The Ivy-Plus is well worth studying, but it has a separate problem: Maxing out on test scores and grades. When, hypothetically, all your applicants are 97th-99th percentile, the info is not as useful as 70th-80th percentile like for a worse school. So you need to use other things. I saw that problem in economics PhD applications from China, and it is a problem for PhD applications generally. A special test designed for people in the SAT 95th percentile plus (instead of the 50th percentile) is needed. Someone could make money creating and running it (an Ivy-plus consortium could do it).
It's true that many students would be matched on academic qualifications if those were the only criteria, but at the same time, switching to an academics-only admissions regime would still increase socioeconomic diversity at schools while curbing elite admissions advantages.
SAT subject tests were a good option for figuring out ability in a more discrete way in the right tail, but they've gone the way of the dodo.
So if I understand from a quick reading, bringing back some form of standardized testing for admission would help the students at the lower-middle end of the admissions spectrum, but “punish” those now at the higher end? That begs the question of whether the Ivy’s really want this. Much to be had from wealthy students and their families in this league—regardless of their current virtue signaling.
Can we get a Cremieux reading list? I am specifically interested in what you recommend in terms of technical material like structural equation modeling, measurement invariance, econometrics, etc--the highly involved statistics that is your calling card.
For another post: You show via the footnoted link that retesting is not very important for the rich/poor divide. Can you write on whether superscoring is a good idea? One thing it does is make a lot of money for the testing agencies (how much of their income?). But it helps by regressing to the mean-- unlucky first-takers can get a more accurate measure by taking it twice. How helpful is it? That depends on the variance in retested scores, which is hard to measure because of the learning effect, but I bet it can be done. For example, between test 5 and test 6, there wouldn't be much learning.
Brings back an old memory from my grad days. My major professor always considered retest scoring effect in admission recommendations (in those days, student admissions were voted upon by all faculty). He considered these in essence negatively and ignored increases of 50 or less points. In short, he expected that increase simply from “learning effect”. I remember being glad in not retesting as I was marginal.
I'm incredibly curious about your timed writing method... Did it ever fail ? Did you have to rewrite the post?
Are you doing this because you think constraints (here time constraints) help you be more creative? I've been thinking about implementing a method like that.
I always enjoy your work! Thanks for the great content.
I'm interested in the idea of "timed writing". Could you give us more details about it? Why you are trying it? What are your initial feelings about the experience? What specific editor are you using?
"One potential reason for this is that educated parents recognize the competitiveness of modern admissions, so they send their kids to schools that hand out higher grades or they broker with schools and teachers to make sure higher grades happen."
How do you know the improvement in grades is due to grade inflation, and not to those same parents pushing their children to work harder, helping them with homework, and getting them tutoring if they are behind in any subject? It's a natural response to the competitiveness of modern admissions, and unlike IQ test scores, it's very doable for most kids to improve their grades.
At my institution and others, yes tuition “milking” is rampant and the standards of admission declining. They make no bones about expanding *total* enrollment. But the Ivy’s? If they expand enrollment at the lower levels of student family prestige and wealth, does that come at the expense of their wealthy, elite, clientele? I’m not privy to insider info there.
"But the Ivy’s? If they expand enrollment at the lower levels of student family prestige and wealth, does that come at the expense of their wealthy, elite, clientele? I’m not privy to insider info there."
I am also not privy to info from there. But as long as the aura of eliteness remains, I suppose everything is copacetic.
One of the biggest threats to quality education in the United States is the destruction of meritorious colleges and even high schools in the name of equity, as is the case currently at the Thomas Jefferson High School in Virginia and recently in Chicago, where a decision to eliminate the city’s 11 selective public high schools was made.
We had just such an incident here. We had *one* prestigious university high school. It was always ranked in the national top 10 high schools. The (then minority) school board voted to demand that the top 10% of every grade school be admitted. It’s only “fair” you know.
Of course, not all grade schools are alike in student body performance and they were warned repeatedly about a decline in achievement. School board did not care. Our one university high school has now dropped out of any national rankings. And then there were none…
I think the best option would be to recognize that specialized-knowledge tests and general-intelligence tests should be used for different purposes. Most Chinese have no need for learning English, so there really is no justification as to why most Chinese should be forced to pass a test demonstrating knowledhe of English.
Specialized-knowledge tests should be used as a means of certifying the testees as having sufficient specialized knowledhe, and general-intelligence tests should be used for finding those among the general populous with sufficient raw intellectual capability to be worth bothering to educate in the fields requiring at least above-average intelligence to have any hope of successfully mastering said field.
There will be exceptions, like savants, and there can be accomodations made for them depending on their needs, but the current system we have is stupidly wasteful and inefficient. There's only a relatively small amount of knowledge and skills that are useful, and practical to teach, to the majority of any populous.
The Ivy-Plus is well worth studying, but it has a separate problem: Maxing out on test scores and grades. When, hypothetically, all your applicants are 97th-99th percentile, the info is not as useful as 70th-80th percentile like for a worse school. So you need to use other things. I saw that problem in economics PhD applications from China, and it is a problem for PhD applications generally. A special test designed for people in the SAT 95th percentile plus (instead of the 50th percentile) is needed. Someone could make money creating and running it (an Ivy-plus consortium could do it).
