RESEARCHER INTELLIGENCE METHODOLOGY v1 0

# HE Researcher-Level Institutional Intelligence — Methodology v1.0

**Status:** Dashboard intelligence layer; non-scoring
**Core ranking:** HE Open Data Ranking 2027 v0.10.6 / Methodology v1.0
**Researcher module:** v1.0

## 1. Analytical purpose
The researcher layer provides institution-level research intelligence by identifying leading contributors associated with each university and measuring the impact of research that can be attributed to that researcher at that institution.

It is not designed as a standalone global ranking of scientists.

## 2. Candidate association
A researcher becomes a candidate for an institution when OpenAlex lists that institution in `last_known_institutions`.

This is interpreted as the most recently observed bibliographic institutional association, not as a verified employment contract.

## 3. Exact institutional attribution
For each candidate author, HE retrieves or scans works where both the candidate author and institution are present. A work is attributed to the author–institution pair only when the author and institution occur in the **same authorship object**.

This prevents false attribution caused by a work containing the target author and the target institution in different authorships.

## 4. Institution-specific measures
For each eligible author–institution pair HE calculates:

1. **Affiliated Publications** — number of exactly attributed works.
2. **Affiliated Citations** — total current `cited_by_count` of those works.
3. **Affiliated h-index** — h-index calculated only from those attributed works.
4. **Affiliated i10-index** — number of attributed works with at least 10 citations.
5. **Recent Institutional Citations** — citations recorded during 2021–2025 to the attributed works, using OpenAlex `counts_by_year`.

## 5. HE Institutional Research Contribution Score (HE-IRCS)
Each component is transformed to a midrank percentile (0–100) within the collected HE Research Contributor Reference Population.

**HE-IRCS =**

`0.40 × Affiliated Citation Percentile`
`+ 0.25 × Affiliated h-index Percentile`
`+ 0.15 × Affiliated i10 Percentile`
`+ 0.10 × Affiliated Publication Percentile`
`+ 0.10 × Recent Institutional Citation Percentile`

HE-IRCS is reported on a 0–100 scale.

## 6. Ordering rule
Researchers are ordered **within their institution only** by HE-IRCS. Ties are resolved using, in order:

1. affiliated citations;
2. affiliated h-index;
3. recent institutional citations;
4. researcher display name.

HE does not assign an HE world scientist rank, country scientist rank, continental scientist rank, or field scientist rank.

## 7. Eligibility and list size
Default eligibility requires at least **3 exactly institution-affiliated works**.

Up to 100 researchers are retained per university. Institutions with fewer than 100 eligible researchers display only the available researchers.

## 8. Candidate prefilter
To keep production computationally tractable, HE may preselect the highest-cited current-associated OpenAlex authors before exact institution-specific calculation.

The default cap is 500 per institution in the module; higher values may be used in production if computational resources allow. The prefilter is disclosed because it can affect recall at the lower edge of the Top 100.

## 9. Lifetime metrics
OpenAlex lifetime citations, lifetime h-index, lifetime i10-index, and lifetime works are retained for context but are not substituted for institution-attributed HE-IRCS components.

## 10. Data limitations
- OpenAlex author disambiguation is algorithmic and may merge or split author identities.
- `last_known_institutions` is a bibliographic association and can lag real-world employment changes.
- Citation practices differ substantially by discipline and career age.
- Institution-attributable measures depend on the completeness of authorship affiliations in OpenAlex.
- A candidate limit can omit a lower-lifetime-citation researcher who would otherwise score highly on institution-specific contribution.

## 11. Ranking independence
Researcher intelligence is explanatory only. It has **zero weight** in the frozen HE Open Data Ranking 2027 score and cannot alter a university's official ranking position.