The University Index — The Prestige–Compassion Gap

Every prominent university ranking measures roughly the same thing. Times Higher Education, QS, and the Academic Ranking of World Universities (Shanghai) all score research output, citations, faculty reputation, employer surveys, and internationalization. They are prestige instruments. They tell a prospective student which institution is most admired by other academics and employers — and they tell that student nothing at all about how the institution treats the humans inside it.

The Compassion Benchmark's University Index — the institution's 8th index, joining countries, the Fortune 500, AI labs, robotics labs, US states, US cities, and global cities — measures the thing the prestige tables omit. It scores the top 100 universities worldwide on the same eight dimensions and the same 0–100 composite used across the benchmark, asking a single question: does this university recognize, respond to, and reduce suffering among the students and staff who depend on it?

That is a different question from "is this a great university," and it produces a different answer. A university can rank first in the world for research while running an active Title IX investigation, while its graduate workers pay more than half their stipend in rent, while adjunct faculty earn near-poverty wages, and while whistleblower complaints go unanswered. The prestige rankings cannot see any of that. This index is built to.

The 100 institutions were selected by composite of the three most-cited global prestige rankings — so inclusion is defensible to any institution ("you appear in the top tiers of THE, QS, or Shanghai") and is never paid for. No university pays for inclusion, exclusion, or score reconsideration; no university received its score before publication; no score was changed in response to outreach or pressure. The index interprets public evidence — counseling-center reports, union contracts, regulator findings, investigative journalism, institutional disclosures — and nothing else.

The headline, stated plainly: the most admissions-competitive universities score mediocre on compassion. Selectivity is not compassion. The rest of this briefing shows what the score actually measures, what earns a high one, and what the flat middle of the table reveals.

The benchmark scores eight dimensions, each on a 1–5 scale, averaged and adjusted into the 0–100 composite. Each dimension is a distinct facet of how an institution behaves toward the people who depend on it. Here is what each one asks in general — and, crucially, what it asks of a university specifically.

Awareness (AWR) — does the institution detect distress before it becomes a crisis?

For a university, this is the mental-health early-warning system: counseling-center capacity and triage, student-to-counselor ratios, wellbeing surveys, and whether monitoring reaches the people most often missed — international students, graduate workers, adjuncts, and first-generation students. Georgia Tech's Zero-Suicide-based "Tech Ends Suicide Together" model and the University of Toronto's annual public mental-health report with a 24/7 multilingual crisis line are strong Awareness signals; detection that surfaces only after a cluster of student deaths is a weak one.

Empathy (EMP) — do people feel genuinely cared about, not merely processed?

For a university, this is whether students and staff from non-dominant groups describe feeling seen, whether an independent ombudsperson exists, and whether the institution acknowledges emotional harm before legal review forces it to. This is the single weakest dimension across the index — most institutions sit at 2.5–3.0 — reflecting how often the felt experience of care is uneven at scale.

Action (ACT) — does identified need produce timely, proportional, effective help?

For a university, this is counseling wait-times, the speed and reach of emergency financial aid, and whether mental-health staffing actually grew to match documented need rather than lagging it. The recurring tension is between services that exist on paper and services that respond fast enough at peak demand.

Equity (EQU) — is care distributed fairly, especially toward those with the least power?

For a university, this is the access-and-completion picture: Pell Grant or equivalent low-income enrollment share, first-generation completion rates, and the treatment of contingent labor — adjunct pay relative to tenure-track, and graduate-worker compensation. It also reaches historical harm: whether the institution acknowledges and repairs its history with excluded groups. Equity is, notably, where the top of this table is strongest — the inverse of the pattern in most of the benchmark's other indexes.

Boundaries (BND) — is helping sustainable and non-exploitative?

For a university, this is the dimension most about workers: graduate-student stipends against local cost of living, health coverage and workload caps in teaching-assistant agreements, and honest consent for prospective and international students about the full cost of attendance. Boundaries is one of the two dimensions that most consistently drags universities down. Across the index, graduate-worker and adjunct precarity is near-universal, and the institutions that improved it usually did so only after a strike.

