Monthly Archives: August 2009

is GMAT verbal arbitrary and subjective?

I received an interesting  GMAT-related email this weekend, and wanted to share the question with anybody who might be interested:

… I don’t see how I can improve my level in verbal. This section looks so arbitrary and subjective. Most official answers in SC, RC and CR are highly arguable. It also explains why most people get a low score in verbal while there is no required knowledge for this section contrary to quant. I don’t understand how people manage to nail this section… I never scored steadily. Yesterday I made a 40 on a prep test, today I made a 20… Is there some magic to understand what test makers expect from the candidates?

There are some great questions in here, some of which, unfortunately, do not have great answers. But I’ll do my best to pick apart some of the topics in here.

First of all, I think that GMAT sentence correction is not at all subjective. In theory, the GMAT tests your knowledge of standard, written American English grammar and usage. In practice, I can see why many of the principles tested in GMAT sentence correction seem completely arbitrary. As somebody who has worked for much of the last decade as an editor and GMAT tutor, I can assure you that the GMAT concerns itself with a lot of rules that would slip right past the vast majority of copy editors. A sizable minority of “wrong answers” on GMAT would be completely fine in most major American newspapers, and an even larger minority of the “wrong answers” would be considered completely acceptable spoken English, at least in some parts of the English-speaking world.

However, I don’t think that the arbitrary topic selections should be confused with subjectivity. If you pick up a good usage dictionary or style guide (Bryan Garner’s The Oxford Dictionary of American Usage and Style is my personal favorite), you’ll find endless collections of rules that most people never think about. The GMAT applies these rules consistently–even if the GMAT tests a seemingly arbitrary subset of those rules.

For most people, the easiest way to make progress on verbal is to very thoroughly learn the major grammar and usage rules that appear on the GMAT. Manhattan GMAT’s sentence correction book is probably the best self-study resource for this purpose; PowerScore’s book isn’t too bad, either. If you struggle to learn grammar from a book, then a good  GMAT tutor could help straighten you out; the GMAT forums also have some fantastic sentence threads with great explanations.

Similarly, critical reasoning and reading comprehension questions in the official GMAT material are extremely consistent. Sure, they can get outrageously difficult, but the test-writers are trying to test your ability to understand the nuance of language, your ability to notice fine details, and your ability to stick to the internal logic of a passage. If you think that answers to official critical reasoning questions are arbitrary, it might be because you’re allowing yourself to “think outside the question.” Once you spend enough time with official GMAT material (and again, a study partner or a good GMAT tutor might be useful), you’ll start to see how the test is rigid and consistent in its application of logical principles.

If you want to improve your reading comprehension and critical reasoning skills without paying a bunch of money for a tutor, I strongly recommend reading The Official LSAT SuperPrep, or at least the parts that describe logical reasoning (LSAT’s version of critical reasoning) and reading comprehension. It’s not necessarily an easy read, but it might help to convince you that the people who write the GMAT are systematic, even if they are kind of evil.

Here comes the rough part.

Unpleasant dose of reality, part 1: I think that it can be incredibly hard to improve your score in reading comprehension. It takes a ton of practice, and possibly a fair amount of guidance. During my long, ugly history of teaching test-prep (SAT, GRE, ACT, TOEFL, TOEIC… including a stint as a TOEFL book writer), I’ve rarely seen anybody make large improvements in reading comprehension with anything less than a herculean GMAT study plan. You could make moderate improvements in a few weeks, but major gains on reading comprehension usually take months. (Occasionally, I can browbeat a smart-but-lazy teenager into doing much better on the ACT or SAT in a few weeks, but that’s not a skill problem–it’s a focus problem.)

Unpleasant dose of reality, part 2: unless you’re using official GMAT (or LSAT or GRE) material, “knockoff” verbal questions actually ARE completely subjective and arbitrary, at least some of the time. I strongly discourage my students from touching any verbal material other than that found in the GMAT official guide, the official GMAT verbal review guide, GMATPrep, and official LSAT books. (Manhattan GMAT is an exception of sorts, but I still don’t think that anybody should rely exclusively on MGMAT for verbal. Manhattan does an absolutely amazing job, but their material is still designed as a copy of the real thing, and like all copies, the material inevitably has its biases and imperfections.)

