SCEA Returns So Far

Colleges and Universities A-Z Yale University
21

<p>As indicated in my previous statistical posting, what follows is a more general summary of Yale SCEA posters who have posted decisions received both those who pledged to and those who did not. Some of the non-pledgers nevertheless were participating in the Yale forum and elsewhere on CC prior to the Dec. 15 release of results and some have posted only after that date.</p>

<p>Methodology</p>

<ul>
<li>As before, we have collected posters’ data from across CC, not just SCEA or Yale threads, in order to get as complete a picture as possible.</li>
<li>Where ACT scores only are reported, we have converted the score to an SAT equivalent according to the recently update ACT-SAT concordance tables. Where both scores are reported, we have used the higher score. Posters that do not report scores are excluded from the SAT statistics.</li>
<li>We have been attentive to collecting AP score data but not so with IB scores. Any AP score reported we included, even if the poster indicated s/he was participating in IB. This should be kept in mind when evaluating the average AP load, as IB students also doing advanced work do not have all or even any of their efforts reflected in the results. Again, posters that do not report scores are excluded from the AP statistics.</li>
<li>This time we have removed the results we had previously “assigned” to 17 pledgers (one has since reported) who have not reported any. Consequently there are 181 posters who have reported results.</li>
</ul>

<p>Results</p>

<p>Participation
At least partial data was collected on 210 posters
181 of these reported admission results (call these reporters)
156 reported SAT or ACT scores (86% of reporters)
110 reported at least one AP score (52% of reporters)
156 reported location (86% of reporters)
156 reported gender (86% of reporters)</p>

<p>Results
30% of 181 reporters were admitted
50% were deferred
20% were denied</p>

<p>Demographics
156 reporters reported gender: 57% male, 43% female
Of these:
33% of males, 27% of females were admitted
48% of males, 57% of females were deferred
19% of males, 16% of females were denied</p>

<p>156 reporters posted state/region or country (86%)
Of these…
22% were from Northeastern states (New England and New York, 7% from CT)
21% were from Atlantic states (MD, NJ, PA, VA, WV)
9% were from Southeastern states (AR, FL, GA, LA, MS, NC, TN)
13% were from Midwestern states (IA, IL, MI, MN, MO, ND,WI)
6% were from Southwestern states (TX)
3% were from Western states (AZ, CO, MT)
17% were from Pacific states (CA, OR, WA; 12% from CA, most of any single state)
10% were internationals (most: 3 from Canada)</p>

<p>Test Profiles
156 Reporters posting SAT/ACT scores, Average and Median SAT (w/ACT equivalent)
Admitted: 47, 2303, 2330
Deferred: 81, 2275, 2290
Denied: 28, 2192, 2190</p>

<p>Score brackets for admitted reporters giving test scores (25th percentile – 75th percentile)
M: 750-800 (median 780)
CR: 740-800 (median 790)
W: 740-800 (median 780)
Total: 2280-2370
Note: ACT-only and SAT-total-only scores are excluded from section brackets</p>

<p>Average AP score and average number of tests per reporter
Admitted: 4.86, 5.01 (36 reports, 67% of admitted reporters)
Deferred: 4.68. 4.49 (58 reports, 64% of deferred reporters)
Denied: 4.34, 4.86 (16 reports, 43% of denied reporters)</p>

<p>Admission results of reporters posting test scores by SAT/ACT score bracket
2390-2400 (22 scores) Admitted: 36%, Deferred: 59%, Denied: 5%
……….(note: ACT 36 = SAT 2390 by latest concordance table)
2350-2380 (18 scores) Admitted: 55%, Deferred: 33%, Denied: 11%
2300-2340 (41 scores) Admitted: 41%, Deferred: 51%, Denied: 7%
2250-2290 (26 scores) Admitted: 19%, Deferred: 61%, Denied: 19%
2200-2240 (16 scores) Admitted: 6%, Deferred: 75%, Denied: 19%
2100-2190 (18 scores) Admitted: 17%, Deferred: 39%, Denied: 44%
Below 2100 (15 scores) Admitted: 20%, Deferred: 40%, Denied: 40%</p>

22

<p>^ I think that is good evidence that if you are not hooked, you "need" 2300+ to have a real chance at admission. Below 2300, the admitted percentage drops from 41% to 19%.</p>

23

<p>^ in SCEA, at least. The pool is much tougher.</p>

<p>Plus, you're only dealing with a very, very small amount of applicants in the first place, and there were 15 more samples for the 2300+ range.</p>

<p>The 2390-2400 range had a 5% less chance of acceptance (according to this) than the 2300-2340 range kids, lmao</p>

24

<p>^ haha makes me think twice about sending my ACT score after all -_-</p>

25

<p>Could you perhaps use the data to make a naviance-like graph? Or perhaps maybe could I have a copy of the data? I can play around with it to make a graph..</p>

26

<p>
[quote]
Could you perhaps use the data to make a naviance-like graph?

