The early questions to worth more right....?

right so it is better to get to the highest pools faster from the beginning than getting it later since we may never reach it. Anyone ever use this "knowledge" or people just try to do all the problems with equal time/consideration?

the first question you get is a average question meaning 50% of the test takers got it right and the other 50% got it wrong. the following question depends on how you answer that question so, its a tree diagram.. you got it right you move up and just for the sake of illustration you get a question which 40% got right and 60% got wrong... and vice versa if you get the second question wrong.. This process goes on but as this quite clearly illustrates if you don't shoot up quickly you will take longer to go up to the same mark you could have reached had you answered the earlier questions. I will post a tree diagram to clarify this to you, but I hope you get the idea.

I get it. I was just making sure it is the case and what was told to me couple years ago havent changed that's all. And also curious if test takers here use that knowledge at all (like using more time on earlier ques.) or not.

yup spend more time on earlier questions compared to the rest... i mean if u looked at it in segments... generally you should spend more time on the first batch of questions, then a little lesser time on the second batch and so on.. but the fun part of this analysis is that he early questions will be easier compared to the later questions if you are answering them correctly so in a manner of speaking you should be spending equal amounts of time if difficulty of questions warrants more time being spent on them.its all quite fuzzy.. so basically your strat should be to get the first few questions RIGHT and get them RIGHT QUICKLY compared to the ending few questions b/c getting wrong or right the last 2 questions everything else cetrius paribus is going to make a very little impact on your score compared to getting the first 2 questions wrong. i think this all has been discussed somewhere on the forum... i really don't know how to make those links otherwise i would search and post it for you.

This is one of the GMAT myths that deserve attention. Unfortunately, it proved to be right - early questions do carry more weight and, in the end of the day, determine your raw score. So, try not to screw up in the first couple of questions, especially try not to make 2-3 wrong questions in a row in the first 10-15 questions. If you do that... retake will be the only option.

Said but true, this just add even more pressure on already stressed test takers.

. Another question... do the levels "reset" after certain points on the test? Like in verbal when it switches from reading to reasoning or the level stays on course the entire exam. I guess in the end this doesnt matter too much since if one cant solve it it doesnt really matter....

At the same time though, if you spend more time on the first 10 questions or so and you don't have time to get to the last 7 or 8 questions, then it hurts you too. Lets say you go 10 for 10 on the first 10 questions, then level off and you're about right for that point. Let's say at this point, question 29, the CAT thinks your score on the tree diagram is going to be a 760, but then you have 3 minutes left for the remaining 8 questions because you took more time on the first 10 questions, more time than you should have, even though you got them right.

Question #30, similar difficulty to #29, you guess and get it wrong.
Question #31, easier than the pervious because you got the previous one wrong. CAT doesn't care how long you spent on the question.
Question #32 is even easier than #31 because now you've gotten 2 questions in a row wrong. If you guess again, you've got a 20% chance of getting this one right, so likely you get it wrong.
Question #33 is easier still since you've gotten 3 questions wrong.

This goes on and on until you hit all 37 questions. My point is that by guessing on a lot of question in the end because you spend a lot of time at the beginning, the CAT sees that you're getting a lot of questions wrong and these are very easy questions because after you get 5 or 6 wrong in a wrong, you're going to get 1+1=? and you won't have time to answer it, or even look at it so you guess 3 and get it wrong. In the algorithm for calculating your score, these questions at the end hurt you big time because you get so many wrong when guessing and they're easy questions because of how the CAT adjusts difficulty.

Another perspective to consider.

Do you guys think that the GMAT prep algorithm and the GMAT algorithm are one and the same? Does GMAT say that they are the same? (or) It just uses the disclaimer that your scores on GMATprep do not indicate your actual performance in disguise? The reason why I say this is, I read on the forums that the first Q is not as easy as the prep software shows it to be. For example a simple exponent problem or range problem shows up in prep. How ever, on the actual test it might be a inequality question. It is very well possible that the inequality question is a easy one, but I am getting skeptical.

