<p>PC ~ I see the problem...I thought there was more data released from Berkeley a few years back in the Moore report, but it probably just shows more inexplicable variability.</p>
<p>Also, as the level of the scores goes up (above 700) the RHS tail of the distribution in progressively chopped off because sometimes the kid who missed three problems gets an 800, as does the kid who missed two, the one who missed one, and the one who missed none...</p>
<p>Also, some tests get repeat sittings much more frequently than others, so those who liked their scores keep them and those who didn't study and retake. Again, I suspect that this leads to a need to get rules of thumbs by major. In that sense, I have much less to offer for your humanities student...that's not where I came from and it's not where my kids seem to be headed, so I pay less attention to that area.</p>
<p>I also don't understand what the Williams data means. And I did not know that SAT IIs were also re-centered when the SAT I was. I'm surprised to hear this, and might almost suspect a type in the header.</p>