Relative Citation Ratio (RCR, NIH)

The Relative Citation Ratio (RCR) is an article-level, field-normalised citation metric developed at the US National Institutes of Health and published by Hutchins et al. in PLOS Biology in 2016. Rather than normalising an article’s citation count against a pre-defined subject category (the approach most other field-normalised metrics take), RCR builds a custom co-citation network for each individual article — the set of papers that are themselves co-cited alongside it — and derives an expected citation rate from the journals appearing in that network. An article’s citation rate divided by that expected rate gives its RCR; the scale is benchmarked so that 1.0 corresponds to the median performance of NIH-funded publications in the same field and time window. NIH distributes RCR through its own iCite tool.

What problem it is trying to solve

Raw citation counts cannot be compared across fields with different citation cultures — a highly cited paper in a fast-moving, large field and a modestly cited paper in a small, slow-citing field can represent similar relative influence within their own communities. RCR was built specifically to let NIH compare citation impact across its enormously disciplinarily diverse funding portfolio (from basic biochemistry to clinical and population research) using a field-normalisation method that adapts per article rather than relying on a coarse, manually assigned subject category.

How a research office encounters RCR

• NIH’s iCite tool surfaces RCR for NIH-funded publications, and some institutions pull RCR distributions as one input (never the sole input) into portfolio-level bibliometric reporting.
• Program staff or research offices preparing an NIH-related report may be asked to include RCR alongside raw citation counts for a set of publications.
• Because RCR is NIH-specific and NIH-benchmarked, it appears far less often outside biomedical and NIH-adjacent research contexts than field-agnostic metrics like citation counts or the h-index.

Documented methodological criticisms

RCR is a genuinely researched, peer-reviewed metric, but it is also a contested one, and a research-assessment reference should not present it uncritically. Points raised in the methodological literature include:

• Co-citation network sensitivity: because an article’s field-normalisation denominator is derived from its own evolving co-citation network, a paper that starts in a small field but is later picked up and heavily cited by a much larger field can see its expected-citation-rate denominator shift as that network grows — a theoretical scenario critics have pointed to as a way RCR’s normalisation can behave counter-intuitively for boundary-crossing or newly interdisciplinary work.
• Small-subspecialty distortion: in very small research subspecialties, citation counts alone may not clearly separate genuinely high-impact from low-impact work, which can let a modest paper in a small field earn an RCR comparable to a substantially more influential paper in a larger one.
• Influence is not importance: RCR’s own developers are explicit that it measures citation-based influence, not importance, intellectual rigor, or research quality directly — a citation-based proxy captures uptake and attention within the literature, which is a related but distinct thing from scientific merit.
• Time-window and citation-convention sensitivity: like most citation-based metrics, RCR’s values are sensitive to how much time has elapsed since publication and to field-specific citation conventions the co-citation approach does not fully eliminate.

This is why CASRAI, in line with the position taken by the San Francisco Declaration on Research Assessment (DORA) and the Coalition for Advancing Research Assessment (CoARA), treats RCR, like any single citation-based metric, as one input to be interpreted with methodological caveats and alongside qualitative expert judgement — never as a standalone, decisive score for evaluating an individual researcher or a funding decision.

How it relates to other terms in this cluster

• ‘Field-Weighted Citation Impact (FWCI)‘ — a comparable field-normalised metric using subject-category normalisation instead of RCR’s per-article co-citation approach; both aim at cross-field comparability but by different mechanisms.
• ‘Article-level metrics‘ — the broader category RCR belongs to, alongside altmetrics and simple citation counts.
• ‘Altmetrics‘ — non-citation-based indicators (attention, mentions, downloads) sometimes used alongside RCR to give a fuller picture than any citation metric alone.

Frequently asked questions

Is RCR only usable for NIH-funded research? RCR can technically be computed for any indexed article, but its benchmark (1.0 = NIH-funded median) and its distribution through iCite are NIH-specific, so comparisons outside NIH-funded biomedical research need explicit methodological caveats.

Should RCR be used in promotion or tenure decisions? Responsible-metrics guidance from DORA and CoARA cautions strongly against using any single citation metric, RCR included, as a standalone or decisive input into an individual’s evaluation; RCR was designed and validated as a portfolio-level research-impact signal, not an individual-assessment tool.

How is RCR different from the journal impact factor? The impact factor is a journal-level metric describing the average citation rate of articles in a journal; RCR is an article-level metric describing an individual paper’s own citation performance relative to its field — they measure different things and are not interchangeable.

References

• Hutchins B.I. et al., ‘Relative Citation Ratio (RCR): A New Metric That Uses Citation Rates to Measure Influence at the Article Level’, PLOS Biology, 2016 (doi:10.1371/journal.pbio.1002541).
• NIH iCite Support, ‘Relative Citation Ratio (RCR)’ documentation (support.icite.nih.gov).