Accountability (ACC) — does the institution own its failures and make genuine repair?

For a university, this is the Title IX / Clery / OCR record (or the jurisdiction's equivalent), and — more importantly — whether the institution self-disclosed and made substantial repair, or only acted under external compulsion. A history of resolution agreements is not itself penalized; the score turns on whether the university acknowledged harm before it was forced to. This is the second dimension that most often holds universities back.

Systemic Thinking (SYS) — does compassion reach root causes, not just symptoms?

For a university, this is multi-year structural investment in student and worker wellbeing (not just crisis response), and willingness to advocate publicly — at institutional cost — for changes in higher-education access, funding, or student protections. Glasgow's and Utrecht's high marks here come from genuine structural reckoning: a decades-long reparations programme, a values-based overhaul of fossil-fuel partnerships.

Integrity (INT) — is the institution's stated commitment to welfare genuine and consistent under pressure?

For a university, this is the endowment-versus-values test (holdings in sectors the institution's stated values oppose), the administrator-compensation-versus-student-resource trend, and whether welfare commitments hold when there is no reputational upside. Integrity is where several elite names score lowest — values acknowledged mainly under external pressure, donor contradictions, or governance failures surfaced by outsiders.

A reader who finishes this section should be able to look at any score in the table and know what it is claiming: not how prestigious the institution is, but how it treats — and can be shown to treat — the people inside it.

Only three institutions reach the Established band (60–80). None reaches Exemplary. The five highest-scoring institutions did not earn their place through selectivity or fame — three of the five are public universities, and the single highest is a Scottish public university that no prestige table places near the global top. What they share is specific, evidenced, often costly action on the dimensions that the prestige rankings cannot see.

Several things are worth naming about this top tier.

First, the lever is action, not exclusivity. Glasgow's lead comes from committing real money to reparations over two decades with no near-term reputational necessity — a non-performative Integrity signal (3.5) and a Systemic Thinking signal (4.0) that almost no peer matches. Brown earns its place by being the most balanced institution in the index: it has no dimension below 3.25, the only institution in the top five of which that is true. UC Irvine earns its place on a single, overwhelming strength — access and social mobility — that the prestige tables actively penalize as a lack of "selectivity."

Second, even the leaders are not exemplary. The top institution scores 62.5, barely into the Established band. Every one of the top five carries a visible weakness, most often Boundaries — Glasgow (2.75) and UC Irvine (2.75) both sit in the bottom half of the scale on how they treat contingent and graduate labor, reflecting sector-wide casualisation and stipend pressure that even leaders have not solved.

Third, the high scorers are geographically and structurally diverse — a UK public university, two private US universities, one public US campus, one Dutch public university. There is no single "type" of compassionate university. There is only evidence of specific conduct.

The five lowest-ranked institutions are all in Mainland China and Hong Kong:

This is the part of the table that requires the most careful reading, because two very different things are happening at once, and they must not be conflated.

The first is documented constraint. The University of Hong Kong is scored at high confidence — the evidence is substantial — and it scores low (29.7) because the record shows a genuine collapse of conduct toward dissenting members of its community: the dissolution of the elected student union that was students' primary representative voice, the midnight removal of the Pillar of Shame without due process, mandatory national-security education replacing commemoration, and a documented decline in academic freedom under the National Security Law (Human Rights Watch, 2024). Its lowest dimensions — Accountability, Systemic Thinking, and Integrity (all 1.75) — reflect a record in which stated academic-freedom values were contradicted by self-censorship under political pressure. This is a high-confidence, evidence-backed low score.

