Chance your own danged self! (Here's how)

Chance Me / Match Me!
1

I am putting this out here because there is such a proliferation of “Chance me” threads and, in my
opinion at least, many of the responses are pretty low quality. I don’t claim to be the world’s
greatest expert, but I have been helping kids from my town pick schools to apply to for the past few
years, and I have read extensively on the subject and played around a lot with the available data, so
I think I have something to offer. The methods I am going to outline here are a) relatively easy to do
on your own and b) the results they yield are probably better than what you are going to get from
random internet strangers.

The first step is to collect the information for the schools you are looking at. By far the most
valuable source is the college’s Common Data Set (CDS) which you usually can find either on the school
web site, or by googling it. Currently, most schools have the CDS for the currently enrolled freshman
class (2014-15) by sometime around the end of the Fall Semester. Most of the information I will be
discussing below is available in Section C of the CDS. Some schools either don’t create a CDS or
withhold certain information. The second best source of information is the IPEDS database which is
maintained by the Department of Education (www.collegenavigator.com) - this overlaps pretty well with
the CDS, but generally lags behind by about one year (most of the data there right now is for the
2013-14 freshman class.) If you want information that is even more current, some schools make it
available on their web sites (check incoming class profiles or something like that.)

The key pieces of information that you need to chance yourself are:

  1. The admit rate (Section C1)
  2. Standarized test scores of the incoming class, usually given in 25th/75th percentile form (Section
    C1), and the SAT/ACT submit percentages (Section C9).
  3. The average GPA of the incoming class (Section C10) and/or the breakdown of the class ranks
    (Section C11).
  4. The chart that breaks the admissions criteria the school uses (Section C7).

A lot of this information is also readily available elsewhere (IPEDS, the school’s website,
CollegeData.com) if you can’t get a hold of the CDS or the data there is incomplete. The most common
piece of missing information is the GPA data, but I will explain how to estimate that below.
If you are chancing yourself for a large number of schools, it might make sense to put all this in a
spreadsheet, but for a reasonable number of schools, it may just be easier to do it by hand for each
school.

Here’s an example:

Source: Vanderbilt University 2014-15 Common Data Set

  1. Admit rate: 13.1%

Total Applicants: 13,058 male + 16,460 female = 29,518
DIVIDED BY
Total admits: 1,981 male + 1,884 = 3,865

  1. Test Scores

SAT
% Submitting: 41%
25th/75th scores
CR: 710/780
MA: 720/800
WR: 680/770
Combined: 2110/2350
(Note: I know it isn’t technically correct to combine the percentiles in this way - however, there is
a lot of cross-correlation in the scores, so it is not a mortal sin against the gods of statistics.)

ACT
% Submitting: 62%
CM: 32/34
MA: 30/35
EN: 33/35

  1. GPA

Average GPA: 3.78
Report %: 99.6%

GPA chart
3.75-4.00: 65.2%
3.50-3.74: 20.2%
3.25-3.49: 10.4%
3.00-3.24: 3.0%
2.75-2.99: 1.1%

Class Rank:
Top 10%: 91%
Top 25%: 96.7%
Top 50%: 99.5%
Report %: 34%

  1. Admission Criteria

Course rigor: 3
Class rank: 3
GPA: 3
Test Scores: 3
Essay: 3
Recommendations: 2
Interview: 1
ECs: 3
Talent/Ability: 2
Character: 3
First gen: 1
Alumni relation: 1
Geography: 1
State Residency: 1
Religion: 0
Race/Ethnicity: 1
Volunteer work: 1
Word Experience: 1
Level of interest: 0

So, that’s the raw data for the school. Now we can do a little additional work to make it more usable.

  1. Eyeballing the figures, it looks like it might be a little easier for a male applicant than a
    female, so it’s worth calculating the admit rates by gender:
    Male: 15.2%
    Female: 11.4%
    Difference: 3.8%

  2. Because there are schools where one test is more popular than the other or because the lower
    resolution of the ACT causes the scores not to correspond perfectly, it can be worthwhile to convert
    all the scores to the same system and combine them. You can find a SAT/ACT concordance on google. In
    this particular case, the SAT combined and ACT cumulative seem to correspond fairly well, so we’ll
    just use the SAT numbers for the rest of the calculations. The big thing to keep in mind is that exact
    scores aren’t really all that important - a 2000 and 2050 on the SAT are essentially the same score.

  3. We want to get a 25th/75th percentile for GPA, which requires a little estimation and some algebra.
    Again, this won’t be perfect, but good enough for what we will use it for. I came up with 3.61/3.92,
    which seems reasonable and close enough.

  4. This section doesn’t really require any additional math, but it gives us a good idea of how the
    school is weighting the various elements of the application. For the purposes of calulating a chance
    of admission, we are interested in the fact that the school weights GPA and Test Scores equally, and
    also that it considers Course Rigor just as important as GPA.

