Dartmouth to require SAT scores again

[budwom]

https://www.nytimes.com/2024/02/05/b...mouth-sat.html

Interesting article. New president who came from Barnard asked some of her profs to assess whether not requiring SAT tests had been a good idea...one of the main findings (does not surprise me) is that test scores were a better predictor than essays, high school grades or teacher recommendations for how kids would perform at Dartmouth. Which leads to the second finding, that in fact many disadvantaged kids hurt themselves by not submitting SAT scores, thinking they had scored too low, when in fact Dartmouth would have admitted many of them.

[PackMan97]

Shocking! Measuring how well you test can be an indicator of how well you spend the next four years being tested.

[Bluedog]

https://www.nytimes.com/2024/02/05/b...mouth-sat.html

Interesting article. New president who came from Barnard asked some of her profs to assess whether not requiring SAT tests had been a good idea...one of the main findings (does not surprise me) is that test scores were a better predictor than essays, high school grades or teacher recommendations for how kids would perform at Dartmouth. Which leads to the second finding, that in fact many disadvantaged kids hurt themselves by not submitting SAT scores, thinking they had scored too low, when in fact Dartmouth would have admitted many of them.

There was a good podcast on this from NY Times too.
https://www.nytimes.com/2024/01/17/p...n-the-sat.html

It's somewhat surprising to me at least that a single test is a better predictor than 4 years of high school curriculum/grades. That shows how disconnected grading in high school is from college (perhaps "Effort" vs. "results") as well as how high schools are so different from on another, and huge grade inflation across the board. But seriously, that's kinda crazy. The podcast also addresses the concern that the SAT is biased towards wealthy students -- suggesting while there is a slight advantage based on ability to spend $$ on tutoring, etc. the advantage is much more muted than one would expect and is actually LESS than the advantage on other aspects of the application like essays, teacher recs, etc. They compare SAT scores to nationalized standardized tests (which wealthy students don't do extra test prep for, because nobody cares about the results on an individual-bases, schools/districts care about it since they get judged based on the aggregate results compared to state and national averages) and show that the advantaged-disadvantaged gap on those tests are very similar to the SAT.

The podcast presents data from MIT which has also re-instituted the test policy as a way to better gauge who can be successfully - and actually admit an even MORE diverse class by having that extra datapoint.

Finally, the fact that elite schools publish SAT/ACT ranges probably scares low-income students from submitting their scores when they are below the range, but they don't take into account the fact that being below by a bit is actually seen as a GOOD thing by the admissions committee as long as they are in shouting distance. They aren't as informed about this fact, so probably withhold scores that would actually help them.

Based on the evidence, I'm in favor of Dartmouth's policy. Without this, there is even more randomness and just "guessing" by admissions as to which disadvantaged students would do well vs. actually having a datapoint that would better help inform that decision. Schools will struggle with it though for two primary reasons: 1.) Concerns that it will reduce their applicant pool numbers (almost assuredly), 2.) Concerns it will subject them to potential litigation based on the recent Supreme Court ruling of not using race in admissions (as test scores are more quantifiable "data", GPAs and the like for top tier candidates across races are nearly indistinguishable). Both of these are addressed in the linked article though, but some schools may come to a different conclusion.

Shocking! Measuring how well you test can be an indicator of how well you spend the next four years being tested.

Yes, but I don't think many intuitively believe that a single test is more indicative of GRADES in college than are four years of GRADES in high school. But that's what the data says. SAT/ACT scores are a better predictor of college GPA than high school GPA.

[sagegrouse]

I spoke to a Duke official two weeks ago who said that Duke is going to collect 4-5 years of data before making a decision. I can't remember when test scores became optional at Duke.

[Bluedog]

I spoke to a Duke official two weeks ago who said that Duke is going to collect 4-5 years of data before making a decision. I can't remember when test scores became optional at Duke.

COVID year. Like all the rest. Made sense that year given not everybody could go to a testing center to take the test. All the schools at the time announced it as a temporary COVID measure, but it has since stuck.

Thanks for the insights.

[CrazyNotCrazie]

https://www.nytimes.com/2024/02/05/b...mouth-sat.html

Interesting article. New president who came from Barnard asked some of her profs to assess whether not requiring SAT tests had been a good idea...one of the main findings (does not surprise me) is that test scores were a better predictor than essays, high school grades or teacher recommendations for how kids would perform at Dartmouth. Which leads to the second finding, that in fact many disadvantaged kids hurt themselves by not submitting SAT scores, thinking they had scored too low, when in fact Dartmouth would have admitted many of them.

