Publish or Perish (Harzing’s citation-analysis software)

Publish or Perish is a free desktop software program, created by Anne-Wil Harzing (Emerita Professor, Middlesex University), that retrieves raw citation records from multiple academic data sources and calculates a set of citation-based metrics from them — most commonly the h-index and g-index, but also more than a dozen other established bibliometric indicators. It is a query and calculation tool, not a database in its own right: it holds no citation data of its own and instead sends searches out to other providers’ catalogs, then processes whatever results come back.

What data sources it queries

As of Publish or Perish 8 (released November 2021), the software can query up to nine underlying data sources: Google Scholar, Google Scholar Profile, Crossref, PubMed, OpenAlex, Lens.org, Scopus, Semantic Scholar, and Web of Science. Access differs by source: Google Scholar, Crossref, PubMed, OpenAlex, and Semantic Scholar are queried without a subscription (Google Scholar has no official public API, so Publish or Perish, like other Scholar-based tools, works against the public search interface); Scopus and Web of Science require the user’s own institutional subscription and API credentials, since Publish or Perish is a client that queries these providers’ data on the user’s behalf rather than a licensed re-distributor of it.

Because each source indexes a different slice of the literature and applies different inclusion rules, the same author’s citation counts, and therefore their h-index, will typically come out differently depending on which source a given search runs against — a general limitation of citation databases discussed further on this site’s h-index entry, not something specific to this software.

What it calculates

From a retrieved set of publication and citation records, Publish or Perish computes more than a dozen established research-performance metrics, including: number of papers, total citations, citations per year, citations per paper, citations per author, papers per author, authors per paper, the h-index and g-index, the hc-index (a variant weighting more recent publications), the hI-index and hI-norm (individualized, co-authorship-normalized versions of the h-index), age-weighted citation rate (AWCR) and its per-author form (AWCRpA), the e-index, and the hm-index. The software does not invent these metrics — it automates the otherwise tedious manual process of pulling a citation list and running the calculation, which is also why understanding the underlying metric (see this site’s h-index entry for a worked-example walkthrough of the calculation itself) matters more than which tool produced the number.

Typical uses

• Assembling a candidate’s own citation record ahead of a tenure, promotion, or grant application, where an institution asks for h-index or total-citation figures alongside a CV.
• Comparing citation performance across several possible data sources for the same author or journal, since the software surfaces the same query run against multiple sources side by side rather than committing a user to one database’s number.
• Scoping the citation landscape of a topic or journal during the early stages of a literature review, using citation counts as one signal (among others) for which papers in a results list are likely to be foundational.

What it is not

Publish or Perish is not a citation database, a persistent identifier registry, or a current research information system (CRIS) — it has no persistent record of its own and produces no institutional identifiers. It also is not an authoritative or standardized metric source: because results depend on which underlying database was queried and when (citation counts change continuously as new citing papers are indexed), a Publish or Perish figure quoted in one context (e.g. a CV) will not necessarily match a figure quoted from the same author’s profile on Google Scholar, Scopus, or Web of Science directly, even though all three ultimately draw on the same kind of underlying data. Responsible-assessment guidance such as the San Francisco Declaration on Research Assessment (DORA) applies to any single-number citation metric regardless of which tool calculated it: a citation-derived indicator like the h-index is field-dependent, time-dependent, and insensitive to highly-cited outliers, and should not be used on its own for hiring, promotion, or funding decisions.

Worked example

A researcher preparing a tenure file runs the same author search against both Google Scholar and Scopus inside Publish or Perish. The Google Scholar search returns a higher total citation count and h-index than the Scopus search, because Google Scholar indexes a broader range of document types (including non-peer-reviewed and grey literature) than Scopus’s editorially curated collection. Neither number is “wrong” — they are answers to two different questions (“how many citing documents did Google’s web crawler find” versus “how many citing documents did Scopus’s Content Selection and Advisory Board-vetted collection contain”) and the researcher’s institution should specify which source it expects before either figure is submitted.

Counter-example

A university-operated CRIS platform (such as Pure or Elements) that ingests a researcher’s publication list directly from Scopus or Web of Science via an institutional data feed is not the same category of tool as Publish or Perish — the CRIS holds and maintains a persistent institutional record over time, while Publish or Perish runs an on-demand query and reports a snapshot result that is not stored or updated afterward.

References

• Harzing, A.W. Publish or Perish, harzing.com/resources/publish-or-perish (official software page, data-source and metrics documentation).