The second is thin evidence, scored conservatively. The four Mainland-China institutions in the bottom five — USTC, Fudan, Zhejiang, and Shanghai Jiao Tong — are scored at low confidence, and the briefing is emphatic about what that means. For these institutions, fewer than five of the eight dimensions have substantive public English-language evidence; much of the documentary record exists only in Chinese, and disaggregated institutional data (counseling wait-times, completion gaps, labor conditions) is not openly published. Where evidence is absent, these institutions were scored conservatively toward the middle of the scale rather than assigned a confident verdict, and per the index's own rule no low-confidence institution may be scored above composite 60.

A low-confidence score reflects an evidence gap, not a proven failure. Shanghai Jiao Tong's index-lowest 22.7 rests partly on documented and serious signals — student suicides reportedly tied to supervisor exploitation, and a system in which accountability arrived through state/ministry intervention rather than transparent institutional self-disclosure — and partly on the simple absence of the public evidence that would let the benchmark score it any more precisely. The honest position, stated on the index and repeated here, is that these are the institutions the benchmark can see least clearly, and their scores carry that caveat openly rather than hiding it.

This distinction is the whole point. The benchmark does not penalize an institution for being hard to observe, and it does not dress an evidence gap up as a confident indictment. It says, plainly, where the record is thin — and it scores conservatively when it cannot see.

The most striking single fact about the University Index is not at the top or the bottom. It is the middle. The field is compressed: 0 Exemplary, 3 Established, 76 Functional, 21 Developing, 0 Critical, with a mean of 46.2 and a median of 46.9 — both sitting almost exactly at the center of the 0–100 scale. Universities, as a category, are neither failing nor leading. They are middling.

And the most famous, most selective names sit squarely in that middle:

Every institution that anchors the global prestige tables lands in the middle of the Functional band here — competent, with significant gaps remaining, and behind a Scottish public university, two public US campuses, and a Dutch public university. This is the prestige gap: the rank that admissions competitiveness buys on a QS or THE table does not transfer to how an institution treats the people inside it.

The mechanism is visible in the dimension profiles. The elite private US universities tend to score well on Equity (Harvard, Stanford, MIT all at 3.75 — generous, well-funded financial aid is something wealthy institutions can buy) and poorly on Integrity and Accountability. Harvard's Integrity score of 2.5 is among the lowest in the top half of the table, reflecting donor-values contradictions and a 2023–24 governance crisis handled reactively under external pressure. Stanford's Accountability and Integrity (both held down) reflect a presidential resignation over research-integrity handling that came only after external investigative pressure. The recurring pattern across the elite tier: strong on what money can provide, weak on what only institutional character can provide — owning failures before being forced to, and holding values when there is no reputational upside.

The other half of the prestige gap is Boundaries. The single most common story in the index is an institution that improved its treatment of graduate workers or adjuncts only after a strike, a lawsuit, or sustained organizing: Yale reformed its mental-health leave policy after a 2023 disability-discrimination settlement and recognized its graduate union only after a decades-long fight; Boston University's grad workers won their raise after the longest US higher-ed strike in a decade; UC campuses raised stipends only after the largest higher-ed strike in US history exposed grad workers paying more than half their income on rent. Compassion that arrives only under pressure scores lower than compassion that is structural — and across this index, worker-facing compassion is overwhelmingly the former.

What the prestige gap means for a reader is concrete. A prospective graduate student choosing between a top-5-ranked private university and UC Irvine is, on the prestige tables, making an easy call. On the question of how the institution will actually treat them — access, fairness, labor conditions, whether their distress will be detected and answered — the call is very different, and in several cases inverted. That inversion is the reason this index exists.