Great. Logic is restored.

If you don't submit a test score, they are logically going to assume that it is not good. I don't know what the assumed score is, but it is probably pretty low. My guess is that they tell some preferred applicants (quarterbacks, kids of billionaires, virtuoso tuba players, those fed through special inner-city recruiting programs) who might have weak scores but are going to get in anyway not to formally submit so as to not lower averages, but for the non-preferred applicant, their chances of getting in without a test score are probably miniscule. I would love to see the breakdowns of their entering classes in terms of those who submitted and those who didn't.

Meanwhile, they are getting inundated with applicants from those who think they can get in without a test score.

I recognize that some people are better test takers than others (I am a good test taker, my older child also is, so I have some bias here, though those who are poor test takers often have the opposite bias), but as part of a overall application package, test scores are a valuable normalizing resource. Yes, tutors can help, but there is plenty of material available for little cost for those who want them.

[sagegrouse]

COVID year. Like all the rest. Made sense that year given not everybody could go to a testing center to take the test. All the schools at the time announced it as a temporary COVID measure, but it has since stuck.

Thanks for the insights.

So, it began with the class entering Aug. 2021? Surely the previous class's applications were done by March 2020.

[Bluedog]

So, it began with the class entering Aug. 2021? Surely the previous class's applications were done by March 2020.

Yes, that's correct. Sorry for the ambiguity. They announced it in June 2020 to apply for the "2020-21 admissions cycle":
https://today.duke.edu/2020/06/state...dmissions-2021

[Bluedog]

Great. Logic is restored.

If you don't submit a test score, they are logically going to assume that it is not good. I don't know what the assumed score is, but it is probably pretty low. My guess is that they tell some preferred applicants (quarterbacks, kids of billionaires, virtuoso tuba players, those fed through special inner-city recruiting programs) who might have weak scores but are going to get in anyway not to formally submit so as to not lower averages, but for the non-preferred applicant, their chances of getting in without a test score are probably miniscule. I would love to see the breakdowns of their entering classes in terms of those who submitted and those who didn't.

Meanwhile, they are getting inundated with applicants from those who think they can get in without a test score.

I recognize that some people are better test takers than others (I am a good test taker, my older child also is, so I have some bias here, though those who are poor test takers often have the opposite bias), but as part of a overall application package, test scores are a valuable normalizing resource. Yes, tutors can help, but there is plenty of material available for little cost for those who want them.

Yes, students coming from "Advantaged backgrounds" without a hook are largely expected to submit test scores (and have a high one -- would want to be ~75th+ percentile of admitted students ideally). It's easy for admissions to simply go to the next applicant where rest of application is similar, but DOES have a good test score and decided to admit that person instead -- plenty of students to choose from in that category so might as well go with the "lower risk" options. For those from disadvantaged backgrounds, the submission rates on test scores is likely noticeably smaller, so much more guessing going on.

[aimo]

Wake Forest stopped requiring SATs in 2008. They were among the earliest to do so.

[Mtn.Devil.91.92.01.10.15]

I guess I'm pushing uphill here - I have personally seen lots of anecdotal evidence that these tests can be "learned" by students that have families and resources to teach and tutor. Now, this may correlate to success in school - being taught and tutored to beat these tests may indicate better study habits and or the resources to do similarly in college - but I don't understand how there isn't some sort of more equitable system for evaluating college readiness.

I had a friend who had mediocre test scores and from 7th grade on spent literally 2 hours a day all summer learning the SAT.

She's a medical doctor now, so it measured sole degree of something. I'm just not sure there isn't a better way to include students who don't have those same advantages.

(Full disclosure - the system worked nicely for me. I didn't prep for the SAT and consequently I managed to miss one question on the verbal)

[Bluedog]

I guess I'm pushing uphill here - I have personally seen lots of anecdotal evidence that these tests can be "learned" by students that have families and resources to teach and tutor. Now, this may correlate to success in school - being taught and tutored to beat these tests may indicate better study habits and or the resources to do similarly in college - but I don't understand how there isn't some sort of more equitable system for evaluating college readiness.

I had a friend who had mediocre test scores and from 7th grade on spent literally 2 hours a day all summer learning the SAT.

She's a medical doctor now, so it measured sole degree of something. I'm just not sure there isn't a better way to include students who don't have those same advantages.

