Old SAT scores compared to redesigned SAT scores.

Applying to College
221

Lots of data from Vassar http://congrats2021.vassar.edu/about/ New SAT scores similar to or less than Old SAT. Note that they included what I think the number of students for each type of test, 693 for New SAT, 357 for Old SAT, 1024 for ACT

New SAT EBRW Math
Mean Score 717 717
Median Score 720 720
Mid 50% Range 680-750 680-760

SAT I (Old SAT) CR Math Writing
Mean Score 724 720 721
Median Score 720 720 730
Mid 50% 690-770 680-770 690-780

ACT
Mean Composite Score 32.5
Median Composite Score 33
Mid 50% Range 31-34

Looks like means and medians are higher for Class of 2021 than for Class of 2020, for both Old SAT and ACT https://admissions.vassar.edu/about/statistics/

222

In this debate/speculation about which test the “elite” students took this past year, I can only share what the guidance counselors at my son’s elite prep school told us. (1) take the test that fits you best (based on PLAN/PSAT or taking each one once); (2) if you want to see the biggest “jump” in scores, it’s easier to do this on the ACT, due to the ability to prep for it more & also take it over & over & over again to superscore.

My son is a bit of a black sheep and/or contrarian, because he decided on the new SAT (he did better on it than ACT) & the ACT really turned him off (too rushed, too many questions, it stressed him out). He also decided to only take the NSAT 2x…once in March & then again in the fall of senior. The second time, his scores went up (60 pts) & he just decided that he was done & that he wasn’t going to waste anymore time/brainpower/stress on it. So, his submitted SAT is NOT superscored (690m/700v). Meanwhile, he had friends taking the ACT every month, strategically focusing on improving certain sections to get a Superscore that really seems contrived to me.

In the end, I do think having a “borderline good” score of 1390 did affect his outcomes in college admissions, to the good & the bad. While the NSAT did not affect his admissions into safety schools, I feel it did result in unpredictable & unexpected results in his match & reach schools. It yielded him a hefty merit scholarship at one match school (U of Miami) which was nice surprise, but also probably contributed to 5 WLs (Wake, BC, NE, Colby & Villanova). I would have thought a 1390 at BC & NE would be low & earn a rejection (but he got WL), while with 1390 he should have been a shoo-in at Nova. (I realize there are tons of other factors, including Nova’s bb success leading to an 20% increase in apps, but he has a 4.1 GPA & tons of ECs, captain of a varsity sport, student govt – just like almost everyone else out there applying). He did not get into any of his “high reach” schools – but DID get transfer options from USC & a Guaranteed Transfer Option from Cornell… In other words, the results are a bit all over the place. I would never have expected him to have the opportunity to attend Cornell (as a soph), but not Villanova…

223

Superscoring is way more common on the SAT (old or new) than the ACT. For example, HYPS do not superscore the ACT. I think Penn is the only Ivy that superscores ACT. Here’s a list: https://www.princetonreview.com/cms-content/superscoredownloadablepdf.pdf

224

@evergreen5

A very important difference is that the Class of 2020 figures are for actual enrolled students, whereas 2021 figures are for admitted students. The 2021 figures for actual enrolled students will almost certainly be closer to the 2020 figures, because the strongest candidates treat Vassar as a safety and many of them will matriculate at higher ranked schools.

These figures Vassar provided are very interesting because they provide the actual number of students submitting each test. The first thing we notice is that the total number of test submissions (2074) is greater than the total number of students in question (1769). Assuming each student submitted at least one score (Vassar is not test-optional, and in fact advises students to use College Board’s concordance tables to determine which of their scores is best*), that means there are 305 “extra” scores from students who submitted more than one test type.

Of particular interest are the 75th percentile scores. When I was looking at schools for my unhooked daughter, this is the main one I would focus on. We can see a gap between the 75th percentile of admittees (2021: ACT 34) and enrollees (2020: ACT 33); this is to be expected. Also, if you look at the 2020 data, the ACT and old SAT percentiles are in perfect alignment with the concordance tables (33 vs 1480). However, the 2021 data show a misalignment between the ACT and old SAT 75th percentiles (34 vs 1550). A 1550 on the old SAT concords to a 35 composite ACT. These old SAT scores are really high (more along the lines of the 75th percentile of an Amherst or Williams) and I consider this another indication that the old SAT cohort was a small group that self-selected for high scorers.

If you got a 1550 on the old SAT, you did it before the new SAT even came out. Many counselors will say: Your score is high enough and you’re done, don’t take the SAT/ACT anymore. Therefore, a lot of these kids wouldn’t have bothered with the new SAT, so the new SAT cohort is deprived of a small but significant population of high scorers. This will depress the 75th percentile for the new SAT at a school like Vassar.

For reference, here is the Academic Rating (AR) cheat sheet that Williams admissions allegedly uses to group applicants according to academic caliber. The SAT scorers are for the old SAT. Unhooked applicants are not admitted below AR 2, and are preferably AR 1.

     verbal     math      composite  SAT II    ACT    AP

AR 1: 770-800 750-800 1520-1600 750-800 35-36 mostly 5s
AR 2: 730-770 720-750 1450-1520 720-770 33-34 4s and 5s
AR 3: 700-730 690-720 1390-1450 690-730 32-33 4s

http://ephblog.com/2010/12/02/academic-rating-at-williams/

Similar rubrics are used at Amherst, etc.

*Vassar’s advice to applicants to use the College Board’s concordance tables:

SAT middle 50% ranges are based on the pre-March 2016 (“old”) SAT. Please Note: The CollegeBoard has indicated that scores for the redesigned SAT are higher than scores from the pre-March 2016 SAT. To determine how your scores from the redesigned SAT compare to these middle 50% ranges, please use the CollegeBoard Concordance Tables to convert your scores.
https://admissions.vassar.edu/apply/answers/#stdscores

225

Median scores for Macalaster, Class of 2021 admitted students:

https://www.macalester.edu/admittedstudents/classprofile/

226

So is new sat harder?

227

nillupa, it depends on what range the scores are in. The new stats are in. A new 1420 equates to an old 1430. A new 1460 to and old 1470, and a new 1510 to old 1540. This is based on the colleges acceptances. No one will ever know if the two pools were much different, but when looking at the stats for who schools accepted, high school class of 2017, the concordance tables appear to be very far off, by about 60 points at the 1450 range, less of a difference for lower scores and higher at the top level. 60 points coming from they said a new SAT would be 50 points higher, and it was actually 10 points lower. These stats are from colleges such as Vassar and Hamilton, who listed acceptance rates for new and old SAT scores. You need to look at very competitive schools with very high SAT scores to see where the concordance tables went wrong, they weren’t off that much for average scores.