On the Subtleties of GMAT Guessing

August 31st, 2009

Nate Burke is a Content Developer at Knewton, specializing in question creation for the Quantitative section of Knewton’s GMAT prep course.

“The best way to win at Russian Roulette is to not play at all. Or you could just have the other guy go first and then run away quickly.” –Unknown

In the heyday of paper-based tests, the cure-all for the common ill of “getting stuck” on a question was simple: Skip the question and simply move on.  The rationale behind this strategy was that work done for other questions on the test might illuminate simple key concepts that were overshadowed by things like early-morning drowsiness, test anxiety, tip-of-the-tongue syndrome, etc, etc,.

Things have changed. The GMAT, like many other standardized tests, is administered on a computer. Though the question formats have remained roughly the same, the switch to computer-based-testing has rendered the “skip the question” strategy obsolete. On a computer-based-test, it is impossible to skip a question. The best that a time-constrained student with a total conceptual block can do is to guess and hope for the best outcome of what is essentially a game of Russian roulette with a 5-chambered revolver.

An ideal strategy for “getting stuck” within the context of a computer-based test is thus constrained to the narrow confines of the maximum time allotted per question. Consider that at the beginning of this time interval, WITHOUT EVEN HAVING READ THE QUESTION, a test-taker has already been granted a 20% chance of answering correctly by guessing randomly. This fact alone has consequences. If the student is still at a random-guess level of confidence at the end of the 2-3 minute maximum-time-per-question-interval, then all that the student has accomplished on this question was to lower his or her score. Every second that passes without progress lowers the score incrementally (and that’s not even accounting for the added adverse effects of things like stress-induced second-guessing, etc. etc.).

What is a student to do then, should he “get stuck?” If the path to the correct answer is obscured and overgrown, how does one trim the hedges?

Read the rest of this entry »

Jumping to Conclusions

August 27th, 2009

Emily Holleman is a Content Developer at Knewton, helping students with their LSAT preparation.

As the name suggests, the Logical Reasoning section of the LSAT tests your ability to digest and understand different forms of reasoning. For this very reason, the majority of LR prompts are presented as arguments. The ability to quickly and easily identify (or “jump to”) conclusions will allow you to deconstruct the argument and improve your score.  Of course, since this is the LSAT we’re talking about, these conclusions are often presented in confusing or distracting ways. Fortunately, you can usually locate the main idea of an argument by keeping your eye out for certain key words and phrases.

Here are some common words and phrases that introduce the conclusion of an LR argument:

  • Thus/therefore/so/hence
  • It appears that/it follows that/it is clear that/there is evidence that/it must be true that

Read the rest of this entry »

Unfortunate news

August 25th, 2009

I was very saddened to hear the news of Stanley Kaplan’s passing today. I met Stanley not long after I started working for Kaplan, in 1991. It is fair to say that Stanley Kaplan is a big reason why promising students from all backgrounds have a shot at admission to top colleges and grad schools. This is both because Stanley invented the test preparation industry (in his basement!) and because he was ever afterward a relentless advocate for merit over status in admissions. He deserves a great deal of credit for turning the tide.  What impressed me the most about him was that even though he had retired from full time CEO duties by the time I met him, he had in no way lost his infectious love of teaching or his enthusiasm for young people trying to improve their lives through education.

Sincerely,

Jose

The GMAT and India

August 24th, 2009

I wanted to take some time to dispel a fascinating—but unhealthy—rumor about the GMAT. This rumor is best summarized by a concerned Knewton student who wrote me the following:

I was speaking to someone in India in the test prep industry about the GMAT and that person seemed very confident that the database of questions for the GMAT test in India is different from the database of questions in the US and Canada (the pool of questions in India being harder). Have you heard anything like this?

And here is how I responded:

You’ve touched upon some topics that interest me a great deal, so let me give you a longer answer.

The test-makers have always used multiple question pools at any one time, even within the United States. This is done for security reasons, so questions can’t be pirated—something I know quite a lot about. (I once reverse-engineered the computerized test and proved the question pool at the time was much too small, and hence susceptible to cheating. The test-makers pulled the exam for months and implemented the changes I and others recommended—adding more questions, rotating pools, etc.) Question pools rotate every few months. Pools used internationally either just came from and/or will soon head back to North America. These pools are all (almost) perfectly calibrated with each other so that your score on one continent will be—within the margin of error—your score on another.

My suspicion is that you’ve heard some kind of urban myth that has its roots in:

  • paranoia that U.S. schools are too full of Indians and want to restrict their numbers
  • chauvinism that Indians are smarter/better at math

Read the rest of this entry »

Thoughts on "Unless Statements" from Alex K.

August 20th, 2009

KH3A conversation with a good friend this morning got me thinking about LSAT prep when she announced:

“I will move out of my apartment, unless my rent is lowered.”

As any good friend would, I translated her announcement: P unless Q.

Her intended meaning, I thought, probably implied two conditions:

1. If I do not move out of my apartment, then my rent was lowered. (“If not P, then Q.”)
2. If I move out of my apartment, my rent was not lowered. (“If P, then not Q.”)

