<p>This issue is actually discussed here - <a href=“The Difficulty With Data | MIT Admissions”>a | MIT Admissions. Personally, I would be skeptical of any school that did not have admissions rate correlate with test scores.</p>
<p>
</p>
<p>It is not that they lie as much as being oblivious to the incongruity of their positions. In articles and interviews, they often recite a couple of romantic viewpoints about what admission should be. They loved to talk about equality in education, a return to the love of learning, a return to a world of leisurely filled summers left to exploration and not resume padding. </p>
<p>That is what the bosses at Harvard and the former Jefe of MIT brought to the table. Yet, at the same time that the defrocked Marilee Jones was decrying the rampant increases in EC and pursuit of “academic badges” her own office greatly expanded the numbers of lines on the applications that were used to list such activities! </p>
<p>The reality between how their own office work (and what they measure and reward) is hardly the same as they HOPE it might be. For instance, Jones could have sent a clear message that they did NOT care to see 12 APs or 10 schools ECs by reducing the list to FOUR and asked students to present the ones that were the most meaningful. </p>
<p>Romantics or forked tongues? It is both with various percentages. </p>
<p>PS Does one really believe that MIT does not look differently at a 700/700/700 than at a 790/800/760? Seriously? </p>
<p>It is not because a school spent a lot in a public blog to describe the process that it carries the truth … all the time! </p>
<p>What is the difference between saying we do NOT reward higher scores, but simply reward the student who happen to have higher scores mixed with great qualifications? Like you do not reward the height of your BB center but reward him for grabbing 20 rebounds a game because he happens to be taller than everyone else. </p>
<p>Bottom line? The higher your scores, the better your odds are. And that is because nobody would think that lower scores actually help you! </p>
<p>billscho - I am not a math expert, but I do see the problem with your understanding of statistics. You are confusing correlation with causation. The higher test scores are not the reason they are getting in, the kids they are letting in simply have higher test scores. </p>
<p>You see, what happens is that kids who aware early on in life that they are academically gifted are either guided or guide themselves toward the actions that most lead to admissions in the most selective schools. </p>
<p>Another way to think of it is that because of their selective reputations, these schools have a higher percentage of applicants with astronomical scores. It only stands to reason that those with the higher scores will tend toward the top as they also would tend to have higher GPA and class rank.</p>
<p>FWIW, the correlation between SAT math 25th percentile and admission rate is -.78.</p>
<p>
</p>
<p>I do. But maybe I am too trusting. At the very least, CollegeBoard even says you can not differentiate between a 700 and a 760. Maybe you don’t trust CollegeBoard either, but it would be weird to go in the direction of overestimating the accuracy of SAT scores.</p>
<p>MIT reps have made different statements about scores in different places. I’ve seen reps say that if your scores are 700+ on each subsection, the scores won’t keep you out of MIT; or reps say that the difference between a 2100 and 2250 is more significant than the difference between a 2250 and 2400, but a 2100 will not keep you out of MIT if the rest of the application is competitive; or SAT subject tests (not SAT I) evaluation have a threshold that varies by the subject and is often near 700. </p>
<p>I do believe that MIT emphasizes scores to a lesser extent than most other highly selective colleges. Various other aspects of MIT’s admission process are also unique such as having a far higher admit rate for similarly qualified female applicants than male applicants, not considering legacy, treating interviews as an important and influential portion of the admissions process, saying the most important criteria for admissions is the personality/character/match – more important than grades, course rigor, scores, etc. They go into more detail about this “match” criteria at <a href=“What we look for | MIT Admissions”>http://mitadmissions.org/apply/process/match</a> .</p>
<p>Naviance, scatterplots, Parchment and other sources do support this lesser emphasis on scores and greater emphasis on holistic criteria. Among male Parchment members with perfect stats (perfect GPA and SAT, with high course rigor), their admit rate to MIT is lower than their admit rate to all other colleges with a statistically significant sample size. I realize Parchment is flawed and mostly self-reported, overestimates chance of admission, etc. Nevertheless, I find it useful this type of imprecise trend.</p>
<p>
</p>
<p>Have you ever considered that the people who throw that sentence around are usually the same who admit not being a math expert. Or the ones who understand nothing about statistics! </p>
<p>
</p>
<p>I DO trust many statements and many studies presented by the College Board. They are often very detailed and benefit from probably having the ONLY useful database on standardized testing in the US. But, it also happens that people accept the statement that fit their own views and dismiss the rest. </p>
<p>The College Board states that repeating the SAT will not improve the scores by more than 30 points. They have the statistics to support that. However, what they do not state, is that the expectation is based on students repeating the test without … having much effort or hiring a professional tutor! Repeating a test with the same preparation should in theory yield a very small improvement. Yet, when students are “coached” by COMPETENTtutors or accept to learn from their own mistakes, the improvements can be massive. </p>
