<p>Here is what I am suggesting: there are lots of people here at CC with lots of experience helping students to take the SAT. And there are issues that seem to come up here again and again. Some of them are controversial but some of them really seem to be settled...and then they come up again. I would like to see if we can agree on a few of these and then put them in the stickies so that new folks will see them right away.</p>
<p>For items that are more controversial, we can use the threads to see if we can come to agreement. If we can't, then no stickie...</p>
<p>So to get the ball rolling, I'll start with this one (and one more)...</p>
<p>THERE IS NO SUCH THING AS AN EASY DAY OR A HARD DAY. </p>
<p>The equating sections take care of that. So when you are considering when to take the test, go by what fits your schedule and by when you have time to be prepared. Give no thought at all to whether the seniors take it, the sophomores, the gifted, the impaired -- none of that matters.</p>
<p>I agree - each test varies in its own way. What makes a day easy or hard, or a test easy or hard, is your mood and your strengths.</p>
<p>I disagree. Everyone has strengths and weaknesses. For me, there is no hard math, but reading might pose a challenge. If the test I’m taking has a hard math and easy reading, I will do very well. However, if the test has an easy math and hard reading, I won’t do as well.</p>
<p>^Yes, it’s certainly true that you might have a day where the reading passages “click” for you more than on another day, but it’s not what I was talking about. </p>
<p>I’m talking about the myths that go around that suggest that some particular test date can be predicted to be easier IN ADVANCE based on some belief about the group who is more likely to test that day. While you may personally find the reading to be challenging on a given day, there was no way to predict that – you shouldn’t think “Oh, November reading sections are always harder because [insert bogus reason here]”.</p>
<p>I am bumping this and hoping we can reach a consensus…</p>
<p>I have another, more controversial issue to hash out next, but let’s settle this one first.</p>
<p>I think it’s pretty well settled that there is no predictable trend, as shown by cheerioswithmilk’s mini-study that you can find on this site.</p>
<p>Let’s get to the more controversial one; sounds interesting…</p>
<p>It is very true that no month is easier than another. I just posted some reasons in another thread.</p>
<p>The typical reasoning goes: Dumber people take the SAT in certain months. Therefore the SAT is easier to score well on in certain months.</p>
<p>Truth: While the first statement may be true, its proposed consequence is not. That would totally defeat the purpose of scaling our scores. SAT scores are meant to be comparable across all test-takers in a given year (and beyond), not just among test takers in a given month. A college wants to see scores from someone from January and those from someone from June and be able to categorically say that one of them is better at what the SAT tests. They can’t do this if test-takers are only compared within their testing pool, that would be a false comparison.</p>
<p>If the CB did want to just compare test takers within their test pool (which they don’t) they could just use percentile scores and be done with it. Scaled scores would be unnecessary. Scaled scores exist because: The CB wants to give students a score that allows them to be compared to all other students who take the SAT, ever. Clearly percentile doesn’t serve this purpose, because in the future a lot more students will take the SAT. Thus the CB’s data is not complete enough to give you your percentile score among all students who ever take the SAT. They can’t see into the future. And so they use the scaled score system instead. They decide the difficulty of a test and they decide the curve <em>before</em> the test is taken, and this is all done quite independently of what the test takers are going to be like in a certain month.</p>
<p>The scaled score system prevents the following hypothetical scenario: Everyone who takes the SAT in June has a last name of Einstein and gets every question on the test correct. Except for you. You get one question wrong in each section. All of the scores are compared to each other and scaled scores are assigned based on percentile performance, and since you did worse than everyone else in the whole month and perform in the 1st percentile, you receive a score of 200/200/200. If you had taken the test in any other month, you would have gotten close to a 2400. But nope, you’re SOL because you took it in genius month. This is the kind of thing the CB can’t allow to happen because it’s against the spirit of standardized testing. This is precisely why they scale scores and why scaled score is totally independent of percentile score.</p>
<p>I’m sorry to argue with people who are agreeing with me…but reach the "expert " consensus, we have to arrive there for the right reasons – and I disagree with parts of the last two posts:</p>
<ol>
<li><p>Cheerioswithmilk’s analysis shows that there is no pattern regarding the chart to convert raw scores to scaled scores. But read the thread to the end and you’ll see why even if there were a pattern, it wouldn’t make one day better for you than another…</p></li>
