A general formula to determine match?

Parents Forum
1

<p>I am a technical person and always trying to put a transfer function on a process. Based on the admission data I have seen on Cc so far, I think one could come up with a general formula. Of course, as in any real life process, there will be outliers.</p>

<p>Stats X admission rate x E/C factor</p>

<p>If one’s gpa and standard test match the 75% of a school, he/she get 0.75; For schools have around 10%, 20%, 20-40%, and 40% or higher admission rate, one could use 1, 2, 3, and 4, respectively as the second factor. An exceptional E/C, such as intel winner or a founder of a national level movement, use a E/C factor of 3, above average E/C, use 2, and avergae use 1.</p>

<p>Any one get below 2 will be a reach, 2.5 - 5 will be a match. </p>

<p>What do you all think?</p>

<p>A stduent with above average E/C and stats at 75% of a school with 10% admission rate will get 2x1x0.75 = 1.5. therfore that will be a reach.</p>

2

<p>The X factor is your particular HS. If kids with your kids stats regularly get rejected at the college, where your kid fits with the mean stats of the school is pretty irrelevant. I think the naviance data is more helpful than the national statistics; they control for rigor of curriculum, grade inflation, etc. a lot more than the mean GPA’s do.</p>

3

<p>I agree with blossom, for example at my high school, kids with 1500/1600 SAT scores and GPAs of 102 or better (don’t know how to translate that to 4.0 scale) have about a 50% chance of getting into Harvard. That’s a lot better than you’d think, keeping in mind that many of those top students are also Intel semi-finalists, Davidson scholars and other cool stuff. Same kids applying to Stanford have a zero percent chance. (Didn’t stop mathson from trying, he got rejected too.)</p>

4

<p>mathmom,
DS1’s school has a similar track record as yours re: Stanford. Princeton, too. On the other hand, there are other top schools that love those same kids. If your school uses Naviance to track data of students and where they apply/are accepted, I am finding that very correlative to DS1’s experience this year. It’s been more accurate than anything else so far.</p>

5

<p>The weakness with this approach is that it tries to objectively describe a process that at many schools has a strong subjective component. Rating fundamentally different ECs is the old apples and oranges comparison. And the method takes no account of “what the school is looking for.”</p>

6

<p>Apart from the factors mentioned by blossom and mathmom, I think there is a lot more refining you have to do.</p>

<p>For example, for a college with a 10% admission rate (HYPS), being at the 75% level in SATs and GPA does not mean anything like a 75% chance of admission. Maybe it means a 20% chance of admission, maybe not even that.</p>

<p>For most schools, even HYPS (although their yield is so high it doesn’t matter much), the stats on their admitted classes will be somewhat-to-significantly higher than the stats on their enrolled classes, since the higher-stats applicants are more likely to have other options and thus more likely to choose one of the others. So, as a rough matter, I generally assume the the 75% level on the enrolled class represents something like the average for the admitted group. That’s probably over-conservative, but you get the idea. Also, I think your function considerably misstates the difference between having stats at the 75% level and having stats at the 25% level. At colleges that accept less than 50% of their applicants, a 75% kid probably has more than three times the admissions likelihood of a 25% kid. But a kid with “above average” ECs probably doesn’t have twice the chance of a kid with “average” ECs (and how about below average?).</p>

<p>You also ought to account for some status factors – underrepresented minority status, legacy status, other strong connection to the college. Those can make a difference, although how much of a difference varies from college to college. Also, geographical issues. At a northeastern college, an applicant from Montana will always have the edge over an applicant from New Jersey with a comparable profile. Geography can be tricky, though. Here in Philadelphia, a very strong applicant probably has a much better than average chance at being accepted at Penn, but an average applicant may have no chance at all.</p>

