<p>West Sidee: "Oh yeah. BTW, the stats for Berkeley is based on best one sitting score. While the private schools use best verbal + best math. You don't have to take anything more than High school algebra to see the difference.
The 25th-75th percentile is more 1240~1250 - 1490~1500 at Berkeley."</p>
<p>You are thus giving 50-60 points to Berkeley due to private schools best verbal + best math
(Berkeley's listed 25th percentle/75th percentile
1190/1440
<a href="http://cds.vcbf.berkeley.edu/content.cfm?formyears=2003-2004§ion=c%5B/url%5D">http://cds.vcbf.berkeley.edu/content.cfm?formyears=2003-2004§ion=c</a>)</p>
<p>Ok, let's do the algebra based on the statistics (<a href="http://www.collegeboard.com/research/pdf/rn05_10756.pdf%5B/url%5D):">http://www.collegeboard.com/research/pdf/rn05_10756.pdf):</a></p>
<p>51% of SAT takers only take the SAT once (Table 1). So for these students, scores at private colleges do not have added advantage.</p>
<p>38% of test takers took it twice and 10% three times (1% 4 or 5 times).
For these remaining 49%, it gets slightly tricky. For the two time takers, if both verbal and math go up (or stay the same), or both verbal and math go down, there is no added advantage for the private colleges. Only if one goes up and one goes down on the second sitting is there an advantage. In the score range for those accepted to Berkeley (so we are not talking about students with less than 430 verbal or math scores), between 40-50% of verbal or math scores go up and 25-35% go down (the majority between 20-70 points either way)(Table 5). Therefore by random chance here, approximately 13.5% (~0.45 x ~0.30) of second time test takers will see one score go up and one go down (arguing that takers who go up in one should go up in the other brings this percentage down). These 13.5% of two sitting SAT applicants (therefore 5% of total applicants (0.135 x 0.38)) will add to the private schools SAT averages but how much? Berkeley uses best sitting. So between the two sittings, Berkeley will choose the one that was higher (meaning the one where the verbal or math was higher than the other verbal or math was down). Berkeley's disadvantage, therefore, is by the amount the divergent verbal or math score goes down. We can give Berkeley the benefit here and count the average decline for test takers of above 580 in either section (since higher scores decline more). On the verbal, of the decliners (again, all derived from table 5), ~9% (~3/34) will decline ~90 points, ~29% (~10/34) ~60 points, and ~62% (~21/34) ~30 points, for an average decline of ~44 points. On the math, of the decliners, ~7% (~2/29) will decline ~90 points, ~28% will decline ~60 points (~8/29), and ~66% (19/29) will decline ~30 points, for an average decline of ~43 points. Therefore, for 5% of the total applicants (the ~13.5% of second time test takers (38%) with divergent (one up, one down) verbal or math scores), private colleges have ~44 points added to their applicants score.</p>
<p>What about the 10% that took the test three times? Mean scores for each subsequent testing show an increase of 7-13 points on the verbal score and 8-16 points on the math score (page 1). So for three time SAT takers you could expect an average gain of 14-26 for verbal and 16-32 for math. So there is a strong possibility that both verbal and math sections were higher on one of the two subsequent sittings (converting it to a two sitting situation for modeling purposes) or that the decline of the verbal or math score of the highest combined test that Berkeley uses is less than the 44 points seen when only two tests are used (for privates to gain more than 44 points the highest gain test would have to have both the highest gain in either verbal and math coupled with the greatest decline of the other score from the other two sittings). Therefore, it is reasonable to use a 44 point advantage for three time takers as well that had divergent scores (~1.5% of total takers (~13.5% of the 10% three time takers).</p>
<p>Thus:
51% took the test once, no benefit
49% took the test more than once, ~13.5% of these had a benefit of ~44 points</p>
<p>~7% of the total (0.135 x 0.49) have an average benefit of ~44 points.
The overall benefit for private colleges' SAT scores is only 3 points (0.07 x 44), nowhere near the 50-60 points you give them.</p>