The SAT matters

[Jayjay]

In order to tell if SAT discriminates against lower socioeconomic status, it seems like there should be some other better objective metric to compare it to. And if there was one, then the colleges should be using that instead.

To me it seems impossible to prove that SAT is disriminatory. Or at least very hard.

EDIT: I suppose one could look at how people of different socioeconomic background but identical SAT scores do in college. If there is a negative correlation between a student's family income or wealth, then the SAT is discriminatory. A positive correlation wouldn't prove that SAT is non-discriminatory though.

 

[fromderinside]

The problem is that the ability to communicate matters and facilities in the US varies with region, ethnicity, income, and family history among other factors. SAT is normalized for IQ and common regional environmental factors.

It is not a one-size-fits-all test since it requires both written and multiple guess responses. Obvious personal conditions need work arounds. As for normalization there are a variety of methods and appropriate instruments available to particular populations.

What it comes down to is how hard in time and money such remedies require. We are not used to spending very much beyond basic instruction to HS students in Public schools as faculty, service, and activity budgets attest.

My recommendation is permit everyone who wants to try higher education receive syllabi from potential institutions from which they can choose those that best fit their preparation rather than have the institution weed candidates. Depending on their rates of successfully preparing students the institutions can set prices for their services tailored to student performance.

 

[Loren Pechtel]

I got a 32 composite on the ACT test and a 96% on the Military Aptitude Test, without any practice, so you may be right. I am definitely smarter than anyone else I know.

I like your thinking! This means that several posters here acknowledge that I’m significantly smarter and better at life than they are!


(Oh, wait, I don’t believe that this test could possibly ascertain that, because so much more goes into our life capabilities. But delightful to walk around knowing they think it does! Gonna polish up my tiara tonight.)

Knowledge is where one finds it. I would never have known that preparing for a test does not help one get a better score, if not for this thread.

A long time ago in a faraway galaxy, I used to work on cars. General Motors came up with a program giving exams to dealership technicians and rewarding the highest scoring. It was a four hour exam of 200 multiple choice questions. The questions were drawn for GM service materials, and training materials. Service materials in those days was all on paper., but training materials were transitioning to VHS tapes.

GM had just come out with a new anti-lock brake system they named ABS6. It was a completely new engineering approach to anti-lock brakes and had nothing in common with previous systems. I hadn't even seen one on a car yet, but I had watched a video which explained it. This was the week before the exam. I was surprised to see a lot of ABS6 questions on the test. It was the entire brake section.

After the exam I walked out into the hall to see a guy in tears. Over the years, he and I had taken many advanced training classes together and knew there were few better than him, if any. He had not see the ABS6 training tape and had no idea what any of the questions were about. Over the past couple years, he an I had scored in the top 100 nationwide, but this time, he didn't make the cut.

This doesn't sound like preparing for a test, but simply doing the continuing education you should be doing. Something new comes along, you spent the time to learn about it, he didn't.

 

[Loren Pechtel]

One of the arguments I remember was that word problems were not situations minorities understood or somethimng that.

In China private tutors have been banned in order to ensure equal education.

My impression is that it's more about avoiding pushing too hard.

The forms of the SAT test quetions and problems are known. You can drill a kid with similar problems.

Yes, and you can do that without expensive test prep classes.

In Japan parents preparing kids for exams is insane, it starts at a young age.

The Moynihan Report circa 1960 looked at poverty and drew wnat was then controversiall conclusions.

One was that black and white school kids performed equally when there were stable families and an income. The second was that welfare was destroying the black family.

I don't know specifically about welfare, but most "racial" effects vanish when you compare apples to apples. They're actually socioeconomic.

There has to be metrics. The liberal solution is if minority can make the bar lower it.

No. The liberal solution is to take on faith that there are no differences, any differences you find are measurement errors or caused by external forces. ("It's a bad school" vs "a school with bad students will inherently do poorly" etc.)

 

[Bomb#20]

All of which means that Politesse was wrong to claim that the article "flatly contradicts" what Trausti wrote.

Let's have a look at what Trausti wrote:

Standard tests are the great equalizer; no matter your sex, your race, your religion, or social class. Thankfully, it looks like we’re returning to that understanding.

