Monthly Archives: July 2009

a ridiculous, costly GMAT myth

Before I make the main point of this blog post, a few disclaimers: 1) I have absolutely nothing against GMAT prep classes; on the contrary, I think that they’re more cost-effective than a private tutor for many students. 2) I have absolutely nothing against any of the major GMAT test-prep companies. In fact, I had an absolutely wonderful experience working for one in the early 2000s. 3) The following is meant to be informative, not bitter or critical.

OK. Just wanted to make sure that I made it clear that I’m not trying to be a jerk here.

But I had to share this with anybody who might be interested. I just started working with a new student here in NYC, and she’s using her GMAT tutor as a supplement to a test-prep course. Her classroom teacher apparently said the following: “If you’re debating between A and D on a GMAT question, the answer is probably A.”

Whoa. Really? This deserves a special place in the Crappy Test Advice Hall of Fame, alongside the old adage about picking C whenever you need to guess.

I would be shocked if the GMAT doesn’t randomize its answer choices. Randomizing might be a little bit tricky on, say, data sufficiency, but I’m pretty certain that A and D are both correct about 20% of the time on GMAT. It’s an awfully sophisticated test; why would they do something stupid like making A the answer most of the time?

As soon as my student said this to me, I probably got a crazed, skeptical look on my face. I started searching for any conceivable explanation… was it on sentence correction, and the teacher just meant to say that people are often overly reluctant to choose A? Was it on data sufficiency, where the same might be true? Nope. My student insisted that they were discussing a plain old math problem solving question.

Again, (here comes disclaimer #4), this is second-hand information, and I’m wondering whether something got lost in translation from teacher to GMAT student to GMAT tutor. But if a major test-prep company is making blanket statements about A (or D or C or any other letter) being a fundamentally better guess than other choices, that’s pretty scary stuff.

exhausting the OGs: a case study

As a private GMAT tutor who presents himself in unconventional ways, I tend to attract unusual students. People who are “average” or “typical” tend to do well in prep classes; there’s no need for them to pay for a private GMAT tutor. Often, I find that I’m the GMAT resource of last resort. I get lots of calls from people who have already done everything they can (self-study, prep courses, perhaps repeated re-takings of the GMAT) and can’t think of anything else, so they call an independent tutor, hoping that I can offer some help.

Usually, I immediately have an idea of what to do for a student. Some people complain about their GMAT verbal scores, but haven’t had a lick of sentence correction grammar training–it’s easy to see that there is, at the very least, a content issue in these cases. Some students reveal major deficiencies in their math preparations, and these are also easy enough to fix. I’m about to start working with a student who has scored 330 and 430 on her first two practice tests, and is in the middle of a prep class; at least as a starting point, she just needs a little bit of one-on-one attention to help her grasp some fundamental concepts and build her confidence. Basically, there’s an obvious starting point with nearly every GMAT student I’ve ever worked with, and I can usually figure that out over the phone.

But when I first came to NYC last month, I met a student who might, in some ways, know more about the GMAT than I do. This guy has done every single official GMAT question (GMAT Official Guide 11th edition, GMAT Official Guide 12th edition, GMAT Verbal Review, GMAT Quant Review, GMATPrep, GMAT Focus)… twice. And he’s taken all of the Manhattan GMAT tests twice. He kept a journal of all of his mistakes. And he’s taken the real GMAT test twice. I couldn’t ask for a more thorough course of self-study.

Strangely, the poor guy (let’s call him Mr. V) is stuck with a sub-650 score, with particularly weak outcomes on the verbal section (low 40s on quant, but he was right around 30 on the verbal). Apparently, Mr. V called two other independent GMAT tutors before contacting me, and both said that they couldn’t do anything to help him. And I can understand why. What can possibly be done to help this guy, when he has apparently left no stone unturned? He has a strong work background, but is aiming for a 700+ GMAT score so that he can have a shot at top MBA programs. Jumping from, say, 620 to 700 is no joke under the best of circumstances, but it’s one hell of a task for a student who has already devoured every retired GMAT question that has ever been published.

