# Interpreting representativeness of reported average test scores when only a subset of admitted students take them?

What’s the best way to interpret SAT/ACT score ranges, or averages, when only a subset of applying students are submitting them? Some of the schools my daughter is looking at publish scores that are higher than hers, but in one case, only 40% of students reported scores. What would this have been if we were back in the good old days and everyone was submitting scores? How can I honestly compare her scores to these ranges or averages, other than to discount their value as a characteristic of an incoming class or an “average” admitted student?
Someone better versed in statistics must have a way to interpolate, with relative accuracy, what these numbers would look like if they represented the population of admitted students, not just a very skewed sample. Of course this assumes that the only kids not submitting are those for whom their scores wouldn’t help them gain admission (too low), or who otherwise might not have had access to the test because of COVID, but the latter seems less of an issue than it was a year or two ago. Thoughts?

2 Likes

I’ve had the same question but no answer. It seems to me TO just lets schools raise the middle 50 scores for everything (and get a lot more money in application fees) but clearly it’s not representative of the full incoming student body (or even the majority in many cases). Assuming tests are more available now, I’d have to assume in most cases, the other 40-50 percent not submitting are lower than the 25percentile or at least lower than median.
I’m not sure how to interpret the data except to take it all with a huge grain of salt.

3 Likes

All that matters is the data the schools give us. Trying to estimate what the full range of scores was isn’t meaningful because the AOs didn’t see the scores for those who applied TO (generally).

First step is to decide whether or not one’s test score adds to their application or not. If it doesn’t, why submit?

Then, look at each school’s mid 50% range and if the score is at or above the 25%ile, one can make the case to submit. But, some schools, like Tulane, tell students in admissions sessions to submit only if at the median score or above, so one has to do the research. Note I am only talking about unhooked applicants.

5 Likes

Obviously, scores would be much lower if everyone was submitting.
It’s actually a cycle that, if it continues, the mean-reported SAT score for every competitive school will be 1500+…
As everyone looks at the reported scores, and if they are not in the top 50-75% of the score range, they don’t submit.

So if in year 1 – the reported 25-75th percentile of a university is 1350-1450…
Then in year 2, nobody below the 25th percentile submits. So it pushes up the entire range… So it’s now 1390-1480…
Then in year 3, nobody below 1390 submits, so the 25th-75th percentile range is now 1430-1510…
Then in year 4, nobody below 1430 submits…

The only thing I can say is, if an applicant is above the most recent 50th percentile, they should submit. If they are above the most recent 25th percentile, they probably should submit, but it’s more questionable.
Below the 25th percentile… Probably need to play the same game as everyone else. Maybe submit if you’re just barely below the 25th percentile, so there is no assumption that you’re way below.

But this is why Test Optional won’t last forever. Either they will go Test Blind or reform the testing. But if you stay test optional forever, then eventually only 10% of applicants will submit scores and they will all be over 1500.

9 Likes

On tours of two different test optional colleges with D the admissions officers suggested that applicants submit scores only if they are at or above the school’s 50th percentile. YMMV.

2 Likes

That’s clearly a deliberate attempt to increase their rankings on lists that consider scores…
If only people above the median submit, then it’s guaranteed that the median will increase every year. In just a couple of years, nobody below 1590 will submit, when they report the median range of 1580-1600. Just 0.1% of people will be submitting by then…

6 Likes

I don’t disagree. Jon Boeckenstedt has explained how that recommendation is mathematically unsustainable (assuming one wants to provide data to inform decision making).

I don’t know all the ranking methodologies, but USNWR (the only one colleges care about) penalizes schools where less than 1/2 the enrolled students submitted scores. And obviously they could change their methodology as things progress.

A change for the 2022 edition – if the combined percentage of the fall 2020 entering class submitting test scores was less than 50% of all new entrants, its combined SAT/ACT percentile distribution value used in the rankings was discounted by 15%. In previous editions, the threshold was 75% of new entrants.

https://www.usnews.com/education/best-colleges/articles/how-us-news-calculated-the-rankings

3 Likes

I believe they said that. But it’s truly an absurd policy which demonstrates likely ulterior motives.