What? According to the article students with a higher socioeconomic background get better SAT results, which means they are advantaged in terms of admission to institutions of tertiary education. SATs manifestly do not equalise social class. Not even by your convoluted attempt to prove Trausti right and Politesse wrong.

It seems some here are using "equalizer" to mean "that which makes equal", while others are using it to mean "that which makes more equal". Neither usage is wrong per conventional English; and pointing out statistics of who scores how high on what test does not establish that one's own side's terminology is better than the other side's.

 

[Loren Pechtel]

Dropping any admissions requirement is necessarily a decision to weigh other factors more heavily. If other student characteristics, such as essays, recommendations, and coursework, are more strongly correlated with family income than test scores are, then dropping test scores actually tilts the playing field even more in favor of richer students. This was the situation that MIT found itself in after it suspended its SAT requirement in 2020. And other schools that dropped standardized tests during the pandemic will soon find themselves in the same straits.

In short, while SATs discriminate against students with a lower socioeconomic background, alternative methods of selection discriminate against them even more. Hence MIT's decision to reintroduce SATs.

Faith-based reasoning--you're assuming a differential outcome is proof of discrimination.

What MIT has found is that softer measurements of ability are easier to game. Nothing has shown that the SAT differences do not reflect actual ability.

 

[Loren Pechtel]

My recommendation is permit everyone who wants to try higher education receive syllabi from potential institutions from which they can choose those that best fit their preparation rather than have the institution weed candidates. Depending on their rates of successfully preparing students the institutions can set prices for their services tailored to student performance.

I'd like to see a system where a condition of public funding is that they make course material available online for any course where this makes sense. (Thus, lab courses wouldn't be available etc.)

 

[Trausti]

All of which means that Politesse was wrong to claim that the article "flatly contradicts" what Trausti wrote.

Let's have a look at what Trausti wrote:

Standard tests are the great equalizer; no matter your sex, your race, your religion, or social class. Thankfully, it looks like we’re returning to that understanding.

What? According to the article students with a higher socioeconomic background get better SAT results, which means they are advantaged in terms of admission to institutions of tertiary education. SATs manifestly do not equalise social class. Not even by your convoluted attempt to prove Trausti right and Politesse wrong.

It seems some here are using "equalizer" to mean "that which makes equal", while others are using it to mean "that which makes more equal". Neither usage is wrong per conventional English; and pointing out statistics of who scores how high on what test does not establish that one's own side's terminology is better than the other side's.

Equalizer in the sense that it’s the metic that can’t be gamed. Standardized tests, especially the SAT, are proxies for IQ. Your born with that; irrespective of your social class. It’s why requiring universal testing reveals poor geniuses that might otherwise be forgotten.

 

[Trausti]

No. The liberal solution is to take on faith that there are no differences, any differences you find are measurement errors or caused by external forces. ("It's a bad school" vs "a school with bad students will inherently do poorly" etc.)

Adopted children exhibit the traits and abilities of their biological parents. So much for external forces.

 

[Jayjay]

IQ is only about 60% inherited. The rest is nurture.

 

[Toni]

My recommendation is permit everyone who wants to try higher education receive syllabi from potential institutions from which they can choose those that best fit their preparation rather than have the institution weed candidates. Depending on their rates of successfully preparing students the institutions can set prices for their services tailored to student performance.

I'd like to see a system where a condition of public funding is that they make course material available online for any course where this makes sense. (Thus, lab courses wouldn't be available etc.)

Which would favor students who had unlimited access to the internet avc who learned well that way.

One thing the pandemic has taught us is that most students, from pre-k through university do NOT learn well by watching on the internet.

 

[Politesse]

One thing the pandemic has taught us is that most students, from pre-k through university do NOT learn well by watching on the internet.

Oh Jesus, no. But we actually knew that well before the pandemic, online success and retention always lagged relative to in person learning, especially for younger grade levels.

 

[Bronzeage]

I got a 32 composite on the ACT test and a 96% on the Military Aptitude Test, without any practice, so you may be right. I am definitely smarter than anyone else I know.

I like your thinking! This means that several posters here acknowledge that I’m significantly smarter and better at life than they are!


