New paper on determinants of college admissions…

[Anonymous]

Anonymous wrote:

We show that under the identification assumption that different college admission committees’ assessments of a candidate’s underlying merit (i.e., the component that predicts long-term outcomes) are positively correlated with each other, comparisons of students who are admitted vs. rejected from the waitlist can be used to identify the causal effect of admission for marginal applicants.

Using this design, we find that being admitted from the waitlist to an Ivy-Plus college increases students’ chances of achieving early career upper-tail success on both monetary and non-monetary dimensions. The causal effects of admission to an Ivy-Plus college are much larger for students with weaker fallback options– e.g., whose colleges in their home state channel fewer students to the top 1% after college. Exploiting this heterogeneity in treatment effects, we estimate that the marginal student who is admitted to and attends an Ivy-Plus college instead of the average flagship public is about 50% more likely to reach the top 1% of the income distribution at age 33, nearly twice as likely to attend a highly-ranked graduate school, and 2.5 times as likely to work at a prestigious firm.

Seems like the assumptions about admission from the waitlist are not warranted. Most waitlists are need-aware. Generally, waitlist acceptance is related to institutional priorities. I think they are reading way too much into that.

--my kid was admitted to an Ivy-Plus off the waitlist this year. full pay, high stats. Was not admitted off the waitlist at other, lower-ranked schools. Not being admitted off the waitlist at the other schools does not imply anything about the merit of his app.

They are trying to do too much in this "study".

College-Specific Analysis Sample. When studying admissions and matriculation at specific colleges Section 3.2), admissions decisions (Section 3.3), and the causal effects of colleges on outcomes (Section 4), we focus on the subset of Ivy-Plus and flagship public colleges for which we have internal application and admissions data. In these analyses, we define the analysis sample as all permanent residents or citizens in the college-specific dataset who submitted a first-year undergraduate application to the college over the years for which we have data who (1) can be linked to the tax data based on their SSNs or ITINs and (2) can be linked to parents in the tax data.

Full pay students often do not include SSNs in the application because they are not applying for need-based aid. My kids did not.

The data come from 1098-T forms that colleges submit for all tuition paying students. For non-tuition paying students, it’s complemented by Pell Grant data. Read before you criticize.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:…to highly selective colleges (8 ivies, Stanford, MIT, UChicago, Duke)

https://opportunityinsights.org/wp-content/uploads/2023/07/CollegeAdmissions_Paper.pdf

The analysis shows that highest income applicants (top 1 percent) have an admission advantage over the average applicant due to: 1) legacy admissions; 2) athletic recruitment; 3) non-academic factors (e.g., private school extracurriculars). In fact, legacy admissions explain about half of the gap between acceptance rates between highest income and average applicants.

Also, attending IvyPlus colleges does improve earnings and leadership prospects after colleges. The authors do a nice job of identifying the causal effect of IvyPlus attendance.

Putting these findings together, the implication is that more socioeconomic diversity can be achieved without sacrificing academic quality by eliminating legacy admissions and athletic recruitment.

Actually it does not say that. Might want to actually spend some time with it before you type.

OP here. I have actually read the paper. What I wrote above is my summary of the findings (rather than a direct quotation). And yes, I am qualified to summarize an economics paper

If that is the case then your summary is incorrect. And I too read the paper and am qualified to evaluate the results.

Just curious: what are your qualifications?

Graduate degree in applied mathematical economics.

The paper is interesting but the conclusions are not what some here are taking away and there are nuances in the results when certain controls are put in place.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:…to highly selective colleges (8 ivies, Stanford, MIT, UChicago, Duke)

https://opportunityinsights.org/wp-content/uploads/2023/07/CollegeAdmissions_Paper.pdf

The analysis shows that highest income applicants (top 1 percent) have an admission advantage over the average applicant due to: 1) legacy admissions; 2) athletic recruitment; 3) non-academic factors (e.g., private school extracurriculars). In fact, legacy admissions explain about half of the gap between acceptance rates between highest income and average applicants.

Also, attending IvyPlus colleges does improve earnings and leadership prospects after colleges. The authors do a nice job of identifying the causal effect of IvyPlus attendance.

