This is important to me as it gives me evidence that: If I see an easy question in a test following a hard one, it doesn't necessarily mean I got the previous one wrong. Also, it explains to people how the experimental questions are used with regard to the overall score a candidate achieves.

The "image" of taking 2 tests at once is a really good explanation.

Thanks!
Jon

Quote

GMAT/MBA Expert

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

by Stacey Koprince » Mon Sep 08, 2008 5:57 am

Yep, I like the idea of "two tests." There are some very minor ways in which the experimentals play into your scored test but not your test score - eg, they're part of the count of different content areas and question types you get (so you wouldn't get a bunch of questions on one obscure content area). But that doesn't actually affect your score, of course.

jsl, thanks for your suggestion on the experimental stuff. We do plan to add experimental capability to our tests, but we've been working all out on some other stuff that will get released later this year and the experimental bit has just been a bit lower on the priority list. (Also, we're hoping to add true experimental capability: that is, actually test our own questions before we use them for "real" on your tests - but your suggestion is a good one just to get things going. I'll forward it on to our algorithm guy as "food for thought.")

Quote

GMAT/MBA Expert

Ian Stewart: GMAT Instructor; Posts: 2623; Joined: Mon Jun 02, 2008 3:17 am; Location: Montreal; Thanked: 1090 times; Followed by:355 members; GMAT Score:780

by Ian Stewart » Mon Sep 08, 2008 7:02 pm

Stacey Koprince wrote: There are some very minor ways in which the experimentals play into your scored test but not your test score - eg, they're part of the count of different content areas and question types you get (so you wouldn't get a bunch of questions on one obscure content area). But that doesn't actually affect your score, of course.

Stacey, I'd be very interested to know- what has led you to think this is true? I have no confirmation either way, but from what I know about how experimental questions are selected, I would be a bit surprised if they influenced the selection of real questions in any way.

Content area is certainly a factor, along with difficulty level, when real questions are selected during a test, but I don't think seeing two experimental questions on probability will reduce the number of real questions you'll see on probability. After all, the real questions need to span the content areas the test is trying to test. Anyway, I'm curious if you have any additional information about this.

For online GMAT math tutoring, or to buy my higher-level Quant books and problem sets, contact me at ianstewartgmat at gmail.com

ianstewartgmat.com

Quote

Post new topic Post Reply

MGMAT CAT Quant harder than PowerPrep

This topic has expert replies

MGMAT CAT Quant harder than PowerPrep

GMAT/MBA Expert

GMAT/MBA Expert

GMAT/MBA Expert

GMAT/MBA Expert

GMAT/MBA Expert

GMAT/MBA Expert

GMAT/MBA Expert

GMAT/MBA Expert

GMAT Course Reviews

Admissions Consulting Reviews