(Oh, wait, I don’t believe that this test could possibly ascertain that, because so much more goes into our life capabilities. But delightful to walk around knowing they think it does! Gonna polish up my tiara tonight.)

Knowledge is where one finds it. I would never have known that preparing for a test does not help one get a better score, if not for this thread.

A long time ago in a faraway galaxy, I used to work on cars. General Motors came up with a program giving exams to dealership technicians and rewarding the highest scoring. It was a four hour exam of 200 multiple choice questions. The questions were drawn for GM service materials, and training materials. Service materials in those days was all on paper., but training materials were transitioning to VHS tapes.

GM had just come out with a new anti-lock brake system they named ABS6. It was a completely new engineering approach to anti-lock brakes and had nothing in common with previous systems. I hadn't even seen one on a car yet, but I had watched a video which explained it. This was the week before the exam. I was surprised to see a lot of ABS6 questions on the test. It was the entire brake section.

After the exam I walked out into the hall to see a guy in tears. Over the years, he and I had taken many advanced training classes together and knew there were few better than him, if any. He had not see the ABS6 training tape and had no idea what any of the questions were about. Over the past couple years, he an I had scored in the top 100 nationwide, but this time, he didn't make the cut.

This doesn't sound like preparing for a test, but simply doing the continuing education you should be doing. Something new comes along, you spent the time to learn about it, he didn't.

My point is, I considered him my equal and perhaps better as a professional automotive technician. He was blindsided by an exam that was heavily weighted toward a new technology which very few people had actually seen. As a result of that exam I got a nice wristwatch and a ring, which is nice, but it doesn't indicate that I was any better at the job than him.

 

[Hermit]

All of which means that Politesse was wrong to claim that the article "flatly contradicts" what Trausti wrote.

Let's have a look at what Trausti wrote:

Standard tests are the great equalizer; no matter your sex, your race, your religion, or social class. Thankfully, it looks like we’re returning to that understanding.

What? According to the article students with a higher socioeconomic background get better SAT results, which means they are advantaged in terms of admission to institutions of tertiary education. SATs manifestly do not equalise social class. Not even by your convoluted attempt to prove Trausti right and Politesse wrong.

It seems some here are using "equalizer" to mean "that which makes equal", while others are using it to mean "that which makes more equal". Neither usage is wrong per conventional English; and pointing out statistics of who scores how high on what test does not establish that one's own side's terminology is better than the other side's.

Can there be such a thing as "more equal" than "equal"?

 

[Hermit]

Dropping any admissions requirement is necessarily a decision to weigh other factors more heavily. If other student characteristics, such as essays, recommendations, and coursework, are more strongly correlated with family income than test scores are, then dropping test scores actually tilts the playing field even more in favor of richer students. This was the situation that MIT found itself in after it suspended its SAT requirement in 2020. And other schools that dropped standardized tests during the pandemic will soon find themselves in the same straits.

In short, while SATs discriminate against students with a lower socioeconomic background, alternative methods of selection discriminate against them even more. Hence MIT's decision to reintroduce SATs.

Faith-based reasoning--you're assuming a differential outcome is proof of discrimination

What MIT has found is that softer measurements of ability are easier to game. Nothing has shown that the SAT differences do not reflect actual ability.

There is a discernible trend showing that students with a higher socioeconomic background get better SAT results. On what grounds do you conclude that the SAT differences reflect a higher actual ability of students with a higher socioeconomic background?

Gaming the system is not the issue. Quoting MIT’s dean of admissions, Stu Schmill, what MIT has found is that "Not having SATs/ACT scores to consider tends to raise socioeconomic barriers to demonstrating readiness for our education." Kathryn Paige Harden elaborates on that for the benefit of those who are a little slow on the uptake:

Dropping any admissions requirement is necessarily a decision to weigh other factors more heavily. If other student characteristics, such as essays, recommendations, and coursework, are more strongly correlated with family income than test scores are, then dropping test scores actually tilts the playing field even more in favor of richer students. This was the situation that MIT found itself in after it suspended its SAT requirement in 2020.

 

[Hermit]

All of which means that Politesse was wrong to claim that the article "flatly contradicts" what Trausti wrote.

Let's have a look at what Trausti wrote:

Standard tests are the great equalizer; no matter your sex, your race, your religion, or social class. Thankfully, it looks like we’re returning to that understanding.