Putting these findings together, the implication is that more socioeconomic diversity can be achieved without sacrificing academic quality by eliminating legacy admissions and athletic recruitment.

Actually it does not say that. Might want to actually spend some time with it before you type.

OP here. I have actually read the paper. What I wrote above is my summary of the findings (rather than a direct quotation). And yes, I am qualified to summarize an economics paper

If that is the case then your summary is incorrect. And I too read the paper and am qualified to evaluate the results.

Just curious: what are your qualifications?

Graduate degree in applied mathematical economics.

The paper is interesting but the conclusions are not what some here are taking away and there are nuances in the results when certain controls are put in place.

OP here. My qualifications: PhD from top3 Econ program (think MIT, Harvard, etc.) with specialization in econometrics. In fact, I took labor economics with one of the authors of the study.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

We show that under the identification assumption that different college admission committees’ assessments of a candidate’s underlying merit (i.e., the component that predicts long-term outcomes) are positively correlated with each other, comparisons of students who are admitted vs. rejected from the waitlist can be used to identify the causal effect of admission for marginal applicants.

Using this design, we find that being admitted from the waitlist to an Ivy-Plus college increases students’ chances of achieving early career upper-tail success on both monetary and non-monetary dimensions. The causal effects of admission to an Ivy-Plus college are much larger for students with weaker fallback options– e.g., whose colleges in their home state channel fewer students to the top 1% after college. Exploiting this heterogeneity in treatment effects, we estimate that the marginal student who is admitted to and attends an Ivy-Plus college instead of the average flagship public is about 50% more likely to reach the top 1% of the income distribution at age 33, nearly twice as likely to attend a highly-ranked graduate school, and 2.5 times as likely to work at a prestigious firm.

Seems like the assumptions about admission from the waitlist are not warranted. Most waitlists are need-aware. Generally, waitlist acceptance is related to institutional priorities. I think they are reading way too much into that.

--my kid was admitted to an Ivy-Plus off the waitlist this year. full pay, high stats. Was not admitted off the waitlist at other, lower-ranked schools. Not being admitted off the waitlist at the other schools does not imply anything about the merit of his app.

They are trying to do too much in this "study".

College-Specific Analysis Sample. When studying admissions and matriculation at specific colleges Section 3.2), admissions decisions (Section 3.3), and the causal effects of colleges on outcomes (Section 4), we focus on the subset of Ivy-Plus and flagship public colleges for which we have internal application and admissions data. In these analyses, we define the analysis sample as all permanent residents or citizens in the college-specific dataset who submitted a first-year undergraduate application to the college over the years for which we have data who (1) can be linked to the tax data based on their SSNs or ITINs and (2) can be linked to parents in the tax data.

Full pay students often do not include SSNs in the application because they are not applying for need-based aid. My kids did not.

The data come from 1098-T forms that colleges submit for all tuition paying students. For non-tuition paying students, it’s complemented by Pell Grant data. Read before you criticize. [/quote
1098-T forms are only for enrolled students. Students who applied, but did not attend, would not have their income included.

How does having a 1098-T form tell the researchers what the student's income level is? What are the mechanics by which personal income data is available to researchers, and should it be, both for those who receive need-based aid and for students who didn't submit financial aid forms? The university does not have a family's financial information for those who didn't apply for aid.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

We show that under the identification assumption that different college admission committees’ assessments of a candidate’s underlying merit (i.e., the component that predicts long-term outcomes) are positively correlated with each other, comparisons of students who are admitted vs. rejected from the waitlist can be used to identify the causal effect of admission for marginal applicants.

Using this design, we find that being admitted from the waitlist to an Ivy-Plus college increases students’ chances of achieving early career upper-tail success on both monetary and non-monetary dimensions. The causal effects of admission to an Ivy-Plus college are much larger for students with weaker fallback options– e.g., whose colleges in their home state channel fewer students to the top 1% after college. Exploiting this heterogeneity in treatment effects, we estimate that the marginal student who is admitted to and attends an Ivy-Plus college instead of the average flagship public is about 50% more likely to reach the top 1% of the income distribution at age 33, nearly twice as likely to attend a highly-ranked graduate school, and 2.5 times as likely to work at a prestigious firm.