So if you’ve taken tests from any sources other than the aforementioned, you should expect your GMAT verbal scores to be completely random. Trust me, it is more difficult to write “real GMAT reading comprehension questions” than you can possibly imagine, and most test-prep companies would never dream of spending the money it takes to produce, test, and polish a truly accurate set of GMAT verbal questions. Focus exclusively on official material for a month or two, and the GMAT will start to look much less arbitrary and subjective.

Thank you for the great set of questions! You’re surely not the only person who has these thoughts, and I wish you all the best with your GMAT studies.

GMAT and GRE test center glitches

Disclaimer: I have no real reason to share this batch of GMAT horror stories, other than to scare you just a little bit for no good reason.

Pretty much everybody faces some serious time pressure on the GMAT quant section, and many people are forced to scramble on the verbal and AWA as well. So if you lose two minutes due to a computer glitch, it’s pretty maddening. Two minutes won’t destroy your GMAT score, but it might make you flustered and cause a cascade of errors.

In the past 12 months, at least five of my private tutoring clients have been affected by GMAT test-center glitches. This past weekend, two of them–one in NYC, one in DC–had issues. In both cases, the dudes working for Pearson VUE (the company that runs the testing centers) had a hard time logging the test-taker back into the system after a break. In both cases, the students lost a couple of minutes for the quant section. Pretty crappy.

In another couple of cases, the GMAT testing center dudes accidentally shortened the test-takers’ breaks by failing to notice when the test-taker was finished with a section. When you’re ready for a break after a section, you’re supposed to tell the computer that you’re ready for a break, and then raise your hand so that you can be escorted out of the testing room. Supposedly, the “escort” didn’t notice when a couple of my GMAT students were ready, and a few minutes passed before the proctor noticed the test-takers’ flailing hands. It isn’t a big deal to have eight minutes instead of ten for your break, but it’s still annoying.

And then there are the computer glitches. In one case, the system crashed during a break, and somehow restarted with several minutes already elapsed in the quant section. (I can’t explain why these problems seem to occur between AWA and quant. GMAT hates you?) Another student faced a really bizarre glitch which prevented him from clicking on certain radio buttons–if I remember correctly, he was literally unable to select most of the answer choices, and had to click “next” with some questions unanswered. He complained at the testing center, but they couldn’t really do anything about it. I don’t think that he even finished the test. In both of these cases, the test-takers called GMAC every single day until they were allowed to re-take the GMAT for free.

Again, I have no real point here. I’m not trying to criticize Pearson VUE; generally, I think that the company does a solid job administering the GMAT, and I’ve had good experiences in their test centers. (The GRE is another matter–last time I took the GRE at a Prometrics test center, I was forced to an old, flickering monitor which made my eyes hurt. By the time I left, I felt like I’d been staring at a strobe light for four hours, which made me a little bit crazy.) As a full-time GMAT tutor who watches students spend craploads of time and money on the MBA admissions process, it’s painful to see people get thrown off by these stupid glitches.  But human and computer errors happen, and all you can do is roll with the punches.  And if the glitches really affect your score, you can always bitch and moan until GMAC compensates you for the errors.

GMAT Focus quirks

For the most part, I’m a fan of GMAC, the company that produces the GMAT. I’ve spent a good chunk of the last decade working as an SAT, ACT, GRE, and TOEFL tutor, and I respect that fact that the GMAT is much more complicated, precise, and nuanced than any of the aforementioned exams. And GMAC is also fairly generous about publishing retired test questions; some of us whine about the GMAT OG’s lack of hard questions, but the official guides for GRE and TOEFL are far worse, offering only a handful of (ancient, moldy) practice questions. So while I generally think that GMAC does its job really, really well, they seem to have been a little bit sloppy with GMAT Focus lately.

In some ways, GMAT Focus is the best of the GMAT practice material. Sure, it’s overpriced ($25 for one test, or $65 for three… basically, you’ll pay almost $1 per question). And I think it’s weird that they sell the tests in groups of three–you can actually take the test four times before questions start to repeat. And you’re given only 45 minutes to complete 24 questions–a little bit odd, since you’ll ultimately need to get used to doing 37 GMAT questions in 75 minutes. But GMAT Focus still contains an impressive bank of difficult, fresh GMAT questions; I’m convinced that GMAT Focus questions are much closer to the real thing than any other practice resource available. If you’re serious about improving your quant score, GMAT Focus is absolutely not optional.

