I am always surprised at the number of 1450 SATs with 36 ACTs at our school. 1580-1600s are so much more rare than 36s.
3 Likes
Again, let’s not get distracted by off-topic comments and make the thread go further off-topic. BMI is unrelated to either SAT or Dartmouth. Additionally, there’s no need to reply to users who are no longer with us.
Interestingly, some casual dives I have done into the data suggest if anything, ACT scores are treated more, not less, generously than the official concordance would suggest, by at least several prominent highly selective colleges.
A plausible hypothesis was offered that this might be a side effect of a sort of affirmative action for applicants from interior states where the ACT is more popular by coastal universities. But in any event, Dartmouth in these materials is purportedly just using the official tables.
+1. My personal network overlaps with a number of Dartmouth CS graduates, all of whom are gifted computer scientists and extremely sharp thinkers. In the early-but-not-crazy-early beginnings of CS departments (think early-mid 1990s), it was just as likely to find CS as an offshoot of mathematics departments as standalone organizations in engineering departments.
I also really appreciated @blossom’s message, which I think articulated well the value of the activity that underpins the extracurricular as an oft-overlooked guide towards success in college. I always talk about this in a simple way - for every student, no matter how bright, there comes a time in their education when just being smart isn’t enough to (get the grade or whatever). What you do then is, I think, a key predictor of success. Fortunately my child who didn’t need to learn that during high school had the implicit skills to succeed when that happened (as I did), and my child who did need to learn that during high school has built the muscles under our roof (and through her outside-of-academic activities).
I also like how this thread is having some of us reflect on all of this. Coming out of HS, I was a catch on paper - great test scores, 3rd in my class (and the only one of us who did anything besides school) - except for the lack of funds, I was a pretty easy admit. And I did fine in college, but I was a pretty shallow thinker, and that exposed itself to me and others as the classes got more challenging and abstract. I figured out pretty quickly that I wasn’t right for the math track (which I didn’t want to be on anyway), and while I could be successful in an engineering education, I was bottom-of-the-top-third successful - like you’d be happy having me on your team, but I wouldn’t be a first-round pick. Fortunately I was self-aware enough to see this and adapt accordingly career-wise, and it’s all worked out (living many lives).
4 Likes
In their review, SAT explained 22% of variance in first year GPA. Of phrased different, the overwhelming majority of variance in first year GPA was not explained by SAT. That leaves a lot of room to improve predictive power with different factors. As a simple example, the author of the Dartmouth analysis recently published a separate review that found adding HS variables (correlated with name of HS attended) increased predictive power up to explaining 65-70% of variance, far more than the ~20% explained by SAT in isolation.
Obviously colleges, should not base admission on name of HS attended. However, there can be other factors that can be used that better incorporate the reasons why name of HS attended is better correlated with first year success than SAT/ACT score. One example, is considering the specific classes the student took and how rigorous those classes were, rather than just looking at HS GPA in isolation. Another example, is looking at the GPA distribution of the the specific HS the student attended, to get a better sense of what the student’s GPA in isolation means. Studies that have considered rigor of classes taken often come to different conclusion than ones that just look at HS GPA in isolation.
For example, the Ithaca study found the following. Including SAT only explained 1% more variance in cumulative college GPA than excluding SAT.
HS GPA + SAT + Rigor + AP hours + … – Explains 44% of variance in cumulative GPA
All of Above with SAT Removed – Explains 43% of variance in cumulative GPA
3 Likes
But are they disappointed? I don’t really see that. (We have seen other studies, or maybe just essays, that have suggested that a disproportionate number of TO students are less prepared to be successful in college, but I can’t recall anything quantitative.) I don’t think Dartmouth is aiming for everyone to have a 4.0, and I don’t think they’re necessarily “disappointed” in their 3.2 students.
I think the message is just simpler - the message going to many* disadvantaged applicants (and/or their counselors and family) about not reporting scores was significantly impacting their ability to be admitted, and Dartmouth wanted to admit more of them, so requiring scores makes them more likely to be able to select them.