I was honest with Mr. V from the start: I’m not sure how well any of this will work, but I gave him a smorgasbord of GMAT options. Ultimately, we decided to do a full run-through of all of the key grammar concepts on sentence correction, as well as a quick tour through the math content to see if we could find any holes in his understanding of the fundamentals. Beyond that, we realized that we needed to squeeze every possible drop of knowledge out of the official GMAT material, and we had to find some way to supplement the verbal content with outside sources.

For sentence correction, I’m forcing poor Mr. V to identify every single mistake in every wrong answer choice. Much of the material in the OGs is far easier than what he saw on the actual GMAT, and I realized that he probably hadn’t really trained himself to find every type of error. I think this approach is driving him nuts, but seems to be helping somewhat. (I’ve given the same assignment to other students, with mixed results.)

On the math section, we’ve discovered that Mr. V is an algebra genius, but that he gets tripped up by some of the logic- and word-based material. Combinations, permutations, conditional probability, and venn diagram questions tend to make him see double. So we’re working through some methods to make him more systematic in his approaches to those questions, and we’re both digging around to find as many additional, 700-level problems as we can for those topics.

The other two GMAT verbal sections, as is often the case, have proven to be more problematic. I’ve shown Mr. V several different tactics for approaching GMAT critical reasoning, most of which involve creating some sort of visual guide for navigating the questions. I’m not sure how well the tactics themselves are working, but we decided to make him suffer through LSAT logical reasoning questions alongside the OG critical reasoning material. At worst, the LSAT material will make the GMAT seem a little bit easier and more straightforward; I hope that he’ll look at the verbal section of the actual GMAT and be a little bit less intimidated by the convoluted language. At best, a steady practice diet of LSAT might actually cure him of his logical errors.

Mr. V’s test date is still a few weeks away, and I’ll probably be almost as nervous as he will be. I’ve certainly helped some unique individuals reach their MBA goals, but Mr. V has forced us to rewrite the GMAT test prep playbook. I just hope that the first draft of the new playbook is good enough to get him the points he needs to achieve his MBA goals. I’m fascinated, and will update this page with his progress, regardless of whether his result ultimately makes his GMAT tutor look good.

comparing GMAT tutoring rates (no, not mine)

I was just thinking about the good ol’ days when I worked for a big test-prep company, and I was curious to see what the going rate might be for GMAT tutoring purchased through similar companies. In case you’re curious, here are GMAT private tutoring rates for the four biggest names in GMAT test-prep, presented in alphabetical order:

Kaplan: rates range from about $133/hr (with purchase of a 35-hour GMAT tutoring package) to $163/hr (if you buy only 15 hours).

Manhattan GMAT: $185/hr if you purchase 25 hours, $215/hr if you only buy 10 hours of private tutoring

Princeton Review: prices range from $150-$450/hr, depending on whether you choose a “standard” GMAT tutor, a “master” GMAT tutor, or a “premier” GMAT tutor.

Veritas: about $164/hr with purchase of a 42-hour GMAT tutoring package, $175 if you go for 28 hours, $200/hr for 14 hours.

These rates were taken from each company’s own website on July 15, 2009. If asked, I searched for GMAT tutors in New York City–it’s possible that rates are different outside of NYC.

I’ll let you draw your own conclusions on this. All I’m going to say is that I earned less than $20/hour in 2002 when I worked as a test-prep tutor and classroom instructor for one of these companies, and I was given very little training by the company. I’m not saying that it isn’t worth the price–after all, these companies usually include a fat wad of books and materials in the tutoring package, and it’s arguably worth the extra money to hire a familiar, “reliable” company. I’m not going to tell you that independent GMAT tutors are necessarily better for everybody, but if you go with one of the big GMAT test-prep behemoths, please shop carefully.

five-minute rule for MBA essays

When I was a teenager, I worked in a crappy chain diner in the Midwest, where I received my first introduction to the “30-second rule.” Over the following decade or so, I heard varying versions of the 30-second rule in bars and restaurants: there was the 5-second rule, the 10-second rule, and–for the slowest and most unethical restaurant staff–the five-minute rule.

