<p>Is the curve better on the October or November SAT? My D is took the March SAT as a junior and will be retaking the test in the fall.</p>
<p>i think it depends on the test each time... depending on how well the general popluation does on it.</p>
<p>if it's easy.. meaning more people score very well then its going to be a harsh curve...</p>
<p>whereas if its extremely difficult then even if you miss a few questions you can still do well.</p>
<p>The tests are normalized such that there should be no difference from date to date.</p>
<p>There isn't a set difficulty for each particular month(that's a myth), i.e. March isn't necessary always the hardest curve, but it did happen to be particularly harsh this year.</p>
<p>Well I kinda heard that in a particular month (is it December or something?) more Chinese people will take the SATs during their hols... And the curve for the maths will be much steeper. Dunno how true it is though, just my $.02. </p>
<p>Good luck!</p>
<p>Err...unfounded claims, anyone?</p>
<p>Much time, money and effort is put into ensuring that a specific candidate will have the same opportunity to earn the appropriate score during any and all testing dates. Period.</p>
<p>that said there are dates when the curve is easier. for example in january you could miss one on the writing MC, write a 9 essay, and still get an 800 on wr overall.</p>
<p>First off, there is no such thing as a curve in the sense that we think about it in school; it does not matter who else is taking the test. ALL test administrations are scaled, and each test scale is equivalent to every other test scale. On an extremely rare occasion, a bad question will slip thru the experimental gauntlet and be thrown out after the fact.</p>
<p>mhmm, don't take the Oct. one... cuz nowadays more and more Chinese are studying for SATs... and Oct.1 to Oct.7 is their National Holidays... as a result, most Chinese will attend the Oct.4 SAT Reasoning test... so Math and Writing score will go steeper cuz they're STRONG in these two fields... most of them got 800 in math and 650 in Writing...</p>
<p>It DOESN'T MATTER WHICH DATE YOU CHOOSE, though many people "know" better.</p>
<p>The "curve" (by which I'm assuming everyone means the mapping from raw scores to scaled scores) is NOT dependent on, say, whether a disproportionate number of strong math students happen to take a particular test.</p>
<p>The curve is used to compensate for changes in the difficulty of the SAT from one test to another. The "variable" sections in each test contain repeated questions from previous tests to distinguish between, e.g., whether a test was too easy or whether a bunch of very good students took the test (both would result in unusually high raw scores).</p>
<p>I'm not saying that there aren't psychological effects due to taking a test with an easy curve vs. one with a hard curve, just that the curve won't reflect that a bunch of very good students took the test.</p>
<p>For more info (very technical), see:</p>
<p>January has historically had the best curves. But like all the above posters said, the test is designed so that everything is equalized. It really doesn't matter.</p>
<p>
[quote]
mhmm, don't take the Oct. one... cuz nowadays more and more Chinese are studying for SATs... and Oct.1 to Oct.7 is their National Holidays... as a result, most Chinese will attend the Oct.4 SAT Reasoning test... so Math and Writing score will go steeper cuz they're STRONG in these two fields... most of them got 800 in math and 650 in Writing.
[/quote]
</p>
<p>This is just fuzzy logic. Regardless of which test date you take, miss one math problem and you will receive a 770. Get them all right and you will recieve an 800. It does not matter if one million others also get them all correct on the math section on the October test -- it will still be an 800. The SAT is NOT CURVED! </p>
<p>
[quote]
January has historically had the best curves.
[/quote]
</p>
<p>This is an urban myth, propagated by an former admissions rep to sell books.</p>
<p>?????
This is no urban myth...January QAS explicitly display the curves. And you're incorrect about the math score. Depending on the test difficulty, CB changes the conversion accordingly, as you can see below...where there was an administration that actually gave 800s to people who missed a math problem. </p>
<p>I'm not too clear on how exactly they determine it, but I do believe the kids taking the test have an input in the curve determination. They are predetermined based on test difficulty, then CB works some voodoo with the "equating" process to ensure equality. There are articles out there that discuss what the equating process entails.</p>
<p>
<p>Jan 2006 CR.........M..........W 67 - 800 54 - 800 E12 or 11 = -1 still 800 66 - 800 53 - 800 Perfect MC and as low as E9 for 800 65 - 800 52 - 780 64 - 780 51 - 760 63 - 760 50 - 740</p>
<p>Jan 2007 CR.........M..........W 67 - 800 54 - 800 E12 and -2 for 800 66 - 800 53 - 780 E11 and -1 for 800 65 - 800 52 - 760 Perfect MC and as low as E9 for 800 64 - 800 51 - 740 63 - 780 50 - 720 62 - 760 49 - 710 61 - 750 48 - 700
</p>
<p>You're right, the curve does change from one test to another.</p>
<p>Suppose you had gotten 1 wrong on math (raw score: 53). You would've gotten different scores depending on the test:</p>
<p>01/2008: 790
10/2007: 780
05/2007: 800
05/2006: 790
01/2006: 800</p>
<p>The Oct 2007 test for example was made a bit too easy, so the curve was "harsh".</p>
<p>However, the true difficulty of the test <em>can't</em> be predetermined ahead of time: the raw scores of the test takers on the variable sections are used to figure that out. The main concept is that a score of 750, say, should reflect the same ability level independent of who took the test and how hard the test was.</p>
<p>don't forget that the test was changed in March 2006, so any tests prior to that time are not comparable. AND, Alg II problems were added in March 2006. Any given test may have 4 Alg II problems (or 5), all of which are considered "difficult" so the curve is not relevant; the number of 'difficult' & 'medium difficult' problems can vary by one or two per test, so it ain't a curve folks, it's difficulty variance.</p>
<p>
</p>
<p>Good summary. Essentially two adjustments are made:
1. For the relative difficulty of the test.
2. For the relative strength of the testing group.</p>
<p>Thus, a student will receive the same adjusted score no matter which test form is taking and who else is taking the test.</p>
<p>@bluebayou: The SAT changed in March of 2005, not 2006.</p>
<p>I guess its all semantics. You might not feel comfortable calling it a "curve," but you have to agree that the conversions change from test to test. And in high school, curves usually mean that a 100% is no longer exactly a 100%. If the highest score was a 70%, then that becomes the 100%, and those who got 50% don't have to worry about failing. The same occurs with the SAT to some degree. A perfect score is not just -0, it can be -1, -2, -3 etc, depending on the test. And "difficulty variance" is just too much of a mouthful :P</p>
<p>Thanks for the explanations by fignewton and BigIs!</p>
<p>Echelon:</p>
<p>sry, my bad on the date. But, I'm still looking for evidence, any evidence, that the "...the relative strength of the testing group." is part of CB's equating.</p>
<p>According to CB:</p>
<p>
[quote]
Every SAT includes a 25-minute section, which doesn't count toward your final score. The test is created so that you won't be able to tell which section this is. It may be a critical reading, mathematics, or multiple-choice writing section.</p>
<p>Since many different forms of the SAT are given each year, the unscored section adjusts for minor differences in difficulty across them. It ensures that a score of, say, 560 represents the same level of ability—regardless of what form of the test you take. This is another important part of making the SAT a standardized test.
[/quote]
</p>
<p>And:</p>
<p>
[quote]
Equating adjusts for slight differences in difficulty between test editions, and ensures that a student's score of, say, 450 on one edition of a test reflects the same ability as a score of 450 on another edition of the test.</p>
<p>Equating also ensures that a student's score does not depend on how well others did on the same edition of the test.
[/quote]
</p>
<p>Do y'all think CB is telling incorrect truths?</p>