No way was it changed such that 20 more correct questions are required to get the same score, even in the middle range. By your "personal experience," do you really mean your opinion?
My opinion, based on having taken all of the NBMEs with extended feedback when available, is that any change was probably minor, on the order of 3-5 points given equal percentage scores. Here's how I come to that conclusion: I took 6 and 7 before they were changed and my scores on 11 and 12 were identical. (I took them in this order: NBME 6=258, 7=264, 11=264, 12=268.)
For all the voodoo about which NBMEs are most predictive, they are remarkably consistent, at least at the upper end. What people don't seem to realize is that they are not trying to predict your Step 1 score in the way the UWorld sim exams are. They are assigning you a score as if you had, on that day, taken an abridged Step 1 exam in the years preceding the release of the NBME form. Depending on your time to exam, the age of the form, how much of the form's content is now included in FA/UW, how you slept the night before, and many other factors, these simulations will more or less accurately predict your actual step 1 score.
Long rant summary: The most recent forms' content more closely mimics the actual exam, but among the more recent tests (6, 7, 11, and 12) there should be little difference in reported score at a given knowledge level. Please, please make the threads about "which NBME is best" go away.
If someone has taken 6 or 7 recently and scored in the 255-265 range and wants to share their percent correct, I'd be happy to compare. We could interject some, you know, facts and maybe salvage this thread. 😉