If you have to ask: whenever an employee drops something in a restaurant, somebody is likely to shout “30-second rule,” meaning that whatever fell on the floor is still “good” if you pick it up within 30 seconds. (Assuming, of course, that no customers or managers are watching.) Silverware? No problem, it can go out to the customer’s table if it’s picked up within 30 seconds. A hamburger fell on the ground? No worries, we have 30 seconds to pick it up and put it back together. A slice of pie? Trickier, but definitely doable if you scoop it up, put it in the microwave, and top it with some ice cream.

You’ll never eat out again, right?

And does this have anything to do with MBA admissions?

Yeah, sort of. It’s midsummer, which means that the most ambitious MBA applicants are beginning to throw all of their spare time into the GMAT. The earliest first-round deadlines are less than three months away, and this is the time of year when you might be thinking, “okay, I’ll spend two months studying for the GMAT, and once that’s done, I’ll focus on my essays.”

I tell everybody the same thing: don’t wait until the last minute to start your MBA application essays. Many of you probably know this, but writing these essays is exponentially more time-consuming than you might think. A set of four, 400-word essays might not sound like much, but it can be unbelievably tough to put your entire life in a flattering, readable, 1600-word package, especially if you’re trying to avoid sounding like an egomaniac. Almost every applicant (and this goes for non-MBA applicants, too) underestimates the amount of time needed for MBA applications, especially if your energy is simultaneously sapped by the GMAT.

If you’re still doing battle with the GMAT and your MBA application deadlines are getting close, I strongly recommend employing the five-minute rule. (Every time you drop food on the floor, wait five minutes before picking it up. It’ll taste great, and build your immune system! Oops, sorry, wrong industry…) Every day, spend about five minutes on your essays, even if it just means brainstorming a little bit, writing a phrase here or there, taking a couple of notes, or crossing something out. Sometimes, you’ll be inspired to do some serious writing, and that’s great. At the very least, you’ll keep yourself engaged in your MBA essays, and they’ll always be somewhere in the back of your head.

This might sound silly, but many professional writers will tell you that they come up with their most brilliant ideas and phrases at the most random times–in the shower, at the gym, walking the dog, or whatever. Their “ghost writer” (i.e. their own subconscious) is doing most of their work, but they have to make sure that they “tell” the ghost writer what to work on.

I’m not going to tell you that you’ll be able to completely finish your essays in five minutes a day, but I guarantee that you’ll make some great conceptual progress if you’re 100% disciplined about looking at your MBA applications for a few minutes every day. By the time you’re finished with the GMAT after (hopefully) a month or two of studying, you’ll at least have a great set of notes and outlines, studded with brilliant phrases. At that point, the task of turning your MBA essays into something truly spectacular will be much, much easier.

what? the test-prep industry isn’t totally awesome?

I just read a great article in the Wall Street Journal that pretty much hits it on the head when it comes to the test-prep industry: It doesn’t specifically mention the GMAT, but it’s still an interesting commentary that applies, in some ways, to everything in the test-prep industry.

Basically, the article quotes a bunch of academic studies that suggest that the average score improvement from SAT and ACT prep courses is minimal: about 30 points on the SAT, and less than 1 point (!!) on the ACT composite. The article, based on some solid reporting from students at Lowell High School in San Francisco, also accuses test-prep companies of rigging their “diagnostic tests” so that they can claim that their students had huge score improvements.

They wouldn’t really do that, would they?

Um, yeah, they would. It’s the oldest trick in the test-prep book, and I’m surprised that parents and students are still duped by promises that a certain private tutor or test-prep class “will raise your score by 240 points… guaranteed!” I’ve been teaching test-prep for nearly a decade now, and I can tell you that there are never any guarantees–some students, for a host of reasons, will never gain 240 points on their SAT score, no matter how brilliant their test-prep course or private tutor may be. (Similarly, there are some people who will never gain 240 points on their GMAT score.) And yes, many companies still rig their “diagnostic” tests so that they can claim that you made a huge score jump.

