@Data10 did the initial conclusion to keep the SAT by the UC system (Regents?) not indicate (or reference) that they determined the SAT was better than grades and that previous studies indicating grades were better did not account for recent increases in grade inflation?
Yes, and the top schools are taking from known schools where they know the GPA isn’t as inflated or they know the rigor is there. Meaningless unless one considers all of the factors. One cannot remove the SAT and rely on GPA alone or remove the GPA and vice versa.
I’m sure for the schools you cited that there is relatively little difference. The Harvard papers from litigation show that kids in fact do not come from ALL schools. They come from known schools and the admissions office knows the schools. Of course, they are not test optional but the optional schools also take many applicants from known schools.
The UC Task Force review recommended a long term plan for UC to cease consideration of existing standardized test scores and instead develop a new assessment system. A quote from the summary recommendations is below:
Regarding why the task force report showed that SAT was more predictive of first year GPA in some of the recent years than suggested by my earlier post, there are a number of factors. The Cal first year GPA study referenced in my earlier post was for Cal State, rather than UC. UC tends to have GPA that is more compressed at the high end than Cal State. For example, a comparison of UCLA admission profile http://www.admission.ucla.edu/Prospect/Adm_fr/Frosh_Prof19.htm and Cal State LA CDS at http://www.calstatela.edu/sites/default/files/groups/Institutional%20Research/pdf/cds1819.pdf is below. UC GPA is far more compressed at the upper end than Cal State.
Enrolled Students 25/75th Percentiles: UC-LA
Unweighted GPA: 3.88 to 4.00
SAT Composite – 1280 to 1510
Enrolled Students 25/75th Percentiles:: CS-LA
Unweighted GPA: ~3.05 to ~3.55
SAT Composite: ~890 to ~1080
The difference between a 3.88 and 4.0 at UCLA is probably less significant than the difference between a 3.05 and 3.55 and CSLA. SAT score doesn’t seem as compressed at UCLA as GPA. HS Grade inflation contributes to this GPA compression, as does the UC admission policy, which has focused more on grades over other factors including scores, in recent years, including things like near guaranteed admission for top x%. The more compressed GPA (or SAT) is, the less predictive those differences become.
One also need to consider what controls the particular studies use. The older UC graduation rate study linked in my earlier post controls for SES, including family income. This is particularly relevant for graduation rate stats because a significant portion of students at UC fail to graduate within 4 years for financial reasons that are correlated with family income. If you do not control for income, then test scores may appear more predictive of graduating on time because you are predicting who is likely to have financial troubles, rather than predicting who is likely to fail out. The newer faculty review did not control for income in many of the key figures, which contributes to the different results.
Specific numbers for 4-year graduation from the newer faculty review study are below. Had they compared SES/Demographics alone to SAT + SES/Demographics, then you could see how much of the low 2.6% variance explained was due to SES correlations and how much was due to academics. Regardless, the predictive ability is abysmal for 4-year graduation, and it’s not much better for GPA + SAT… .still only reaching 7.5% of variance explained.
Variance in UC 4 Year Graduation Rate Explained By:…
SAT Alone – 2.6%
GPA Alone – 6.3%
GPA + SAT – 7.5%
GPA + SAT + SES/Demographics – 12.0%
I expect rather than graduation rate, you were referring to first year GPA, which did show that SAT alone was more predictive than first year GPA in 2012 and 2015, but not in earlier years. The increased UW GPA compression compared to SAT compression in more recent years likely contributed to this effect. Other studies I have seen do not show this result, including ones for selective colleges that have a highly compressed GPA, so I’d consider this UC task force study is more the exception than the rule. It’s also unclear why the Geiser UC study numbers for GPA in graduation differ so much from the task force UC study, when comparing similar (older) years. This makes me suspect there is a relevant difference in methodology that would be good to understand before drawing conclusions, such as the noted SES control differences.
@Data10
I would very much like to see research on the predictive value of SAT and ACT vs. GPA for students majoring in STEM (engineering in particular).
A high GPA in certain “easier” majors is not uncommon. However, I would predict that for difficult STEM majors, the added predictive value of an SAT/ACT would be higher for STEM than for non-STEM majors. Same with 4 year graduations rates.
I took a moment to look this up after typing the first part.
" The differential validity results show that across majors, the SAT is most predictive of GPA in the STEM fields"
https://files.eric.ed.gov/fulltext/ED563124.pdf
We can all speculate on why this is.
