<p>tokenadult:</p>
<p>I come to a different conclusion to your post #11, likely bcos I'm a cynic. :D</p>
<p>CB is literally a data-mining and publications factory, and, of course, a master at marketing. They have the data. You have asked for it. They have ignored your requests. Thus, my plausibility is that it must show something; and that "something" could embarass their big customers, starting with a certain college in Cambridge. If the data showed nothing, they would publish it, no?</p>
<p>Oh, I'm sure they have a report generation cycle at College Board just as at most huge organizations that generate lots of data. Their table on the average score changes found upon retesting </p>
<p>showed class of 2004 data for a few years, then showed class of 2007 data BRIEFLY right after the class of 2007 reports were released, then reverted to the 2004 table again while the College Board website was extensively redesigned, and now again shows the class of 2007 data. College Board is a big, lumbering bureaucracy, and I'm sure release of score data has to go through several layers of approvals even for aggregate data like the data you and I are curious about. But, yes, if your curiosity continues, write to College Board and tell them that this issue is frequently debated online, and maybe by the time class of 2008 data come out (that would be August 2008), there will be a report showing highest single-sitting scores for each test-taker, and another report showing highest "superscores" for all repeat test-takers. That would be interesting reading for several of us.</p>
<p>
</p>
<p>Been there, done that, as I think I already posted to this thread. </p>
<p>Anyone who has personal experiences with superscoring to relate is welcome to post to that thread. I only took the SAT once. Most of the repeat test-takers I know personally (who are mostly middle-school-aged students taking the SAT as part of a Talent Search program) improve or stay steady in all their sections as they retake, so that superscoring has no effect on them. </p>
<p>What I'm disagreeing with, of course, is the estimate that the effect of superscoring is a "50 to 75 point differential" in actual practice among actual groups of students who apply to differing colleges. Exactly how big a difference superscoring makes is an empirical question, with an empirical answer specific to each college's applicant pool. But I doubt it is that large.</p>
<p>Well, I totally agree with you. I'd like to see the numbers, too.</p>
<p>Here's another thread on here with a similar theme that I just recently posted to:</p>
<p>How about we look at each person's postings as of today's date and time.</p>
<p>The following posts have been made:
(First is the person's name--then the difference between their superscore versus non-superscore, single sitting result):</p>
<p>Historymom: insufficient information
‘rent of 2: +50
dvm250: insufficient information
Elf220: 0
CurrySpice: not applicable
Brillar: 0 (but only CR and Math given)
NewEngland SocSciMan: not applicable
Theprez: +80
Bearpooh: not applicable
SeniorSlacker: +30
Somedumbnoob: not applicable
Ducktape: stated that superscored was better, but didn’t give scores
Akahmed: will count as 0, but not specifically stated
2blue: not applicable
kluge: insufficient information
makin it rain: insufficient information
gk23: not applicable (scores not from a real test)
fa-la-la-lena: 0
deutsch: 0
historymom (again): not applicable
moewb: not applicable
bulletandpima: 0
treesnogger: +20
historymom (again): not applicable
dte: 0
orchestramom: +40
kikib1028: 0
Debbie427: +10
Hellohelloduh: 0
Maas0890: +50
Calcruzer: +40</p>
<p>Including me, the average of the 16 who actually gave results is:</p>
<p>8 had tests where there would have been no superscore change, 8 would have changed</p>
<p>Overall average of those 8 whose scores would be higher if superscored: 40
Overall average (includes all 16 responders): 20.00 point difference between superscore and single sitting score.</p>
<p>Of course, this isn't scientific at all, either--but it would support the argument that I'm too high and you're too low--and may be the most accurate of anything I've seen so far. (Doesn't chi squared analysis say we need 30 totally random totals to arrive at a 95% confidence interval?)</p>
<p>And you make a very reasoned argument also when you point out that knowing the overall is nice--but knowing how it affects each school's numbers is probably what people really want to know when considering which schools to apply to and attend.</p>
<p>A good AP statistics course would tell that you need to make sure your sample correctly represents your population, and that your statistics correctly describe your sample. </p>
<p>Technology</a> Innovations in Statistics Education </p>
<p>By that I mean that if we want to know why, say, Berkeley shows a different score profile from Penn, </p>
<p>College</a> Search - University of California: Berkeley - Cal - SAT®, AP®, CLEP® </p>