Don’t get me wrong: score increases of 240 points happen on the GMAT, but they’re exceedingly rare. Familiarity with the test will definitely improve your score, and a thorough review of basic test content is always productive. A great private GMAT tutor will help you with the psychological aspect of testing, and any good teacher will help you to get a grip on the trickiness inherent in the test. But everybody is different, and there are limits to what prep courses and private GMAT tutoring can achieve for any given individual.

All I’m saying is that it’s good to be wary of the claims and statistics. The lively forum debates help keep GMAT test-prep companies honest, but there are still plenty of snake-oil salesmen in our midst.

why MBA rankings are (partially) BS

When I was a teenager, I used to hate talking about college admissions with my father. He was absolutely convinced that there was, somehow, a big difference between the #12-ranked university and the #14-ranked school, based solely on US News & World Report’s annual list. On multiple occasions, I’ve told him that he’s completely full of crap.

Now, I’m going to tell you why he’s full of crap.

First of all, keep in mind that MBA and undergraduate rankings are based on completely arbitrary formulas concocted by random journalists. (Click here for some commentary on the formulas themselves.) I’m not saying that the basic components are necessarily flawed (who would disagree with the notion that salary increases, reputation, and student selectivity are good indicators of the quality of an MBA program?), but there’s some randomness in the way that any particular list might choose to quantify and weight these measures. They should always be taken with a grain of salt.

And then there’s my favorite indicator of the stupidity of MBA rankings: the contradictions among the various lists. Here are a few of my favorites:

Dartmouth (Tuck): #1 MBA program according to Forbes, #3 in The Economist, but #12 in Business Week

Southern Methodist: #18 in Business Week (WTF??), #47 in US News and World Reports, and unranked in The Economist or Financial Times

Carnegie Mellon (Tepper): #15 in US News & World Reports and #24 in the Financial Times, but #5 (and higher than Harvard!) according to a Wall Street Journal survey of recruiters

Berkeley (Haas): #4 in The Economist, #7 in US News, #10 in Business Week, #16 in the Financial Times

Yale: #9 in Financial Times, #10 in US News, #15 in The Economist, #24 in Business Week

Northwestern (Kellogg): #3 in US News and Business Week, but #10 among US schools in the Financial Times… and only #21 globally

Would anybody out there really believe that Kellogg isn’t even among the top 20 MBA programs globally? Would any of you put Tepper above Harvard on any list? Would any of you dream of putting Southern Methodist above Yale, as Business Week did?

All I’m saying is that these MBA rankings should be taken with a big, fat grain of salt. If you’re thinking that you want to go to a “top 10 b-school” or a “top 20 MBA program,” make sure that you don’t rely on just one of these silly, arbitrary lists.

my quest for 800… er, 790

I was just thinking about my previous entry about how hard it is to score an 800 on the GMAT, and realized that I was arguably being a little bit too impersonal about it. For what it’s worth, I’m definitely gunning for an 800 next time I take the test. Is that a realistic goal? Probably not–I’m a precise test-taker, but I’m not sure that I’m quite precise enough to get an 800 on the GMAT. But I’d be lying if I said that I wasn’t going to try.

(Though I might screw myself over in a small way if I manage to get an 800. Test-takers who earn an 800 are banned from taking the GMAT for five years, which means that I’d miss out on my annual visits to the testing room. Boy, that would suck.)

And just to recap, my GMAT tutoring career began with a subpar practice test when I worked for a large test-prep company, and I slowly have worked my way through the 700s. My most recent GMAT score was a 780 (51Q/47V). If there’s room for me to improve, it’s probably in the verbal section. I would love to hit a 51 in verbal, but I’ve been humbled enough times already, and I don’t think that I’m the guy who can pull that off.

So what would a GMAT tutor do to improve on his own verbal score of 47? The same things, basically, that I ask my students to do once they’ve ripped through all of the GMAT official guides. Lots and lots of LSAT practice to build skills and stamina for the hardest critical reasoning and reading comprehension questions, and then some polishing of my grammar and usage, based largely on (fun!) resources such as style guides and grammar manuals. It never hurts to go back through official SC questions with a fine-toothed comb, but I think we all agree that the hardest stuff in the official guides and GMATPrep tests do not even begin to prepare you for the real thing… especially not if you’re shooting for an insanely high GMAT score.

