# GMAT Tip of the Week: Think Like the Testmaker

In this week’s GMAT Tip, our instructor shows how thinking about the test as a whole can accelerate your understanding of its individual parts, and more importantly how that can help you study efficiently and effectively.

At Plymouth-Salem High School in the 1990s, a chemistry teacher by the name of Mr. Barnes was a divisive character.  He may not have been anyone’s absolute favorite teacher (read: he never brought in candy, showed movies, or held class outside, the three cornerstones of favorite-high-school-teacherdom) but he was most certainly some students’ least and a beloved figure for others.  He challenged students with rigorous standards and assertive discipline.

If you bought into his style early, you likely realized that he was a great teacher particularly for those who planned to go on to college and beyond.  He didn’t sugarcoat difficult topics or alter his test schedule or homework assignments to make  things easier for athletes or the prom-bound to keep up with extracurriculars.  His quote as it pertained to your responsibility as a student and studier was always “sometimes you have to rob from Peter to pay Paul”.  And that’s likely your challenge as a GMAT student: sometimes as you study for the GMAT you’ll have to rob from Peter to pay Paul.  The trick is — who is Peter and who is Paul?

What Mr. Barnes (and other great teachers) means by that quote is that time and attention are limited resources and it’s just a fact of life that you will need to make decisions about how to allocate them.  In his world, if you want an A in chemistry class you may need to sacrifice an extra-credit project in your English class (or just an episode of your favorite TV show).  As it pertains to the GMAT, you only have so much time and energy to devote to work, fitness, social relationships, b-school research, and GMAT study.  And breaking down the GMAT and its  topics, you’ll have to make some decisions on whether to rob from Peter (maybe quant) to pay Paul (maybe Critical Reasoning).  What you need to know about that decision is this:

Most people make that decision poorly.

There are several examples of choosing the wrong subjects to study.  Often those who excel on and enjoy the quantitative section will go off in search of the hardest possible math problems they can find, enjoying the process of solving challenging number properties and data sufficiency problems while they put off that undesirable task of learning sentence correction.  Others will take practice test after practice test, monitoring their scores without ever digging deeper to determine how they can best improve those scores.

But most common among the poor uses of limited GMAT study time is the repetitive study of facts, formulas, and grammar rules (particularly idioms).  Students recall the experience of cramming for tests in high school and college (perhaps even for one of Mr. Barnes’ chem tests) and replicate that process, poring through flashcards and margin-notes to fill their heads with knowledge.  But in doing so they miss this crucial fact: the GMAT isn’t testing your knowledge.  And we have proof.

As we posted yesterday, representatives from GMAC have explicitly said that the GMAT is a quantitative reasoning and verbal reasoning test (with their own underlines on reasoning) and have even offered the quote:  “We’re not testing your geometry and algebra knowledge.  If you’re really lacking in that knowledge you’ll get some questions wrong, but we’re using those topics as way to test your higher order thinking.”

What’s more, when you consider the role of the GMAT scoring algorithm, its job is to continue with each question to hone in on its estimate of your ability level.  The algorithm essentially:

1) Constantly updates and compiles statistics so that it knows which items can differentiate between users at different levels (say, a 640 examinee from a 660).

2) Chooses and assigns you items that will help create an estimate of your ability level (e.g. “you’ve answered a 550 correctly, a 650 correctly, and a 700 incorrectly…let’s see if you’re above or below 670 with the next item)

3) Continues to get more specific with its items to hone in on your ability level, obviously readjusting for “false positives/negatives” (if you happen to miss a 500 question early in the test, the exam will within a few questions realize that you’re actually better than that and start feeding you tough questions again)

We’ve mentioned here before that the scoring algorithm works a lot like a more sophisticated game of 20 Questions. It’s essentially asking:

Are you above 500? (YES)

Are you above 600? (YES)

Are you above 650? (YES)

Are you above 700? (NO)

Are you above 670? (YES)

Are you above 690? (NO)

Are you above 680? (NO)

Are you above 660? (YES)  —> Score is 670 (NOTE – this is a simplification…the GMAT takes more questions than this to do so, to account for your correct guesses and the occasional silly mistake that could skew the data, but ultimately this is what the algorithm is doing).