(On many - since this comes up sometimes - it is difficult for us outsiders to quantify the “diamonds in the rough” or whatever we want to call them [I find that phrase weirdly dehumanizing], but even 50 of them nationwide, which feels like very few students, would be a lot for one school. That would be almost 5% of Dartmouth’s incoming class - bigger than their Questbridge match. You don’t need a million students in this category nationwide. You really only need a few hundred.)
6 Likes
I’m more surprised that they admitted any students with a 3.2 HS gpa.
That is incorrect; just the opposite. Dartmouth has long offered a Bachelors of Engineering (ABET) in a 5th year at Hanover (full financial aid for the 5th year), after one earns the BA. Dartmouth also participates in a 3-2 program offering students from other schools teh opportunity to earn a BE in Hanover.
3 Likes
Specific numbers for different classes are below. I realize Pell, low income, first gen and similar are not the same as disadvantaged, but it doesn’t look like there has been a notable decrease since switching to test optional. Instead Dartmouth’s “welcome class of …” news stories state they hit all time record number of Pell and first gen kids shortly after switching to test optional. Dartmouth seems able to admit disadvantaged students from a wide variety of HSs both with and without tests. It’s possible that while the number of disadvantaged students is similar to previous classes, the performance of disadvantaged students is worse than in previous classes, but I didn’t see that claim made in any of the documents.
2019 – 15% Pell, 18% Low Income, 15% First Gen, 57% Public HS, 932 different HSs
2020 – 17% Pell, 20% Low Income, 15% First Gen, 59% Public HS, 909 different HSs
2021 – 17% Pell, 24% Low Income, 16% First Gen, 56% Public HS, 910 different HSs
2022 – 17% Pell, 24% Low Income, 15% First Gen, 54% Public HS, 956 different HSs
2023 – ??% Pell, 27% Low Income, 14% First Gen, 55% Public HS, 971 different HSs
3 Likes
Both scores are 99th percentile, so by definition, 1% of students score in the range for both tests, so a 36 and a 1570+ occur with the same frequency. Just because you see some scores that aren’t both 99th doesn’t mean that there are less with that SAT score.
I don’t see how Dartmouth or any of the popular, highly selective schools has the time or resources to examine each file “holistically.” According to their most recent common data set, they had 28,841 applicants in the most recent cycle. They have fewer than 30 admissions professionals. Even if just one admissions staffer looked at each file, that’s still about 1,000 per staffer. And more than one is supposed to look at each file. You’d have glazed eyes in no time. They need the SAT back to help them whittle down the pile into something more manageable, and decrease the time they spend on each file.
5 Likes
They crank through a lot per day in the high season, but reportedly they are implementing a fast tracking protocol where only some will get traditional full reads. As mentioned above, they are using an AI/Big Data modeling approach to the first cut, and while they could do that without tests, apparently they believe tests at least sometimes help that process.
Dartmouth, like many other schools, may hire external readers. Some part-time, some full-time, pay range $17-$25/hour or so. This is how a school like UCLA to take another example, can say they read apps a minimum of two times. UCLA receives the most apps of any school (146K for Fall 2023 first year enrollment). Regardless, a full time staffer should easily be able to read several thousand apps per cycle.
AI, at the schools that use it, will cut down on the need for readers.
For the institution where I read apps (highly rejective, holistic admissions), I’ve read 1500 apps in the last several months. I read every word of the app and attachments. This isn’t rocket science.
7 Likes
Use of AI is conjecture, on the Dartmouth admissions podcast they clearly said they don’t use AI and they didn’t add any supplemental staff despite the increase in applications. Also if they require tests, the number of applications may go down to 2018-19 levels which would be 22000-23000. This may enable them to read the apps more thoroughly.
1 Like
This is very evident when I work with international students trained to deal with complex essays, not multiple choice or standardized tests. When your exam is a 4hour essay where you must construct an answer to “How is knowledge geopolitical?” Or “Why are we all something else thab we appear to be?” With detailed, structured 3-part outline that will then be elegantly written with specifc (expected) evidence of close reading and broad general culture… you do very poorly on the SAT.