<p>Basically, if you define a “match” as at least a 50% chance of admission, you probably have to exclude the most selective colleges categorically, or at least de-weight them more than you have. (Actually, you have probably de-weighted them enough to ensure that no one is ever a match there.) Sure, a Mexican-American Intel finalist/student body president with 1600 SATs and a 4.0 GPA with a challenging curriculum probably has better than a 50% chance of being accepted at Harvard. But I don’t know who else has.</p>

<p>I also think you need to account for some obvious things. An applicant whose resume does not scream math-science is not going to be a match for an engineering school, regardless of SATs, etc.</p>

<p>In the end, the whole “match” concept seems to have limited utility, especially for the strongest students. All of their “matches” are “reaches”, with some very limited exceptions, and some of those exceptions are widely believed to discriminate against applicants who look like they are probably headed to HYPS (see: “Tufts Syndrome”). In any event, it’s a pretty limited universe, and you don’t need an algorithm to identify it. Word of mouth does just fine.</p>

7

<br>

<br>

<p>Very interesting, I have 2 questions - what is the sample size? Over the past say 5 years, how many kids applied and were rejected at Stanford, AND, secondly, why? has anyone questioned why? Does Stanford think the kids are going to Harvard or MIT instead, so they don’t bother accepting them?</p>

<p>It is strange to me to hear those stats, because in our small area (total population in 2 counties of about 400,000 tops), there are probably 25-50 Ivy caliber students (that is kids with the gpas and the SATs and the ECs) each year, half of those don’t apply to any selective schools, but the acceptance rate of the other 10-20 is pretty high, probably at least 50%, maybe even 75%, and they go to a wide variety of schools - I know of all the Ivies except Cornell, Stanford, MIT, etc. I don’t know of any university that particularly “shuns” any high school. Maybe it is a geography or novelty thing - it is so rare for someone to apply to Stanford, that if he makes a good case , he gets admitted.</p>

8

<p>The Ivy League (maybe others too) use an Academic Index to calculate a number that seems to correlate somehow with admissions. It can be found on the CC site:</p>

<p>[The</a> Academic Index - Ivy League Admissions Key?](<a href=“http://www.collegeconfidential.com/academic_index.htm]The”>http://www.collegeconfidential.com/academic_index.htm)</p>

9

<p>I have a technical education also. But I’ve found technical disciplines to be little help in human endeavors. (What was that line in The Boxer? “A man hears what he wants to hear and disregards the rest.”) As the designated Parent-in-Charge-of-Risk-Mitigation it was my job to be relatively certain that there would be at least two college acceptances for each of my daughters. And any technical person will tell you that “being sure of two” mean “being likely at many.”</p>

10

<p>

We’re an east coast school and substantially fewer kids apply to Stanford than HYP and MIT. Unfortunately the scattergrams no longer seem to be up at our school’s website or I no longer have access to them, I think there were around 12 applications to Stanford vs. 40 to Harvard. In four years two kids have been accepted at Stanford, both with SAT scores around 1250, weighted GPAs around 95. There were a few kids with worse scores/grades, many with better. In the Harvard case, all the acceptances came from the top quarter of the scattergram. Of the Stanford acceptances, I later found out one was an underrepresented minority and related to a NYC politico, the other was an athlete and a legacy. I was pretty sure looking at the Naviance scattergram that something like that was going on, but only found out for sure after the fact. It’s possible that Stanford is so busy accepting kids from the wealthy suburbs in our county that they don’t realize that the top kids at our school are great candidates too, or it may just be a statistical blip. Still it made it clear to me that Harvard was probably a match not a reach for my son and that Stanford was highly unlikely.</p>

<p>I was under the impression that the Academic Index is used mostly for determining how academically unaccomplished the sports teams are allowed to be at the Ivies. Am I wrong?</p>

11

<br>

<br>

<p>Maybe Stanford has developed a relationship with one or more GCs at those county schools, and just accept those kids - “Bingo, one accepted, check off this metropolitan statistical area”. The acceptance to Stanford here was unusual enough that it got mention in the paper.</p>

