Researcher-level metrics are where decisions get personal — hiring, promotion, recognition, and recruitment all draw on them. They are also where bad data does the most damage: a misattributed paper or a missing identifier can distort a researcher's measured performance in either direction. This article explains the two researcher-level metrics most commonly used — the h-index and Field-Weighted Citation Impact (FWCI) — and the discipline required to use them responsibly.
The h-index
The h-index is defined simply: a researcher has an h-index of h if at least h of their outputs each have at least h citations. It balances productivity (output count) and impact (citations per output) in one number. A researcher with many papers but few citations has a low h-index; so does a researcher with one highly cited paper and nothing else. Only sustained productive impact pushes it up.
The h-index is widely used because it is intuitive and resistant to single-paper outliers. It is widely criticised because it is age-biased (it can only grow with time), discipline-biased (citation cultures vary), and database-bound (an h-index computed from Scopus differs from one computed from Google Scholar or OpenAlex). For comparing researchers, those caveats matter.
Field-Weighted Citation Impact (FWCI)
FWCI normalises citations against the world average for the same field, year, and document type. A FWCI of 1.0 means citations match the field-average expectation; 2.0 means double; 0.5 means half. Unlike the h-index, FWCI is field-aware by construction — it is the most defensible single number when comparing researchers across disciplines.
FWCI has its own caveats: it depends on the field categorisation used, the citation database, and the document-type classification. Outputs with too few citations to be statistically meaningful should not drive conclusions.
Why both metrics fail on poor data
Name ambiguity is the single largest source of researcher-metric distortion. Common names produce false merges (two people counted as one) or false splits (one person counted as several). Affiliation changes lose attribution. Outputs published under previous name forms are missed. Without persistent identifiers — Scopus Author ID and ORCID — every researcher-level metric is a guess. We cover this in detail in Scopus Author ID and ORCID Explained.
What good practice looks like
- Reconcile identifiers first. No researcher-level metric is meaningful without ORCID and Scopus Author ID linkage to the institutional record.
- Combine, do not collapse. Use h-index and FWCI together, with output type diversity and collaboration evidence — never as a single number.
- Field-aware always. Cross-discipline panels should default to field-normalised metrics; otherwise, citation-heavy fields are systematically rewarded.
- Career-stage aware. Early-career researchers cannot match the h-index of senior colleagues; comparing them on h-index alone is meaningless.
- Align with DORA. Use journal-level metrics for journal context, not as a proxy for the individual researcher.
How a RIMS supports responsible use
A RIMS reconciles author identity across global sources, computes researcher-level metrics on the resulting clean record, and presents them alongside the wider evidence base — output diversity, collaboration, altmetric signals, and societal impact. Decision-makers see the numbers, but in context. The companion pillar guide places this within the broader metrics architecture; our JIF article covers why journal-level metrics are not researcher-level metrics.
Frequently asked questions
What is a good h-index? Field- and career-stage dependent. A mid-career researcher with h=20 may be exceptional in one field and average in another.
Is FWCI better than h-index? For cross-discipline fairness, yes. For sustained-productivity signal, h-index adds something FWCI does not.
How does our institution improve researcher metrics? Start with identifiers and reconciliation. Most apparent under-performance is missing or misattributed output, not absent research.
Where to start
Discover RIMS reconciles author identity across Scopus, OpenAlex, ORCID, Crossref, and Scimago, so researcher-level metrics — including h-index and FWCI — are computed on a defensible record. The metric still needs to be interpreted responsibly, but the foundation underneath it is trustworthy.