A good example of this would be a French kid who scored 23 on the Math ACT. The university wisely decided to ignore the score (figuring that if she got As in calculus, linear algebra, and discrete math, she could do the work) and admitted her to her chosen major, Engineering, where the student did extremely well while being an athlete - if you know student-athletes majoring in Engineering you know how hard that is. Not only that but the student did so well in OPT that the company fought tooth&nail for an H1B (again, if you know what that means you understand what a big deal that is.)
Another example, this time from Newark. A student heard there’d be some kind of test the following week because one of her “class” periods was working in the administrative office. Being inquisitive, she inquired and got a brochure for it despite the GC telling her it wasn’t for them at the school, just sth the state required but it meant nothing, they didn’t need it to graduate HS or go to CC. It was the SAT. The student did terrible on it even though she was a top student because all she had was that brochure you get and a week to read through it; then she decided to prepare more and retake the SAT, despite the discouraging 1st score. She found her way to CollegeConfidential, got some advice, etc., and while her score was nothing exceptional, it was MUCH higher than her school’s average - let’s say, 1240 v. 760 at her school. This student applied to 4-year colleges, got in with scholarships, and did well in college because of her intelligence and grit.
I don’t read Dartmouth’s statement as saying they admitted students who weren’t qualified - they have enough qualified applicants and enough indicators that no one gets in who can’t succeed, and many who don’t get in would have succeeded too.
I read it to mean that TO artificially inflated average scores to sth ridiculous like the 25-75 spread being 1490-1580, scaring away good applicants who never applied.
Further, being TO made students like the Newark kid not submit scores, which made it impossible to distinguish this inner city kid with As, from another inner city kid with As but from a school that had sent them students through Posse or QB.
Their reasoning, as I read it, is that requiring scores will bring the score spread back to something reasonable since it’ll include all scores, not just high-scoring submitters. As a result, it’ll create a positive loop.
Second, they can be very transparent about contextualized reading - which I thought was obvious but some reactions on this board made me reconsider, too: SAT/ACT scores are heavily correlated with socioeconomic class. Now Dartmouth can say something like: if you’re from a more privileged background, we expect 1500. But if you’re from a less advantaged background, 1400 is fine, and if you’re from a school that sends fewer than 50% students to a 4-year college we want you to score 300 or 400 points more than your school’s average.
Most importantly, the last 2 points can be easily disseminated to guidance counselors all over based on their school profiles, to identify what Dartmouth would consider “diamonds in the rough” in their context. It’s a short, clear message for overworked GCs.
(You’ll notice also that students enrolled in a national exam system can submit these scores or their predicted scores instead of the SAT/ACT.)
10 Likes
Yes, seems simple for D to get the message out. OTOH, on counselor list servs/FB pages there are any number of counselors who work with these groups of students who are saying one school changing their testing policy will not change their recommendations to their students for testing (many, sometimes most of these students, don’t take these tests and don’t need to).
Other counselors are saying they wouldn’t recommend an institution that requires test scores to their students just based on the school’s [perceived] ignorance of the issues facing many of these students at these high schools. I don’t personally agree with everything I’m hearing, but that’s not important…what’s important is that there are counselors not open to receiving D’s message.
This is an important point for international students…D is not SAT/ACT test required for them, they are test flexible.
1 Like
My eyes hurt just reading that!
1 Like
Same. Some of the feedback is spicy indeed.
1 Like
My biggest question about the Dartmouth decision - and I don’t have a quarrel with it as it seems well thought out on their end - is that I really, really question the likelihood of required testing resulting in finding more talented kids from challenging backgrounds. Prior to TO many (if not most) weren’t applying in the first place, add to that the fact that few lower SES kids take the test to begin with (old data says 25% but that is likely lower now, post TO) and, in my view (which could be wrong) this is more likely to reduce the number of kids applying from under-represented groups than it is to increase it.
1 Like