<p>Sadly, I think Hunt is right. The difference between a 2300 and a 2400 shouldn’t matter, but there’s evidence to suggest that it does at the elite schools. For example, Brown (which provides fairly detailed admissions rates by SAT score): admit rate of 22.2% for 800 CR compared to 14.5% for 750-799; admit rate of 18.2% for 800 M compared to 13.3% 750-790; admit rate of 22.4 for 800 W compared to 14.9% for 750-790.</p>
<p>Very similar for Dartmouth: 34.4% for 800 CR compared to 16.3% for 700-790; 20.8% for 800 M compared to 13.7% for 700-790; 32.3% for 800 W compared to 14.9% for 700-790.</p>
<p>that’s a slippery slope argument: next you know my 2260 is just as good as a 2300 which is just as good as a 2400. =)</p>
<p>
</p>
<p>i agree with this quote, but realize that you can miss a few questions and still get a 2400 =]</p>
<p>as for whether a person with a 2300 should retake the sat…well it depends. my friend who received a 2290 on the sat retook it and got a 2370 the next time. basically, if you know you can do better, take it again. if you feel like you maxed out, don’t get cocky</p>
<p>It’s not a slippery slope of an argument. Everyone knows the difference between 560 and 600 is larger than the difference between 760 and 800. Mathematically this is true if you realize that there is about a 4 question difference between 560 and 600 but a possible 1 question difference between 760 and 800. You get penalized a lot for being marginally less than perfect. </p>
<p>And the thing about post 81 is just correlational. There’s no (or little) causation there. The possible confounding variables are enough to dismiss causality in the relationship.</p>
<p>It would certainly be nice if one could be pretty much guaranteed admission with a 2400 SAT or 36 ACT coupled with a #1 class rank – the current “holistic” system leaves pretty much every student wondering if they’ve done enough, even if they barely have time to sleep at night. Of course, one might then need to set a limit on how many times one can take the SAT/ACT and stop this silly super-scoring trend.</p>
<p>While I still consider a 2400 vastly superior to a 2300, I wonder if a once-taken 2300 should be worth nearly the same as a 5-times-taken, then super-scored 2400. This is an issue we have not yet addressed here: how do we account for those who have deep enough pockets and a strong enough obsession to become professional test-takers?</p>
<p>Our admissions folks (who are very good, and have backgrounds in college admissions offices) say that taking the SAT more than three times will generally be seen as a negative. Adcoms want to see kids spending their time elsewhere. Kids at our school do VERY well on the PSAT/SAT. (We generally have about 50% of the graduating class at the Commended level or above.) The admissions folks strongly suggest that kids not take the SAT more than twice, and for many, once is sufficient.</p>
<p>A student of that caliber will more likely than not get into more than one elite University, such as HYPSM+Other Ivies+Duke+Chicago+Top LACs. It always happens.</p>
<p>LoremIpsum was referring to the lack of a guarantee and, in turn, the anticipatory uneasiness – some of which may be justified because, well, it doesn’t always happen.</p>
<p>So… I got a 2300 with 800 writing and 800 reading, and 700 math.
Colleges like math.
Should i retake?
I’m afraid I won’t be able to ace writing and reading again, although they do come fairly naturally to me…</p>
<p>Predictably, threads like these largely become warehouses for propagating the ancient, imperishable myth that scores above 2XX0 (with the first X not being a “4”) won’t make any difference in admissions, particularly for the universities that are so pervasively revered on this message board. </p>
<p>
</p>
<p>This is the same slavish argument, and the truth is that you aren’t aware of any data-verifying, a-causal models that qualitatively support your opinion. If you actually look at the data, the associated graphical transformations, and analyses of these “confounding variables” you’re forging as “just correlation” you wouldn’t dispel the reality of a steepening contingency of admissions probability on test scores with such totally implausible and transparent dismissals. For many universities’ data sets, you will observe a weaker or decelerating dependence on class rank or GPA once within the upper echelons. But that’s not true for standardized test scores. Other admissions entities, or these “confounding variables,” as you would inaptly call them, such as work ethic, dedication, extracurriculars, reference letters, work experience/volunteering, institutional interest, essays, GPA, grades, and so forth are much more difficult to evaluate and compare, so the notion of diminishing relevance, and even flat contingency in some aspects, may apply. It’s incredibly untenable to assert that largely subjective criteria are the causal components driving the “correlation” beyond whatever “2XX0” you think is the magical cutoff where supposedly the score attained no longer matters. </p>
<p>Moreover, there is not some underlying list of factors causing all this “correlation” data. Standardized test scores comprise the only reliable barometer that uniformly applies to all applicants and meaningfully disperses the pool across an objective scoring spectrum. Before perpetuating the hollow nonsense about “correlation” and “confounding variables” it would be nice to have a more fundamentally sensible understanding of precisely what you’re arguing for or against. </p>
<p>Good grief, mifune, Never use a long word when a short one will do, and never use a dozen when one will suffice. You aren’t penning an SAT essay here, judged on length rather than content.
You’ve also set up a strawman. The question does not concern ‘any score of 2xxx’ compared to the Holy Grail 2400, but more specifically, 2300 and above compared to 2400.
I fail to grasp the basis of your seeming scorn for the terms correlation and confounding variable.
Standardized test scores are hardly a uniform barometer.