2

Ok, FINALLY, we are ready to chance. Here’s our first candidate:

Male
SAT scores (CR/MA/WR): 700/700/700
GPA: 3.80
Course Rigor: 4.00 (explained below)

For this example, it is all fairly straightforward. The overall SAT number is almost exactly right on
the 25th score for Vanderbilt, so his score for the standardized test component is 25%.
To see how his GPA stacks up, we have to estimate his course rigor, which I do on a 3.00 to 4.00
scale. This is obviously a somewhat subjective judgement and feel free to tinker with the formula
anyway you like. I look at the class schedule through the junior year, and award .05 points for each
honors/AP/IB/dual enrollment class taken + I add any AP scores 3 or over (divided by 100) and I cap it
4.00. Because some student don’t have the opportunity to take AP/IB classes, I make the floor 3.00.
In this case:
Honors classes: 8 (times .05 = .40)
AP classes: 6 (times .05 = .30)
AP scores: 5 + 5 + 5 + 5 + 5 + 5 = 30 (divided by 100 = .30)
So I rate his course rigor as 3.00 + .40 + .30 + .30 = 4.00 (This value is always capped at 4.00)

We see from the chart above that Vanderbilt rates Course Rigor equal in importance to GPA, so we simply average his GPA (3.80) and Course Rigor score (4.00) to get an overall score of 3.90, which is almost dead on the 75th percentile GPA estimated above, so we give him a 75% score for grades.

So we have:
Test Scores: 25%
GPA: 75%
which we average (50%) to estimate that this fellow would land right in the middle of the Vanderbilt
freshman class academically. To translate that into a chance of admission, we figure out what quintile
he lands in, and then get the factor from the following table:

Percent: Multiple
1-20%: .25
21-40%: .5
41-60%: 1
61-80%: 1.5
81-100%: 2.0

Then multiply by the overall admit rate (13.2) to get our chance% = 13.2%
If we want to be extra fancy, we might adjust by the difference in the male/female admit rates, but
there are issues with those rates (you can’t just assume that the two groups are equal in terms of
credentials or intended majors or many other factors), so I usually use just 1/4 the value of the
difference, which would be +.95, which raises our candidate’s chance to 14.1%.
Obviously, the numbers in this example were chosen to simplify the calculations, but everything is
just simple linear algebra, so I’m not going to insult your intelligence by spelling out the exact
formulas for everything. I also think that if you disagree with some the constants I am using
(especially for stuff like estimating course rigor, or translating the percentile rank into the
factors to multiply by the admit rate), go ahead and come up with your own. I am constantly tinkering
with this stuff - what I have works reasonably well for me, but I am not saying it is perfect by any
means.

The big takeaways that I think are important:

  1. Use BOTH GPA and Test Scores to rank yourself - there is lots of evidence that colleges weight
    these about equally, and you should too.
  2. The school’s admit rate is critically important and has to be the basis for your calculations.
  3. Obviously, there are many other factors to consider (see chart above) and they can make all the
    difference in the world to your application - but these are the objective factors that can give you
    the rough odds.

Anyway, if anyone actually reads through this whole thing, I hope that you find it helpful!

Good luck, and may the odds be ever in your favor.

3

Flaws in the method:

  1. In the form above, there is no attempt to incorporate hooks (URM, legacy, recruited athlete, whatever) - these are obviously critical components if you fall into one of those categories. I do try to take these into account in the spreadsheets that I use, but I left that stuff out for simplicity.

  2. You should not just rely on your overall test scores to judge your chances - you should look at the component scores and how they stack up. In general, if you are in the bottom 25% of any category, your chance are probably reduced.

  3. Your intended major definitely impacts the relative weights of your component scores - if you are an English major, your CR / WR scores should be weighted more heavily than your MA score.

  4. For specialized programs or sub-colleges within a larger university, the incoming test scores and the admit rates may be substantially different than the overall figures for the institution. The canonical example is engineering schools, which usually have much more stringent requirements than the overall numbers for the school. (Examples: Software Engineering at CMU, Stern Business at NYU, the Colleges at Cornell) For engineering schools, you can often get the data from websites, and you should if you are applying to those programs.

A different kind of example is an Acting Major at CMU - admission is going to be much more about your audition than it will be about your grades or test scores.

  1. Public universities have often have different admit rates for out-of-state students, and you should use those when applicable.

There’s plenty more to consider, and this being the internet, I expect people will point out plenty of more flaws here in this thread.

I do think the framework is pretty good, and that the results returned are in general better than you get from many of the commercial web sites that attempt to do this.

4

Interesting. Always wonder how you were coming up with you %s!!

5

Great method!