If Admissions officers want to make an informed decision, why would they ever DISCOURAGE submission of test scores? Why would they DISCOURAGE the submission of any information?
Imagine if an Admissions Director told people, “we are going to give you an option of which grades to submit… you really should only submit the A’s”

The Admissions Director knows their process is selective. So why would you actively discourage applicants from giving information that would help make the selections?
And if you know that 50% of the students you admit have a score below 1400 (or whatever score), then why would you tell those types of students not to submit the information?

As I said – in another post, I suspect an ulterior motive. The schools want to raise their Score averages that get published. Great way to raise your average — By only having the top scores even get submitted.
So if the 25-75 is 1350-1450 with a 50th percentile of 1400… And you have a fantastic applicant with a 1390, top rigor, top grades, top ECs, top recommendations, top essay… But… if the school admits you, then that lowers their median SAT! On the other hand, if you didn’t submit the SAT, they could admit you without risking reducing their median SAT!

It’s a stupid game.

Either go test blind, or require testing. Imagine telling students, “you don’t have to submit any teacher recommendations unless you can get a great one.”

7 Likes

So the irony is that – If the goal is to make test scores count less, this could make test scores even MORE important.

With a lower percentage of students submitting scores, schools will have to have a much higher acceptance rate among those who do submit scores, in order to meet the 50% threshold.

So for mathematical simplicity –
Imagine a school with a 25% acceptance rate. Imagine only 25% submit test scores, but they are all really high test scores. That 1380 didn’t get submitted because their 2022 median was 1400.
So to preserve their USNWR rankings, they will have to maintain an acceptance rate of 50%+ for those that submitted scores.
So while that 1390 would have had an excellent chance of admission in the “old days”… now that score isn’t even submitted. And since you didn’t submit, you’re in the lower-acceptance-rate pile.

The effect – A high score helps even more than the old days!

3 Likes

My guess on motive is \$\$\$\$ - \$75 per application fee. They can increase applicant pool and make gobs of \$ but if they’re still accepting scores, they can push up scores and make even more applicants think they’re exclusive the following year.

Is there data anywhere on how many applicants were accepted from TO pool versus those who submitted?
I agree it’s not sustainable though.

1 Like

40% applicants submitting scores is still a heck a lot of scores for competitive schools.
My philosophy is unless the school explicitly states that scores will not be considered, look at their historical non-covid data and submit if in range.

The USNWR penalty applies if fewer than 50% of enrolled students (not admitted) applied without test scores. I expect USNWR is ultimately going to change their methodology. In the current methodology, test blind schools are ‘penalized’ as well (and for awhile USNWR refused to rank test blind schools, but now they must).

U.S. News again ranks “test blind” schools, for which data on SAT and ACT scores were not available, by assigning them a rankings value equal to the lowest test score in their rankings. These schools differ from ones with test-optional or test-flexible admissions for which SAT and ACT scores were available and were always rank eligible.

It continues to be too bad that anyone cares about USNWR rankings.

I would not recommend students do this if the result would be different at schools that are directly stating which scores to send (like Tulane and the other schools that say only submit if at the median or higher, based on last year’s numbers).

Yes, some schools have published this data…and they generally say correlation is not causation. Meaning the test submitters may have had relatively stronger apps in many ways…rigor, number of core courses, etc etc etc.

Yes-- but that doesn’t explain why they would actively tell kids not to submit scores.

If everyone plays the game and decides to submit or not based on the test score range of a test-optional college, that range is guaranteed to shift dramatically higher each and every year. It wouldn’t be useful at all soon, even if it’s still somewhat useful now.

2 Likes

What’s especially confusing is when schools list standardized scores as very important but then say don’t submit. I hope at some point soon they go back to testing required frankly.

4 Likes

Deleted.

At Boston University, In the last pre-test optional year, median SAT was about 1420. In the first test optional year, it was about 1460. They are reporting this year, second cycle of test optional, an average of 1496. (Not clear if that’s mean or median).

It’s arguable whether BU is even a T20 school, but they are now claiming the practically the same SATs as Harvard and Yale.

7 Likes

What about the student who doesn’t test as high as their GPA and academic rigor? Without the SAT they may qualify for a larger scholarships at some schools.

2 Likes

Looked at CDS again and the percent submitting data point is for enrolled students not admitted or applying. So in the last cycle for Norrheastern, only 42% of enrolled students submitted scores. Previous cycles it was around 85-90%. Which means what - that TO applicants are getting in as much or more than those submitting? Am I interpreting that incorrectly?