• What would lift the field. The ceiling is low — nothing reaches Exemplary — because the dimensions that cap universities (Boundaries and Accountability) require structural change, not spending. The fastest plausible upward movement would come from an institution that pairs the elite tier's financial-aid generosity with Glasgow's structural reparative work and Brown's enforceable labor protections. No single institution currently does all three; the one that does would clear the Established band decisively.
• The institutions on the cusp. Utrecht (59.4) and Princeton (57.8) sit just below the Established threshold; a single verified improvement on a weak dimension — Princeton's Integrity or Accountability, Utrecht's still-medium evidence base — would move either into the top tier. They are the natural ones to watch.
• The labor signal. Because Boundaries is so consistently the drag, the index is unusually sensitive to graduate-worker and adjunct organizing. The wave of grad-union contracts since 2023 (Brown, MIT, Stanford, Cornell, Johns Hopkins, Columbia, and others) is the clearest force currently moving university scores upward — and the institutions that resist it longest are the ones whose Boundaries and Accountability scores will lag.
• The evidence frontier. The low-confidence cluster at the bottom of the table is the index's largest open question. As more disaggregated institutional data becomes available in English — or does not — those scores should be revisited. Movement there will reflect evidence availability as much as institutional change, and the briefing's central discipline (distinguishing thin evidence from proven harm) should hold through any such revision.
• What would falsify the prestige-gap thesis. The thesis weakens if a cohort of the most-selective institutions begins to lead on conduct — if elite selectivity and high compassion scores start to coincide. The current record shows the opposite, but the grad-union wave gives the elite private tier a concrete path to close part of the gap on Boundaries specifically. Whether they take it is the pattern's natural test.

• Canonical scores (ground truth): site/src/data/indexes/universities.json — the 100-institution roster, every composite, band, dimension vector, and confidence flag cited above (Glasgow 62.5, Brown 60.2, UC Irvine 60.2, Utrecht 59.4, Princeton 57.8; Harvard 52.3, Stanford 54.7, MIT 56.3, Oxford/Cambridge 53.1; University of Hong Kong 29.7, USTC 29.7, Fudan 28.1, Zhejiang 28.1, Shanghai Jiao Tong 22.7; band distribution 0/3/76/21/0; mean 46.2, median 46.9) reconcile exactly with the index data.
• Per-institution scoring rationale (the "why"): research/university-index/scores-batch-{1,2,3,4,5}.json — the dimension-by-dimension evidence notes and cited public sources for every institution discussed, including Glasgow's reparations programme, Brown's grad-worker contract and Slavery & Justice report, UC Irvine's first-gen/social-mobility record, Utrecht's social-safety office and fossil-fuel policy correction, Princeton's no-loan aid and leave-policy reform, the elite-tier Integrity/Accountability weaknesses (Harvard, Stanford, MIT), the grad-labor strike narratives (Yale, Boston University, UC), and the conservative low-confidence scoring of the China/Hong Kong cluster.
• Framework / dimensions: site/src/data/dimensions.ts (the eight dimensions, their subdimensions and anchor ladders, and the canonical band vocabulary) and docs/PRDUNIVERSITYINDEX.md §5 (the university-specific signal map: AWR = mental-health early-warning; EQU = access/Pell + completion gaps + adjunct pay; ACC = Title IX/Clery/OCR; BND = grad-student/adjunct labor; INT = endowment-vs-values), which this briefing's §2 summarizes for a general reader.
• Independence / method: docs/PRDUNIVERSITYINDEX.md §3 and §8 (entity selection by composite of three public prestige rankings; the confidence-flag logic; the permanent prohibition on paid inclusion, pre-publication review, or score changes under pressure).
• External evidence cited in the scoring record (not re-fetched here): Human Rights Watch, "We Can't Write the Truth Anymore: Academic Freedom in Hong Kong Under the National Security Law" (2024) <https://www.hrw.org/report/2024/09/24/we-cant-write-truth-anymore/academic-freedom-hong-kong-under-national-security-0>; University of Glasgow reparatory-justice programme <https://www.gla.ac.uk/postgraduate/taught/reparatory-justice/>; AFT, "Brown University Graduate Workers Win Union Contract, First in Ivy League" <https://www.aft.org/press-release/brown-university-graduate-workers-win-union-contract-first-ivy-league>; UC Irvine first-generation graduation data <https://news.uci.edu/2024/06/10/49-percent-of-uc-irvine-bachelors-degrees-earned-by-first-generation-students/>. No web-search budget was drawn down for this briefing; all external links are those already on record in the per-institution scoring files.