Well, with that 20 Questions analogy in mind, think about another game of 20 Questions in the more common sense.  Say that you were trying to guess an actress that I was thinking about (maybe Minka Kelly).  You’d ask questions like:

1) Is this person female? (YES)

2) Is this person alive? (YES)

3) Is this person famous? (YES)

4) Is this person famous for business or government? (NO)

5) Is she an entertainer? (YES)

6) Is she a musician? (NO)

7) Is she a television actress? (YES)

Now, to this point your questions have all taken one tone – you’re narrowing in on this person’s station in life…female, alive, entertainer, TV actress.  Each question is relevant to the last – that’s how you narrow in your estimate of the person I’ve selected in this game.  So there’s no way  that you should ever ask an eighth question like:

8) Do you like Pepsi?

That question is irrelevant to your quest.  It’s a wasted question, and you only get to ask 20.  So, similarly, you wouldn’t at this point ask:

9) Has she ever dated someone with a barbed wire tattoo on his left bicep?

That question is far too specific and out of scope to be helpful.  And that’s what an obscure-knowledge type question at this point on the GMAT would do.  To this point on the GMAT, the scoring algorithm has started to figure out an assessment of your abilities to think reasonably and logically, to think critically, to solve problems effectively.  An obscure-knowledge-based question doesn’t add any value to that assessment in any way – the thought process is too out-of-scope, the relevance to the previous or next question just isn’t there…it wouldn’t make any sense for the GMAT to so quickly change the scope of its questioning. In order for the CAT algorithm to perform to its ability – remember, it only gets 41 verbal questions and some of those are unscored and experimental (for the purposes of gathering that ability-level data) so it may only get 34-35 questions to assess your ability. And on questions that are way, way above your level you still have a 1/5 probability of answering correctly in spite of yourself, so the algorithm needs to account for that.  It cannot waste questions!

So know this about the GMAT:  there has to be a common thread between questions, and that thread is logical reasoning.  And as you study, you should force yourself to emphasize that thread because it’s scalable – logically reasoning through sentence correction sharpens that part of your brain for problem solving and critical reasoning. You should note that the “misplaced modifier:

While studying for the GMAT, I think you should focus on logic more than you focus on knowledge.

Isn’t so much grammatically wrong…it’s just horrendously inaccurate.  While studying for the GMAT, YOU should do these things…I’m not studying for the GMAT.  You are.  And the GMAT will reward you for seeing that inaccurate meaning that’s also present when verb tenses are tested (could those things really happen in that order?) or comparisons are made (can you really compare those two things?).  There are, indeed, grammatical rules that you should know, but remember that on a reasoning test they can’t be arbitrary.  The GMAT can really only test you on the rules that have a logical basis to them.  They have no incentive – and actually have a huge disincentive since a wasted question is a wasted opportunity – to test you on “either you know it or you don’t”.  The GMAT is a test of higher-order thinking, and on Bloom’s Taxonomy of Educational Objectives (more on that in a blog post next week) “remembering” is the lowest level of thinking, well below “apply”, “analyze”, and “create” -the objectives that the GMAT wants to reach.

Which brings us to Peter and Paul (sorry, Mary).  As you study from the GMAT you will sometimes need to steal time from Peter to pay Paul.  You’ll need to allocate study time, and remember that while it might feel natural to study by memorizing obscure information, the GMAT almost by definition has to test you on logical reasoning.  So use your time that way – study how you think, how you can apply logic to each rule (on the math side, too) to better master it and employ it.  Focus on how you apply that logic and knowledge and not just on knowing it.

As an homage to Mr. Barnes and excellence-demanding teachers everywhere, we won’t sugarcoat this test. The GMAT is hard – it has to be to serve its purpose as a sorting mechanism for graduate school.  But one of the items that makes it hardest is that people don’t know how to study for it.  Well, now you know.  Think like the testmaker – they’re forced to use logic in creating a test that assesses what it seeks to test.  So use that logic to your advantage; you know what they’re testing.

Getting ready to take the GMAT soon? Take a look at our GMAT prep courses and start your journey toward a great score. And, as always, be sure to find us on Facebook and follow us on Twitter!