I don't mean to state the obvious: there is a verbal problem with the GS tests. To be more precise, it's a 2-3 out of 7 passage "problem."
Technical things are the same (number of words per passage, passages per test, questions, timing, highlighter, whatever) and, though I could get shot for saying this, the score is quite predictive (on average, within about 1 point for VR). The issue is that the "experience" feels different because there are some passages that are too straightforward compared to the real exam.
Here is how it plays out: the bell curve assures a lower scaled score for the same raw score that you would have on an AAMC test. The reason is due to a large pool of students scoring higher on the raw score but the mean scaled score stills end up being between 8 and 9 (as it is with the real test).
So the summary: "easy question" for GS VR usually means "easy passage" which results in, by the nature of the bell curve, a difficult curve.
We wanted to have the new passages online by early Spring but I think that another month will be more likely. Those initial scores will certainly NOT be predictive!
On April 3rd, 2010, I will not say that GS VR is your best non-AAMC choice for realistic verbal passages!! I would also not say that you just wasted an hour of study time. I would say that the experience is not predictive enough of the real MCAT and that must change.