Seems like the assumptions about admission from the waitlist are not warranted. Most waitlists are need-aware. Generally, waitlist acceptance is related to institutional priorities. I think they are reading way too much into that.

--my kid was admitted to an Ivy-Plus off the waitlist this year. full pay, high stats. Was not admitted off the waitlist at other, lower-ranked schools. Not being admitted off the waitlist at the other schools does not imply anything about the merit of his app.

They are trying to do too much in this "study".

College-Specific Analysis Sample. When studying admissions and matriculation at specific colleges Section 3.2), admissions decisions (Section 3.3), and the causal effects of colleges on outcomes (Section 4), we focus on the subset of Ivy-Plus and flagship public colleges for which we have internal application and admissions data. In these analyses, we define the analysis sample as all permanent residents or citizens in the college-specific dataset who submitted a first-year undergraduate application to the college over the years for which we have data who (1) can be linked to the tax data based on their SSNs or ITINs and (2) can be linked to parents in the tax data.

Full pay students often do not include SSNs in the application because they are not applying for need-based aid. My kids did not.

The data come from 1098-T forms that colleges submit for all tuition paying students. For non-tuition paying students, it’s complemented by Pell Grant data. Read before you criticize.

DP. Curious. 1098-T forms are only for enrolled students. Students who applied, but did not attend, would not have their income included.

How does having a 1098-T form tell the researchers what the student's income level is? What are the mechanics by which personal income data is available to researchers, and should it be, both for those who receive need-based aid and for students who didn't submit financial aid forms? The university does not have a family's financial information for those who didn't apply for aid.

[Anonymous]

I feel like Stanford has already run the numbers on this.

They're keeping legacy even though it's costing them state funds via cal grants.

They cut and then reinstated a couple non-revenue producing sports like sailing after alumni went nuts.

Either, someone has some dirt of Stanford trustees and they're especially effective. Or maybe they dont know how to run the numbers. But I suspect it's neither. Stanford is just a business and they care mostly about the money -- they dont care about the best and brightest. They'll take the money and close enough to best and brightest

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

We show that under the identification assumption that different college admission committees’ assessments of a candidate’s underlying merit (i.e., the component that predicts long-term outcomes) are positively correlated with each other, comparisons of students who are admitted vs. rejected from the waitlist can be used to identify the causal effect of admission for marginal applicants.

Using this design, we find that being admitted from the waitlist to an Ivy-Plus college increases students’ chances of achieving early career upper-tail success on both monetary and non-monetary dimensions. The causal effects of admission to an Ivy-Plus college are much larger for students with weaker fallback options– e.g., whose colleges in their home state channel fewer students to the top 1% after college. Exploiting this heterogeneity in treatment effects, we estimate that the marginal student who is admitted to and attends an Ivy-Plus college instead of the average flagship public is about 50% more likely to reach the top 1% of the income distribution at age 33, nearly twice as likely to attend a highly-ranked graduate school, and 2.5 times as likely to work at a prestigious firm.

Seems like the assumptions about admission from the waitlist are not warranted. Most waitlists are need-aware. Generally, waitlist acceptance is related to institutional priorities. I think they are reading way too much into that.

--my kid was admitted to an Ivy-Plus off the waitlist this year. full pay, high stats. Was not admitted off the waitlist at other, lower-ranked schools. Not being admitted off the waitlist at the other schools does not imply anything about the merit of his app.

They are trying to do too much in this "study".

College-Specific Analysis Sample. When studying admissions and matriculation at specific colleges Section 3.2), admissions decisions (Section 3.3), and the causal effects of colleges on outcomes (Section 4), we focus on the subset of Ivy-Plus and flagship public colleges for which we have internal application and admissions data. In these analyses, we define the analysis sample as all permanent residents or citizens in the college-specific dataset who submitted a first-year undergraduate application to the college over the years for which we have data who (1) can be linked to the tax data based on their SSNs or ITINs and (2) can be linked to parents in the tax data.