12

<p>our school, also northeast like mathmom’s, is very large and competitive and according to 4 years’ worth of data shown on Naviance, the admission rate for Stanford is 8% and Harvard is 16%. Harvard has twice the number apply, but that shouldn’t make a difference–should it? The acceptees at Stanford have been–and this is hearsay–athletes of a certain sport, but athletes with very high grades and test scores. Princeton and Yale accept 8% and 9% of our applicants (Princeton likes a certain sport, too, but there again the athletes are at the tippy top of the academic chart). Dartmouth and Cornell are 23%, Columbia and Penn are 19%, Williams is 18%, WashU is 23%. Lots of kids apply to all of those, so the numbers probably are pretty good at predicting, once you look at accepted GPAs and SATs.</p>

<p>More interesting stats…UNC-CH is 6% and MIT is 6%, too. And Caltech is 50%. Seems high? Only two kids in the last 4 years–Caltech is very self-selecting, isn’t it? One was accepted. The rule here is not to look at acceptance rate but average GPA accepted.</p>

<p>So numbers don’t–and never will–always tell the story. There will always be several X factors: school, location, makeup of individual college classes.</p>

<p>But it’s fun to speculate, certainly keeps us busy while waiting. I remember when I was pregnant with the first I poured over research and old wives’ tales about sex of babies born a certain time of year/day (and also doing things holding a needle on a thread over my belly and seeing if it spun around–scientific things like that). And then there are people coming up to you saying, “It’s a boy!”</p>

<p>I think I’m off track. But I’m just agreeing with blossom that one’s local Naviance data is more meaningful than a nationwide equation but that neither will reveal the secret to college admissions. And statistics didn’t stop my kid from doing a little magical thinking during the college app process.</p>

13

<p>" . . . the stats on their admitted classes will be somewhat-to-significantly higher than the stats on their enrolled classes, since the higher-stats applicants are more likely to have other options and thus more likely to choose one of the others. So, as a rough matter, I generally assume the the 75% level on the enrolled class represents something like the average for the admitted group. That’s probably over-conservative, but you get the idea."</p>

<p>Yeah, that is probably too conservative:</p>

<p>Amherst Class of 2011 </p>

<p>Yield: 40%</p>

<pre><code> Accepted Enrolled
</code></pre>

<p>CR 690-790 670-770
M 670-780 660-760
WR 680-770 670-760</p>

<p>That is, the 75% figure for enrolled students is well above the average for accepted students. The accepted students stats are higher than the enrolled – but not that much higher. (The data do show, not surprisingly, that accepted student with scores in the 750-800 range enroll at a lower rate than those with scores in the other ranges.) </p>

<p><a href=“https://cms.amherst.edu/media/view/33730/original/61st+Revised+SSR.pdf[/url]”>https://cms.amherst.edu/media/view/33730/original/61st+Revised+SSR.pdf&lt;/a&gt;&lt;/p&gt;

14

<p>

love this one. :slight_smile:
btw, what’s “local Naviance data”, is this the long list shown up in some HS profile that how many kids addmitted to what colleges in 4 years duration? But there is no detailed gpa or standard test score info shown in the list.</p>

15

<p>anotherNJmom: our school’s naviance site has graphs for each college but also has charts with data for the last four years. These are even more interesting than the graphs, as they show GPA averages and SATs for Applied (early and reg), Accepted (early and reg) and, even more interesting, the lowest and highest GPAs and SATs of those applying/accepted, so you can see if you are way off before you apply. Since it’s online in our Naviance section, I assumed it’s part of every school’s Naviance data. Gee, I would hope so. Maybe it’s not part of Naviance itself but what our school does. Not sure…</p>

16

<p>A lot of schools don’t have Naviance, including ours. Our high school has a book showing the SAT, GPA, class rank, and number of AP courses for each student that applied to each college, along with status of attend, accepted, waitlisted, denied. The data runs to many pages for certain popular schools and of course is not sortable unless you hand copy all the numbers, take them home and put them into an Excel spreadsheet.</p>

