I can’t think of one thing in a college application where there is not a correlation to wealth, and that should not be surprising. Better schools, better grades, better test scores, better ECs.
4 Likes
It’s interesting that AOs are fully capable of viewing GPAs, ECs, essays, etc. in context but when an SAT score pops in front of them they lose all their senses and perspectives.
4 Likes
I am actually not assuming that they are seeing other things in context. in fact, I know that they are they are not.
As for everything else, no, they don’t really see them in context. All once has to do is to look at the boost the wealthiest applicants who attended well funded private schools get from their achievements compared to the lack of such a boost for low income kids (non academic ratings of AOs at elite colleges demonstrate that).
Moreover, the AOs haven’t looked at test scores in context before, so why should we expect them to do so now? Looking at the 2023 Chetty article, AOs did not rank low income kids higher, when controlling for test scores, which means that kids from low income families were ranked, in academics, the same or lower as kids from high income families with the same test scores.
This is an enormous thread (2603 posts). If this has been covered. My apologies.
Interesting insight from the former deputy undersecretary of the Department of Education for the United States. Test optional artificially inflated schools’ SAT scores as only students who scored at or above the published college mean submitted their score. This, in combination with the overall upward migration of standardized test scores, led to inflated mean SAT scores.
So with the return of mandated scores, It’ll be interesting to see what happens to the average SAT score at these schools. Application #'s and mean scores will both likely decline.
1 Like
I suspect more kids with high SAT scores will erroneously believe it’ll help their applications to highly selective schools.
1 Like
Is it an erroneous belief?
A highly selective school isn’t going to take an applicant with no meaningful ECs and a 1580 over someone with a stellar total profile and a 1550…BUT…for 2 students with great total profiles, and one has a 1580 and the other has a 1420, you don’t think it can help?
I think there are so many kids who score above 1500 that it wont really help them (but will hurt the kids who are 1420).
Yes, it may eliminate some competition but not to make a meaningful difference. For example, 10% of the kids in my daughter’s graduating class scored 1500+ (every single kid in the top 10).
As I said earlier, the one who scored 1600, his best acceptance was UNC. He had very good ECs. Maybe his essay/recs weren’t as good but the SAT was probably the least significant data point in his application.
Once you clear that 1500+ hurdle, it probably doesnt matter much at all.
2 Likes
Yeah. Kids should really stop trying to do their best. It doesn’t matter. The enemy of good is great.
They’ll still try because they think it matters. But we know unless your FGLI, a high SAT score wont make much of a difference.
Unless you think it will?
I know that colleges don’t want students to think anything matters. 4.0 GPA, great letters, great essays, exceptional EC’s, 1600 score. Meh. And we wonder why kids are so stressed these days. You really should read the article. It talks precisely about how that kind of one-sided approach where only the College matters leads to student agony.
I guess the take home is that no one should strive for academic greatness nor perfection. And if, you’re just not destined to be the #5 pole vaulter in the country, too bad for you too, you should just quit.
Great message for our kds.
For most high achieving kids, it’s all about opportunity cost. IMO, many kids waste too much time studying for the SAT when it’s not that relevant. If they’re not studying for the SAT they’re usually doing something else - not sitting around underachieving.
4.0 GPA/1550+ SAT kids, with tons of ECs/volunteering/leadership.
It’s about allocation of resources (time and money) , not about settling for second best ot not putting in their best effort.
Resources are limited - pick your battles. That’s actually a pretty important message to kids.
1 Like
Yes, but now you are talking about an either-or rubric which really doesn’t apply to high performing students and kids who often have the highest grades and the most involved EC’s.
Before, you were talking about how kids naively would think doing well on the SAT would matter. Which it might. Colleges have basically said as much especially for underprivileged kids.
And as to time management, the prep it takes for a 1450 kid to get to 1550 is literally nothing compared to varsity sports, an extra AP class or becoming even the #50 trombone player in the country. A smart kid would figure that out presumably.
1 Like
Yes. That’s why I mentioned FGLI. It may matter for them.
Before this gets too conversational, I’d just point out that not all underprivileged students are first generation nor low income. (FGLI) as you state. Adverse context can happen to anyone. Do you have any data that colleges limit contextual consideration only to first generation students?
At this point, you are disagreeing because I stated first generation and low income instead of underprivileged.
Ok. I’m done.
A lot has happened since 2019. Students who endured online courses which were hit or missed may be disillusioned by school, leading to a drop off of college applicants. Whereas those who decided to go with the SATs are more likely to have gone all in to get the higher score (i.e the true believers). In the middle is a group of students who are dealing with a lot of uncertainty, from the whole FAFSA fiasco to less socialization as a result of the pandemic.
