<p>@Pizzagirl Not sure how to respond to your rain/umbrella analogy except to say I usually don’t mind getting wet. The 4%/8% numbers were yours. I believe the actual real numbers in the example were 1 in 4 vs 1 in 8. Based on your comments and analogy, I think we look at the question fundamentally differently. You seem to look at it as only slightly decreasing your chances of something unfavorable occurring. I look at it as significantly increasing your chances of something favorable happening. Just the way I look at things.</p>
<p>@calmom “Because that’s the way the common data questionnaire is set up.” The CDS does show 25th percentile and 75th percentile numbers, but does not show the type of admit data I was referring to. The data these 13 schools share is not required and completely voluntary to my knowledge.</p>
<p>@intparent My oldest son just graduated from college and I started coming to this site frequently when he was a Junior in HS like many other people, and have continued since then. Given your welcoming response, I have no idea why I haven’t posted more … will have to give that some thought. </p>
<p>(I just chose Berkeley because was the first result I got in Google for a search for “Common Data Set”)</p>
<p>Look PAST the first table with the 25th/75th numbers at the two tables under the words:
“Percent of first-time, first-year (freshman) students with scores in each range:”</p>
<p>(Of course the writing test scores aren’t reported in the older forms as that test did not yet exist). </p>
<p>The reference you cited, Barrons – is drawing it’s information from the Common Data Set. Nothing extra is being provided to them by the colleges.</p>
<p>Here are some statements about test scores and college admissions I believe to be true:</p>
<ol>
<li><p>Very few kids are excluded from consideration because of low test scores. Selective colleges are willing to consider applicants with scores much lower than their averages, and they want to encourage such applicants to apply, both because it helps their numbers, and because they want to find that 1-in-100 diamond in the rough that they will want to admit notwithstanding mid-600 scores on all SATs. That doesn’t mean test scores are irrelevant above the qualification level.</p></li>
<li><p>Good test scores are helpful for applicants, but no one is admitted to a highly selective college based on test scores. Perfect test scores and a mediocre application means a rejection. Perfect test scores and an OK application means a rejection. High test scores are a plus factor for a good application. That’s about it. They are especially valuable for athletes and legacies, where the college has to defend itself against charges that it lowers its standards.</p></li>
<li><p>Colleges don’t pay attention to relatively small differences in test scores. One, small differences don’t predict anything for a small pool of individuals. Two, the colleges themselves compromise test scores by officially superscoring – so that when they report their numbers everyone is above average, just like in Lake Woebegone. Which means they know that a 30 or 40 point difference could simply be a function of practice, not any quality they care about.</p></li>
<li><p>If colleges let their average test scores get too high, it interferes with other things they care about, like good sports teams and racial diversity. You can’t have a whole class of people where 80% have 800 SATs and there are 20% athletes with 650-700s. Everything works better if there is a mix.</p></li>
<li><p>It’s a mistake to look at the difference between top scores and somewhat less than top scores. Because of the long tail problem, there are lots more kids with 800s on any SAT section than 790s or 780s. Some of those kids with 800s would have scored 900 or 1000 if the test had been scaled that high; others were 750 kids on a good day (or were 750 kids two tests and 100 hours of test prep ago). It’s pretty easy for the colleges to tell one group from the other from the rest of their applications. The 900-1000 types get accepted at a high rate. The lucky 750s get accepted at about the same rate as the other 750s.</p></li>
<li><p>Standardized tests are a good way to compare people from different backgrounds and different schools, but the same scores don’t necessarily mean the same thing. One kid’s 2300+ means that he was within 30 points of the equivalent scores his parents and sibling got. He was meeting expectations. His friend’s 2200+ reflects an amazing personal achievement against all sorts of odds – neither parent or any older siblings, or indeed any cousin, had finished high school, much less gone to college, and his home life was conducted in Spanish. As between the two kids, no rational person would assign a lot of meaning to the first kid’s 100-point test score advantage.</p></li>
</ol>
<p>@Calmom The data you’re referring to in the CDS is not the data I was referring to. The data in the CDS gives a distribution of scores for attending students, not admit rates by scores for applying students.</p>
<p>An example of the school published data I mentioned:</p>
<p>Actually if you are buying say 12 at least partially independent lottery tickets for admission, these differences are huge. 12 attempts (applications) at 8% each chance of at least one success is vastly better than 12 at 4% success rate. </p>
<p>I would hope that colleges don’t think differences of 30 points are significant. The SAT just isn’t that reproducible.</p>
