Thanks, Kinji! And if I haven't said it before, congrats on that 760!
Sahil - really good point on getting the last 15 right. Couple thoughts there for you:
-We've been pretty fortunate to work with the former president of the International Association of Computer Adaptive Testing on our exams, and with a nod to the A/B/C parameters in that one article I can tell you that we have confirmation that the algorithm is doing what it's supposed to, even if it doesn't look like it to the naked eye at all times.
-In this case (I have your test up right now), missing 4 out of 5 from questions 16-20 hurt you, and then from that point on you had a good amount of catching up to do.
-I can also confirm from your test that the error margin at the end of that section is well within the acceptable error margin that the IACAT folks have set out, so it's a pretty valid score.
Now, what's tricky about really any non-GMAC practice test is that there's often a diminishing-impact effect as you get to the later questions, more so than on the official exam. I was just talking with a few instructors over coffee this morning about this. GMAC shifted around 2006-07 to more heavily emphasize "content balancing" (which categories of skills - algebra vs. applied word problems vs. probability, etc.) than just letting the algorithm always determine the best possible question for you based on CAT parameters. (For example, at question #6 the best possible question to serve you based on CAT parameters might be a Data Sufficiency algebra problem, but the system is controlled more by content so if it's programmed to give you a Problem Solving word problem, you'll get the best possible PS word problem at that point, not the best possible quant question overall)
The effect of that is that on the official exam you'll find that the last few questions matter a little more than they sometimes do on practice tests, which usually let the CAT algorithm run a little more freely. On a lot of practice tests, the higher-value questions get served early and so the later questions don't carry as much weight; because the official test is expertly content-balanced (and have a deeper pool of high-quality questions than most practice tests), the later questions are much more similarly weighted. The impact on your *overall score* shouldn't be much different, but you may find on practice tests that you don't get enough credit for getting the last 5 or so questions right, and even more detrimentally the last 5-10 questions are "more forgiving" (getting more of them wrong doesn't hurt your score as much as it will on test day). So as always, use practice tests to help you determine how to improve and to get a good feel for your score range, but try not to overanalyze the question-by-question metrics of the scoring algorithm. That's just energy that isn't really pointed in the right direction.
Anyway...with not too many days until your exam, that description is probably a little bit of overkill. Keep working on shoring up weaknesses and on verbal, particularly, I'd attack the problems in the GMAT Prep Question Pack and the
OG (or just search by GMATPrep tag in the forums here).
Glad you're enjoying the videos and getting a good feel for using logic. Keep me posted and good luck with that impending test day!
close
14 Apr 2015, 05:40
but it has helped me...)