Unfortunately, the GMAT Focus questions are no longer all that unique. Many of them appear in the 12th edition of the official guide, and I have a funny feeling that even more will show up in the next edition of the Quantitative Review guide. So you might be paying $1 for questions that you’ve already purchased in another book. That’s really annoying, especially if you’re already paying for a GMAT tutor.

Even more annoying: the GMAT Focus might not be as representative of the real test as I once thought. For a long time, it seemed that the GMAT Focus would give you a pretty good idea of the composition of the quantitative section–specifically, GMAT Focus included a lot more tricky logic, combinatorics, and “creative algebra” questions than the official guides, and GMAT Focus seemed to be an extremely accurate representation of the actual test. Sadly, something might have shifted, at least a little bit. In the past month, several of my students (all with quant raw scores above 40) whined that the real test didn’t have any of the tough combination/permutation/probability questions that appear on GMAT Focus. It makes me wonder whether the GMAT Focus is becoming somehow less indicative of the actual test. Again, I find that a little bit irritating–if it’s actually true.

Here’s the strangest report of all: one of my GMAT students swears that he saw a GMAT Focus question on the real exam last weekend. He claims that the question was absolutely identical, with the same numbers and everything. He didn’t share the details of the question (that would be illegal), but I find that pretty alarming if it’s true. Is GMAC getting sloppy? Was the “retread” GMAT Focus question tossed back into the real test as some sort of experiment, or as a statistical control for other questions? Or was my student just hallucinating after a long, stressful few hours of test-taking?

Either way, it’s interesting. I still think that it’s worth spending $90 for all four GMAT Focus tests, but I couldn’t really blame a budget-conscious GMAT student who comes to a different conclusion.

fun GMAT toys

I should have posted this a long time ago: the good people over at GMAT Club have created a GMAT score estimator. You can punch in your Manhattan GMAT, Kaplan, GMATPrep, and Princeton Review scores, and they’ll give an estimate of your score on the actual GMAT exam, based on a model that they’ve developed. You can find the GMAT score estimator here. I’m not sure how well it works, but it’s a brilliant idea.

I’ve had far less luck with the GMAT score calculator created by 800Score.com. The idea is that you can punch in your raw scores from the math and verbal sections of the GMAT (on a 0-60 scale), and the website will calculate your composite GMAT score on a 200-800 scale. I’ve tried the calculator a few times using my students’ raw scores from GMATPrep and actual GMAT tests, and the calculator always seems to be off by a little bit.

More recently, I punched in my own scores from the real GMAT. According to the calculator, a quant score of 51 and a verbal score of 47 will give you a composite score of 350, in the 7th percentile. Hilarious. Still a fun toy, but the execution could use some work.

GMAT Verbal Review 2nd Edition

Fresh off the presses: our friends at GMAT just released the 2nd edition of the GMAT Verbal Review. This time, they’ve switched to a sexy blue theme. Much nicer than the lavender on the 1st edition.

Aside from the color change, not a whole lot is different. Most of the verbal questions are the same, but they supposedly removed 75 questions and replaced them with 75 new ones. But actually… they removed 82 questions and replaced them with 82 new ones. You get 7 more real GMAT questions than they promised! Isn’t that exciting?

If you’re looking for some extra verbal practice material, it probably wouldn’t hurt to have an additional 82 “official” questions, especially since the book is so cheap (about $13 on Amazon). But do the lords of the GMAT give us any special hints in the new edition? Are there any systematic changes that might hint at a new composition of the actual test questions?

Unfortunately, not really. I spent a good chunk of time poring over the new questions, and I can’t say that all that much is substantially different. It’s not as if the new GMAT verbal guide is suddenly covered with, say, parallelism questions. It is, however, covered with balsamic vinaigrette. At least my copy is. (A little salad accident happened while I was working through the book. That was my punishment for ordering a salad, I guess.)

Anyhow, here’s a breakdown, by question type:

Reading Comprehension

Questions removed from the 1st edition: 13, 22-28, 55-63, 91-97, 98-105 (32 questions removed)

Questions added to the 2nd edition: 1-6, 18-23, 45-49, 64-70, 91-97 (31 questions added)

Random edits: a few typos were fixed (missing punctuation on 1st edition #40 and #51), indentations were added at the beginning of each passage, and line references were changed as a result (#48, #50, #74, #76, #102). Yes, I’m that anal. Want me to edit your MBA application essays?