<p>I once found on the SAT site an interesting document that addressed score changes upon retake and included information on very high scorers. Unfortunately I did not save it and as far as I can tell, the college board has removed it from their site. It’s my possibly faulty recollection that only about 20% of kids scoring 800 were able to reproduce that on a re-take, and it was quite common for them to score in the mid 700’s. Perhaps someone from this site knows the table I am talking about and has a copy of it. Not sure why the college board would want to stop disclosing this information. I think there may be a similar document available now but the information on very high scorers has been removed.</p>
<p>I think some of this comes down what we think is the difference between a 700 and an 800. Those who want to think that there is a cut-off point of 700 also want to think that there is no real difference between a 700 and an 800. And that, I think, depends on what colleges think the SAT measures, and how well they think the test does in measuring it. Does a higher score on the verbal test, for example, correlate to a better ability to write and to read, a better vocabulary, and so forth? If what the colleges are looking for is demonstrated competence, but they don’t believe that the SAT is the right way to show mastery, then a simple cutoff point would make sense–anything above a 700 means that the kid has a solid grasp of English essentials, and we’ll leave it to his grades in English class and his essay to show just how good he is. If the colleges actually think that a 780 on the CR means that the kid is probably a better writer or reader than the 700 scorer is, they’ll take the 780. If the 780 is in fact indicative of writing skill, that skill is probably borne out in the other evidence, making him a better candidate. </p>
<p>As in so many things, it’s probably a mixture–higher scores do look better, if the kid hasn’t taken the test six times, but they aren’t everything. And a lower score may be perfectly adequate, as long as the kid has enough other things going for him. But the Should I retake? asker really wants to know something concrete, in a process that seems so fluid, and the bottom line is that there is no known quantity.</p>
<p>Collegeboard reports SAT scores in a range of 30 points on each side, suggesting that even they don’t think a difference of 30 points is meaningful.</p>
<p>But 30 is not a hundred. Even sixty is not a hundred. I think it is entirely reasonable that a 750 is seen by adcoms as essentially the same as an 800; I’m just not sure if a 700 is. It may be; if the cutoff concept is the true one, it is; but in that case I start to wonder why the test is graded as it is. Maybe it should just be high pass, pass, fail, or some other system, to eliminate all the heart-burning (and the profit centers for test-prepping, and the inherent social discrimination).</p>
<p>An 800 is a LOT more impressive than a 700. The Montana rancher with 2400 will do a LOT better in admissions than the Montana ramncher with 2100, or 2200.</p>
<p>I personally believe that a 1400 is looked at quite differently from a 1600 (and a 31 from a 36), but is a 1550 seen as different from 1600 or 35 from a 36 (as some people here assert)? Maybe at some schools, but at most schools, I don’t believe that the data supports such an assertion. Furthmore, at some schools, a 1550 or 35 from a first-time single sitting is probably more impressive than a superscored 1600 or 36 from taking 5 tests.</p>
<p>I have to think that a 750 is viewed as almost equal to a 790, but neither is an 800. I think a 2400 will catch every admissions officer’s attention, even as they plow through applications with 2300+ scores. That doesn’t guarantee acceptance, of course. </p>
<p>Of students scoring 680-720 as juniors, 21% dropped 50 or more points in CR and 29% dropped 50 or more points in writing upon retaking as seniors. 15% dropped 50 or more points in math. I can believe that they forgot certain math concepts, but do 17 year olds become worse readers or writers in a few months? I do recall from the chart I saw before that the kids with higher scores than 720 were more and more likely to score lower upon retake as their initial score increased. Not surprising.</p>
<p>I suspect that woogzmama is right. Even though we know that there’s no real difference between a 780 and an 800, and that if the kids took it again there’s a good chance the scores would be swapped, the 800 candidate is going to seem more impressive even though it’s probably only one question different. Maybe the admissions officers have learned not to be impressed by it.</p>
<p>I do believe there is a difference between 700 and 800. If you look at the chart, only 7% of students were able to increase their score by 80+ points–which may or may not even have gotten them to 800. That was true in all three tests. If 93+% of kids cannot raise their score to 800 (or at least 760 for the kids starting at 680) in a retake, I think it’s a real difference. Whether you think it’s an important difference is another matter, as the scores at the high end are as much about being careful and pacing oneself as they are about knowing the material. </p>
<p>If you look at the scoring charts (eg, the one for their official practice test at <a href=“SAT Practice and Preparation – SAT Suite | College Board”>SAT Practice and Preparation – SAT Suite | College Board) in CR you can only miss 2/67 questions and still get an 800; missing 9/67 gets you a 700. In math and writing you need to be perfect to get 800; missing 6/54 gets you a 700 in math and missing 5/49 gets you 700 in English. (I’ve ignored the deductions for wrong answers for simplicity, so let’s just assume the student skipped those questions). Overall, the student with straight 700’s missed 20 questions and the student with straight 800’s missed at most 2. That seems pretty significant to me and can’t be explained by careless error, accidental wrong answer entered, misread or misunderstood question, poor exam pacing etc, in the way that a 770+ score can.</p>