<p>College</a> Search - University of Pennsylvania - Penn - SAT®, AP®, CLEP® </p>
<p>then we'd want to sample students who apply to those colleges.</p>
<p>Yes, I totally agree--which is why the previous thread I posted statistics for is not totally valid for our discussion of Ivy league schools. (like I said, it is not scientific)--and as you pointed out earlier, the data is not currently available to arrive at a conclusion and to prove or disprove one way or the other.</p>
<p>I'll wait for collegeboard to put out the statistics--if they ever do.</p>
<p>Thanks, Calcruzer, for the clarification. So will you join me in writing to the College Board to ask them to publish the aggregate superscore table for all test-takers? :) You'll have my thanks and bluebayou's (I venture to guess) if College Board publishes those data.</p>
<p>tokenA:</p>
<p>the problem with CB's data (post 44 and others) is that we do not know if these are separate students, so, as a math guy, I'm surprised you'd even post it.....as a cynic, I'm guessing that the data request will fall on deaf ears bcos it will show what some of us think... :)</p>
<p>bluebayou, while you may be correct that CB has something to hide with respect to this particular issue, I know several social scientists who study education issues and they are constantly frustrated by the refusal of CB to release any data about any subject to anyone.</p>
<p>I wouldn't conclude anything based solely on the fact that they are keeping the numbers close to their chests; that is just the way they do business.</p>
<p>
</p>
<p>No, we do know that, although it is said in the body of the SAT National Report rather than in the table </p>
<p>itself. (I think I may have learned this from PapaChicken--it was some other CC participant, for sure.) The table shows the highest single-sitting score for each distinct individual in class of 2007 who took the SAT. The previous year's table showed the highest single-sitting score for each distinct individual in class of 2006. </p>
<p>College Board could (and in my opinion should) produce a comparable table showing the "superscored" composite scores of each individual test-taker (without student identifying information, of course). As midmo points out, there are other data that College Board could release that would be helpful in resolving some even more controversial issues, but precisely because those issues are controversial, I suppose College Board is not in a hurry to release those data. But write 'em a letter and see what happens. Over the years, it has been easier and easier to gain data on certain (usually less contentious) issues than it used to be from the College Board, just by Web browsing rather than by having to request a print publication.</p>
<p>While I join others in the wish that CB was more permissive with the release of its data, I also lament that colleges likewise sometimes aren’t very transparent about what their data means. In particular, I am referring to those institutions that claim to admit based only on a single SAT sitting, but actually report their data (whether on their CDS or to USNWR, IPEDs, etc) on a per section basis and reflects super scoring. </p>
<p>In other threads, I have drawn the analogy that this is similar to the financial records of many companies where they have numbers for operating purposes and a second set of numbers for tax purposes. They use whatever set of numbers suits their institutional need. Barring an explicit statement from a college that their CDS or other reported numbers are drawn from the composite admissions, single sitting only, score, I will always conclude that the reported data is superscored.</p>
<p>my superscore and my one-take score differs by 110 points.</p>
<p>How 'bout this proposition: What if your one-and-only SAT I score (so far) is
CR 800 W 800 M 680. You think that with some focussed studying you can raise the M score by at least 50 pts, maybe more. If you can totally rely on superscoring, should you re-take the SAT, and leave all sections relating to CR and W undone, and ONLY do the M sections? It would certainly be less stressful than doing all of the test. Would you do that? Assume for the moment that all the schools you want to apply to superscore.</p>
<p>I am still of the belief that the best way to do well on the SAT is to be relaxed and "on" for all three sections. That's what was reported by many of the students who replied in the </p>
<p>thread. </p>
<p>I have read, since I prepared the earlier FAQs, that College Board (not the colleges themselves) looks askance at students who have, say, </p>
<p>200 800 200 </p>
<p>one time and </p>
<p>800 200 800 </p>
<p>some other time. Those students are often investigated by College Board on the warranted assumption that they may have used more than the allotted amount of time on some of the test sections. DON'T CHEAT ON YOUR TEST. Do the test honestly, within the set time limits. Do it without worries about the effect of a (possible) small drop in scores on one or two sections, because most colleges are quite in earnest about superscoring.</p>