Full pay students often do not include SSNs in the application because they are not applying for need-based aid. My kids did not.

The data come from 1098-T forms that colleges submit for all tuition paying students. For non-tuition paying students, it’s complemented by Pell Grant data. Read before you criticize.

DP. Curious. 1098-T forms are only for enrolled students. Students who applied, but did not attend, would not have their income included.

How does having a 1098-T form tell the researchers what the student's income level is? What are the mechanics by which personal income data is available to researchers, and should it be, both for those who receive need-based aid and for students who didn't submit financial aid forms? The university does not have a family's financial information for those who didn't apply for aid.

“We obtain data on children's and parents incomes from income tax returus (1040 forms) and third-party
information returns (e.g., W-2 forms), which contain information on the earnings of those who do not file tax returns. We measure income in 2015 dollars, adjusting for inflation using the consumer price index (CPI-U).
Parental Income. Our primary measure of parental income is total household-level pre-tax income. In
years in which a child's parent files an income tax return, we define household income as the Adjusted Gross Income reported on the 1040 tax return. In years in which a parent does not file an income tax return, we define household income as the sum of wage earnings (reported on form W-2) and unemployment benefits (reported on form 1099-G) for all parents linked to a child. In years in which parents neither file tax returns nor receive information returns, household income is coded as zero. Chetty et al. (2020) show that these income definitions yield an income distribution similar to that in the American Community Survey (ACS) under the same income definitions.”

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

We show that under the identification assumption that different college admission committees’ assessments of a candidate’s underlying merit (i.e., the component that predicts long-term outcomes) are positively correlated with each other, comparisons of students who are admitted vs. rejected from the waitlist can be used to identify the causal effect of admission for marginal applicants.

Using this design, we find that being admitted from the waitlist to an Ivy-Plus college increases students’ chances of achieving early career upper-tail success on both monetary and non-monetary dimensions. The causal effects of admission to an Ivy-Plus college are much larger for students with weaker fallback options– e.g., whose colleges in their home state channel fewer students to the top 1% after college. Exploiting this heterogeneity in treatment effects, we estimate that the marginal student who is admitted to and attends an Ivy-Plus college instead of the average flagship public is about 50% more likely to reach the top 1% of the income distribution at age 33, nearly twice as likely to attend a highly-ranked graduate school, and 2.5 times as likely to work at a prestigious firm.

Seems like the assumptions about admission from the waitlist are not warranted. Most waitlists are need-aware. Generally, waitlist acceptance is related to institutional priorities. I think they are reading way too much into that.

--my kid was admitted to an Ivy-Plus off the waitlist this year. full pay, high stats. Was not admitted off the waitlist at other, lower-ranked schools. Not being admitted off the waitlist at the other schools does not imply anything about the merit of his app.

They are trying to do too much in this "study".

College-Specific Analysis Sample. When studying admissions and matriculation at specific colleges Section 3.2), admissions decisions (Section 3.3), and the causal effects of colleges on outcomes (Section 4), we focus on the subset of Ivy-Plus and flagship public colleges for which we have internal application and admissions data. In these analyses, we define the analysis sample as all permanent residents or citizens in the college-specific dataset who submitted a first-year undergraduate application to the college over the years for which we have data who (1) can be linked to the tax data based on their SSNs or ITINs and (2) can be linked to parents in the tax data.

Full pay students often do not include SSNs in the application because they are not applying for need-based aid. My kids did not.

The data come from 1098-T forms that colleges submit for all tuition paying students. For non-tuition paying students, it’s complemented by Pell Grant data. Read before you criticize.

DP. Curious. 1098-T forms are only for enrolled students. Students who applied, but did not attend, would not have their income included.

How does having a 1098-T form tell the researchers what the student's income level is? What are the mechanics by which personal income data is available to researchers, and should it be, both for those who receive need-based aid and for students who didn't submit financial aid forms? The university does not have a family's financial information for those who didn't apply for aid.