<p>So the TCB is not incorrect; but only correct for the cases they want to present … and those are the 80/90 percent of the Joe Six Pack students who do not care much for a higher score. On the other hand, the people who make a living improving the scores of students by hundreds of points know … how effective a test prep can be when done well! </p>
<p>As far as difference between a 700 and a 760, I would cling to my perhaps erroneous opinion that schools do routinely make a distinction between a 2100 and a 2280! And when odds are in the low single digits, I think that subtle differences do … make a difference. </p>
<p>Now as far as trusting an adcom? I probably would give the benefit of the doubt as long as they are not lying and hypocritical devil worshippers who fabricated their resume. Humm, pretty clear who that’d be at MIT!</p>
<p>One kid’s safety is another kid’s reach.</p>
<p>For some kids, all schools are a reach (low stats & poor).</p>
<p>For other kids, all schools are a safety (super hooked: e.g., dad is POTUS).</p>
<br>
<br>
<p>Oh Lord, I am dying from laughter! - another quote of the day!</p>
<p>^
You are correct – Hillsdale, and a couple of others, would not consider the family of the current POTUS a shoo-in. The ROW … are likely to find it hard to reject any of them. </p>
<p>
</p>
<p>@Torveaux I am not a math expert either. I am just a scientist and use statistics a lot on my daily work. We do use correlation as a suggestion of causation all the time in our publication though. Perhaps they are not looking for higher scores, but they are looking for best candidates who happen to have higher score. The reality is, the best candidate is less likely to have weaker scores and that is what the correlation is telling. When the best candidate is correlated with higher score, it also suggests having a weaker score would make it less likely to be the best candidate. Statistics is not a direct proof of a statement, it is just an assessment of the validity of a statement versus possibility of something happens by chance.</p>
<p>I just wanted to make the point that correlation statistics CAN prove causation. It depends on the research design. There is a difference between correlational statistics and correlational research designs. Correlational research designs can’t prove causation but correlation statistics CAN prove causation when used in an experimental research design. …for those who care…</p>
<p>OK, if all of this actually made sense then those kids with the 800/800/800 would get into every school to which they apply. That is simply not the case. I don’t know exactly where the line is for each school. 700 may be low, but I would be shocked if they ever looked differently at a 760 versus and 800. I know some schools do not use the writing for admissions at all. At some point, they stop looking at the scores and move on to other things. Either way, you cannot expect for any of the highly competitive schools to be a safety for anyone. (excepting the aforementioned FDOTUS?)</p>
<p>Where do you draw the line between Match and Reach? Is is <50% chance = reach?</p>
<p>No, the admission rate is higher for 800/800/800 but still not 100%. So they won’t get into every school they apply. I usually put 20% or less as reach.</p>
<p>Um, no, you have an invalid assumption, @Torveaux (and also ignore another issue).</p>
<p>Firstly, schools vary by how much they value test scores. Secondly, many schools care about their yield (or want kids who have them as their first choice).</p>
<p>There are some private elites (and many publics as well) who, I’m positive, would take every 800/800/800 kid they can get their hands on if they were assured that those kids are 100% sure to attend, no matter how unimpressive or down-right sucky they may be in other areas (assuming no massive character red flags like robbing a gas station on their record). Other schools are more holistic. Different schools weigh different things differently. Simple concept.</p>
<p>I’d really like to see those SAT/GPA plots (but only for ED candidates) at various schools.</p>
<p>I kind of do think that a less than 50% chance of admission is a reach. Certainly, a 30% chance of admission is not as much of a reach as a 5% chance–but you could apply to a bunch of 30% chance places and get into none of them.</p>
<p>And it’s really not quite accurate to talk about odds and chances, anyway. There is really no random element in the selection process–other than, perhaps, the makeup of the overall pool (and that isn’t really random, either–it’s just out of anyone’s control).</p>
<p>^That’s why I have no patience with trying to figure out exact odds or worrying about the definition of match vs reach. If you get rejected from all the schools you thought were matches, all that matters is what your safeties are. You never know what will happen. The year my son applied to RPI it got published in Newsweek as one of the 25 new Ivies. The number of kids who applied went up by a lot, and the acceptance rate plummeted. I think it was still a safety for our kid, but someone who thought it was a match for them, might have been shut out by it’s new popularity. I thought my oldest had a better than 50% chance of getting into Harvard. I was probably wrong, all I know is he did get in. But we never, ever considered it anything but a reach. </p>
<p>There is no exact definition in terms of numbers. A safety is a school you are absolutely (well you can never really be a 100% certain, but close) going to get into based on acceptance rate, test scores, etc. You need to apply to a couple of these. Then apply to schools you are interested in, given you can afford, etc., without a specific ratio of matches and reaches, but you should try to include a few schools, where possible, with a moderate admission rate, where you would be near 50th percentile of students. Reach is anything below 20%, regardless of scores, IMO. </p>
<p>I know three kids with near perfect SATs. Two are at Harvard and one is at Yale. And a perfect ACT I know is also at Harvard. There’s apparently only 500 perfect scores per year, so I doubt any US citizen with a perfect SAT score ever gets rejected from anywhere. Overseas is a different story because cheating is so prevalent.</p>