In everyday usage, we intend most “unless” statements to be translated in this way, as biconditional statements.

For LSAT purposes, however, only the first conditional statement (“if not P, then Q”) is a valid translation of the original statement: “P unless Q.” If she stays in her apartment, then we can be sure that her rent was lowered. The necessary condition, lower rent, must have occurred considering the outcome that she did not move. If she moves out of her apartment, however, I can’t be certain that her rent was not lowered. She could have decided to move for an unrelated reason, like a fire or an infestation of beetles.

Read the rest of this entry »

Data Sufficiency — Nurture over Nature

August 17th, 2009

Nobody likes Data Sufficiency questions. They’re like the Pigeon Pose in yoga — unnatural, and therefore uncomfortable.

The good news is everyone feels this way. No one is born a good Data Sufficiency question-taker. You become good, but you have to work at it.

Turn this to your advantage. Data Sufficiency questions are very coachable. If you’re planning to study for the GMAT for 3-4 months you should actually be HAPPY that Data Sufficiency is on the test because you can gain a competitive advantage.

Try this question:

Is the positive integer k a multiple of 54?

1. k is a multiple of 6.
2. k is a multiple of 9.

Check out the answer to this sample Data Sufficiency question. Click on problem #2.

Knewton gets some great press

August 13th, 2009

We’re all really excited about a few features that came out this week.  We knew our customers were happy, but it’s also rewarding to see that our hard work is generating some buzz.



BusinessWeek. In the magazine’s bi-annual review of the best GMAT prep courses on the market, the former CEO of Kaplan says “I hate to say it, but Knewton has rendered every other test prep company totally obsolete.” The article concludes that “Knewton, with its all-online format, is convenient for just about anyone and might be a particularly good solution for those who travel often or like to spend many hours online.” And that “By eliminating costs, such as rent and supplies for classrooms, and the travel costs you’d have to spend to get to the class, Knewton is also one of the most affordable options available.” Read the full piece here.

FastCompany. Anya Kamenetz, in her article about how “Edupunks” are transforming education, quotes our CEO, Jose Ferreira, as saying “The Internet disrupts any industry whose core product can be reduced to ones and zeros.” The article went on to mention Knewton as one of the companies cutting down big trees in the online education forest. Read the whole thing here.

In yet another piece, FastComapany listed Knewton as one of the “5 Startups to Watch.” Citing that we “promise to customize content for each student, down to the concept level, with integrated assessment tools.” Check out the full list here.

A note from Tony Poor — the UI intern who should never leave

August 10th, 2009

tpoor

Four months ago, I answered a job posting for a summer internship at a test prep company. At the time, I had no idea that the company’s goals aligned so perfectly with mine.

As a student of human-computer interaction, I am passionate about making technology easier to use–bridging the gap between the brain and the system. The grand vision of my entire field is, essentially, technology that invisibly supports the goals of its users.

Here’s the kicker: Knewton is that vision, plus more. Anyone can make an e-learning product ‘easy to use.’ But adaptive learning — delivering a truly personalized education based on a student’s habits — is revolutionary. Through this single mechanism, Knewton is not just creating an easy-to-use product; it’s revolutionizing the usability of education as a whole. All in a completely invisible package. You don’t have to know that something behind the scenes is constantly tailoring lessons to your individual needs and learning habits. It just is.

I live by the mantra that the technology should support the user; why should we settle for educational practices that remain cruelly inflexible? No. Knewton doesn’t think so; that just won’t do. And I’m with them 100%.

Customize your LSAT practice with “Create a Quiz”

August 6th, 2009

qbot2

There are many ways to prep for the LSAT, but all the experts agree on one piece of advice: practice, practice, practice.

The LSAT tests a lot of skills. How can you make sure your practice program is right for you? Knewton has a new solution: “Create a Quiz.”

Create a Quiz is an interactive study tool that lets you tailor your practice tests to fit your needs. Looking for extra Logic Games work? Design a quiz that tests your Selection and Absolute Ordering skills. Logical Reasoning section giving you trouble? Run through a quiz of Assumption, Parallel Reasoning, and Strengthen/Weaken questions.

You can make as many quizzes as you want, in as many combinations as you can imagine. The quiz tool draws on nearly 2,000 real LSAT questions, and it shows all your results so you can track your progress. Additionally, every question features a detailed explanation written by our team of experts.

Knewton’s Create a Quiz makes LSAT prep personal and adaptive. We give you all the LSAT sections, all the question types, all the explanations—you just decide how you want to use them.

How B-School Rankings Affect Future Earnings Potential

August 4th, 2009
[youtube=http://www.youtube.com/watch?v=nuMOHnSouik&hl=en&fs=1&] Knewton CEO Jose Ferreira talks about selecting the business school that is best for you. He offers his expert opinions on the publications with trustworthy rankings, the correlation between GMAT scores and starting salaries, and the business school application process. Does US News and World Report have the best rankings, or does the Wall Street Journal? Jose will share his thoughts on those questions and others.