Espenshade’s data uses a range of 1560-1600 for the top bin of data which is equivalent to 2360 to 2400. There are also the curious dips in admissions rates in some of the higher bins at some schools. MIT RD admit rates , for example, are slightly higher for 1450-1500 bin than the 1510-1550 bin. A similar dip exists for Yale RD data. Avery’s data is similar.</p>
<p>That’s very funny. You complained about another poster’s writing earlier in this thread (who frankly didn’t deserve it), then conceded to the pointlessness of the tactic, and now you proceed to irrelevantly chide about another’s. </p>
<p>
</p>
<p>Your charge is relatively baseless because I wasn’t, as a matter of course, referring to the thread title; I was referring, in general, that any opinion stating that the SAT score becomes irrelevant after any arbitrary 2XX0 (2100, 2250, 2300, etc.) is a load of complete rubbish. On the contrary, at the most selective universities, the top end of the scale is frankly where the most meaningful distinctions can be made among plausible candidates, hence the steepened contingency. </p>
<p>
</p>
<p>The quote to which I responded in #91 claimed that there were “confounding variables” causing the accelerating, nonlinear dependence of test scores on admissions probability, leading admissions-to-test score figures to be merely “correlation[al],” which is totally false.</p>
<p>
</p>
<p>That’s incorrect. Three of the four institutions depicted graphically [on RPR, page 8] show slight, flat, or reducing SAT sensitivity at the mid-ninetieth percentile ranges, with Princeton’s showing a marked dip across the 93rd to 98th. But the MIT data in the RPR and the annual aggregated figures presented on [it’s</a> website](<a href=“http://www.mitadmissions.org/topics/apply/admissions_statistics/index.shtml]it’s”>http://www.mitadmissions.org/topics/apply/admissions_statistics/index.shtml) don’t demonstrate that behavior. MIT admissions shows a uniformly escalating, nonlinear curve as scores approach the upper percentiles and it’s something that becomes more apparent as SAT performance tables are disaggregated by separate demographic groupings, namely, by ethnicity, gender, and domestic or international status. That is, the aggregated tables presented on university websites and studies that don’t control for such conceal the SAT’s influence. It follows that as scores increase, both international representation and U.S. admission increase, as well, with international admittance rate being a fraction of the domestic rate.</p>
<p>Princeton’s data set, on the other hand, reveals reducing sensitivity to test scores in the mid-ninetieth percentiles, with steepening dependence on scores above the 98th, which is uniformly true for the four universities represented on page eight of the RPR. The effects can also be seen in Espenshade’s regression studies (primarily from pre-1995 data) on SAT admissions effects.</p>
<p>One can also see the effect of disproportional preferences of individuals with SAT scores above 1500 (out of 1600) in Espenshade’s study on preference categories in a paper more commonly cited during discussions on other forms of preference (e.g., affirmative action, and athletic and legacy preferences). The influence of test scores rapidly heightens at the very top percentiles; there isn’t a flat cutoff or a reducing sensitivity. </p>
<p>Also, below are some percent-admittance categories the RPR released from various SAT-SAT Subject Test performance combinations from a sample size of roughly 1,400 applicants:</p>
<p>1400-1490 SAT I, 700-790 SAT II – 34% (percent admitted)
1500-1590 SAT I, 700-790 SAT II – 50%
1500-1590 SAT I, 800 SAT II – 71%
1600 SAT I, 700-790 SAT II – 74%
1600 SAT I, 800 SAT II – 84%</p>
<p>Moreover, I take it that you now don’t dispute the legitimacy of the RPR study. Months ago, you deviously attempted to discredit what I stated in another thread on this topic by saying that the RPR had never been published and never passed peer review, both of which are totally false. One could claim that Avery et al only had access to exterior data by monitoring individual applicants throughout the admissions process. But it’s in no way a methodological knock and Avery et al reached the same fundamental test score conclusions as Espenshade’s regression models (which were formed directly from institutional data) and the linear statistical modelings presented in the Early Admissions Game. There’s consilience among them, and any competent statistician with access to the available data can draw the same test score-to-admissions causality determinations.</p>
<p>After looking over these studies (which were previously referenced by a poster named siserune -coincidental suffix?) I actually changed my opinion about the importance of the SAT in admissions decisions. My main concern was about the vintage of these data, but this report from Brown seems to confirm that, at least there, you receive a significant boost merely by recieving an 800 -</p>
<p>It’s hard for me to believe that the overall resume of your typical 800 scorer differs significantly from that of a 750-790 scorer. So the only variable left is the actual test score.</p>
<p>Granted, these aren’t strictly controlled studies, but I was at least marginally convinced that there is at least some benefit to higher test scores across the spectrum.</p>
<p>Not that this means one should bother to retake a 2350 - there’s no guarantee of improving your score anyway. And not that I think this is necessarily a good thing. In some instances I think carelessly missing one problem on the Math section could drop your score 30 points - and I’m not sure that’s really indicative of much of anything. I just think this is the way things are.</p>
<p>2400 is better but by little. Getting a 2400 as opposed to 2300+ really means the student with the 2400 answered a few more questions right than the student with the 2300.</p>
<p>S2 is one of the many who wonders if he should retake even though he has a high score (2360). His 240 psat makes him think he has a shot at a 2400 and after reading this thread I’m starting to think there could be a slight benefit if he is able to bump up his score. He has received plenty of advice against it. He only took it once and didn’t do a lot of prep so I don’t think he has necessarily hit his max. Also, do school really look at a 36 composite ACT as highly as a 2400 SAT?</p>
<p>^^^Despite my recent conversion on the “SAT score verses admissions rate” question, I would still not retake a 2360. And if I had a 2360 and a 36 ACT I would submit both scores and forget about it.</p>
<p>The difference between 2360 and 2400 really is negligible. Even by CC convention, 2350+ is about as safe as you can get with your SAT score. I implore you not to let your son retake.</p>