What is the median difficulty level of the GMAT Advanced Qui

[tomslawsky]

These questions are difficult, no douby, but there is deff. a "spread" and some are medium hard, some are VERY HARD. What is the approximate median level of this book on a basis of 1-51. I think it is slightly retarded to speak of the scoring on each sub section in terms of the 0 - 800 scale, given the possible huge ranges of each sub sections that can lead to identical overall scores. Before I get the canned "correct" answer that MGMT doesn't have access to the official scoring scale and therfore, cannot grade these questions, I will say that I'm 100% certain that the authors have a great approximation of where a question alls on the 0-51 scale; if not then this type of book would be impossible to write. So, cough it up, what is your best guess.

[jnelson0612]

Tom,
I can't even begin to make estimations on the scale of 0-51. I think in terms of question difficulty levels of 300-500, 500-600, 600-700, and 700-800. This is how we categorize all of our CAT questions as well as the Official Guide questions in our OG quant companion book.

Having looked at this book, I think it's safe to say that at least half of the questions are 700-800, with lower percentage of 600-700 mixed in. Hope this helps!

[tomslawsky]

Thanks, but sorry, this doesn't make sense to me. The intro to the book states that I should be at least 75 percentile before staring this book. However, since I am already awesome at verbal, I average 48 V sub-score. So, what you are telling me is that just over half of these advanced quant study guide questions are written at a 38 Q sub-score and the remainder are at a even lower than 38 because a 38 Q is all I need to score a 700 overall. Therefore, in order to be a 700 level foe me, the quant questions in this book are pretty much centered around a 38, because combined with my V 48 score, this would yeild a 700 - 710 overall score.

Doesn't this classification seem just a bit illogical to you? Dare I say that were the system put to a test, in say a critical reasoning GMAT question, or an argument essay, a prudent test taker would be able to pick your rating system of classifying sub section scores by using overall score numbers apart? If I were to write a review of this book on, say Amazon warning potential buyers that the average sub score of the questions in this book is a 38, you would have no recourse, right? In my case, this must be true?

[jnelson0612]

Hi Tom,
I think you are misunderstanding what the book says. It says it "should only be used by students already achieving at least a 70th percentile math score on practice exams." See this link: http://www.manhattangmat.com/storeitems ... loc=forums Your overall score is not a good gauge as to whether this book is appropriate for you; you should only consider your current math percentile score.

I'll be honest with you--I've taught GMAT for many years and I just don't conceptualize individual problems on the scale you are suggesting, nor was I involved in the writing of this book. I've alerted Stacey, Ron, and Tim to this thread and if any of them have more information I'm sure there will be another post soon.

Thanks!

[tim]

Tom, what part of Jamie's phrase "at least half" led you to conclude "just over half"? And what would possibly make you think that just because YOU only need a 38 on quant to get a 700 that this means Jamie's estimate of a 700 level problem translates universally to a 38?

As far as Jamie's classification system seeming "illogical", I'll point out that there is pretty much a one-to-one correspondence between raw scores and percentiles, and then between percentiles and 200-800 scores. How then can you conclude that one makes more sense than the other, other than to conclude that Jamie's method of putting things into a more intuitive scale makes more sense?

I won't bother picking apart the rest of your argument because I think I've made my point. My only request is that if you review our book on Amazon you include all the detail you've given above so people can put your comments into the proper perspective. And do yourself a favor - stop trying to pick a fight with Jamie, and use that time instead to work on your critical reasoning skills.

[tomslawsky]

Tom, what part of Jamie's phrase "at least half" led you to conclude "just over half"? And what would possibly make you think that just because YOU only need a 38 on quant to get a 700 that this means Jamie's estimate of a 700 level problem translates universally to a 38?

As far as Jamie's classification system seeming "illogical", I'll point out that there is pretty much a one-to-one correspondence between raw scores and percentiles, and then between percentiles and 200-800 scores. How then can you conclude that one makes more sense than the other, other than to conclude that Jamie's method of putting things into a more intuitive scale makes more sense?

I won't bother picking apart the rest of your argument because I think I've made my point. My only request is that if you review our book on Amazon you include all the detail you've given above so people can put your comments into the proper perspective. And do yourself a favor - stop trying to pick a fight with Jamie, and use that time instead to work on your critical reasoning skills.