Useful conclusions: none, really. I thought it was interesting that they removed a pair of random questions (1st edition #13 and #65) while keeping the rest of the passage intact, but I don’t think that leads us to any stunning conclusions about the makeup of the test. The five new passages include two science readings, two “business-y” readings, and one social science-y passage about W.E.B. DuBois. These replaced four passages, including one science reading, two business-y readings, and one social science-y passage about Florence Nightengale. As a tutor who spends way too much time with these books, I have to admit that I’ll miss Florence.

Brutally honest conclusion #1: anybody get the feeling that GMAC released this new edition just to sell books?

Brutally honest conclusion #2: I’m not serious about missing the Florence Nightengale passage.

Critical Reasoning

Questions removed from the 1st edition: 1, 2, 5, 8, 13, 15, 19, 22-23, 29, 35, 42-43, 46, 54, 55, 59, 60, 66, 68, 77, 82 (22 questions removed)

Questions added to the 2nd edition: 1, 3, 6, 9, 13, 17, 21, 25, 29, 33, 37, 40, 45, 48, 51, 54, 59, 62, 65, 70, 74, 78, 81 (23 questions added)

Random edits: underlining was removed from questions 24, 26, 27, 49, and 79 in the 1st edition. Not that you care, but it looks like some editor at GMAC had to work overtime once or twice.

Useful conclusions: actually, I do see a few semi-significant changes in the critical reasoning part of the test. There are now two (wow!) of the “boldfaced” questions, compared with one in the 1st edition. In the 1st edition, there were six “paired” questions attached to a single passage; all of those have been removed, or at least converted to single questions.

More generally, I think that GMAC has been experimenting with a broader range of question stems for critical reasoning. Once upon a time, CR didn’t go much beyond some basic “strengthen” and “weaken” questions; other phrasings (see 2nd edition #70 and #78) are a little bit more common than they used to be. This doesn’t radically alter the test, but it’s interesting nonetheless.

Sentence Correction

Questions removed from the 1st edition: 3, 14, 17, 18, 21, 24, 27, 28, 33, 36, 38, 43, 47, 50, 52, 54, 55, 58, 59, 69, 74, 75, 77, 81, 83, 87, 89, 90 (28 questions removed)

Questions added to the 2nd edition: 1, 5, 8, 11, 14, 17, 22, 27, 31, 36, 42, 45, 50, 53, 57, 61, 65, 69, 73, 79, 84, 88, 92, 96, 100, 104, 108, 112 (28 questions added)

Useful conclusions: none, really. For the past couple of years, I’ve theorized that the GMAT is trying to make their sentence correction questions more “test-prep proof” by inserting more questions that involve some sort of tricky logic, unusually difficult comparisons, or funny forms of parallelism (i.e. false parallelism traps or opaque parallel structures). I also think that we’re starting to see more idioms, and fewer questions that can be solved just by knowing simple grammar and usage rules.

Although my students continue to tell me that the sentence correction questions on the real test are much harder than in the GMAT official guides, the 2nd edition of the Verbal Review gives us pretty much nothing. The new 28 questions don’t seem to be significantly harder than the 28 that were deleted from the 1st edition. The last few new ones aren’t exactly a walk in the park, but they’re still nowhere near the difficulty level of some of the evil stuff I’ve seen on the real test.

So there you have it. If you’re being really aggressive about your GMAT prep, you might want to buy both books, and then use the question lists above to cherry-pick the non-redundant questions out of one of the books. Other than that, there’s no particular reason to think that the 2nd edition offers anything all that special. If you’re confident with your verbal skills, there’s no need to race to the local bookstore for the 2nd edition–if the 1st is already in your hot little hands, you probably won’t need to bother with the 2nd.

my shortcomings, in plain view

I pride myself on being a brutal realist when it comes to GMAT tutoring. If I think that I can’t help somebody improve his GMAT score, I immediately tell him. If a student isn’t doing her homework, I’ll tell her to stop wasting her money on a tutor. If I think that a student can’t possibly achieve his GMAT goal, I’ll find a polite way to say so.

