<p>I’ve read various things on these boards about how the curve can vary from month to month, test to test, year to year, etc. But for one test, how much does it actually change? </p>
<p>For example, the 2007-2008 edition of the Kaplan US History prep book says that on a “recently administered” test, you could miss ten and still get an 800. Is this going to remain fairly consistent (i.e. this year you can miss nine or eleven and still get 800) or is it common occurence that it changes significantly?</p>
<p>Well the SAT subject tests are a different affair, and moreover an affair I know little about. But, in terms of the SAT I. . . .</p>
<p>The variance in the curve isn’t actually that significant, especially when you start getting questions wrong. Allow me to illustrate:</p>
<p>I’ll be comparing scaled scores from October 2005 (O05), May 2006 (M06), and January 2007 (J07) resulting from answering 3 questions incorrectly and from answering 30 questions incorrectly:</p>
<p>63 CR: 770 (O05), 780 (M06), 780 (J07)
30 CR: 500 (O05), 500 (M06), 500 (J07)</p>
<p>50 M: 720 (O05), 720 (M06), 720 (J07)
17 M: 430 (O05), 420 (M06), 420 (J07)</p>
<p>Writing is kind of weird. . . I’ll get back to you on that one.</p>
<p>But, anyway, as you can see, the curve doesn’t really make that much of a difference. There is a somewhat significant difference in a 64 CR score specifically (sometimes it’s 800, sometimes as low as 770), yet getting a 64 is pretty rare, considering you’d have to omit 1 and get 2 wrong.</p>