[/quote]
</p>

<p>No, because the 2nd axis of a Naviance scatter-plot, GPA, is (1) not consistently available, (2) not consistently reported (W vs. UW, rounding practices unclear), and (3) not reliably standardized (different weighting schemes, different grade curves, different demographics).</p>

27

<p>Descartesz,
I know this is a lot to ask, but will you be doing the same thing for RD?? It would be great to have the data to compare the two pools and look at the age old question of whether or not those accepted EA have higher stats than RD applicants.</p>

28

<p>It would be interesting, wouldn't it? I think I will follow-up on the SCEA defers next April (assuming they cooperate at a level to extract some kind of meaningful data by posting their RD results) but I won't be doing the RD pool. It was enough to follow the 200+ potential participants in various SCEA threads and I had something of a personal interest: my S was in the pool.</p>

<p>Anyone else want to track the RD results? This is what I did:
-- I had a spreadsheet with poster name, stats links, test scores (SAT/ACT, SAT Subject, and AP), GPA, EC free text, state/country, region, gender, and a few other things of note.
-- On SCEA threads, I looked for participating posters and researched their stats or old posts for their stats as they came in (saving the link or links to their stats in the spreadsheet). It took several weekend hours to do this initially and about 10-15 minutes a day thereafter to maintain it.
-- The spreadsheet was set up to maintain most of the aggregate statistics (count and averages). Once finished, I sorted copies of the spreadsheet in various ways to get bracket breakdowns.</p>

29

<p>If it isn't too much trouble, what was the admit rate for international students? Canadians?</p>

30

<p>Great idea about following the deferred EAs into the RD round. </p>

<p>I would be willing to take a shot at RD, but I'll likely take a slightly different approach to reduce the collation time involved. I'll PM you and see if you think it will work.</p>

31

<p>Yes, please do follow the deferred candidates to the RD round! Yale doesn't make available how many applicants from the deferred pool are actually admitted in RD, so it would be helpful for next year's class to have the data. </p>

<p>And Descartez thank you so much for doing this in the first place. :)</p>

32

<p>
[quote]
If it isn't too much trouble, what was the admit rate for international students? Canadians?

[/quote]
</p>

<p>16 internationals, all of whom reported results: 25% admitted, 50% deferred, 25% denied.
13 of these reported test scores, averaging 2183.
14 of these reported gender, 10 male, 4 female
1 of the 3 Canadians reported admission, others deferred.</p>

<p>I have one poster indicating internationality and admission, but no particular country nor any other data.</p>

33

<p>Descartesz, wonderful number crunching! How did you decide what the SAT bands should be? Would love to see a graph using each score point or smaller bands so we could see exactly where the real peak is.</p>

<p>The stats certainly suggest that perfect scores might be an admissions disadvantage, but I found myself wondering how many of those scores were super-scored. Do you think the admissions folks are discriminating against students they think are focused too much on standardized tests or is there is a correlation between the highest scores and something else in the applications of those students?</p>

34

<p>I picked the bandwidths that I thought were broad enough to give a double-digit n and narrow enough to be interesting. They already appear to be subject to both noise and bias (how many rejected applicants bother to announce their results and theirs scores?), so I am disinclined to make them narrower.</p>

<p>At its root, of course, self-reported data from unsolicited providers is of marginal utility, but I do think the results say at least as much about CC as about Yale admissions. I do believe, for example, that the alleged bias toward postings from successful admits motivated to share their success is at least countered by postings from highly qualified but unsuccessful applicants seeking consolation and affirmation from each other.</p>

35

<p>Descartez – In the event that you are bored, I have compiled a dream list. Obviously, the work would be extensive so do or don't do what you wish. Also, on a side note, I would be more than happy to do the crunching but it is your data. If you would like assistance, you can PM me. As for the list:</p>

<p>1) ACT vs SAT – I understand that you converted the ACT to SAT using conversion charts and used those numbers in your statistics but there has been much debate about the ACT and is it respected and truly utilized the same as the SAT in the admissions process. With that in mind:
a) for those with ACT as their best score, how do their admissions statistics compare to those applicants with SAT as their best score</p>