My schedule is still funky as I adjust to my new lifestyle in NYC, so I don’t know when I’ll have the time to re-take the test. Last year, I took it on about 24 hours’ notice, and I’ll probably do something similar this time around: when the mood strikes (and my schedule allows), I’ll jump on the first GMAT appointment I can find. As soon as I do that, I’ll post a few entries and share my experiences.

Brutal SCs

One of my favorite new GMAT students in NYC asked me to take a look at a set of verbal questions that have been floating around the internet for awhile.  The file is called “Brutal SCs”, and it contains 70 GMAT sentence correction questions.  They’re supposed to be brutal, I guess.  And they are… just not in the way that the author (or compiler?) apparently intended.

I’d be lying if I told you that I analyzed every single question in the document with extreme care, but I did take a reasonably thorough survey of the first 25 sentence correction questions.  I found major typos in four questions, and a fifth question had a misuse of the phrase “due to” in the non-underlined portion of the original sentence.  These mistakes may or may not have impacted the answer, but they definitely made me question the quality of the resource.

More importantly, several of the answers seemed to be just plain wrong.  At least two questions (#21 and #22, if you happen to have a copy) are very clearly flawed, and I’m deeply suspicious of several others.  In short, at least 20% of the questions in this resource contain mistakes of some sort, and I suspect that this figure is an understatement.

The bottom line:  stay away from Brutal SCs, and find another way to practice your GMAT sentence correction skills.

how hard is 800?

Before I started to tutor in NYC, one of my old students asked me how long she would have to study to get an 800 on the GMAT.  I told her that she would be crazy to even think about trying.

I meant her no disrespect at all.  She is about to start her MBA program at the University of Chicago, so it’s safe to say that she’s an extremely talented woman.  But achieving an 800 on the GMAT is ridiculously tough.

Let’s start with some stats from the official GMAT site:  fewer than 50 students (out of roughly 250,000) score an 800 each year.  If you’re keeping score at home, this means than an 800 score would fall somewhere above the 99.98th percentile.  And that’s assuming that there are close to 50 people who get a perfect GMAT score–the figure could, in reality, be much smaller.

Now, let’s talk about the torture that GMAT will put you through if you start to get close to a score of 800.  (For the sake of illustration, I’m going to talk mostly about the quantitative section here.  The same ideas apply to verbal, but are trickier to explain.)  On the GMAT math section, formulas are of limited value on difficult questions.  You obviously need to know all of the fundamentals, but it’s easy enough for GMAT writers to cook up an evil question that defies formulaic thinking.

At the highest levels, there are some extremely wicked questions.  Last time I took the GMAT, I ran into an absolutely brutal series notation question that destroyed me.  I spent eight minutes on it, and I still had no idea how to solve the stupid thing.  It was so complex that I couldn’t even remember the question after the test.  And believe me, I was trying like crazy to wrap my head around it, so that I could study it later. Nope.

So basically, the GMAT writers are brilliant enough to lay waste to almost everybody (including arrogant, Stanford-educated GMAT tutors with nearly 10 years of experience), at least some of the time.  I might have gotten 35 questions correct on the quant section last time around (though I’m sure that I made some dumb mistakes somewhere… I can say that I was confident in 35 of my answers, and definitely missed at least two questions), and I scored a 51.  But the GMAT test can still stump me whenever it wants to, apparently.

And let’s talk about that raw quantitative score of 51.  I’ve never heard of anybody getting higher than a 51, and I’ve been stuck on that score for a long time.  Strangely, it’s theoretically possible to get a raw GMAT quant score of 60. (Visit for an unsatisfying discussion of this topic.)  But I’m pretty sure that I understand what it means to get something higher than a 51:  you need to get those evil questions right, somehow.  I suspect that the impossibly hard GMAT quant questions would have kept coming last year, even if I’d managed to get that series notation beast right.

I’ll go back in to take the test again in the next few months, but the odds that I’ll ever get to 800 remain slim.  There’s absolutely no room for silly errors, and I would have to somehow slug my way through the dirtiest of GMAT questions.  You won’t, of course, find any truly brual, 800-level questions in any official GMAT material, including the GMATPrep and GMAT Focus tests.  So there’s no way to practice, really, once you’re beyond a certain level.

