Tests are back, where’s the data?

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:The SAT is now a 2-hour, very preppable, superscorable test of math through Algebra 2, basic grammar/punctuation, and vocabulary in context. Median scores are declining because kids in the middle and below are learning less in high school. Scores at the top are rising because motivated kids with access to resources can easily master the content or just get lucky after taking it six times. It's a test mildly useful to validate high school grades, especially in math. It has little to do with whether kids can write or read and understand more than a paragraph or two at a time.

Caltech: "Test scores are predictive of success, even into students' sophomore and junior years"

Everyone is misunderstanding me. The SAT is better than nothing. Students without near-perfect math scores are going to struggle at Caltech. But requiring it isn't going to produce radical improvements in student quality, especially in reading and writing skills. We also don't have a lot of data about the digital SAT, which is a significantly different test. The trend toward taking the SAT three times at a minimum and often 5+ times is also, while not completely new, continuing to rise.

That's why Caltech looks at individual sections, not the composite. That's an important point overlooked by many including you.
They don't let kids with 800 math 730 verbal to be a bucket A kid. That will be a bucket A + bucket C.

SAT is an aptitude test. Prepping helps for sure but there is a limit, especially for reaching bucket A. The majority of kids being prepped and taking the test multiple times still would not reach bucket A in both categories.

You may argue that SAT is not hard enough and you wish it's harder. That's another topic, and college board has revamped the test multiple times in history so it's not unlikely they will do it again (as seen in August test).

The reality is not that many can achieve double bucket A at present, and it's a holistic review process so score isn't the only thing they look at.

NP. Clarification, the SAT stopped being an aptitude test many years ago. It moved away from aptitude gradually over a number of rewrites. Anything measuring aptitude, like the old analogies section, College Board removed quite some time ago. The test is currently (supposedly) a test of academic skills. While inherent aptitude will impact academic skills, the test is not a direct measure. Every student will have their individual potential high score, based in part on aptitude and in part on how well their academic skills have been developed in school. That said, I do not think the current digital test is a high-quality product. Scoring is too inconsistent; perhaps that's a topic for another thread.

Hmmm … I don’t think you understand what “aptitude” means.

The "A" in SAT once stood for Aptitude, and the College Board changed it to "Achievement" in the 1990s.

[Anonymous]

I agree with OP. Where is the damn data!

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:The SAT is now a 2-hour, very preppable, superscorable test of math through Algebra 2, basic grammar/punctuation, and vocabulary in context. Median scores are declining because kids in the middle and below are learning less in high school. Scores at the top are rising because motivated kids with access to resources can easily master the content or just get lucky after taking it six times. It's a test mildly useful to validate high school grades, especially in math. It has little to do with whether kids can write or read and understand more than a paragraph or two at a time.

Caltech: "Test scores are predictive of success, even into students' sophomore and junior years"

Everyone is misunderstanding me. The SAT is better than nothing. Students without near-perfect math scores are going to struggle at Caltech. But requiring it isn't going to produce radical improvements in student quality, especially in reading and writing skills. We also don't have a lot of data about the digital SAT, which is a significantly different test. The trend toward taking the SAT three times at a minimum and often 5+ times is also, while not completely new, continuing to rise.

I agree about the test. The test should be made much harder to really allow separation at the upper end, and you should get two tries.

Even CalTech doesn’t see the need to distinguish between a 780 math and an 800 math. So I don’t see how there can possibly be any demand for a test that allows even more separation at the upper end.

My guess is that has more to do with margins of error on the test scores than because there is no benefit to getting more tail end data.

the 780 gets you to like 3 standard deviations from the mean, I bet they would love to get 4 or 5 standard deviations like you used to be able to get with the test back in the 80's

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:The SAT is now a 2-hour, very preppable, superscorable test of math through Algebra 2, basic grammar/punctuation, and vocabulary in context. Median scores are declining because kids in the middle and below are learning less in high school. Scores at the top are rising because motivated kids with access to resources can easily master the content or just get lucky after taking it six times. It's a test mildly useful to validate high school grades, especially in math. It has little to do with whether kids can write or read and understand more than a paragraph or two at a time.