<p>In terms of the formula idea, I have a concern with how GPAs are calculated by schools. While the common data set asks for the information on a 4.0 scale, it is clear that some schools have used weighted GPAs. For example, the UVA average is listed as 4.05.
“C12. Average high school GPA of all degree-seeking first-time, first-year (freshman) students who submitted GPA: 4.05”
If you were applying from out of state, your high school’s weighting system may be entirely different and therefore the GPA would not be comparable. I’m sure the admissions folks are smart enough to make the proper adjustments, but I don’t know how the out of state applicant could. And, if UVA shows their data this way, I have to believe other schools do as well. There have been numerous threads demonstrating that weighting in one school district is not at all comparable to the weighting process in another district or state.</p>

17

<p>Neumes,
I thought someone posted once that a school can buy various levels of packages from Naviance. Like your school, we seem to get the whole kit and kaboodle of data. Our school has eight years of data, divided up into the old 1600 SAT (2005-prior) and current 2400 SAT (the last two years). </p>

<p>Our data is compiled into larger groups and sorted into interesting charts (i.e., where else kids who applied to School X also sent apps), but there are also the scattergrams where one can see each applicant (anonymously) plotted by SAT and GPA. Having the 2400-based data as a separate scattergram is especially helpful, as it captures the new scoring methodology as well as the increased competitiveness of the pool in the past couple of years.</p>

<p>This was very helpful in seeing which schools were truly safeties, reaches, and don’t bothers – with this caveat – within the context of this particular school, what it has to offer, and the students who attend.</p>

18

<p>WesDad:</p>

<p>I’ve seen those numbers from Amherst before. I’m not sure how characteristic they are, though. (1) Amherst only admits about 850 kids, total, and only about 700 RD. (2) A huge percentage of Amherst’s enrollees have to be athletes. (3) Amherst is ultra-selective, so that numbers in general probably mean less there than almost anywhere else besides HYPS.</p>

<p>That said, in terms of numbers of kids, there is a pretty meaningful difference between a 790 CR score and a 770 one, or 780-760 M. In the class of 2007 SAT figures, there were about 9,200 kids with 790 or better for CR, and 16,600 with 770 or better. For math, the spread is 17,200 (780+) and 28,900 (760+). In other words, almost twice as many kids in the applicant pool may be at or above the 75% level for enrolled students as for accepted students. And, at the 25% level, there are as many students nationally with 670-680 CR as there are 750+ (about 30,000), and there are almost as many students with 660 M (25,700) as there are with 760+ (see above). I think the practical difference represented by those 200 points is not insignificant.</p>

19

<p>If anyone wants to see what a typical Naviance scattergram looks like here is a link [url=<a href=“http://www2.newton.k12.ma.us/~brad_macgowan/naviance]Naviance[/url”>http://www2.newton.k12.ma.us/~brad_macgowan/naviance]Naviance[/url</a>]. Just type the passwords into the guest slot.</p>

20

<p>I’m sure DadII’s formula has flaws because any formula will. And I imagine he well realizes that.</p>

<p>But I think any proposals for some sort of quick-and-dirty quantifying mechanism are worth looking at, because I know we all want some kind of shortcut - as reliable as possible - to figuring out which schools will be in the reach/match/safety for our kids’ applications.</p>

<p>I know my S’s school had a guideline for stratifying a kid’s list - I don’t have it at hand, but it used SAT scores only and was something along the lines of +60 points over the benchmark score (don’t recall if this was the 75% mark or mean) is a safety, within -60 to +60 of mean is a match and -60 below is a reach. Now this was a few years ago and is obviously flawed. But the school has good results with kids’ admissions to all levels of selectivity.</p>

<p>My point is that people are out there using <em>something</em> to make a first cut at reach/match/safety. Everyone knows final decisions are holistic at lots (not all) schools. But different proposals for a numbers-driven method for the first cut are worth looking at, imho.</p>