Methodology FAQ #13 2026

Every year, over 6.4 million internationally mobile students enroll in higher education institutions worldwide, according to the UNESCO Institute for Statistics. At the same time, the International Education Ombudsman in Australia reported a 14% increase in formal complaints about misleading course information in 2025 alone. These figures underscore a hard truth: choosing a university is a high-stakes decision, and the data landscape is often fragmented.

This FAQ serves as our annual transparency report for 2026, detailing exactly how UniReview builds its institutional profiles. We do not rank universities. Instead, we surface structured comparisons across teaching quality, employability, research output, and student experience—so you can build your own decision framework without relying on a single composite score.

What Data Sources Power the 2026 Profiles?

Our methodology rests on a multi-source verification model. No single dataset determines an institution’s profile. We integrate five core categories of information, weighted by recency and jurisdictional authority.

First, we pull from government statistical agencies, including the Australian Department of Education’s QILT surveys, the UK’s Higher Education Statistics Agency (HESA), and the Integrated Postsecondary Education Data System (IPEDS) in the United States. These provide granular, student-level data on satisfaction, continuation rates, and graduate outcomes.

Second, we license bibliometric data from Scopus and Web of Science to map research output and citation impact. Third, we incorporate professional accreditation registers—such as ABET for engineering, AACSB for business, and the General Medical Council for clinical programs—to verify program-level quality. Fourth, we analyze student visa grant and rejection rates published by immigration authorities, including the UK Home Office, Australian Department of Home Affairs, and IRCC Canada. Finally, we conduct annual surveys with international student cohorts, generating over 12,000 responses in 2025.

How Are the Four Pillars Weighted?

Each institutional profile is structured around four pillars: Teaching Quality, Employability, Research Strength, and Student Experience. We do not assign a single overall score. Instead, each pillar carries its own weighting logic, transparently disclosed on every profile page.

Teaching Quality draws 40% of its signal from student satisfaction data (QILT SES, UK NSS, etc.), 30% from student-to-staff ratios, and 30% from continuation and completion rates. Employability leans 50% on graduate employment outcomes at 6 and 12 months, 30% on employer reputation surveys, and 20% on internship and work-placement density. Research Strength is driven by field-weighted citation impact (40%), research income per FTE academic (30%), and the proportion of staff holding doctoral degrees (30%). Student Experience combines international student support services, accommodation quality, and cost-of-living indices—all normalized against city-level benchmarks.

Why Don’t You Publish a Single Ranking?

Composite rankings often mask trade-offs that matter to individual students. A university with exceptional research output may have a poor student-to-staff ratio. An institution with strong graduate salaries might lag in student satisfaction. Collapsing these dimensions into one number forces a value judgment that should belong to the student, not the publisher.

We observed this problem acutely in 2024, when a major global ranking elevated several institutions due to a methodology shift favoring sustainability metrics, causing double-digit rank swings unrelated to teaching or employability. Our approach avoids such volatility. Each pillar stands alone, allowing users to filter by what matters to them—whether that’s post-study work visa eligibility, clinical placement hours, or research supervisor availability.

How Do You Handle Data Gaps and Small Sample Sizes?

Data gaps are inevitable, particularly for smaller institutions or specialized colleges. We apply a minimum reporting threshold: if a data point draws from a sample of fewer than 23 respondents, we suppress it and display “insufficient data.” This threshold aligns with the statistical standards used by the Australian Government’s QILT team.

For missing government datasets, we do not impute. Instead, we flag the gap and provide context—such as whether the institution is exempt from reporting or newly established. In 2026, we introduced a data freshness indicator on every profile, showing the collection year for each metric. Any figure older than three academic years is grayed out and labeled “for historical reference only.”

What Quality Controls Are in Place?

Every data point undergoes a three-stage validation pipeline. Stage one is automated: our ingestion scripts check for format consistency, outlier values exceeding three standard deviations, and cross-source mismatches. Stage two involves human review by a team of six research analysts, who manually verify flagged anomalies against primary sources. Stage three is institutional feedback—universities are invited to review their draft profiles annually and submit corrections with documentary evidence.

In 2025, this process caught 47 significant data errors before publication, including a misreported international student enrollment figure that would have overstated an institution’s diversity metric by 22%. We also maintain a public changelog, updated quarterly, documenting every material revision.

How Often Are Profiles Updated?

We operate on a continuous update cycle rather than an annual “release day.” Government datasets are refreshed within 48 hours of publication. Survey data updates quarterly. Bibliometric data rolls in on a six-month cycle. This means a profile viewed in March 2026 may look different from one viewed in September 2026—reflecting the most current evidence available.

FAQ

Q1: How is “employability” measured differently from graduate salary?

Employability is a multi-indicator construct. Graduate salary contributes, but we also weight employment rate at 6 and 12 months post-graduation, employer reputation survey scores, and the proportion of programs with mandatory work-integrated learning. A university with moderate salaries but a 94% employment rate can score higher on employability than one with high salaries but a 72% rate.

Q2: Do you adjust for cost of living in different cities?

Yes. The Student Experience pillar incorporates city-level cost-of-living indices sourced from Numbeo and government consumer price data. Accommodation costs, transport, and typical food expenses are normalized so that a university in London and one in Sheffield can be compared on a like-for-like basis. This adjustment accounts for roughly 15% of the Student Experience pillar weight.

Q3: Can a university request a correction to its profile?

Absolutely. Every institution receives a 30-day review window annually. Corrections require documentary evidence—such as audited enrollment figures or regulatory filings. We log all accepted corrections in our public changelog. In 2025, 83% of requested corrections were accepted; the remainder were rejected due to insufficient evidence or conflict with government data.

Q4: What happens if an institution refuses to participate?

Non-participation does not block a profile. We rely on publicly available, authoritative data—government statistics, regulatory filings, and bibliometric databases—which do not require institutional consent. We note non-participation on the profile so users can interpret the data with full context. Approximately 12% of profiled institutions did not respond to our review invitation in 2025.

参考资料

• UNESCO Institute for Statistics 2025 Global Education Digest
• Australian Department of Education 2025 QILT Student Experience Survey
• UK Higher Education Statistics Agency 2025 Student Record
• Scopus 2025 Bibliometric Database
• International Education Ombudsman Australia 2025 Annual Report