国产第一福利影院草草

Here is a thought experiment. Take two high school seniors, both with SAT scores around 1480. One comes from a family earning $30,000 a year. The other comes from a family in the top one-tenth of one percent, earning several million. Both score identically on the test that elite colleges use as their primary academic filter. Under a system that rewards merit and ignores background, they should attend selective colleges at roughly the same rate.

They do not. Not even close. According to a landmark dataset from economists Raj Chetty, David Deming, and John Friedman, which links tax records, SAT scores, and college enrollment for millions of American students: the student from the top 0.1 percent attends an Ivy Plus college at 2.5 times the rate of an equally-qualified student from the bottom fifth of the income distribution. At other highly selective private colleges, the ratio is nearly 4 to 1.

This analysis uses the Pipeline dataset, which tracks application, admission, and enrollment patterns across 139 of America鈥檚 most selective colleges, broken down by 13 income bins ranging from the bottom 20 percent to the top 0.1 percent. Every rate in this dataset is adjusted for standardized test scores, which means the gaps we observe cannot be explained by academic preparation. They reflect something else entirely: the accumulated advantages of wealth at every stage of the admissions funnel.

The Hockey Stick: Where Income Meets Access

When you plot the relative attendance rate at elite colleges against family income, a distinctive shape emerges. From the bottom quintile through the upper-middle class, roughly the 60th to 90th percentiles, the curve is essentially flat or gently rising. Students in this broad middle attend elite colleges at or slightly below the rate their test scores would predict. Then, somewhere around the 95th percentile, the curve begins to rise. And above the 99th, it explodes.

Figure 1: The Hockey Stick

Figure 1: Relative attendance rate at four types of selective college by parent income percentile. All rates are test-score adjusted. A value of 1.0 means income has no effect on attendance. Source: Chetty, Deming & Friedman (2023).

The pattern varies dramatically by institutional type. At the Ivy Plus, the eight Ivy League schools plus Stanford, MIT, Duke, and Chicago, students from the top 0.1 percent attend at 2.5 times the expected rate, while bottom-quintile students attend at 0.98 times. At other elite private colleges, the curve is even steeper: top-0.1-percent students attend at nearly 4.0 times the expected rate, while bottom-quintile students attend at just 0.59 times.

But there is one striking exception. At highly selective public universities, flagship state schools like Berkeley, Michigan, and Virginia, the curve inverts. Bottom-quintile students slightly over-attend (1.02 times expected), while top-0.1-percent students dramatically under-attend (0.67 times). The richest students are apparently choosing private institutions over even the strongest public universities, creating an unusual fairness dividend at selective publics.

At other highly selective privates, students from the top 0.1% of family income attend at nearly 4脳 the expected rate. At highly selective publics, the richest students under-attend. Same test scores, opposite outcomes.

Where the Gap Opens: Application vs. Admission

The genius of the Pipeline dataset is that it decomposes the overall attendance gap into two stages: application (do you send your scores and apply?) and admission conditional on application (once you apply, do you get in and enroll?). This decomposition reveals where wealth actually exerts its influence.

Figure 2: The Ivy Plus Pipeline

Figure 2: Relative application rate, admission+yield rate, and overall attendance rate at Ivy Plus colleges by parent income. All rates are test-score adjusted. The gap between the application line and the attendance line represents the admission advantage. Source: Chetty, Deming & Friedman (2023).

At Ivy Plus colleges, the application rate is surprisingly flat across the income distribution. Bottom-quintile students actually apply at a slightly higher rate than expected (1.04 times), and even the top 0.1 percent applies at only 1.20 times the expected rate. The application gap is modest. The admission gap is where wealth provides its decisive advantage.

The relative admission-plus-yield rate, which captures both whether a student is admitted and whether they enroll, tells a starkly different story. Bottom-quintile students who apply are admitted and enroll at 0.95 times the expected rate: slightly below parity. Students from the 60th to 90th income percentiles fare similarly. But above the 97th percentile, the admission advantage escalates sharply. Students from the 98th鈥�99th percentile are admitted and enroll at 1.15 times the expected rate. At the 99th鈥�99.9th percentile, it reaches 1.39 times. And at the very top, the top 0.1 percent, the admission-plus-yield advantage is a striking 2.04 times.

In plain terms: a student from the richest one-thousandth of American families who applies to an Ivy Plus college is admitted and enrolls at roughly double the rate of an equally-qualified student from the bottom fifth. The application pipeline is essentially open. The admission pipeline is where the velvet rope operates.

What drives this admission advantage? The researchers identify three primary mechanisms operating at Ivy Plus schools. Legacy preferences boost attendance by wealthy applicants, since high-income families are more likely to have alumni parents. Athletic recruitment, particularly in sports like rowing, fencing, sailing, and lacrosse, disproportionately draws from wealthy families who can afford years of elite coaching. And nonacademic credentials, including leadership roles, extracurricular achievement, recommendation letters from well-connected adults, tend to be inflated by the resources available to wealthy families, even when underlying academic ability is identical.

The Ivy Plus Scorecard: Not All Elites Are Equal

The Pipeline dataset provides enough granularity to compare individual Ivy Plus schools, revealing remarkable variation within the most elite tier of American higher education. Some schools show large wealth advantages in attendance; others have effectively neutralized the effect of family income.

Figure 3: The Ivy Plus Scorecard

Figure 3: Relative attendance rate for bottom-20% vs. top-1% students at each Ivy Plus school, test-score adjusted. Stars indicate schools where bottom-20% students are over-represented relative to top-1%. Source: Chetty, Deming & Friedman (2023).