(Full disclosure - the system worked nicely for me. I didn't prep for the SAT and consequently I managed to miss one question on the verbal)

The data suggests that while the test can be somewhat "learned" and have money affords some advantage in that capacity, the advantage is actually much less than other parts of the application (e.g. teacher recs, extra-curriculars -how much money does it cost to be on travel sports?? etc.) In addition, there is a lot of free test prep out there for those dedicated students. So, contrary to what one might expect, the data suggests SAT/ACT scores skew less to the wealthy than other parts of the application -- and each student is still evaluated in the context of their environment so disadvantaged students get an extra "feather in the cap" should they do well. And the data also suggest that all the test prep doesn't actually do all that much in aggregate -- the test gaps in state/national standardized test is very similar despite no student paying for extra tutoring geared specifically towards those tests.

[Bostondevil]

I don't for one minute trust the statistical acumen of the average Dartmouth professor. Associating anything with grades at an Ivy League school which have been shown to be the worst as far as grade inflation is concerned - also a problem. One of the things Ivy League students do is purposely take easier classes to boost their GPA, so, associating test scores with grades at Dartmouth would need to be weighted by difficulty of course. We don't know what kind of financial incentive Dartmouth may have to reinstate such tests as well. Having gone through the application process 5 times now, I am award the the College Board is a racket designed to milk money out of upper middle class parents. Kids are encouraged to apply to at least 10 schools now, complete with application fees, and sending one's test scores to more than 3 schools costs extra.

There is also this report on how Ivy League students avoid challenging classes because the GPA is the goal, not learning. https://www.townandcountrymag.com/so...rading-system/

If Dartmouth really wanted to give disadvantaged high schoolers a leg up, they would switch to random admissions. If they are unwilling to do that, at least get rid of athletic admits which predominately go to upper middle class and/or white students. (How many disadvantaged youth fencing teams do you know of? Golf? Lacrosse? If you doubt me about privilege Ivy League sports, pick a random Ivy League team, go to the online roster, and check where the kids went to high school.)

[CrazyNotCrazie]

The data suggests that while the test can be somewhat "learned" and have money affords some advantage in that capacity, the advantage is actually much less than other parts of the application (e.g. teacher recs, extra-curriculars -how much money does it cost to be on travel sports?? etc.) In addition, there is a lot of free test prep out there for those dedicated students. So, contrary to what one might expect, the data suggests SAT/ACT scores skew less to the wealthy than other parts of the application -- and each student is still evaluated in the context of their environment so disadvantaged students get an extra "feather in the cap" should they do well. And the data also suggest that all the test prep doesn't actually do all that much in aggregate -- the test gaps in state/national standardized test is very similar despite no student paying for extra tutoring geared specifically towards those tests.

Agreed. And no one is saying that it should be the sole determining factor, or even close. But it is a valuable tool when used intelligently and in combination with the other factors. With years of experience, I think most admissions offices know how to do this.

If you know you need to do well on the test and the resources are available (they are), then you put in the time to prepare. Admissions offices are smart enough to know that a kid who is working several jobs to support their family and also playing several varsity sports might not have as much time to prepare and will adjust expectations accordingly, within reason. It is a holistic package and people who get too fixated on one part of it, particularly the "equity" of that factor, or losing sight of the overall process.

[bundabergdevil]

https://www.nytimes.com/2024/02/05/b...mouth-sat.html

Interesting article. New president who came from Barnard asked some of her profs to assess whether not requiring SAT tests had been a good idea...one of the main findings (does not surprise me) is that test scores were a better predictor than essays, high school grades or teacher recommendations for how kids would perform at Dartmouth. Which leads to the second finding, that in fact many disadvantaged kids hurt themselves by not submitting SAT scores, thinking they had scored too low, when in fact Dartmouth would have admitted many of them.

Interesting. I worked for a few high end test prep companies for a bit. I’ve always heard the number 1 predictor of test scores is household income.

Not sure what to do about or how to square that.

[CrazyNotCrazie]

Interesting. I worked for a few high end test prep companies for a bit. I’ve always heard the number 1 predictor of test scores is household income.

Not sure what to do about or how to square that.

I live in NYC and have an 8th grader applying to high school. We have selective public high schools here (Stuyvesant, Bronx Science, Brooklyn Tech, etc.) that are among the best in America. Admissions is based solely on the outcome of a test. Nothing else. Upper middle class kids (including my son) have tutors who help. But Stuyvesant, which consistently draws by far the kids with the highest scores (there is a matching process) has a huge percentage of kids who are poor Asian immigrants. They know it is their ticket out so they devote countless hours to studying for it. There are also free tutoring programs available throughout the city for those who seek them out. Many complain that the test discriminates but the fact that so many of these kids do so well on it disproves this to me.

