Doesn’t mean there was causation.
You are looking at the only data point (out of many) available to you, and thus assume that one was “the” deciding factor - when it could also be possible that anyone over 15xx was in the colleges same “Top Academics” bucket, but your school’s top students also consistently managed to show various external awards, balance deep involvement in causes/interests for several years,… - which was what actually stood them apart.
My point about Harvard/Yale is that we get so few kids in per year (usually only 1) that our naviance data wouldn’t allow me to confidently say 40 point differences are meaningful/not meaningful. Kids going to HYP always have something “more” that is pretty substantial.
In my experience, it’s different for Penn. Lots of applicants. Lots of admits. I know the admitted kids and their EC’s. And I would advise a kid applying from my HS that their stats better be extremely high. Again, no one under a 1550 has gotten in.
Cornell is much more of a mixed bag – with each college having very different admit rates and HUGE discrepancies between ED and RD. I really can’t emphasize enough the enormous difference between the ED admits and the RD admits at our school. When you apply appears to be much more important than 40-50 points on the SAT.
All this to say that Naviance data can be illuminating for what individual colleges value – especially for colleges that get a ton of applicants from your high school.
Maybe – but there are lot of applications/admits. NOT saying SAT was determining factor (there are also a lot of 1550+ kids who were rejected.). But the data suggests 1550 – for our school – is a threshold. No data suggests it’s anything less (for our school).
Since the process is so opaque, Naviance data is the best we’ve got. Especially if you personally know the kids. Which all leads me to my opinion that Penn cares more about super high test scores (in my community) than other top schools…
Penn isn’t some generic behemoth.
Wharton admits look different than Arts and Sciences in my neck of the woods. No, Arts and Sciences isn’t admitting by major, but a kid with a passion for ancient history and culture who got a 5 on the AP Latin exam junior year (so maxxed out on their HS’s offerings) but is learning Sumerian and Akkadian on their own does NOT need the same SAT scores as the typical Wharton admit.
My observation is that there is MUCH more variability at “Penn” than at some other highly rejective schools so you need to know much more than GPA and scores to have any insight into what’s going on. I know several nursing school admits who would not have had the stats for Arts and Sciences to study bio… but their obvious passion for a hands-on career in health care, and the volunteer and work experience that went with it, made the stats somewhat irrelevant. Could they do the work? Yes. That’s table stakes for nursing- and you don’t need 800’s for that.
Pity the poor Huntsman applicant who looks at generic data on “Penn admits” at their HS which does not break out early applicants (if you are a legacy, alums have been told- “your kid needs to apply early or we don’t consider legacy”), the various programs and schools, who had which EC’s, etc. I tell everyone who asks me- “assume you aren’t getting into Huntsman. What’s plan B?” and I’ve never been wrong. None of them have gotten in…
Agree that conventional wisdom/generalizations get thrown out the window for specialized programs. But the vast vast majority of our school’s high stats applicants are for Arts& Sciences or Engineering (and are not learning Akkadian). 
We recently had a kid get into UCLA’s new “Music Industry” program who was in the bottom half of the HS class and had an 1100 SAT (UCLA is, of course, test-blind). Not a URM/FGLI. But that makes sense for specialized programs – SATs measure verbal/math skills. Those aptitudes are not as important for all fields.
I think there’s way more noise in the data, none of which we’re ultimately privy to, to provide any useful conclusions. And I’d say that that includes those applicants where you believe you “know the kids.” You don’t know their applications. You don’t know their essays. You don’t know their LORs.
The Naviance spreads are so HS-specific. At our large, diverse public school where 75% go to college and about 10% go to sub-20% acceptance rate schools, the test scores for the highly selectives are much less predictive than the GPAs. Uniformly high weighted GPAs (4.3+, which is about the cutoff for top 10%), test scores as low as 1100s (presumably not submitted). Yes, acceptances cluster in the upper-right, but pretty much every highly selective school has a trail of green checks to the left. Just underscores how important high school context is in these conversations.
Penn seems to have a high score cutoff at our school. There are a lot of applicants as we are in the region. More wiggle room in GPA than test score.
I apologize in advance for the length of this, but something was bothering me, I think I figured out what was bothering me, and I would be interested in thoughts from other people.
OK, so part of what was bothering me was the Yale AO’s reference to the SAT Math score. And although this could have meant nothing, it seemed odd to me in the sentence before he referred generally to both the SAT and ACT, and then he only specifically mentioned the SAT Math score.
And then as I was looking at SCOIR (although not for Yale since we don’t have Yale loaded), I was beginning to get this feeling that despite the official concordance being marked out on SCOIR, I was not 100% confident the SAT and ACT results were lining up neatly.
OK, so I looked at their CDS, specifically with an eye to their SAT Math subscores versus ACT math scores, and, well, it was pretty eye-opening. What I am going to do is report here the HYPSM SAT and ACT 25/50/75, both composite and then Math subscore, and I will put the official concordance in parens:
MIT
SAT Comp 1520(34)/1550(35)/1570(36)
ACT Comp 35(1540)/35(1540)/36(1590)
SAT Math 790(35)/800(36)/800(36)
ACT Math 35(780)/35(780)/36(800)
Princeton
SAT Comp 1510(34)/1540(35)/1570(36)
ACT Comp 34(1500)/35(1540)/35(1540)
SAT Math 760(34)/790(35)/800(36)
ACT Math 32(720)/34(760)/35(780)
Stanford
SAT Comp 1500(34)/1540(35)/1570(36)
ACT Comp 33(1460)/35(1540)/35(1540)
SAT Math 770(35)/790(35)/800(36)
ACT Math 32(720)/35(780)/36(800)
Harvard
SAT Comp N/A
ACT Comp (not comparable)
SAT Math 760(34)/790(35)/800(36)
ACT Math 32(720)/34(760)/35(780)
Yale
SAT Comp 1470(33)/1540(35)/1560(35)
ACT Comp 33(1460)/35(1540)/35(1540)
SAT Math 760(34)/780(35)/800(36)
ACT Math 32(720)/34(760)/35(780)
OK, observations.
