POSI Citation Impact (PCI) & POSI Citation Score (PCS)
PCI and PCS are two open, auditable citation-impact indicators, published as POSI's contribution to the open/alternative-metrics tradition — not a replacement for the Journal Impact Factor or CiteScore, but openly computed options alongside them. Real PCI data is now computed for 992 journals in the curated Global Benchmark (2023–2024 window) — see the table below.
PCI does not yet determine an official POSI Citation Rank, Citation Percentile, or Citation Quartile. No POSI-R-* release has been produced (POSI-R-1.0-SPEC.md), and PNCI — needed to rank within a PSC category — has not been computed for this run. The table below is real, computed PCI data, sorted by value, not an official ranking.
OpenAlex 2-Year Mean Citedness (shown on Citation Reports) is a different, source-level preview figure and is not PCI.
PCS is an independently reported Crossref-based citation indicator and does not determine Citation Rank, Citation Percentile, or Citation Quartile.
Real PCI Data (Curated Global Benchmark)
Two Independent Measurements, Not One Blended Score
Web of Science and Scopus are independently curated databases that report independent numbers — a journal's JIF and CiteScore routinely disagree, because they're computed from different underlying collections. POSI follows the same principle rather than inventing one composite "open impact score": PCI is sourced entirely from OpenAlex (a 2-year window, JIF-shaped), and PCS is sourced entirely from Crossref (a 4-year window, CiteScore-shaped). The two are never averaged or blended — POSI publishes both, side by side, and lets the disagreement between them be visible rather than papered over.
Adoption Is an Institutional Choice
POSI does not lobby for PCI or PCS to replace the Journal Impact Factor or CiteScore everywhere. Whether an institution treats an open metric as sufficient for its own evaluation, funding, or accreditation purposes — in place of, or alongside, a proprietary index — is a policy decision made by that institution, not by POSI. What POSI controls is supply: publishing the numbers, the formulas, and the underlying open data, so that an institution which chooses openness as a criterion has something verifiable to adopt. POSI is a young, emerging data platform; PCI and PCS's relevance grows only as far as institutions choose to rely on them.
Why an Open Metric
DORA (2012)
The San Francisco Declaration on Research Assessment explicitly warns against using Journal Impact Factor as a proxy for the quality of individual articles or researchers. PCI is published with the same caveat, expressed as a per-journal citation-window statistic, not a quality certification.
Altmetrics manifesto (2010)
The founding altmetrics proposal argued that scholarly impact should be measured with open, web-native, auditable signals rather than closed, subscription-gated indices. PCI applies that principle specifically to citation counting.
Leiden Manifesto (2015)
Recommends that quantitative research evaluation support — never substitute for — qualitative, expert assessment, and that the data and methods behind any metric be open to verification. PCI's methodology section exists to satisfy that verifiability requirement directly.
One metric among several
PCI sits alongside MQS, PQF, IRS, and CVI on every POSI journal record — deliberately not blended into a single score, so no one number stands in for the others.
PCI vs. Journal Impact Factor
| Dimension | PCI (POSI) | JIF (Clarivate) |
|---|---|---|
| Underlying data | Open citation graph (OpenAlex, aggregating Crossref) | Proprietary Web of Science Core Collection |
| Formula | Fully published — see methodology below | Historically undisclosed in full detail |
| Reproducibility | Anyone can recompute it from the open OpenAlex API | Not independently reproducible without a WoS license |
| Access to the number | Free and public on every journal record | Requires a Journal Citation Reports subscription |
| Journal coverage | Any journal with an OpenAlex source record | Limited to WoS Core Collection-indexed journals |
PCS vs. CiteScore
| Dimension | PCS (POSI) | CiteScore (Elsevier) |
|---|---|---|
| Underlying data | Crossref DOI registry (is-referenced-by-count) | Proprietary Scopus database |
| Window | Trailing 4 calendar years, published formula | Trailing 4 calendar years, undisclosed sampling internals |
| Reproducibility | Anyone can recompute it from the open Crossref API | Not independently reproducible without Scopus access |
| Access to the number | Free and public on every journal record | Free to view, but underlying data requires a Scopus subscription |
How PCI & PCS Are Computed
PCI: citable-items and citations, OpenAlex-sourced — real data now covers the curated Global Benchmark
PCI[Y] = citations received during Y to citable items published in Y-1 and Y-2, divided by citable items published in Y-1 and Y-2 (PJR-SPEC.md § 5-6) — a full, exhaustive OpenAlex fetch, no article-sample cap. A real run now covers the curated Global Benchmark (990 of 993 journals) — see the PCI table above. POSI's own Core Collection isn't covered yet (too young for a real 2023-2024 window). No POSI-R-* release has been produced (POSI-R-1.0-SPEC.md) and PNCI hasn't been computed this run, so this real data still does not determine an official POSI Citation Rank, Citation Percentile, or Citation Quartile. OpenAlex's own summary_stats.2yr_mean_citedness is a different, source-level figure shown elsewhere as a preview indicator — it is not PCI.
PCS: one source, Crossref — real, frozen-spec-compliant data now published
PCS 1.0 (PCS-1.0-SPEC.md) is the mean is-referenced-by-count across a journal's articles published in the trailing 4 calendar years, computed directly from Crossref via full cursor pagination — no article-sample cap. A real run now covers the Core Collection and curated Global Benchmark, and is the only PCS figure shown anywhere on the site, including Citation Reports and individual journal pages — see the PCS table below. Like PCI, PCS is independent of OpenAlex and never averaged with it, and it does not determine Citation Rank, Percentile, or Quartile.
DOI/Crossref article counts ≠ citation counts
Where Crossref is used to verify how many articles a journal has published (the "Articles" figure on journal records), that is a separate use from the citation-count role above and is never double-applied.
One DOI per cited work
A dataset or preprint on Zenodo (or similar) under its own DOI is supplementary evidence, not a separate citable record, unless explicitly linked to the article DOI as a version/supplement.
No proprietary blending
POSI does not run a combined citation-counting pipeline that merges OpenAlex and Crossref into one number. Each metric is displayed with its own source attribution — full detail and rankings at Citation Reports.
Appropriate Use
Reasonable uses
- — Comparing journals within the same subject area over time
- — A funder or library building or reviewing a journal reading/subscription list
- — A journal editorial board tracking its own citation trend
- — One input, alongside PQF/MQS/IRS/CVI and expert review, in an institution's own multi-factor evaluation policy
Misuse POSI does not endorse
- — Judging an individual researcher, article, or grant application by the citation average of the journal it appeared in (the ecological fallacy DORA was written to stop)
- — Using PCI or PCS as the sole or primary criterion in hiring, promotion, or tenure decisions
- — Treating a single number as a substitute for reading the work itself
This split follows DORA's own recommendations. See Responsible Use Notice for POSI's full policy.
What PCI & PCS Are Not
- — Not a JIF or CiteScore substitute. Neither is calculated or peer-reviewed by Clarivate or Elsevier, and neither is directly comparable to the corresponding proprietary value.
- — Not a quality certification. A high PCI or PCS reflects citation volume within a fixed window, not editorial rigor, peer-review quality, or research integrity — those are covered separately by PQF, MQS, and IRS.
- — Not a substitute for expert judgment — see Appropriate Use above.
- — Not exhaustive. PCS covers Core Collection and the curated Global Benchmark; PCI currently covers the curated Global Benchmark only, not Core Collection yet (see the PCI table above). See Citation Reports for the OpenAlex 2-Year Citedness preview shown for Core Collection in the meantime.