How We Classified

The classification begins with the full universe of American higher education: approximately 7,000 degree-granting institutions in IPEDS. From that universe, 1,607 colleges meet the baseline criteria: granting bachelor's degrees or higher, not open-admission, public or private non-profit. Of those 1,607, 121 institutions had undergraduate admit rates below 30% in either 2023 or 2024. 31 are excluded under the rules listed with the classification, among them military academies and institutions with anomalous single-year spikes. The remaining 90 form the list. The classification then determines, for each institution, whether its low admit rate reflects deliberate behavioral choices that concentrate wealth and engineer exclusivity, or genuine demand for a limited number of seats.

Why the behavioral requirement matters

A school can enroll few low-income students and charge them high prices because it lacks the resources to do otherwise. That is a problem, but it is not the same problem as a school that actively cultivates donor relationships, games rankings systems, or structures its admissions calendar to advantage wealthy applicants. The behavioral categories capture intent; the outcome categories capture result. Both matter, but classification as Highly Rejective requires evidence of at least one behavioral choice.

On the wealth preference ratio

The ratio uses Opportunity Insights (2023) unweighted relative attendance data: how many times more likely is a student from the top 1% of family income to attend a given institution than a student from the bottom 60%, with no adjustment for test scores. The unweighted measure is used deliberately. Controlling for test scores would assume that standardized test performance is a neutral signal of academic preparation, independent of family income. Children of the wealthiest 1% of Americans are 13 times more likely to score above 1300 on the SAT than children from low-income families. Adjusting for scores would remove from the measurement much of what is being measured. Opportunity Insights →

Eight schools are not in the Opportunity Insights 139-college dataset: Colorado College, Pitzer College, Tulane University, Denison University, Grinnell College, Babson College, Harvey Mudd College, and Rhode Island School of Design. For these schools, wealth concentration is assessed using Pell Grant percentage only. These schools are marked with (~) in the full data view.

On the adjusted net price

Raw net price is adjusted by the institution's tuition dependence ratio: raw net price × (tuition dependence / 50%). A school that derives 75% of revenue from tuition and charges $10,000 to low-income families is treated differently from a school that derives 20% of revenue from tuition and charges the same amount. The 50% normalization sets the midpoint: schools with more non-tuition resources (endowment returns, research revenue, gifts) are expected to use them to reduce cost for low-income students.

Tuition dependence is tuition and fees as a share of the revenue a school has to work with, counting 5% of its endowment each year in place of that year's investment return.

On Test Scores

Test policy is tracked as contextual evidence, not a standalone classification factor. High test expectations alone do not determine rejectivity. Caltech, for example, also enrolls exceptionally high-scoring students, but explicitly acknowledges a principle many ranking systems ignore: precision at the extreme high end can be overstated. Its admissions process groups top SAT scores into evaluation buckets rather than treating every marginal point difference as meaningfully distinct. MIT's admitted math profile clusters in that same wealth-correlated edge of the testing distribution.