What are we in third grade now? The only thing missing is "Jamie's rubber, you're glue" I suspect you'll deactivate my account now, but at least I have a little time to respond:

As per YOU picking me apart: Let's see- Why I think that YOUR classification is illogical? The GMAT score, on the 200 -800 scale is a "non moving target". The skill level to reach a 600, 700, 800 is the same today as it was 10 years ago. This is true with ALL standardized tests, with the exception of when the test companies advertise that they are going to deliberately change their scoring curve. Reference the SAT about 15 years ago, when they "reentered" the test.

However, for whether reason, the PERCENTILE associated with each score is a MOVING TARGET. In other words, even though a 40 Quant represents the same skill level today as it did 10 years ago, the percentile can, and in the case of the GMAT does change significantly. Twenty years ago, a 700 was 97th percentile, today the same score (representative of the same skill level) falls in the ~90th percentile. The testing skills haven't changed, but the testing population has changed. As I've said before, there is no indication that the GMAT is making the test easier, therefore, the testing population is changing towards possessing a higher skill level. Therefore, using percentiles to represent the "absolute" skill level that a score represents is completely illogical. The &75th percentile mandated in the intro to MGMT Advanced Quant could be a 47 today and a 50 in 5 years and, if the test does not change, but the testing population, for whatever reason shifts to be less quantitative, that same 75th percentile could represent a 43 sub score.

On to my main point, using an overall score to represent the skill level of a sub score would only be valid if both of the subscores shared the same distribution curve (same average, median and standard deviation and range), AND if the final score shared the same shape distribution curve. Only then, can the scales be used interchangably, without modification. Further, since the GMAT subsections DO NOT share identical distribution curves, and the overall score is the result of the combination of each of these curves, and yields a curve that is distinct from either of the constituent curves. This lends to a scenario whereby there is a LOT of different combinations of sub scores that lead to the same "overall score". The quant component of the final distribution curve is independant of the verbal portion. It makes as much sense to refer to a sub score on the GMAT as a 200-800 scale as it does to refer to a GRE or SAT sub score on the "1600 scale".

A more simple analogy would be this: Suppose a physical chemistry class required solid mathematical skills as a prerequisite. Would the chemistry department do well to reference that a student have taken calculus 3 and an overall 3.0 GPA before taking physical chemistry? What if the student has a 3.5 overall GPA, but failed calc 1,2 and 3? Wouldn't it be more logical for the department to require that a student take calc 3 and have at least a 3.0 average in their math class. The overall GPA COULD be indicative of high math grades, but there are many cases of students with a high GPA and low math grades. To conflate the two would be nonsensical.

To conclude, classifying GMAT sub scores on the 200 - 800 scale is not a valid technique. There are several reasons for this. First, the sub score is a "moving target", and GMAT data has shown that this percentile moves quite significantly over the course of 5 years, the "validity" time frame given by the test writers. Secondly, since the subscore curves and the final distribution curve are not all identical, you can't just refer to one in terms of the other. For these reasons, classifying a sub score on a 200 - 800 score is invalid for GMAT usage.



I do want to add that I think that all of the MGMT material is, by a large gap, superior to anything else out there, I’m not desputing this, or arguing to the contrary. My only issue is with the 200 -800 classification of sub scores. I have laid out my argument as to why it is not a reasonable scoring system. It is what it is, and deactivating my account, removing the post and implicit ad homonym pejoratives can’t change it.

PS- to think that using a 200 - 800 scale is "more intuative" to anyone who is serious about the GMAT seems to me to be on the boarder of a canard. I mean, really, is a 0 -51 scale THAT hard to grasp, especially to your target audience, that possesses an IQ > 125 - 130?

[tomslawsky]

Tim,
by the way this is from your Bio:

"Tim's level-headedness, cool sense of self and ability to thrive on the idiosyncrasies of the classroom experience make him an ideal candidate for tutoring."

And a student review
"Tim never makes the students feel "dumb" if we don't choose the correct answer or if we ask any questions. I don't feel like he is an instructor only, but also a peer. He is very approachable."