There are going to be a lot of kids that are going to fall between the cracks.
Different colleges use SAT score in different ways. Some colleges admit based on simple formula that includes test scores. If stats are above a threshold, admission is almost guaranteed. Others use SAT as something like a basic filter to cut down on applicants. Once you get past the basic stat filter, then they focus on the rest of the application and largely ignore scores. Others use the test score in the context of the full application, with a subjective interpretation that is applied throughout the admission decision. It’s difficult to generalize.
One of the few colleges that we have specific numbers for is Harvard, due to the lawsuit analyses. Harvard rates applicants on scale of 1 to ~5 in a variety of dimensions – academics, ECs, personal, athletics, LORs, alumni interview, etc. Harvard’s internal analysis found that 70% of the variation in academic rating could be explained by a simple formula using GPA and SAT/ACT score. This fits with the description of the academic rating categories in the reader documentation, which are quoted below. A portion of the category listings are below (almost mentions other things beyond what is quoted)
Academic
- “Near perfect scores and grades (in most cases combined with…”
- “Top grades and, SAT and SAT Subject tests: mid 700 scores and up… possible local, regional or national level recognition in academic competitions”
- “Excellent grades and SAT and SAT Subject tests: mid-600 through low-700 scores”
- “Respectable grades and low-to mid-600 scores on SAT”
- “Modest grades and 500 scores on SAT and subject tests”
The external 3rd party analysis had similar findings. Demographics including low income could explain 16% of variation in academic rating. When they added in academic index (2/3 test scores), it jumped to explaining 55% of variation in academic rating. Other rating categories did not follow this pattern. For example, demographics + AI score only explained 6% of variation in EC rating. For first LOR, it was 8%. Only academic rating seems especially well correlated with test scores.
The analysis also found that both a better academic rating and a better academic index increased chance of being admitted, after controlling for demographics, including being low income. For example, increasing academic index from 3 to 2 without any change in other ratings categories increased chance of admission by a factor of 7x on average. Increasing from 2 to 1 increased chance of admission by 15x. Of course this is far from the only factor in admissions or even the most influential factor in admission decisions. In terms of loss of r^2 when removed, the most influential analyzed components of the application on the admission decision were from highest to lowest as follows. Removing the academic rating caused a 17% loss in the model’s explanatory power. Removing just SAT score and not the full academic rating, would be less.
- School Support Ratings (LORs + interview) – 50% loss
- Personal Rating – 19% loss
- Academic Rating – 17% loss
- EC Rating – 13% loss
The lawsuit data went thru HS class of 2015 (Harvard class of 2019.) Harvard RATED applicants at that time in the way you describe. That’s 9 cycles ago, which is a very long time in college admissions. My point is all we know wrt to Harvard admissions is the process they had in the years analyzed in the lawsuit…we have no idea if those conclusions are applicable to the Harvard admissions process today, or their competitors (at any point in time.)
I don’t have direct knowledge that Harvard still uses the 1-5 scale to rate the applicants on the same dimensions they did in the lawsuit. Certainly the process looked different during the test optional years (which would impact the data analyses/correlations as not all applicants applied with test scores), and given a 45% or so growth in apps from class of 2015 to class of 2024, I expect they’ve made changes in how they read and rate apps. @skieurope do you know? Of schools I am familiar with, none have a process that resembles what they were doing in 2015.
Beyond those comments, I agree that different colleges use test scores in different ways. I doubt anyone disagrees with that, but of course CC always surprises.
The quote I listed above for the 1-5 academic rating SAT groupings was from the admission reader guidelines for fall 2019 – just before COVID. They’ve been using this type of 1 to ~5 shorthand rating system for many decades (historical basis relates to why athletics has its own rating category), so I’d be very surprised if they suddenly abandoned it. However, I agree that the use and possible influence of test scores has no doubt changed since the lawsuit, particularly in a test optional environment.
What is the source of your data?
Arcidiacono’s last year of data was from HS Class of 2015/Harvard Class of 2019.
I use a number of data sources for my analysis. The most important of these is the
admissions data produced by Harvard containing selected anonymized data on
individual applications for the 2014 to 2019 admission cycles.14
Footnote 14: The dating of the admission cycles refers to when the applicant would typically graduate from Harvard should they be accepted and complete their studies in four years. Hence the actual application dates are generally five years before the date associated with the
admissions cycle.
1 Like