<p>
</p>
<p>As MIT’s own Dr. Perelman showed, the longer the essay, the higher the writing scores. So I will omit that iffy metric. <a href=“http://www.nytimes.com/2005/05/04/education/04education.html?_r=1[/url]”>http://www.nytimes.com/2005/05/04/education/04education.html?_r=1</a></p>
<p>SCORE/M/R Percentile Ranks For 2009:</p>
<p>800 99 99 730 97 96 660 91 88
790 99 99 720 97 95 650 89 85
780 99 99 710 96 94 640 88 83
770 99 98 700 95 94 630 85 81
760 99 98 690 94 92 620 84 79
750 98 97 680 93 91 610 82 77
740 98 97 670 92 89 600 79 74</p>
<p>Many MIT classes are math-intensive, so the admits self-select on this criterion. Based on the information provided, MIT’s admit pool has 75% above the 98th percentile in Math, and 75% over the 91st percentile in Reading.</p>
<p>cf. Harvard College
</p>
<p>Harvard’s admit pool has a 75% above the 94th percentile both in Math and Reading. Very consistent, nice PR work. Someone has thought this over.</p>
<p>More important for our discussion: two of the most selective, and arguably most challenging academic institutions in the world admit a quarter of their class below these levels.</p>
<p>Harvard is socially smart enough not to provide further breakdowns (so reverse engineers cannot pinpoint the exact constitution of a Harvard class). MIT is too wicked smart to care, so they provide data (so reverse engineers waste time on such nonsense instead of improving their overall candidacy).</p>
<p>How meaningful is all this? Minutiae aside, one clear pattern emerges: MIT shows that all their admits came from the top 15% SAT Math, and top 20% of SAT Reading. It would surprise me greatly if MIT applied less academic rigor in their selection process than the nation’s top prep schools, and I assume they certainly will not accept an applicant if they are unconvinced she could handle the workload. I would venture a guess that SAT M/R ~Top 20% also holds for most Ivies and selective LACs.</p>
<p>There you have it: anecdotal evidence that some ~20% of the SAT population (10% of overall population?) could handle such academic workloads, assuming motivation and opportunity. However, how does this translate into earlier SSAT scores?</p>
<p>Well, it is doubtful that the overall SSAT pool of sixty thousand would be less selective than the overall SAT pool of two million. Thus the above top 20% (at least) would carry over to the SSAT. Also, there is quite some crowding on the high end of the SSAT, skewing the overall percentiles. Both tests are easy, but while the SAT casts a wide net, the SSAT does not. Therefore the SSAT as a test may perform best mid-range, and in raw scores; not with “scaled” scores (which nobody cares about) and definitely not with percentile ranks of a small, elite pool, further divided by grade and gender into ever-smaller test populations. Indeed, doesn’t the SSAT intend to diagnose absolute academic preparedness<a href=“as%20opposed%20to%20ranking%20an%20elite%20pool%20from%20#1%20to%20#60,000%20on%20a%20curve”>/U</a>?</p>
<p>Now it is a different question altogether what the school needs when assembling the class. This is one piece of information, and that only in the area of academic preparedness.</p>