And if I make a mistake, I’ll admit it immediately. I’m not really into hiding.

I’ve had a surprising number of hits on my original post about Mr. V, one of my students here in NYC who has worked like crazy to raise his score. Basically, he’s seen every single useful GMAT question at least twice. He took the exam twice, and couldn’t crack 640. I’ve tutored him for the past six weeks, and I was absolutely convinced that he’d made some great progress. He was nailing some of my toughest math and sentence correction questions, he was holding his own on the LSAT material, and his final GMATPrep test was well into the 700s.

It took some creativity to put together a good program for him, and I thought that I’d be able to write a gloating post about how well my odd schemes worked. Unfortunately, nothing worked as well as we’d planned. Mr. V took the GMAT last weekend, and his verbal score actually went down. I was pretty shocked and humbled by that. I thought that we had really made him better, and it just didn’t happen. His math score improved, but his composite score went sideways.

Mr. V has been more than gracious, and doesn’t blame me at all for his lack of improvement. (Actually, he posted a glowing endorsement of me on his blog. Thank you for that, good sir.) But I can’t help but feel a little bit responsible. Sometimes, even a dedicated tutor’s best efforts just don’t quite cut it, and I don’t want to run from that fact.

For what it’s worth, Mr. V didn’t see anything surprising on the exam. He said that it was pretty much exactly what he expected: the critical reasoning and reading comprehension were hard but reasonable, the math was predictably tough, and the sentence correction didn’t contain any grammatical surprises. If anything, we suspected that the sentence correction idioms got the best of him–he said that he was unsure whether he chose the right phrases on a number of questions. That, coupled with a little bit of bad luck, probably did the damage on the verbal section. And unfortunately, I didn’t advise him to memorize hundreds of idioms before taking the GMAT. (Nor would I ever recommend memorizing more than a few dozen of the most frequently used idioms.)

Unless Mr. V suddenly gets shy and asks me to stop, I’ll continue posting occasionally about his progress with his MBA quest. He has chosen not to re-take the test, but we’ll keep working together in an effort to make his MBA applications sing. The point of all of his GMAT labors was to get into a great MBA program. And if he can pull that off, the disappointment of the score won’t really matter at all.

snarky Yale college advice, 1975 edition

This has absolutely nothing to do with the GMAT, MBA admissions, or GMAT tutoring. Just warning you.

I was wandering around NYC with some visiting friends this weekend, and we decided that it would be fun to take them to Strand Bookstore, which is one of my favorite places in New York. Eighteen miles of books (supposedly), many of which are piled on sidewalk carts for $1 each. If you’re from somewhere else and you come to NYC, the store is worth a visit.

My bizarre little $1 treasure for today was tattered paperback copy of The Insiders’ Guide to the Colleges, written by The Yale Daily News in 1975, long before college rankings became such a huge national obsession. I was curious to see how top schools were perceived more than three decades ago.

I’m not sure that I learned all that much, but I had a good laugh. The writers are snobby and snarky, and definitely have some Ivy League bias.

I reflexively opened to their commentary on Stanford, my alma mater: “There is a certain trendiness in the air which often stifles any serious attempt to approach an academic problem. The school’s California provincialism can be extremely irritating.

“One manifestation of the student body’s provincialism is their penchant for calling the school ‘the Harvard and Yale of the West,’ or even for terming Harvard and Yale ‘the Stanfords of the East.’ The analogies simply aren’t true.

“…There are also a huge number of students who only care about getting good grades… and another, almost equally large percentage who aren’t really interested in doing anything…. If you want the best education (and the most heterogeneous student body) available anywhere in the country, look to the Ivies.”

Lest they be accused of east-coast bias, the goofballs who wrote this book crapped all over plenty of other schools. In their commentary on Tufts, they said this: “Everybody knows, of course, that it is the goal of most high school students in the Northeast to go to college in Boston. That way you can get a lot of hippie, loose-moraled girls if you are a guy, or a lot of radical committed free guys if you are a girl… and you can get a good deal of dope to tide you over the bad times.”

Fordham: “Fordham University is a Roman Catholic institution, and therein lies its problem.” (!!)

Davidson: “The fact is, many of the students have never been north of the Mason-Dixon line, and occasionally those in the administration and faculty act as if they haven’t either–and what’s more, they don’t care.”