Caltech: "Test scores are predictive of success, even into students' sophomore and junior years"

Everyone is misunderstanding me. The SAT is better than nothing. Students without near-perfect math scores are going to struggle at Caltech. But requiring it isn't going to produce radical improvements in student quality, especially in reading and writing skills. We also don't have a lot of data about the digital SAT, which is a significantly different test. The trend toward taking the SAT three times at a minimum and often 5+ times is also, while not completely new, continuing to rise.

That's why Caltech looks at individual sections, not the composite. That's an important point overlooked by many including you.
They don't let kids with 800 math 730 verbal to be a bucket A kid. That will be a bucket A + bucket C.

SAT is an aptitude test. Prepping helps for sure but there is a limit, especially for reaching bucket A. The majority of kids being prepped and taking the test multiple times still would not reach bucket A in both categories.

You may argue that SAT is not hard enough and you wish it's harder. That's another topic, and college board has revamped the test multiple times in history so it's not unlikely they will do it again (as seen in August test).

The reality is not that many can achieve double bucket A at present, and it's a holistic review process so score isn't the only thing they look at.

The OP asked why professors are still complaining about students' lack of preparation at test-required colleges. My point is that many students who have weaknesses in the higher-level reading and writing skills the professors care about are nonetheless capable of getting a 1500+ on the SAT. I am actually in support of reinstating standardized testing in admissions. I am pushing back against the OP's apparent belief that if tests aren't perfect, they are useless. But it does no good for test proponents to overstate their value.

I don't think any one is overstating their value. The value is its tight correlation with academic success. Exactly that. No one is making statements not backed up by evidence.

Evidence after evidence shows that "Test scores are predictive of success, even into students' sophomore and junior years". Bucket A performs better than bucket B, and bucket B performs better than bucket C.

You are quoting a statement about Caltech, which is not typical even among elite institutions. The statement you quote says that there is a persistent correlation but does not quantify that correlation, so we have no idea how "tight" it is. In any event, I think colleges should require tests and am completely unsurprised that students who cannot score above the 98th percentile (which 780 is for math) with potentially multiple attempts on a test of fairly basic math skills will struggle at Caltech. This has nothing to do with whether the average Dartmouth student can write a competent research paper without ChatGPT, something that I can pretty confidently predict will not be much affected by a return to test-required.

Caltech are hardly the only ones to rediscover 100 years of psychometric science

https://opportunityinsights.org/wp-content/uploads/2024/01/SAT_ACT_on_Grades.pdf

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:The SAT is now a 2-hour, very preppable, superscorable test of math through Algebra 2, basic grammar/punctuation, and vocabulary in context. Median scores are declining because kids in the middle and below are learning less in high school. Scores at the top are rising because motivated kids with access to resources can easily master the content or just get lucky after taking it six times. It's a test mildly useful to validate high school grades, especially in math. It has little to do with whether kids can write or read and understand more than a paragraph or two at a time.

All the more reason a score about a certain threshold is important. It is a metric to prevent admitting students who just aren't prepared. It is just not that hard of a test and signals you have the minimal skills needed to succeed.

^ also, I know plenty of kids who did not prep at all and did very well. It is extremely preppable for the motivated, but prep is not required for those who are prepared through reading, school, etc.

Disagree. The SAT is in fact a pretty easy test. Anyone who paid attention in HS freshman and sophomore year should be able to do well IF it just asked questions like a normal test does. The part that makes it difficult is that it asks so many trick questions filled with word games. That's not how they should be testing students just to prove that they deserve to be in college. But this is how the multi billion SAT prep industry thrives, and how College Board makes a ton of money as a result of students taking the test multiple times.