<p>an admissions rep at an Ivy reported that a kid received an 800 on Math, but high 600 on CR. Upon retake, he scored an 800 on CR, but 200 on the math. He was rejected not bcos they assumed he cheated, but bcos he had an unfair advantage the second time, i.e., napping during math while other folks were stressing their brains.</p>
<p>"I don't know why people would be taking the SAT multiple times if they weren't expecting a better score through superscoring."</p>
<p>I think it's worth mentioning again that "superscoring" only benefits you if you go up on one section and down on another. If you go up (or stay the same) on all sections of the test, that just becomes your best single sitting score. But...because the admissions process is so subjective, there's no way you can put too much credence into a statement like, 'We see all your scores but only consider the highest ones." While that policy may mean that adcoms will not state--even to each other--that candidate A is better because he got 2400 in his first sitting as compared to B with a superscored 2400--I think it's very difficult to believe that many people wouldn't think that A is superior to B, and that will be part of the overall impression of the two. But this is true of anything that adcoms will know, even if it technically is not supposed to be considered.</p>
<p>Allow me to repeat myself: </p>
<p>Colleges have given up trying to distinguish one-time test-takers from two-time or three-time or even four-time test-takers, because that wasn't useful information to the colleges. There are a number of reasons for that. </p>
<p>1) The colleges have utterly no way of knowing who spends all his free time practicing taking standardized tests and who takes them "cold." </p>
<p>2) The colleges are well aware that students who have actually taken the tests sometimes cancel scores, so they have little incentive to give students bonus consideration if the students submit only one test score. </p>
<p>3) The colleges are aware that students who take the admission tests at middle-school age, who are numerous, do not have their earlier test scores submitted by default. </p>
<p>SAT</a> Younger than 13 </p>
<p>Hoagies</a>' Gifted: Talent Search Programs </p>
<p>Duke</a> TIP - Interpreting SAT and ACT Scores for 7th Grade Students </p>
<p>4) Colleges are aware that the majority of students who take the SAT at all take it more than once. </p>
<p>5) Colleges are in the business of helping students learn, and they don't mind students taking efforts to improve their scores. They know that students prepare for tests. </p>
<p>
<p>These arguments make sense to Mr. Fitzsimmons [dean of admission at Harvard], who said, “People are going to prepare anyway, so they might as well study chemistry or biology.” He added that “the idea of putting more emphasis on the subject tests is of great interest” to his group.
</p>
<p>Colleges treat applicants uniformly now by considering their highest scores, period. </p>
<p>
</p>
<p>Well, in answer to tokenadult's earlier question, I'll write the collegeboard and ask them to put out this data, but I doubt my one letter (or even two or three from others on the site) will carry much weight. </p>
<p>I believe bluebayou and midmo are correct when they say the information is unlikely to be released. The think collegeboard is more likely to respond to one or more college(s) requesting that this data be released to the public than to a few individuals.</p>
<p>As for Hunt's comment that schools probably look at all the test scores and take them all into account somewhat:</p>
<p>It is usually the situation at most schools that the adcoms only look at what is on the application--and in many cases, only the best scores are asked to be put on the application. At most of these schools, these numbers are verified for accuracy by clerks whose job it is to compare the submitted test transcript to the test score reported by the applicant--but in most cases, the test transcript (showing all test results) is often never even given to the admissions committee--but rather kept in the "backup" file.</p>
<p>Hey guys. I was wondering if the score ranges for accepted students posted on colleges' websites are superscored scores or single-sitting scores. Any ideas?</p>
<p>I have only one (mild) disagreement with what tokenadult is saying, and that is that "colleges" do not read applications or look at scores--individual human beings do that. If the college has a procedure whereby the adcoms only see the superscored highest SAT score, that is one thing, but if the adcoms see all the scores and are told only to consider the highest as a matter of policy, that is something else. I believe (perhaps because I am cynical) that the results at the margin would be different in those two approaches.</p>