It's true that many students would be matched on academic qualifications if those were the only criteria, but at the same time, switching to an academics-only admissions regime would still increase socioeconomic diversity at schools while curbing elite admissions advantages.
SAT subject tests were a good option for figuring out ability in a more discrete way in the right tail, but they've gone the way of the dodo.
So if I understand from a quick reading, bringing back some form of standardized testing for admission would help the students at the lower-middle end of the admissions spectrum, but “punish” those now at the higher end? That begs the question of whether the Ivy’s really want this. Much to be had from wealthy students and their families in this league—regardless of their current virtue signaling.
All institutes of higher learning have discovered the monetary benefit of lowering admissions standards. And have no shame in doing so.
If you wrote this in less than an hour you're a machine. AGI has arrived.
Can we get a Cremieux reading list? I am specifically interested in what you recommend in terms of technical material like structural equation modeling, measurement invariance, econometrics, etc--the highly involved statistics that is your calling card.
I could work on it!
For another post: You show via the footnoted link that retesting is not very important for the rich/poor divide. Can you write on whether superscoring is a good idea? One thing it does is make a lot of money for the testing agencies (how much of their income?). But it helps by regressing to the mean-- unlucky first-takers can get a more accurate measure by taking it twice. How helpful is it? That depends on the variance in retested scores, which is hard to measure because of the learning effect, but I bet it can be done. For example, between test 5 and test 6, there wouldn't be much learning.
In general, superscoring is quite bad for accurately representing student ability. I could write something on that in the future.
Brings back an old memory from my grad days. My major professor always considered retest scoring effect in admission recommendations (in those days, student admissions were voted upon by all faculty). He considered these in essence negatively and ignored increases of 50 or less points. In short, he expected that increase simply from “learning effect”. I remember being glad in not retesting as I was marginal.
I'm incredibly curious about your timed writing method... Did it ever fail ? Did you have to rewrite the post?
Are you doing this because you think constraints (here time constraints) help you be more creative? I've been thinking about implementing a method like that.
I always enjoy your work! Thanks for the great content.
How is it possible to have a 4.0 GPA and a low class rank, if 4.0 is the highest possible GPA, which is what determines class rank?
I'm interested in the idea of "timed writing". Could you give us more details about it? Why you are trying it? What are your initial feelings about the experience? What specific editor are you using?
Thanks a lot!
"One potential reason for this is that educated parents recognize the competitiveness of modern admissions, so they send their kids to schools that hand out higher grades or they broker with schools and teachers to make sure higher grades happen."
How do you know the improvement in grades is due to grade inflation, and not to those same parents pushing their children to work harder, helping them with homework, and getting them tutoring if they are behind in any subject? It's a natural response to the competitiveness of modern admissions, and unlike IQ test scores, it's very doable for most kids to improve their grades.
At my institution and others, yes tuition “milking” is rampant and the standards of admission declining. They make no bones about expanding *total* enrollment. But the Ivy’s? If they expand enrollment at the lower levels of student family prestige and wealth, does that come at the expense of their wealthy, elite, clientele? I’m not privy to insider info there.
"But the Ivy’s? If they expand enrollment at the lower levels of student family prestige and wealth, does that come at the expense of their wealthy, elite, clientele? I’m not privy to insider info there."
I am also not privy to info from there. But as long as the aura of eliteness remains, I suppose everything is copacetic.
This was to Realist’s comment below.
One of the biggest threats to quality education in the United States is the destruction of meritorious colleges and even high schools in the name of equity, as is the case currently at the Thomas Jefferson High School in Virginia and recently in Chicago, where a decision to eliminate the city’s 11 selective public high schools was made.
We had just such an incident here. We had *one* prestigious university high school. It was always ranked in the national top 10 high schools. The (then minority) school board voted to demand that the top 10% of every grade school be admitted. It’s only “fair” you know.
Of course, not all grade schools are alike in student body performance and they were warned repeatedly about a decline in achievement. School board did not care. Our one university high school has now dropped out of any national rankings. And then there were none…
Is data on AP tests available? They measure real knowledge. Combining them with a pure IQ test for applications would be a good idea.
I think the best option would be to recognize that specialized-knowledge tests and general-intelligence tests should be used for different purposes. Most Chinese have no need for learning English, so there really is no justification as to why most Chinese should be forced to pass a test demonstrating knowledhe of English.
Specialized-knowledge tests should be used as a means of certifying the testees as having sufficient specialized knowledhe, and general-intelligence tests should be used for finding those among the general populous with sufficient raw intellectual capability to be worth bothering to educate in the fields requiring at least above-average intelligence to have any hope of successfully mastering said field.
There will be exceptions, like savants, and there can be accomodations made for them depending on their needs, but the current system we have is stupidly wasteful and inefficient. There's only a relatively small amount of knowledge and skills that are useful, and practical to teach, to the majority of any populous.