I have a strong feeling that the score is not moving as you answer the questions. It is not being calculated. If it is calculated on every answer, you would get the score with out delay. Yeah computers dont need that much time. It actually computes at the end of the exam on how many easy/hard questions you got wrong and how many easy/hard Q's right. So if you got 2 out of the first 7 wrong, you are pretty much set up for a lower score as it is difficult to maintain the speed and accuracy towards the end. In one way, it seems to me that the penalties for answering first 10 questions incorrectly grows exponentially as the number of incorrect questions increases. OTH, the penalty for getting questions wrong in a latter stage reduces drastically. So can we conclude that GMAT is saying "Do your easy and medium hard questions right and if you screw up, you have very very little chance? "

Neither MCAT nor LSAC made its test a CAT. MCAT is a CBT and LSAC is still paper based. GMAT and GRE are not comparable. GMAT serves a specific purpose ie business school and GRE is for any graduate program under the sun. IMO, the CBT is a over kill and skewer of actual performance and luck has its role because every one gets different questions. How can we say that XYZ scored in the 90th percentile if his fellow test takers did not take the exact same questions? The research behind the whole thing saying that every question is reviewed by 10 or so test writers and answered by 1000+ test takers ( I am just quoting numbers as example) to conclude that it is a valid percentile does not really hold much water in my opinion.

What is the accuracy of the scores? We can say almost very confidently say that some one who has taken the test today and scored 250 will not score 750 in 2 days. How ever, what is the number of points by which a test taker's score can fluctuate in two takes given that he is taking two different tests? Is it 30 or 40 points? There must be some thing like that right? Has GMAT ever published this?

Read the https://en.wikipedia.org/wiki/Graduate_Management_Admission_Test#Total_Score%20Total%20scores section on wikipedia.

I always thought that the first Q is medium in terms of difficulty, but wiki says other wise. I hope the guys at wikipedia are smart enough to be writing such important pieces of information

Thanks icandy.

I hope it is true, but I would not put much factual faith into it - probably a guess (a good guess) but only a guess.

I don't know for algorithms, but scores from GMATprep are pretty indicative of ones real score. Maybe algorithms are totally different but they do score in the same fashion, so people from GMAC have done pretty good job here.

My real exams have always started with some pretty straightforward exponent problem and two of my three ended with some number line DS. My GMATprep and real exam experience was almost same with some advanced questions similar to those from GMAT Focus.


They should move as you work through the test, simply because the test is adaptive. You can confirm that by taking GMATprep. If you answer randomly, questions will be much easier than some monsters occuring when you hit 10 in the row.

They should move as you work through the test, simply because the test is adaptive. You can confirm that by taking GMATprep. If you answer randomly, questions will be much easier than some monsters occuring when you hit 10 in the row.[/quote]

Let me put it this way. I used the words calculated and moving interchangeably. I should not have. Yes, the difficulty of the next question is changing based on the response of the correct question, but it is not like, you are given 600 to start with and then score goes up and down.

Once you complete all q's the algorithm will compute the scores, some thing similar to a weighted average. Got 2 wrong out of 15 med, 1 wrong of 5 easy. and 7 wrong of 17 hard.

That's an interesting thought...I'm not sure what to think. I can see where they might not want to use the same method in GMATPrep, but on the other hand, the scores do seem to be very predictive.

As far as the difficulty of the first question, I had an algebra question that I thought was fairly easy, once nerves subsided (to say it was easy assumes I got it right which I may never know )

there might be a way to prove it if one takes the same test with answer keys and try different ways to get X numbers of questions wrong - the first X wrong or the last X wrong etc....

I've done preps where I've gotten 5/10 wrong initially and another 10 and still gotten a 50Q

If you screw up the easy questions but nail the hard ones it won't count against you

Hard questions on the GMAT:

Geometry DS
Combinations
Permutations
Inequality DS
Odd/Even DS
Ordering questions
Venn DS

that's pretty much the hardest I've encountered, very few people nail those consistently
If you can do that, and screw up all the easy PS you'll still get a 49-50Q

The only difference between 50-51 is how many questions you get right wrong (51Q is usually 0-3/4, 50 is 3-15).

I'm pretty sure you can't get 15 wrong out of 37 questions and get a 50Q score.

Definitely. You can score 49 with 15 wrong questions, at least in GMATprep (I've done that on several occasions), but for 50 you have to make no more than 10 mistakes.

wow. Would this apply to Verbal too? Because that would really help(at least make me feel better )