I’m sure that very few of you are really gunning for an 800 (it won’t, in all honesty, really help you get into an MBA program), but I find it fascinating that the writers of the GMAT make it so tough to get a perfect score.

nobody reads your essays, part II

Most GMAT students I’ve met really don’t worry all that much about the Analytical Writing (AWA) portion of the test, but I occasionally meet somebody who is absolutely neurotic about this part of the GMAT. In my opinion, there’s rarely any need to spend more than a token amount of time on AWA.

First of all, one of the main reasons why MBA programs care about the GMAT is because of rankings. Unfortunately, US News & World Reports will look less favorably on a b-school if the program has low average GMAT score, so admissions committees are forced to obsess over your GMAT composite results. (Of course, there are plenty of other reasons for admissions committees to care about your GMAT score, but that’s a story for another day.)

Of course, the AWA section has absolutely nothing to do with your composite score, and it therefore has no impact on MBA rankings. You don’t want to completely screw it up, but there’s no reason to believe that your AWA score of 5.0 would keep you out of Stanford or HBS. The adcom might start to wonder if your AWA score is, say, 2.5, but I see no reason to worry if your score is reasonably high. They’re looking for business leaders with decent communication skills, not novelists.

So what does it take to earn a “reasonably high” score? Not much, really. I’ve seen some really, really bad writers earn scores of 5.0 or 5.5. They took advantage of the fact that nobody really reads your GMAT AWA essay.

As you probably know, GMAT essays are graded by a computer and by a human; if there’s a discrepancy between the two scores, then an additional human reads the essay. Obviously, the computer’s exact scoring system is a closely guarded secret, but I’m convinced that the program primarily looks for structure using basic keyword algorithms. From there, it probably counts things like the average number of words per sentence and the average number of letters per word, and then it checks for grammar, usage, variety of words used, and perhaps spelling. I suspect very strongly that structure is the most important part of the AWA scoring system; spelling might not even count at all. (Again, I’ve seen some horrid spellers get really good scores.)

The “human,” for his part, is ultimately looking for the same things as the computer when he grades the GMAT AWA: structure, clarity, correctness, richness of language, etc. But this human probably doesn’t actually read your GMAT essays any more closely than the computer.

Again, imagine that you’re the poor slob who gets paid $20/hour (as of a year or two ago, GRE and TOEFL graders earned $20/hour; I don’t know whether the rate is similar for GMAT) to read bazillions of these AWA essays, most of which are tragically bereft of interesting content. Are you going to read every single word? Hell, no. You’ll skim the introduction, maybe read the first sentence of each paragraph, and then maybe skim the conclusion. (And if anybody asks, you’ll tell her that you read every single word with great care.)

When I wrote about the people who read (or skim) your MBA essays, I claimed that you should strive to make your essay interesting–you want them to read it carefully, right? For the AWA, you don’t necessarily care if the human grader (or the GMAT computer) actually reads your essay. You just want to earn a decent score, with a minimum amount of exertion. If you can conserve energy during the GMAT AWA section, you’ll be much sharper when you battle the (far more important) quant and verbal sections.

So focus on structure on the AWA, even if that means that you produce an essay that would, under normal circumstances, seem like a crappy, contrived piece of writing. Don’t beat around the bush: every paragraph should start by clearly stating the main point of the paragraph. The entire AWA essay should start with a crystal-clear thesis statement. Somebody should be able to read the first sentences of each paragraph, and still basically understand the entire essay. Just make sure that you leave a few minutes to proofread your AWA essay, simply to avoid committing any egregious grammar, usage, and spelling errors. (If this means that you don’t write a terribly long essay, that’s fine–quality matters more than quantity.)

Honestly, it’s almost that simple. If you lay out your argument in a completely clear, five-paragraph format, you’ll be in good shape, especially if your writing mechanics (grammar, usage, etc.) are basically solid. You might ultimately write an essay that is about as much fun to read as a phone book… but fun isn’t really the point of the GMAT, is it?