Using a Knee jerk statement like WHAT PART of X did you not understand is level headed?, Wasn't intended to make me sound dumb? Is being very approachable?

[tim]

Tom,

It's going to be okay. No one is deactivating your account. In fact, we're here to help. I mainly want to help you see that there are more important things to spend your time on than worrying about which of two scales to use.

In the hope that this thread will be of some use to other people as well, I'll point out that this is a very frequent concern students have - just what is important and what isn't when studying for the GMAT. Raw scores and question difficulty levels (other than relative difficulty levels, which both our scales are adequate to assess) are at the bottom of that list, right alongside things like trying to game the system by analyzing question difficulty during the test. It's just not worth worrying over, and particularly not worth fighting over. Let it go.

Best of luck with your test preparation; I hope to see you around on the forums with questions on more relevant topics, and I look forward to helping you with those.

[StaceyKoprince]

Wow, looks like I missed an exciting discussion. :)

Yes, it doesn't literally "make sense" to discuss an individual question as, say, a 700-level question, because the 3-digit scores are a composite of quant and verbal achieved my some magic formula that has never been made public.

But most people know that 700 is about the 90th percentile. So when we say something is a 700-level question, that's short-hand for saying it's around the 90th percentile. That's easier than saying this is a 49 or 50, or 38, or 42... because then all of my students would have to pull out conversion lists to figure out what that means. So the 1-51 scale... not an easy scale with which to have a conversation.

The most accurate of all would be to couch things in terms of percentiles - this is a 75th percentile question - but most students aren't like you. They'll then ask me to give it to them on the 3-digit scale. I agree that the percentile rating is what they should really care about, but a lot of people don't know how to assimilate that information.

So. Just think of the 3-digit answer to this as a proxy - an easy way to discuss the *relative* difficulty of different problems. Not an actual absolute number.

[tomslawsky]

vin·di·cate"‚ "‚/ˈvɪndɪˌkeɪt/ Show Spelled[vin-di-keyt] Show IPA
verb (used with object), -cat·ed, -cat·ing.: to uphold or justify by argument or evidence: to vindicate a claim.

http://dictionary.reference.com/browse/vindicate


It looks as if my theory on the difficulty level of the GMAT is in allignment with the GMAC's Claims, VINDICATING (see above) my claim that it is illogical to refer to percentile scores, rather than absolute scaled scores when referring to difficulty level on the GMAT.

From Manhattan GMAT Blog:

http://manhattangmat.us2.list-manage.co ... 2fcbed43a2

"So what do these numbers means? And how do they change over time? The scaled scores don’t actually change:

"A Quant score of 43 in 2002 represents the exact same level of ability as a Quant score of 43 does in 2011."

The percentiles, though, are more narrow in scope:

"Percentiles report ability relative to test takers from the past three years, calculated from a rolling average, and thus can change over time."

As the pool of test takers changes over time, the percentile rankings associated with specific scaled scores can change. If more highly-educated people start to take the exam, for example, that shifts the relative rankings of the entire pool - and that is exactly what has been happening over the last 5 to 10 years. The GMAT’s worldwide expansion has allowed a greater pool of highly-educated testers to access the test, so the quality of the entire testing pool has been shifting "upwards.""

[StaceyKoprince]

Hi, Tom, hope you're doing well.

I stand by what I said before: that most students won't really know what I'm talking about when I say "this question is a 42." What is the "absolute level of ability" needed to score 42, 10 years ago or today? Nobody outside of those at GMAC can answer that.

So if we're going to have an easy conversation, we're going to use percentiles or even the 3 digit scores as a proxy for that.

On top of that, while it's true that the percentiles change and the scaled scores still represent the same "raw ability," it's also the case that the schools change what they'll accept over time. It isn't the case that schools that accepted a 42 10 years ago will still accept that same "raw-ability" score today. They want the best they can get out of the available pool, so their requirements for "raw ability" actually do change as the percentiles change. If I want to get into a school that wants 90+ percentile people, then I need to be able to answer questions at that level.

We might be having an apples-to-oranges discussion here. I'm talking about having a practical conversation that's precise enough for the needs of most students (to get in). It sounds like you're talking about having a technically-correct academic conversation. There's room for both of those activities on the planet. :)