What am I missing? you’d think in the Covid environment colleges would want test scores as an equalizer.
Our low SES urban high school/district has awful plans for the fall with kids in school. Basically 2.5 days at school on a block schedule; and 2.5 days at home with worksheets. I think the amount of education is going to be drastically pared down. 
How many other schools will have less quality education in the next year or so? Why wouldn’t a national test score be of valuable information in these times?
Following up from the STEM comment:
“…This study, similar to others, found that engineering students, for example, experienced more stringent grading standards…research has found that courses in the humanities, social sciences, education, and business tend to be more leniently graded than courses in the physical sciences or engineering.”
https://files.eric.ed.gov/fulltext/ED563124.pdf
For a future thread perhaps.
Everyone likes to say that TO helps full pay students but most top schools are need blind. D21’s list is mostly high matches (not big reaches) and all of them are need blind. So being TO doesn’t seem like any sort of boost for our full pay family. I never know what to believe about test blind. Schools INSIST that financial aid and admissions are completely separate in this case but so many seem to doubt that.
Test Optional will leave fewer seats to those who have test scores. Those students with low test scores , those who decide not to take the tests, those who could not take the tests who would not even be in consideration for admissions to certain colleges now have a fighting chance, and some of them may be admitted when the ordinarily would not have. Given the same number of seats, there are now some taken that would have been available.
I have a college grad who benefited greatly. He had test scores Well below the lower 25% line. Through Fairtest, he simply never reported them. I don’t think he would have been accepted at a number of his schools had he reported them.
@homerdog If I recall correctly, your daughter has very good rigor and top grades from a competitive HS but she has not been able to attain SAT or ACT scores to match? In that case it seems like your daughter would definitely benefit from TO.
We have a friend whose son is similar (although his AP scores are quite low, with grades of A) who is now considering to apply to much tougher colleges than he would have before the trend of TO.
Fingers crossed that you’re correct! She’s considering ED but less and less so because we are watching the virus and not sure we will be able to visit those schools she’s considering. Might have to do RD. Selfishly, I hope fewer kids apply ED so that there are more spots left for the RD round. Her plan is not set in stone. She will likely make a decision about ED based on if we can see schools.
@cptofthehouse is correct.
TO leaves fewer slots for students with test scores.
Accepted TO kids are not going to be the kids needing a lot of financial aid.
www.■■■■■■■■(dot come)
“2 REASONS WHY NEED BLIND ADMISSIONS IS A FARCE”
The Need Blind Admissions Lie
Were you told that a college to which you’ll be applying has a need blind admissions policy? That is, the college touted — maybe at an information session or during a tour — how they don’t factor in your family’s ability to pay the full cost of tuition when deciding upon whether or not to offer you admission? If so, it’s important that you know you’ve been told an untruth. A cold, hard untruth. Colleges are not need blind as they may suggest so freely to applicants and their parents. Now there are a couple schools that are straightforward — they’re open and honest about being need aware. But all of those colleges that claim to be need blind, they too are need aware. They’re just not as honest about it.
And, yes, we can hear your wheels turning. Maybe you’re saying to yourself, “But that’s what the admissions officers told me. They told it to the whole roomful of students and parents.” Oh, we know. Admissions officers are human beings. They tell untruths, too. They’ve been telling this particular untruth for many years. Some admissions officers have been suggesting their school is need blind for so many years that maybe, just maybe, they’ve even started believing it. But we’ve hired enough former admissions officers at our nation’s most elite universities to know that most aren’t that susceptible to spin. Most knew well and good that their school wasn’t actually need blind — but they said an applicant’s ability to pay didn’t factor into the admissions equation anyway. After all, doing so was their job.
How to Know with Certainty Colleges Are Not Need Blind
But we know you’re still a bit skeptical. You’ve been to a number of information sessions. You’ve heard so many admissions officers claim their school was need blind. You want proof. Well, we’ll give you the proof
– If colleges were truly need blind, then why is there a question on just about every applicant’s supplement asking if the student needs financial aid — yes or no? Why wouldn’t this question be on a separate document that admissions officers aren’t privy to when evaluating an applicant’s file? Wouldn’t that be easy enough to fix?
– Colleges rely on tuition dollars. Not every college has an endowment the size of Harvard’s. In fact, no school has an endowment the size of Harvard’s. And no school — not even Harvard — wishes to dip into their endowment to fund students’ educations. They want to invest their endowments so they’ll grow rather than shrink. Tuition dollars are essential to keeping colleges running. If a school admitted a class in which virtually everyone needed financial aid (which is a risk they’d be taking if they were truly need blind), they’d risk dipping into their endowment and becoming non-operational in time.