Composite: not too bad, although I think at the 75th level, the ACTs seem a little low at YPS, a little high at MIT. Nothing too off, but still, this may explain why it was seeming to me that, say, a 35 was doing the same work as higher 1500s SAT composites when in theory it should not be. Except maybe at MIT.
Math subscore: here I think the wheels kinda come off. At the 75th, HYP are low. At the 50th, HYPM are low. And then at the 25th, HYP are 2 ACT scores low, and Stanford is 3!
This is a kinda big deal to me. Being a more or less full ACT score low is already pretty significant. Being 2 or 3 low, that’s really quite significant.
OK, so, maybe now I get why the Yale AO singled out the SAT when it came to the predictive Math subscore. Maybe they do not see the ACT Math as similarly predictive. If that was a widely-held view, that could help explain what I am seeing in the CDS as well.
Anyway, pretty interesting to me.
To me, this is consistent with kids sorting themselves into tests based on their strengths, with STEM kids favoring the SAT and humanities kids favoring the ACT. (You didn’t report the verbal numbers, but I looked at a couple and the ACT looked higher here.) S24‘s school is particularly strong in the humanities and relatively less strong in STEM. The ACT is disproportionately popular at his school, relative to its national share or its share at top schools. And S24‘s school is not in the Midwest, the traditional stronghold of the ACT.
So the SAT may be more relevant for predicting success in STEM because the kids who submit it are the kids taking the STEM classes.
By the way, this raises a question I‘ve had in the back of my head for a while: to what extent do the AOs look at subscores as opposed to composites? For example, S24 has a verbal score that is high at any school, and a math score that is at or just below the 25th percentile at the schools you list (to which he is not applying). His potential academic interests are highly uncertain and include both a humanities subject and a science. Will colleges just look at the composite without digging much deeper? (I‘m sure an engineering school or a school like MIT will care most about math, but I‘m talking about less specialized liberal arts schools.)
I think self-selection/sorting on the part of test-takers may explain part of what @NiceUnparticularMan detailed. Could also be some weird thing too like, maybe in places where the ACT is more common, the advice to not submit scores that aren’t above some threshold e.g. 50th%ile hasn’t spread as widely or taken hold to the same degree for whatever reason, and so the left tail hasn’t compressed as much as it has for the SAT, and yet these schools still want/need to take kids from all over the place.
Interestingly that is not necessarily the pattern at our HS. Some of the STEM kids see the ACT Science section as basically free points, and generally our kids just try both and then focus on their best.
But I am sure that self-selection could be playing a large part in this.
What the Yale AO said implied they would specifically look at the SAT Math subscore for kids interested in STEM majors. It was a quick line, though, and really raises more questions than answers. Like, what if you are a kid with both a STEM interest and a non-STEM interest (which is very much a Yale thing). Is there any equivalent for non-Math subscores and non-STEM interests? And so on.
But my guess is it might be rather specific. Like, if you focused on STEM classes in a HS where Yale does not necessarily trust the rigor, and you apply to Yale with what looks like a firm commitment to STEM, they might want to see a high SAT Math in cases like that. Which these days is probably a lot of cases!
And some states use ACT for graduation requirements. Mine did equally well on the practice tests for both but since the ACT was a grad requirement, she just took that.
Our school CGs told students if they had unbalanced sub scores to retest.
Totally fair. I’m not sure it is either, seemed like a weak signal at best, but I would hope that the year-over-year timeframes could at least be comparable, and I would hope that the decision to return to testing (which creates a burden on students) would happen more quickly, on average, than the decision to move away (which does not). But I’m sure bureaucratic cruft is the most powerful force in this conversation!
That seems like a promising hypothesis.
Here are the SAT/ACT submitted percentages for all these colleges:
MIT 78/32
Princeton 60/25
Stanford 49/23
Harvard 55/25
Yale 59/29
Based on the data I have seen, I believe those CDS SAT:ACT ratios are significantly lower than in the coastal regions that house those colleges. And I could see a “middle America” preference in admissions turning into a sort of ACT preference by proxy.
Note that there seems to be very substantial variation in the SAT/ACT ratio within a region. Given that some schools tend to send disproportionate numbers of students to selective schools, I‘m not sure that aggregate regional shares will be particularly reliable. For example, at S24‘s humanities-focused school in the northeast, the most recent SAT/ACT shares are 42/58. At Stuyvesant, another excellent high school in the same geographic area, the shares are 90/10. I tried to find shares for some other schools for some additional data points, but many schools do not make this information publicly available.
In my kids’ NYC publics, SAT dominated bc kids could easily take it for free in school in March (on SAT day). Nothing similar for the ACT.
At our public school in MA, the majority of kids take the SAT. The ACT has increased in popularity compared to when I was a kid (rarely available here) but remains a lot less common. S24 would have tried the ACT if he hadn’t done well with the SAT but we never got that far.
Wow, this is a great analysis. At the 75th H-P are already at 800, can’t go any further and what I find interesting is that the Math Subscore is a perfect 800 for nearly every school and thanks to the Yale Admissions Officer for confirming that.
This topic was automatically closed 180 days after the last reply. If you’d like to reply, please flag the thread for moderator attention.