<li><p>The “curve” [yes, it’s actually NOT a true curve…but people here continue to call it that!] is NOT established before the test is given. It can’t be: you need the results of the equating sections to determine the talent level of the testing group so that you can compare that to their results on THIS test date and use the comparison to establish the raw score-scale score conversion. </p></li>
</ol>
<p>If you think about it, it’s actually pretty impressive that they can write test after test and have the conversion charts vary to such a small extent.</p>
<p>BOTTOM LINE: Do not let concerns about the difficulty of a particular test date influence your planned testing schedule.</p>
<p>Well first of all I don’t see how what you said about cheerioswithmilk’s analysis conflicts with what I said, since all I said was that there is no predictable trend. </p>
<p>Anyways, I’m not sure if I understand what you said in your second point. According to what I’ve read and makes sense to me, the curve is set before the exam is released. This is possible because each section has been tested in a previous exam’s experimental section. The CollegeBoard is then able to gauge the difficulty of individual questions and overall sections by the end of the process.</p>
<p>Bottom line: agreed.</p>
<p>@DumbAndLethal is right, the “curve” is determined beforehand from the experimental sections. It does a pretty good job of controlling for variability in difficulty so test date really shouldn’t matter. Even if it is the case that tests in October are generally “harder” the curve accounts for that. If there were a pattern based on difficulty of the exam I would consider taking easier exam if you’re less silly mistake prone (-1=760 math sections are killer) and tougher exam if you’re mistake prone (where -1 could equal 790). I think general trend is that fall exams more difficult and spring exams easier w/ harsher curves. There is a site out there that has the curves listed for every exam so it might be worth investigating.</p>
<p>While individual items are pre-evaluated when they appear on an equating section, there is no way to “pre-test” an entire test. And the test AS A WHOLE will have a difficulty level that can only be determined by comparing the results on the test with the results on items in the equating section that have a “pedigree” so they can be used to cross-reference. Otherwise, how can you know if you have put together too many hard questions or not enough? It is AS A WHOLE that the test has a difficulty level, not just item-by-item.</p>
<p>This has been hashed out quite thoroughly in the past here at CC. I believe it is even stated quite clearly on the college board site that the so-called curve is established based on the results of the equating sections [and look at that name – they don’t call it the “pre-evaluation” section. They call it the equating section because that is precisely what it is for]. I’ll try to find the exact reference – that’s been posted here in the past as well.</p>
<p>I guess it is OK if we disagree about this as long as we are clear about the bottom line. I came to whatever understanding I have of this issue by reading other posts here. There are others here who have explained it better in the past. Maybe they can chime in and we can get this issue closed and stickied.</p>
<p>A good reference for anyone who is wondering about how SATs are equated is:
[this</a> College Board white paper](<a href=“http://professionals.collegeboard.com/profdownload/pdf/rn14_11427.pdf]this”>http://professionals.collegeboard.com/profdownload/pdf/rn14_11427.pdf). If you are really, really interested, try “Equating Test Scores” by Samuel Livingston ([pdf</a> here](<a href=“http://www.ets.org/Media/Research/pdf/LIVINGSTON.pdf]pdf”>www.ets.org/Media/Research/pdf/LIVINGSTON.pdf)).</p>
<p>Pckeller knows what he is talking about. Although the ETS knows how hard each question is individually (each question is pre-tested on an equating section), any new SAT is a previously unused arrangement of these questions.</p>
<p>I’ve said the following many times so I’m basically just copying and pasting from my previous posts. Why not once more for posterity:</p>
<p>The curve is <em>not</em> adjusted so as to make the average score 500 or any other particular number. Likewise, the number of people receiving 800s is not fixed or predetermined.</p>
<p>The curve is <em>not</em> determined before the test is administered (although the difficulty of each question independently is known beforehand). See papers above.</p>
<p>The curve does not depend on or account for the quality (or lack thereof) of the people taking the test in a given month. No one should worry about who is taking the test with them: it doesn’t matter. Take the test when it suits your schedule best.</p>
<p>I don’t know much about the specifics of equating; do you suggest that it’s a coincidence that the October SATs have, without exception, had extremely difficult critical reading sections compared to those of other QAS months? (I’m not insinuating that it’s not)</p>
<p>From my own experiences, it seems that it is severely harder to score 800 on the tests where there is a -3 800 curve, because most students aiming for 800 don’t leave questions blank anyways. Having been analyzing my repertoire of QAS tests this last week, I’m becoming slightly weary of taking the test in October given this trend in CR.</p>
<p>Reassurance?</p>
<p>It does not matter whether the test is easy or hard, it will be corrected by the curves.</p>