“We obtain data on children's and parents incomes from income tax returus (1040 forms) and third-party
information returns (e.g., W-2 forms), which contain information on the earnings of those who do not file tax returns. We measure income in 2015 dollars, adjusting for inflation using the consumer price index (CPI-U).
Parental Income. Our primary measure of parental income is total household-level pre-tax income. In
years in which a child's parent files an income tax return, we define household income as the Adjusted Gross Income reported on the 1040 tax return. In years in which a parent does not file an income tax return, we define household income as the sum of wage earnings (reported on form W-2) and unemployment benefits (reported on form 1099-G) for all parents linked to a child. In years in which parents neither file tax returns nor receive information returns, household income is coded as zero. Chetty et al. (2020) show that these income definitions yield an income distribution similar to that in the American Community Survey (ACS) under the same income definitions.”

How do the researchers have access to the tax forms? That is my question. This is not public information, nor do colleges have it for students who did not apply for aid.

(As a separate question, do families applying for aid agree to give access to their tax information to third parties? I guess I shouldn't be surprised if they do, but that doesn't say much for privacy.)

[Anonymous]

Anonymous wrote:I feel like Stanford has already run the numbers on this.

They're keeping legacy even though it's costing them state funds via cal grants.

They cut and then reinstated a couple non-revenue producing sports like sailing after alumni went nuts.

Either, someone has some dirt of Stanford trustees and they're especially effective. Or maybe they dont know how to run the numbers. But I suspect it's neither. Stanford is just a business and they care mostly about the money -- they dont care about the best and brightest. They'll take the money and close enough to best and brightest

To some extent this is true.

This also highlights how important it is for high stats unhooked kids to apply to 10+ reaches. Stanford may let you go, but Stunfool might just pick you up.

[Anonymous]

Anonymous wrote:Of course. But powerful people propping up these institutions WANT legacy admits to continue. Athletic recruits are a financial concern.

How is recruiting for the Harvard women's water polo team a financial concern? The entire audience is related to one of the players by blood or wants to by marriage.

[Anonymous]

Anonymous wrote:I’m having a hard time believing that eliminating athletic recruitment will change things unless they really mean eliminating most athletics.

Even if they tell coaches they can’t recruit, if they are going to field a squash team or a fencing team as examples, then they will still look for kids with squash and fencing championships who are going to come from the top 1%.

Without the recruiting preference, the demographic of squash and fencing team members would change.

[Anonymous]

Anonymous wrote:

Anonymous wrote:…to highly selective colleges (8 ivies, Stanford, MIT, UChicago, Duke)

https://opportunityinsights.org/wp-content/uploads/2023/07/CollegeAdmissions_Paper.pdf

The analysis shows that highest income applicants (top 1 percent) have an admission advantage over the average applicant due to: 1) legacy admissions; 2) athletic recruitment; 3) non-academic factors (e.g., private school extracurriculars). In fact, legacy admissions explain about half of the gap between acceptance rates between highest income and average applicants.

Also, attending IvyPlus colleges does improve earnings and leadership prospects after colleges. The authors do a nice job of identifying the causal effect of IvyPlus attendance.

Putting these findings together, the implication is that more socioeconomic diversity can be achieved without sacrificing academic quality by eliminating legacy admissions and athletic recruitment.

This is a great paper, and really interesting. Thanks for sharing! However, it does not say the bolded. They do not weigh in on "academic quality" of the student in college. The only measure they use is future earnings of the student. The problem with that measure is that I know many students from wealthy backgrounds who do not chase high-income careers, but choose public service, arts, or research careers instead. This does not mean they were not the engaged or intellectually curious students who contributed to academic quality of the college. It just means that they didn't earn as much later on.

1000%!

Also, there are plenty of kids from wealthy backgrounds that end up in "high income careers" simply because of their connections. connections that would happen whether they went to Salisbury or Harvard.