<p>2) AP scores - So far from the looks of the data, the average number of AP tests for accepted, deferred AND denied rounds to a 5 and the average score rounds to a 5. I would love to see statistics that show a breakdown by number of AP taken by applicant. These dream stats would include:
a) a range of the number of AP tests taken by applicants, ie # of AP, number of applicants and maybe the associated average score
b) more importantly I would love to see (this would be HUGE) a breakdown by number of APs taken, scores and acceptance, deferred, rejected accordingly.</p>

<p>3) Any observations or impact of rank?</p>

<p>4) Observations on SAT II submissions?</p>

<p>5) Lastly, i am very curious about the impact of geography on admissions. There are many threads that ask that question. With that in mind, analysis of the admissions statistics with the data grouped by your 8 regions would be very interesting and then more extensively applying any of the above other requests to those smaller regions.</p>

<p>Again, I do apologize even suggesting any of these, but there might be some intriguing results. Thanks for your work. It is interesting to see.</p>

36

<p>Also, here is D's data. She (nor I) did not pledge to report. If you would like to add the info here are her stats:
Yale SCEA deferred
female
SAT 2080 (680 Reading, 690 writing, 710 math) - single test
ACT 35 ( 34E, 36M, 36R, 33S) 2 times, w/ superscore still 35 (34E, 36M, 36R, 35S)
SAT II Bio 780, Math II 790, Chem 720, Math I 740
GPA 4.0/4.0 UW, 6.26/8.0 W
Rank 1/178
AP 2: Bio 5, Chem 4
Massachusetts
Interviewed twice (on campus and alumni)</p>

<p>Also, one more creative idea. Again, only if you are REALLY bored. Use the academic calculator, and find our the index for each applicant and determine the statistics for that. </p>

<p>Any statistics involving interviews? There is often a debate about the significance of interviews.</p>

<p>Good luck.</p>

37

<p>UPDATE: I'm going to start an RD pledge thread soon and ask for some basic demographics and stats. I hope to get it going tomorrow and will sticky it until results come out.</p>

38

<p>Smoda61 --</p>

<p>ACT vs. SAT
I really don't think I have enough data to draw any conclusions about this. The number of ACT scores I have is in the dozens and there are far too many confounding factors. One example: Yale draws a disproportionate (compared to the population distribution of the U.S.) from the northeast and a disproportionate number of students from private schools, which are more heavily attended in the northeast. The northeast has a high SAT-participation rate. So one might expect that there will be an apparent preference for SAT admissions that is, in fact, not due to a favoritism towards the SAT itself but, rather, to geographic factors and academic backgrounds.</p>

<p>That said, I compiled this data for reporters who had SAT/ACT equivalent scores of 2300+.</p>

<p>** Admits **
SAT-based scores = 31
ACT-based scores = 5 (i.e., 5 where ACT was either only or highest score)</p>

<p>*Defers *
SAT-based scores = 31
ACT-based scores = 9</p>

<p>*Denies *
SAT-based scores = 6
ACT-based scores = 0</p>

<p>Totals:
5/14 reporters with ACT-based equivalent scores >= 2300 admitted
31/68 reporters with SAT scores >= 2300 admitted</p>

<p>You really need much more data and much better statistical techniques to determine any kind of test-preference.</p>

<p>AP scores
If I understand you correctly, you want to see counts and admission status breakdowns based on the number of AP tests taken. I'll see what I can do.</p>

<p>Rank
Rank is so often unavailable or not reported that I did not track it, so I can provide no help here.</p>

<p>SAT II's
In general I recorded up to the 3 highest SAT Subject tests reported, but always recorded the Math 2 score if available. It is, therefore, likely that there was actually more participation in some of these tests than my records indicate.</p>

<p>Of 47 admits:
33 Math 2 scores: Median 800, Mean 783
17 Chem scores: Median 790, Mean 774
16 USH scores: Median 800, Mean 785
15 Bio scores: Median 780, Mean 772 (M and E combined)
13 Lit scores: Median 770, Mean 764
8 Physics scores: Median 770, Mean 770
7 World History scores: Median 790, Mean 787</p>

<p>and a few others.</p>

<p>Geography
I'll see what I can do with geography.</p>

39

<p>I wish you could see the spreadsheet this data is on. It's monstrous. </p>

<p>The unfortunate thing about these data is that information given by pledged reporters varies greatly. This itself creates a self-reporting bias, so be somewhat skeptical of the scoring averages/medians.</p>

40

<p>Thanks for all your hard work, Descartesz and nickknack! It's really interesting.</p>