6

Note that some of the data on CDS is for enrolled freshmen instead of admitted student. For admission statistics, you may need to dig into the school’s website to find out.

7

^^^^ Yeah, I should have stated that properly - I use the enrolled freshman numbers, because they are widely available, despite saying admitted students up in the big post.

8

TL:DR

Are you near the 25th percentile for test scores and grades? Admit rate / 2 i= your chance
Are you near the midpoint of the 25th/75th for test scores and grades? Admit rate = your chance
Are you near the 75th percentile for test scores and grades? Admit rate * 2 = your chance

9

I am vainly going to bump this, because it has been buried in this section by all the “Chance me” posts.

I don’t expect very many people to read it all the way through, but if even a small percentage of the “Chance me for HYPMS” kids read the post directly above this one, perhaps a few of them will find their answer without posting.

10

Well done.

11

Good info. I do wonder why so many people make Chance Me threads and seem willing to rely on the opinions of anonymous CC posters (most of whom are other HS students) instead of reading the college’s own website and looking at its CDS.

12

^^^^ #10

Me too, that’s why I posted this.

  1. There is no reason that they can’t do it themselves
  2. I wanted to be able to point to this when people asked me about my estimates, so they could see where they come from.
13

Also - when kids put very detailed information on these boards, you know that there are admissions officers at elite schools trying to figure out who is who and getting a feel for who is serious about their school to get the yield down.

If I were a top 25 college, I would hire someone whose full-time job was to patrol these forums and figure out info on applicants based on cross referencing personal info/ test scores/ etc.

They are doing it - I feel certain of that.

14

^^#10 I’ve not only wondered that, but wondered about the “I’ll chance back”. If you know enough to chance someone else, can’t you chance yourself?

I guess that mostly they are looking for encouragement from others (even if they are “random internet strangers”).

I usually ignore those threads (mostly because I don’t feel I have the expertise to answer). The closest I’ll come is if I see a kid applying to a school my kid was accepted to then I’ll say something like — “My kid was accepted with similar stats as yours”.

15

@NickFlynn This is awesome. Thanks for sharing. I do feel like this would be very useful at the middle and lower ends of the spectrum. Do you think it might underestimate at the higher end? For example, in the case of Vandy, if your kid was a perfect 4.0/4.whatever and a 2400 SAT, the usual 10-15 AP/Honors classes, the max admit percentage would be around 26-28%. Seems like the admit data would suggest that this kid is probably more like a 80-90% lock. Especially for a stat grubbing school like Vandy. And yes I understand Vandy is a “reach for everyone” but they can’t be turning away too many of this type.

It would be great if there was a way to normalize admit rates. A 13% admit rate at Vandy has to be quite a bit different from a 13% admit rate at Columbia or Penn. One would assume that differences in self selection would show a vastly different applicant pool even though the admit rates are similar.

16

@Nerdyparent you make a good point. I was just reading a thread from someone asking to be chanced for “good but not great” schools to be his matches and safeties (CMC, Dartmouth and Tufts were on the list).

@gluttonforstress you make a good point too. Pitzer and Harvey Mudd both have 13% acceptance rates but they have very different applicant pools.

17

I think the comment in #12 is a little paranoid. I doubt many schools have the resources to track this kind of stuff. Putting the Harvard logo all over your Facebook page does probably have a non-zero chance of slightly reducing your chances at Yale and Princeton though…

18

@gluttonforstress I think you raise some good points in Post #14

To the first point (about Vanderbilt, and other very stats-aware schools) - I do think there is something to this, but I haven’t been able to come up with a decent and fair way to incorporate this kind of factor in a coherent way. I do think something real though.

I actually think that a lot of seemingly “perfect applicants” get rejected for real reasons most of the time (poor recommendations, essay that set off the “jerk detector” in adcoms, etc…) I actually am putting together a post about this which I should have up here within a day or so.

On the second point, again, you have a good point. I have actually fooled around with a method to “normalize” admit rates to compensate for some of these factors, but again, like the “stats-aware” factor, I’m not fully satisfied with the solutions I have come up with so far.

Both good points though.

And, again, I don’t think this is complete system (yet), but more a good framework to build on. The fact is that we can never really know the answer, all we can really do is estimate - BUT for decision making (making sure all your schools aren’t reaches, you have a safety, etc), that is good enough.

19

I think @Nerdyparent was making a point about all the identifying info that often gets put in a “Chance Me” thread. Sometimes in an effort to show they have a hook or to convey how impressive their EC’s are, the poster puts in a lot of identifying info. Particularly if the person is applying to a LAC which has a smaller applicant pool, I would recommend the student avoid making it obvious that he or she considers the school a safety or back up. Maybe it’s a bit paranoid but if it were my kid, I’d say don’t even take that risk. Yet another reason to use the OP’s method and chance yourself.

20

Cool!