|
More Stories |
Guessing on the GMAT, Part 2 Posted by ian on Thu 09 Oct 08 at 7:24pm In part one, we analyzed a pool of real GMAT questions, and concluded that it's quite likely that each answer choice has an equal probability of being the correct answer on GMAT math problems. But what about special question types? Some claim that the correct answer to math questions that begin with "Which of the following..." is more likely to be D or E than it is to be A or B. We're talking about questions like the following: Which of the following has exactly three positive divisors? A) 24 B) 25 C) 26 D) 27 E) 28 (the answer is B, incidentally). Those that claim that D or E is more often the correct answer to this kind of question suggest that the GMAT test designers are trying to force test-takers to do more work; test-takers are more likely to start from the first answer choice and work downwards, so if the correct answer is at the end, the question will be more time consuming. Therefore, the suggestion goes, it's best to start at the bottom and work upwards when testing each answer choice. Well, is this really true? It's difficult to find a lot of real GMAT questions of this type to analyze. We looked at every math question of this type in the Official Guide and on five real GMAT paper tests- there were only 31 such questions in total, ignoring duplicates. We tallied how often each answer choice is correct to 'which of the following...' GMAT math questions: A was correct on 5 questions B was correct on 5 questions C was correct on 9 questions D was correct on 3 questions E was correct on 9 questions The sample is too small to make any reliable conclusions, but the differences above are not statistically significant. The answer was A or B ten times, and was D or E twelve times, a marginal difference, and well within the margin of error for a sample this small. If there is any advantage to starting from answer choice E and working upwards on these kinds of questions, it appears to be only very slight at best, and it seems more likely to us, knowing how much care is taken in GMAT question design, that there is no advantage at all. If we're able to compile more data, we'll update our conclusions. |
|
 |  |  |  |  |  |  | 
|