So, there will be fewer slots for the high testing, middle class and lower middle class kids with TO. There weren’t many now, but this small number will decrease further.
Specific numbers for the linked College Board report linked in your post are below for different majors. I am listing as r^2 without range correction, to have a consistent format with the other studies that have been mentioned in this thread. Note that these numbers do not have controls for anything else, including SES. The differences between STEM and non-STEM seem small to me.
Variance in College GPA Explained by SAT alone – STEM Majors
Biology – 19%
Math/Physics – 19%
Engineering – 14%
Computer Science – 11%
Variance in College GPA Explained by SAT alone – Non-STEM Majors
English + Foreign Language – 18%
Education – 18%
Humanities – 15%
Social Sciences – 14%
Rather than STEM vs non-STEM, the more significant difference seems to be declared vs undeclared. Undeclared students only had 8% of variance explained… much less than any declared major.
The recent UC task force studies also separate by major, but I only see r^2 by major for the combination of SAT + GPA, not SAT alone. A summary is below. Note that the order is very similar to the college board study. Biology/Life Sciences has the greatest predictive ability by a small margin, probably due to the pre-med classes. The predictive ability for engineering/CS is similar or slightly below non-STEM majors.
Variance in UC Cumulative GPA Explained by GPA+SAT – STEM Majors
Life Sciences – 30%
Math/Physics – 26%
Engineering/CS – 24%
Variance in UC Cumulative GPA Explained by GPA+SAT – Non-STEM Majors
English + Foreign Language – NA
Education – NA
Humanities – 26%
Social Sciences – 25%
I believe what you are missing is that – so far – it has proved impossible for a very large number of prospective applicants to get a test score. Full stop.
It depends on the college, but in general I’d expect TO kids at a particular college to have a lower average family income than test submitters and have a higher rate of getting significant FA… This was the pattern observed in many colleges that have gone TO in the past. For example, some specific numbers from the Bates 25 years of test optional report that is linked earlier in the thread are below.
Kids Who Received FA at Bates – 57% Applied Test Optional
Kids Who Did Not Receive FA at Bates – 43% Applied Test Optional
In addition to being more likely to receive FA, other differences between test optional kids at Bates and test submitters from the report include the following. These groups are also associated with test scores being a relatively weaker portion of the application, so I don’t find the relationship surprising. For example, women average higher grades than men at every academic level from elementary school to college, yet don’t average higher combined SAT scores, so SAT scores are more likely to be a relatively weak point on the application for women than for men.
Test Optional kids were more likely to be female
Test Optional kids were more likely to be URMs
Test Optional kids were more likely to be domestic (not international)
Test Optional kids were more likely to be from less urban states. Some of the states with the largest portion test optional are Maine, the Dakotas, Nebraska, Iowa, Utah, New Mexico, Arkansas, and Kentucky.
Thank you @Data10 . I appreciate your insight. Good points.
Doing a bit more digging on TO and financial need of accepted applicants, I found the following:
UNDERSTANDING THE IMPACTS OF THE TEST OPTIONAL ADMISSION POLICY
Yuko Mulugetta, Ph.D.
“…Hiss and Franks found that non-submitters are more likely to be first-generation college students, all categories of racial minorities, women, Pell Grant recipients, and students with learning differences (LD).”
“Furthermore, the study pointed out that 29% of white students also
opted out of test score submissions, only 2% below the overall average of 32%.”
“Interestingly, the study discovered a bimodal income distribution among non-submitters; on one hand, the financially needy group that consisted of first-generation, minority students and Pell Grant recipients, and on the other, the no need group who did not request financial aid…”
Let’s consider the “bimodal income distribution” for TO applicants.
There is much overlap between URM, first generation, and Pell Grant students. (see ).
.
(Per UMass: “In Fall 2018, First Generation students comprised 24% of the undergraduate population. Within this group of First Gen students, many are from an underrepresented minority group (URM), are considered low-income based on receipt of a Pell Grant, or both. Specifically, 51% of First Gen students are Pell recipients, and 26% are URM. Eighteen percent of First Gen students at UMass are both URM and Pell recipients.”)
Who then are the non-URM students using TO? Many must be full-pay students.
Conclusion: the least likely to benefit from TO are middle class, non-URM, non-LD applicants (male more than female).