What? According to the article students with a higher socioeconomic background get better SAT results, which means they are advantaged in terms of admission to institutions of tertiary education. SATs manifestly do not equalise social class. Not even by your convoluted attempt to prove Trausti right and Politesse wrong.

It seems some here are using "equalizer" to mean "that which makes equal", while others are using it to mean "that which makes more equal". Neither usage is wrong per conventional English; and pointing out statistics of who scores how high on what test does not establish that one's own side's terminology is better than the other side's.

Equalizer in the sense that it’s the metic that can’t be gamed. Standardized tests, especially the SAT, are proxies for IQ. Your born with that; irrespective of your social class. It’s why requiring universal testing reveals poor geniuses that might otherwise be forgotten.

It's not a matter of gaming the metric. Unless poverty is an indicator of IQ, the SAT results indicate that wealth tends to favour applicants from higher socioeconomic backgrounds. The article in the OP links to a study titled Socioeconomic Status and the Relationship Between the SAT® and Freshman GPA: An Analysis of Data from 41 Colleges and Universities, which comes up with a meaningful difference (ranging from about .25 to .40) between family background and SAT performance.

 

[Bomb#20]

What? According to the article students with a higher socioeconomic background get better SAT results, which means they are advantaged in terms of admission to institutions of tertiary education. SATs manifestly do not equalise social class. Not even by your convoluted attempt to prove Trausti right and Politesse wrong.

It seems some here are using "equalizer" to mean "that which makes equal", while others are using it to mean "that which makes more equal". Neither usage is wrong per conventional English; and pointing out statistics of who scores how high on what test does not establish that one's own side's terminology is better than the other side's.

Can there be such a thing as "more equal" than "equal"?

Let's say your left tank has 90 gallons but your right tank is down to 50 gallons because you've been crossfeeding your left engine from the tank in the right wing. So you flip the switch and start having the left tank pump fuel to both engines, and after a while the left tank has gone down to 70 gallons while the right tank still has 50 gallons. The left tank used to have 40 gallons more than the right tank but now it only has 20 gallons more. So the fuel levels are more equal than they were, but they still aren't equal. That's what "more equal" means. It doesn't mean some animals are more equal than others.

While you were crossfeeding the right engine from the left tank in order to bring the levels closer to each other, some people would call what you were doing "equalizing" the fuel levels because because it made the levels more equal. Other people wouldn't call it "equalizing" the fuel levels because 70 still isn't equal to 50. It's a matter of taste in terminology.

 

[Hermit]

they still aren't equal.

70 still isn't equal to 50.

Thanks for your lesson, and I agree. Either something equals another, or it does not. If you compare two things that are not, try "less different". "Equal" is like being pregnant or unique. A woman can't be a little bit pregnant and an explanation can't be a little bit unique. 2+2=5 is not a little bit correct. Some things are just binary. Oh, and "binary" is another one of those things.

 

[steve_bank]

The basic requirements for math, reading, and writing have not changed since the 50s.

If you score poorly on SATs I do not see how you can handle an engineering or science curriculum. Business math includes calculus.

The result is remedial math courses before taking regular courses.

Part of the problem may be the expectation of having all kids graduate high school with equal capacity to do college level work.

 

[Bomb#20]

Standard tests are the great equalizer; no matter your sex, your race, your religion, or social class. Thankfully, it looks like we’re returning to that understanding.

Thanks for your lesson, and I agree. Either something equals another, or it does not. If you compare two things that are not, try "less different". "Equal" is like being pregnant or unique. A woman can't be a little bit pregnant and an explanation can't be a little bit unique. 2+2=5 is not a little bit correct. Some things are just binary. Oh, and "binary" is another one of those things.

That's a perfectly legitimate position to take on English usage. On the other hand, telling Trausti "The proper term to use is to say that standard tests are the great less-different-izer." lacks a bit of the rhetorical punch of telling him "That statement is flatly contradicted by the linked article." Making a stink about Trausti's choice of terminology because one is a definition-stickler and sweeping under the rug the minor detail that one has no substantive disagreement with him about what the linked article shows looks an awful lot like playing lame semantics games as a justification for making the perfect the enemy of the good.