Responsible research assessment

1. What our journal tools measure

• Journal Audit's readiness checks, the metadata completeness score and the publisher report measure what a journal's public record holds: DOIs that resolve, reference lists, abstracts, licences, ORCID iDs, affiliations, funder information, full-text links, archiving, and the governance and ethics pages a journal publishes. They use no citation figure.
• The Journal Health Card in DOI Doctor records what a journal has achieved — records completed, policies published, archiving in place — in words and dates, with the public source of each. One of its lines is the journal's h-index in OpenAlex.
• We show no Journal Impact Factor, no quartile and no self-styled "impact factor". The Journal Trust Rating treats a self-styled impact factor on a journal's site as a red flag, and we tell journals plainly that it counts against them with the indexes they want.

2. Where citations appear, and where they count

• Shown as information, read from OpenAlex, dated and linked to the source: a journal's works, its citations, citations per article and its h-index (DOI Doctor's journal pages, GetIndexed's citation report); its i10-index (DOI Doctor); and, in DOI Doctor's indexing dossier, OpenAlex's 2-year mean citedness — not an impact factor, but a journal-level citation average of the same kind. Scholar Shield shows researchers their own works with the citations OpenAlex records. The Research Integrity Desk shows the citations OpenAlex records for each retracted paper of an institution (NIRF's 2025 framework deducts for them), and its journal sheets put the Trust Rating's and the Risk Radar's citation readings beside UGC's citation parameters (5.4, 8.1 and 8.2).
• Counted in GetIndexed's checklist for a journal, about itself: the share of its articles never cited (40% or less is marked well, up to 70% a warning, more a problem) and whether its articles of the last three years are cited at all — because Scopus and Web of Science weigh citations when they select journals.
• Counted in the Journal Trust Rating: 4 of its 100 points (section 3).
• Counted in Risk Radar as warning signs: a journal whose most-cited recent articles are cited mostly by the journal itself (up to 15 points of risk), or by one other journal (up to 10).
• We never sell citations, and no figure on our pages is an impact factor of ours.

3. A journal's grade — and where it counts for papers and people

• The Journal Trust Rating grades a journal from A to E (U when the public sources say too little) on 100 points in five parts: identity and governance 25, editorial integrity 25, technical standing 20, preservation and access 15, reach and citation 15. In the last part, 4 points come from citations: up to 2 for the citations OpenAlex records (100 or more: 2; any: 1) and up to 2 for the journal's h-index (10 or more: 2; 5 or more: 1). The rest of that part is a DOAJ listing (6), works recorded in OpenAlex (up to 3) and other verified listings (2). The rating measures a journal's practices and record; it says nothing about the quality of any single paper, and our rating pages say so.
• The Research Integrity Desk, which universities and colleges use for their faculty's publications, uses that grade in two places, and each institution sets both.
• The policy: the lowest journal grade the institution accepts, from A only to no minimum (it starts at A–C). A paper in a journal graded below it is marked non-compliant in the faculty record.
• The integrity-weighted output: each paper is weighted by its journal's grade — A 1.0, B 0.8, C 0.5, D 0.2, E 0, and 0.5 for a journal not yet rated — unless the institution saves its own weights; setting every weight to 1 counts every paper the same. A retracted paper weighs 0. The figure is given for each person, each department and the institution, labelled as a Smart Scholars index, not a NAAC or NIRF figure.

This is a journal-based measure applied to papers and to people. DORA's first recommendation asks that journal-based metrics not be used as a stand-in for the quality of individual articles, to assess an individual's contributions, or in hiring, promotion or funding decisions. We state it here so that an institution using these figures decides with that in view.

4. Researchers

• Scholar Shield shows researchers their own public record — their works, their identifiers and the citations OpenAlex records — so that they can correct and complete it. It makes no score of the person.
• The one figure per person in our tools is the Research Integrity Desk's integrity-weighted output above: the person's verified papers, weighted by their journals' grades. It is not a reading of the papers themselves.
• When we choose partners, advisers or reviewers for our own work, we look at what they have done, not at where it appeared.

5. The data behind our pages

Every figure on our pages is read from a public, open source — Crossref, OpenAlex, the ISSN Portal, DOAJ — when the page is made, with a link to the record it came from. The metadata of our own Registry is published under CC0. We support open research information because an assessment nobody can check is not one we are prepared to make.

6. Keeping this true

This statement describes the tools as they are on the date at the top. Our release tests check it against the code: every weight, point and threshold above is read from the tools' own source, and the release fails if one differs; the five sites are also read for the words "impact factor" presented as a measure of ours. A change to a tool that would make a line here untrue changes this page in the same release, with a new date. Questions and objections: mou@smartscholars.in.