Yes, wealth is definitely a factor. But a lot of other things go into it and by age 18 wealth intersects with many other factors. Don't throw out the baby with the bathwater.

The process is far from perfect. And we all bring different perspectives (often driven by our and our children's strengths and weaknesses) that influence these perspectives - I admitted my biases above. As I suggested above, using scores while knowing some of their strengths and weaknesses overall is a positive.

[Bluedog]

Agreed. And no one is saying that it should be the sole determining factor, or even close. But it is a valuable tool when used intelligently and in combination with the other factors. With years of experience, I think most admissions offices know how to do this.

If you know you need to do well on the test and the resources are available (they are), then you put in the time to prepare. Admissions offices are smart enough to know that a kid who is working several jobs to support their family and also playing several varsity sports might not have as much time to prepare and will adjust expectations accordingly, within reason. It is a holistic package and people who get too fixated on one part of it, particularly the "equity" of that factor, or losing sight of the overall process.

Yes, this is also a key point that is emphasized in the podcast. Definitely not the most important factor, but just another datapoint in conjunction with the rest of the application.

[Bluedog]

I live in NYC and have an 8th grader applying to high school. We have selective public high schools here (Stuyvesant, Bronx Science, Brooklyn Tech, etc.) that are among the best in America. Admissions is based solely on the outcome of a test. Nothing else. Upper middle class kids (including my son) have tutors who help. But Stuyvesant, which consistently draws by far the kids with the highest scores (there is a matching process) has a huge percentage of kids who are poor Asian immigrants. They know it is their ticket out so they devote countless hours to studying for it. There are also free tutoring programs available throughout the city for those who seek them out. Many complain that the test discriminates but the fact that so many of these kids do so well on it disproves this to me.

Yes, wealth is definitely a factor. But a lot of other things go into it and by age 18 wealth intersects with many other factors. Don't throw out the baby with the bathwater.

The process is far from perfect. And we all bring different perspectives (often driven by our and our children's strengths and weaknesses) that influence these perspectives - I admitted my biases above. As I suggested above, using scores while knowing some of their strengths and weaknesses overall is a positive.

Of course, there has been a lot of blowback on NYC's selective enrollment high schools basing on test scores in recent years and having small enrollments of Black and Hispanic students:
https://www.nytimes.com/2019/08/26/n...gregation.html

Seeing similar murmurings here in Chicago of the desire to go away from selective enrollment schools based on test scores (although in Chicago, the test scores required have different ranges based on the socioeconomic conditions of the area you live in Chicago -- it used to be based on race, but that was ruled unconstitutional at the federal level).

Although universities, as stated, wouldn't rely on only test scores in making decisions, and it wouldn't be even the most important criteria so is definitely a different beast.

No easy answers in all this.

[bundabergdevil]

I live in NYC and have an 8th grader applying to high school. We have selective public high schools here (Stuyvesant, Bronx Science, Brooklyn Tech, etc.) that are among the best in America. Admissions is based solely on the outcome of a test. Nothing else. Upper middle class kids (including my son) have tutors who help. But Stuyvesant, which consistently draws by far the kids with the highest scores (there is a matching process) has a huge percentage of kids who are poor Asian immigrants. They know it is their ticket out so they devote countless hours to studying for it. There are also free tutoring programs available throughout the city for those who seek them out. Many complain that the test discriminates but the fact that so many of these kids do so well on it disproves this to me.

Yes, wealth is definitely a factor. But a lot of other things go into it and by age 18 wealth intersects with many other factors. Don't throw out the baby with the bathwater.

The process is far from perfect. And we all bring different perspectives (often driven by our and our children's strengths and weaknesses) that influence these perspectives - I admitted my biases above. As I suggested above, using scores while knowing some of their strengths and weaknesses overall is a positive.

I don’t think the system is without merit. The wealthy enjoy a lot of privileges. It is what it is. Don’t have great answers but researchers have consistently found material differences in score performance based on family income levels.

[sagegrouse]

Interesting. I worked for a few high end test prep companies for a bit. I’ve always heard the number 1 predictor of test scores is household income.

Not sure what to do about or how to square that.

Ah, yes! The rewards of a meritocracy.