For anyone who wants to wade into the weeds on this, check out
www.nacacnet(dot org) /globalassets/documents/publications/research/defining-access-report-2018.pdf
The specific numbers for portion test optional by financial need group from figure 40 of the NACAC report are below. This is for an aggregate of 21 test optional colleges. The largest difference between the 2 groups occurred in the high need category. High need was the was the only group for which test optional kids were more represented than test submitter kids. Both groups were far from uniform and included a kids with a wide range of income levels.
Financial Need of Enrolled Students at 21 Test Optional Colleges
Test Optional Kids – 36% High Need, 25% Moderate Need, 34% No Need
Test Submitter Kids – 28% High Need, 28% Moderate Need, 38% No Need
High Need = EFC is 0-25% of cost to attend
Moderate Need = EFC is 26-75% of cost to attend
No Need = Little demonstrated need for FA, includes those who do not apply for FA
It’s also worthwhile to note that there was a wide variation among the 21 different colleges, although at all 21 of them the average EFC for test optional kids was lower than test submitters. In general, the discrepancy in average EFC need between submitters and non-submitters became smaller as selectivity increased (and a smaller portion of low income kids applied), although there were many exceptions.
“Test Optional will leave fewer seats to those who have test scores.”
I disagree on this one, if you have a test score that’s at the median of the college’s number, it will be an advantage, even it’s a slight one over someone that didn’t submit a test. Test optional doesn’t mean test blind, they’ll consider submitted scores which means only high test scores. They could easily fill the class with those students.
"Those students with low test scores , those who decide not to take the tests, those who could not take the tests who would not even be in consideration for admissions to certain colleges now have a fighting chance, and some of them may be admitted when the ordinarily would not have. "
Many test optional colleges have pretty high averages and a large percentage that submit scores, I don’t think it’s going to change a whole lot. Maybe the ALDC category that came up in the Harvard lawsuit - athletes, legacies, development, child of faculty/staff, don’t have to submit scores now.
Do we know this is actually the case? I have not seen any data that says “x” percent of college applicants have not had an opportunity to take the SAT/ACT at least once. The HS students I’m familiar with all start taking the standardized tests early in their Junior year.
The other point is that there is currently many opportunities to take the tests now (July - Nov) so my recommendation is to sign-up for one of these test dates to be safe.
Remember that the colleges that decided to go test optional this year did it to get the maximum number of applicants this year but I’m not convinced that not providing a test score doesn’t matter.
I’m currently reading Educated by Tara Westover. In a nutshell, it’s the memoir of a girl who was brought up in rural Idaho by a Morman family that didn’t believe in public schools, the medical profession (including doctors, vaccinations, hospitals, etc.) and whose patriarch more than once (including leading up to Y2K) believed that the world was about to end. Tara wasn’t home schooled - she wasn’t schooled at all. In the book, she sets her sites on “escape” and targets BYU as her escape route. She buys the prep book and self-studies for the ACT - her first score a disappointing 22. She keeps studying, teaching herself algegra and trig, and scores a 28. Accepted to BYU, she goes on to Cambridge on a Gates, and then Harvard for a PhD. I couldn’t help but think about Tara reading this thread. Even with her “story,” would she have gotten in with that 22 or would she still be on that Idaho mountain?
I would hate to tally up how much we spent on test prep (books, some private tutoring, an SAT class, etc.) for our twins (HS 2014). It was a lot. One improved +230 from first SAT test to the third. One stayed about the same, doing better on the ACT. Both had interesting resumes to go along with grades and test scores but we played the game. For all kids applying - URMs, Donut Hole Kids, Priviledged legacies - I’m glad standardized tests scores are ONE factor but also glad they are not the only factor.
“So who will benefit from this test-optional phenomenon?”
“research suggested… that affluent students with suboptimal scores have a better chance of gaining admission to schools that would have normally been out of reach. Colleges love high-income students and they love them even more if they don’t come with mediocre tests scores that would drag down the institution’s standardized test averages.”
Per the June 29, 2020, “Getting Into Elite Colleges May Just Have Gotten Easier: Many elite colleges and universities are not requiring the submission of SAT or ACT scores for the 2021–2022 admission season.”
www.wealthmanagement.com/college-planning/getting-elite-colleges-may-just-have-gotten-easier (Lynn O’Shaughnessy is a nationally recognized higher-ed speaker, journalist and the author of The College Solution, an Amazon bestseller)