I took the GMATPrep test 1 twice and got VERY different scores in quant, infact a lower score the second time when I think I performed better.
attempt #1: 12 mistakes, raw score 48
wrong answers: 3,4,5,7,13,16,18,25,26,33,36,37
one streak of 3 consecutive mistakes (and that too pretty much in the beginning), 7 mistakes in 1st half, 5 in 2nd . 4 mistakes in first 10.
attempt #2: 12 mistakes, raw score 43
wrong answers: 1,6,9,11,13,15,16,17,24,28,36,37
only one streak of 3 consecutive mistakes, 8 mistakes in 1st half, 4 in second. 3 mistakes in 1st 10.
The ONLY possible reason I can think of is that experimental questions screwed up the scores - getting no exp. qs right on 1st attempt and getting all of them right in the 2nd one.
But it would still be very unfair as it could easily swing the final score by upto 50 points.
I'm happy to share my 2¢
Scoring on the GMAT is a tricky thing. The GMAT, of course, employs Computer Adaptive Testing, so the score consists of much more than simply how many right vs. how many wrong. What matters most is the difficulty level of the questions you got right vs. wrong ----- questions of any difficulty could appear anywhere during the GMAT, so problem number doesn't give us any information. What makes a different is how hare the GMAT considers any particular question --- and, of course, that's proprietary information, so the best we could do is guess at the relative difficulty of any particular question.
Here's a blog article that explains this a little more:http://magoosh.com/gmat/2012/computeriz ... mat-score/
Let me know if you have any questions.
Magoosh Test Prep