Many lower income students at T25 schools struggle, to fit in and take full advantage of the perceived perks of being at a T25. They are not spending spring break skiing in Europe or xmas break in Europe with their friends.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:…to highly selective colleges (8 ivies, Stanford, MIT, UChicago, Duke)

https://opportunityinsights.org/wp-content/uploads/2023/07/CollegeAdmissions_Paper.pdf

The analysis shows that highest income applicants (top 1 percent) have an admission advantage over the average applicant due to: 1) legacy admissions; 2) athletic recruitment; 3) non-academic factors (e.g., private school extracurriculars). In fact, legacy admissions explain about half of the gap between acceptance rates between highest income and average applicants.

Also, attending IvyPlus colleges does improve earnings and leadership prospects after colleges. The authors do a nice job of identifying the causal effect of IvyPlus attendance.

Putting these findings together, the implication is that more socioeconomic diversity can be achieved without sacrificing academic quality by eliminating legacy admissions and athletic recruitment.

This is a great paper, and really interesting. Thanks for sharing! However, it does not say the bolded. They do not weigh in on "academic quality" of the student in college. The only measure they use is future earnings of the student. The problem with that measure is that I know many students from wealthy backgrounds who do not chase high-income careers, but choose public service, arts, or research careers instead. This does not mean they were not the engaged or intellectually curious students who contributed to academic quality of the college. It just means that they didn't earn as much later on.

1000%!

Also, there are plenty of kids from wealthy backgrounds that end up in "high income careers" simply because of their connections. connections that would happen whether they went to Salisbury or Harvard.

Many lower income students at T25 schools struggle, to fit in and take full advantage of the perceived perks of being at a T25. They are not spending spring break skiing in Europe or xmas break in Europe with their friends.

The paper looks at other outcomes too, not just income. Public service, etc.

[Anonymous]

Anonymous wrote:

Anonymous wrote:I’m having a hard time believing that eliminating athletic recruitment will change things unless they really mean eliminating most athletics.

Even if they tell coaches they can’t recruit, if they are going to field a squash team or a fencing team as examples, then they will still look for kids with squash and fencing championships who are going to come from the top 1%.

Without the recruiting preference, the demographic of squash and fencing team members would change.

How? They would still recruit kids who played squash and fenced and would give added weight to champions much the same they give added weight to a piano player that wins prestigious awards or a Math Olympiad kid.

The squash and fencing champs would all most likely be wealthy.

[Anonymous]

Anonymous wrote:

Anonymous wrote:I wonder how much of this result is driven by the way they group the colleges.

They grouped colleges like this:
* “Ivy Plus”: Harvard, Yale, Princeton, Stanford, MIT, Brown, Penn, Duke, Chicago, Columbia, Cornell, Dartmouth
* “Other selective privates”: Northwestern, Hopkins, Georgetown, Rice, Vanderbilt, Notre Dame, CMU, WUSL, Emory, CalTech, NYU, USC
* “Flagships”: UCLA, Cal, Michigan, UT Austin, Florida, UGA, UNC, UVA, Ohio State

What would happen if you reorganized those lists?
* If you broke out HYPSM as their own group, would the rest of Group A still show different results from Group B?
* Would Group B improve if you removed the big urban schools (NYU and USC), which are obviously distinct from the others in many ways?
* How much would Group C improve if you replaced Ohio State with Georgia Tech?
* If you broke out the elite tech schools (MIT, CalTech, CMU, Ga. Tech) as their own group, would that group do better or worse than HYPS? Would it do better than Chicago?

The way the paper is presented seems designed to convince you that an anthropology major from Columbia is destined to earn more money than a premed from Hopkins or an aerospace engineer from Georgia Tech, but I’m dubious.

And why are some colleges even listed/included in study given already passed laws cannot do legacy:

California: In 2024, California passed a law banning private colleges from giving preference to applicants with alumni or donor connections.
Colorado: In 2021, Colorado was the first state to ban legacy preferences for all public colleges and universities.
Illinois: This state has outlawed the practice for public universities.
Maryland: This state became the first to pass a ban affecting both private and public colleges.
Virginia: In 2024, a bipartisan law passed that banned legacy and donor preference at all public colleges and universities.

The paper wasn't trying to determine the effect of legacy admissions. it was trying to identify determinants and causal effects of admission to highly selective colleges.
It wasn't just the legacy admissions' effects non admissions