How to get 12 questions wrong on the GMAT and still earn a Perfect 800

Hi livelyangel, we were able to conclude the exact number of "pretest" / experimental questions on each section by using the ESR (Enhanced Score Report) "percent correct by quartile" information, derived from analyzing the reports of hundreds of GMAT club members, to determine the correct denominator (actual number of counted questions) for each quarter of the test.

https://gmatclub.com/forum/couple-quest ... l#p1767360

Don't want to sound rude but in my opinion this information isn't too useful to a test-taker. There's no way to identify which is an experimental question and which isn't. Hence as test-takers, we need to approach all questions with the same type of focus.

It's OK--I don't take it personally. Obviously, the title is meant to be somewhat tongue-in-cheek: I agree that there is little *strategic* benefit to knowing this information when you are taking the GMAT: you have to assume that every question counts, and try to get every question right. Yet I would argue that it should somewhat affect your pacing / guessing strategy / overall aggressiveness level to know that there is a >13% chance that any question you are stuck on doesn't count--it helps you realize that no one question should make or break your score, unless you allow this to happen by spending too much time (>2 or 3 min), or way too much time (>4 min), on a single question.

From a psychological perspective, though, knowing the high number of pretest / experimental questions helps you better understand why GMAT scores can vary so much on test day. This data also helps explain why it makes sense to be persistent, and to give yourself time for a couple of extra attempts if you have a lofty score goal.

In other words, if official GMAT scores were a poll, then the "margin of error" would be higher than normal because of all the experimental questions. If I contact 90 people to ask for their opinions, but only record the opinions of 67 of them (the real GMAT), then of course the results of the poll will be less accurate than had I counted all 90 opinions (GMATPrep).

-Brian

Update, 1/16/18: Recent ESR information that has come to light has suggested that the GMAC is not being fully transparent about the experimental questions, and that while they are technically "unscored," they might still affect the overall difficulty of the test and the section scoring algorithm. You will see in the example linked above, for example, that some students who answer 100% of Verbal questions correctly still only earn a V50 score instead of a perfect score of V51. If so, then what other explanation can there be, than the fact that the experimentals do matter?

This news helps explain why, with the hundreds of thousands of GMATs taken each year, why no one has managed to luck his/her way into a perfect 800 score. And in a weird way, it seems fair. When taking the GMAT, we have to assume that all the questions count, and still try our best on them, so shouldn't they at least count in some minor way?

My only real concern is that the GMAC has not been fully honest about these "pretest" questions and their role in the test: for decades, they have assured test-takers that the experimental questions are "unscored." However, that's not exactly surprising news.