Harvard: “The college atmosphere sometimes seems to ruin those personalities that weren’t warped to begin with.”

Columbia: “…Columbia is unbeatable. But the decision whether to take that beating should be made very carefully.”

Hilarious, right? Of course, the Yale dorks who wrote the book included a glowing four-page review of Yale itself. Not exactly an unbiased piece of writing, but highly entertaining. I’d love to see somebody write a similar book about MBA programs now–can you imagine the lawsuits?

you got GMAT-ed

I’m toward the end of my second full month of tutoring in NYC, and it’s been interesting to see how my GMAT students here differ from the gang that I taught in DC. I had an amazing range of students in DC, including a couple of people who started at or above 700, quite a few others who started in the high 300s or low 400s, and pretty much everything in between.

My first group of NYC GMAT students, on the other hand, consists almost entirely of ass-kickers, gunning for something close to a 700. I’ve met two guys who have gone through every official GMAT question twice; both of them diligently kept notebooks of all of their mistakes, and still didn’t get the scores they wanted. Pretty much everybody else I’ve met has graduated from a Kaplan, Veritas, Princeton Review, or Manhattan course. With only one exception, all of my students have consistently scored in the 600s on the real GMAT, and in the 600s and 700s on practice test.

So all of these guys know the GMAT really, really well. Generally, they start by asking questions about the tough stuff–combinatorics, hard rate problems, set theory, conditional probability, and the most vicious of sentence correction problems. Hardly any of these guys are the least bit worried about their algebra or arithmetic skills when they call me.

I have a few little files of “easy” GMAT questions, mostly consisting of basic algebra, arithmetic, and geometry. Eighth-grade level math, at most. Over the past few weeks, I’ve given the set to most of these veteran GMAT warriors.

How did they do? Well, one GMAT student got a perfect score, which is exactly what he should have done if he wants to get a 700. (I’m very proud of you, Mr. V. May the GMAT gods grant you a 750 and admission to the MBA program of your dreams.) Everybody else got ripped to shreds, missing somewhere between 15% and 30% of the questions. That would be fine on medium-to-hard (say, 600- or 700-level) GMAT questions, but nobody with dreams of admits from Harvard or Stanford should miss this kind of stuff.

Here are a few examples of reasonably easy questions that caused problems:

1. You drop a ball from a height of 16 meters. Each time you drop the ball, it bounces to a level half as high as its starting point. If you catch the ball after the fourth bounce, how far did the ball travel?
(A) 30 meters
(B) 31 meters
(C) 40 meters
(D) 41 meters
(E) 45 meters

2. If x and y are two-digit integers such that x < 40, which of the following is closest to the maximum possible value of xy ?
(A) 400
(B) 1,600
(C) 4,000
(D) 16,000
(E) 40,000

3. If the numbers 13/24, 9/16, 1/2, 2/3, and 5/8 were ordered from greatest to least, the middle number of the resulting sequence would be
(A) 13/24
(B) 9/16
(C) 1/2
(D) 2/3
(E) 5/8

I’m not sure that any of these questions are much more than 500-level questions, but some good GMAT math students screw these up regularly, just because they don’t read carefully, or because they rush through their calculations. Sometimes it seems that arithmetic and algebra questions cause 600+ scorers to immediately think “this is easy, I’m going to destroy this question in 25 seconds so I can move on to harder stuff”… and then they make a dumb mistake. I call this phenomenon “getting GMAT-ed.”

The GMAT writes questions specifically to test your precision, and they’re trying to get you to fall into traps by being imprecise or careless. If you don’t count the bounces carefully, you won’t get (E) for question 1. If you jump to conclusions on #2 (i.e. by misreading an inequality sign or by missing the word “two-digit”), you might not know that the answer is (C). There’s nothing magical about question #3, but it’s easy to get overconfident and make a calculation or comparison error. The answer, incidentally, is (B). Easily 30% of my 600+ students have missed a version of that last one, even though it’s just a simple numerical comparison.

The moral of the story? If you read a GMAT math question and you think that it’s easy, watch your back. Don’t let overconfidence get in the way of your GMAT score. Check your answer twice–it’s always worth spending an extra 15 seconds to make sure that you haven’t done something silly that can damage your GMAT score.