Dartmouth shows the largest wealth gap: top-one-percent students attend at 2.59 times expected, while bottom-quintile students attend at just 0.72 times, a 3.6-to-1 ratio between the richest and poorest equally-qualified applicants. Duke (2.39 versus 0.62) and Penn (2.27 versus 0.67) follow closely.

At the other end of the spectrum, Harvard and MIT stand out as the only two Ivy Plus schools where bottom-quintile students are over-represented relative to top-one-percent students after controlling for test scores. At Harvard, bottom-quintile students attend at 1.78 times expected, higher than the 1.54 times for top-one-percent students. At MIT, the pattern is even more dramatic: 1.18 times for the bottom quintile versus 0.95 times for the top one percent. MIT is the only Ivy Plus school where the wealthiest students actually under-attend.

These differences are not random. Harvard and MIT have both invested heavily in need-blind admissions, aggressive financial aid for low-income students, and outreach programs targeting high-achieving students from underrepresented backgrounds. MIT鈥檚 admissions process, which is famously focused on STEM aptitude and problem-solving rather than extracurricular polish, may also be less susceptible to the nonacademic credential inflation that benefits wealthy applicants elsewhere. The results suggest that institutional choices about admissions philosophy can meaningfully counteract the structural advantages of wealth.

Dartmouth: top-1% students attend at 3.6脳 the rate of equally-qualified bottom-20% students. MIT: bottom-20% students actually over-attend. Same league, different universes.

The Public University Counterpoint

While the elite private colleges show pronounced wealth advantages, highly selective public universities tell a fundamentally different story. At these institutions, schools like UC Berkeley, the University of Michigan, the University of Virginia, and UNC Chapel Hill, bottom-quintile students attend at 1.02 times the expected rate, essentially at parity. More remarkably, students from the top one percent attend at just 0.88 times expected, and top-0.1-percent students at a strikingly low 0.67 times.

This inversion has a straightforward explanation: the wealthiest families are choosing private elite schools over even the best publics. But the consequence is that selective publics end up being the fairest tier in the entire system: the only institutional category where income, after controlling for test scores, does not predict attendance. For a low-income student with strong test scores, a flagship state university may offer not only comparable academic quality but a far more level playing field.

This finding connects directly to the mobility rate analysis from earlier in this series. Highly selective publics have the highest median mobility rate of any tier (2.23 percent), combining meaningful bottom-quintile enrollment with strong success rates. They are not only fairer in admissions; they are more effective at translating that fairness into economic mobility.

What This Means for Students and Families

The Pipeline data carries different implications depending on where a student sits in the income distribution.

For high-achieving low-income students, the data contains both a warning and an encouragement. The warning: the admissions system at many elite private colleges embeds structural advantages for wealthy applicants that test-score parity does not overcome. Legacy status, expensive extracurriculars, and polished application materials create headwinds that are difficult to see from the outside. The encouragement: the application rates suggest that information and aspiration are not the primary barriers: low-income students with strong scores are applying to these schools. And at specific institutions, particularly Harvard, MIT, and the selective publics, the admissions process appears to have genuinely neutralized wealth advantages.

For middle-class and upper-middle-class families, roughly the 60th to 95th income percentiles, the data is arguably the most sobering. These students are the most under-represented at elite colleges relative to their test scores. They lack both the financial aid eligibility that supports the poorest applicants and the legacy and donor advantages that boost the wealthiest. At Ivy Plus colleges, students from the 60th to 80th income percentiles attend at just 0.78 times the expected rate, the lowest of any group.

For the wealthiest families, the data confirms what many have long suspected: immense wealth translates into admissions advantages at the most selective private colleges even after academic ability is held constant. Whether this represents the intended design of these institutions or a systemic failure of their admissions processes is a question the data raises but cannot resolve.

Caveats

Several limitations merit attention. The Pipeline data covers SAT and ACT takers from 2011, 2013, and 2015 cohorts, and the admissions landscape has shifted since then, most notably with the widespread adoption of test-optional policies following the COVID-19 pandemic. Whether test-optional admissions have reduced or amplified the wealth advantage is an open question that this data cannot answer.

The test-score adjustment, while powerful, controls only for SAT and ACT performance. It does not account for differences in GPA, course rigor, or other academic metrics that may vary with income. If wealthy students have systematically higher GPAs conditional on the same test score, perhaps because of grade inflation at elite high schools, and some portion of the admission advantage may reflect academic factors the data cannot observe.

Finally, the admission-plus-yield measure combines two distinct processes: being admitted and choosing to enroll. Wealthy students may be more likely to enroll at elite schools conditional on admission because they face fewer financial barriers. Some of the 鈥渁dmission advantage鈥� may therefore be a 鈥測ield advantage鈥� driven by affordability rather than admissions preference.

The Bottom Line: Merit Has a Price Tag

The Pipeline data makes visible a truth that American higher education has long preferred to leave unspoken. The admissions systems of the nation鈥檚 most selective colleges are not meritocracies in any straightforward sense. They are systems in which merit, as measured by standardized tests, is a necessary but insufficient condition for access, and in which wealth provides advantages at every stage, from the extracurriculars that pad an application to the legacy connections that tip an admission decision to the financial security that makes enrollment possible.

The most important finding may be the variation. Not all elite colleges are equally skewed. Harvard and MIT have demonstrably flattened the wealth curve. Highly selective publics have inverted it entirely. These institutions prove that the wealth advantage in admissions is not a law of nature. It is a policy choice. And policy choices can be changed.