Which databases can I use to find researchers on a topic?

Open options include OpenAlex , Semantic Scholar and Europe PMC ; Scopus, Web of Science and Dimensions are licensed and often available through a university library. Add ORCID to disambiguate names, and the funding databases CORDIS , NIH RePORTER and Gateway to Research to see who holds current funding rather than who published three years ago.

How can I tell whether someone is genuinely an expert?

Look at the last three years of output on the specific question rather than at career totals, since citation counts measure accumulated attention and advantage older work. Check that they still work on the topic, look for retractions or corrections where the stakes justify it, and read their competing interest declarations, funding sources and advisory roles if independence matters. Journal prestige is a property of a journal, not of a person, which is the reason initiatives such as DORA exist.

How do I find an independent expert for due diligence or journalism?

Independence is a separate search from expertise and usually the harder half. Exclude co-authors and collaborators of the people whose work is in question, exclude anyone funded from the same source, and check public competing interest declarations, advisory roles and patents. Editorial boards, learned society committees and institutional press offices are often faster than a database, because they are lists the field has already vetted.

Who is missing when I map a field this way?

The same people from every source: researchers publishing in languages other than English, fields whose main outputs are books or reports rather than journal articles, conference-only communities, industry work that never leaves the company, and practitioners whose expertise is in doing rather than in publishing. Absence from your map is not evidence of absence from the field, and treating it as such is the most common error in this kind of exercise.

Does the research expert finder rank experts by authority?

No. It orders results by proximity to the document you uploaded, not by standing in the field, and it uses no citation count, h-index or journal ranking to do it. Someone deep in a narrow corner of your topic can appear above a much better known researcher whose work is one step further away. It also shows no contact details and contacts nobody. It answers who has published work close to this document, and nothing more.

How to Find Experts in a Research Field

Mapping who works on a subject is a research task with a method. Here is the method, the sources that actually answer the question, and the honest limits of every one of them.

Whoever needs it, the question is the same: who actually works on this. A literature reviewer needs the groups behind a body of work, a journalist needs someone who can speak to a claim, an organisation doing due diligence on a technology needs to know whether anyone independent has looked at it, and a researcher entering a new area needs to know whose work to read first. This page is the method: what expertise means precisely enough to search for, the sources that answer it, how to trace a field through its citations, how to tell expertise from visibility, and what none of it can tell you. The event-specific version is in how to find conference speakers, and the partnership version is in how to find research collaborators.

Define the expertise before you search for it

Expert is not one thing, and most failed searches fail here. Deciding which kind you need takes two minutes and changes every source you would use.

Topic expertise: someone who works on this specific object, such as this disease, this material, this population, this policy instrument. Found in the recent literature on the object itself.
Method expertise: someone who knows the technique regardless of what it is applied to. Method specialists are systematically missed by topic searches, because a methods paper rarely uses the vocabulary of the field it serves.
Context expertise: someone who knows the setting, such as a country, a regulatory system, a clinical service or an industry. Frequently found outside academia, in agencies, hospitals and companies, and largely invisible in bibliographic databases.; Independence, which is a separate axis from expertise and often the actual requirement. For due diligence, journalism and evaluation, the useful person is one who knows the work and has no stake in it, which rules out co-authors, collaborators and anyone funded by the same source.; Write the question you want answered in one sentence before searching, including the object, the method and the setting. That sentence is the query, and it will not be three keywords.

The sources that answer the question

There is no register of experts. What exists is the published record and a set of registries built for other purposes, and knowing what each one covers is what stops you mistaking its silence for absence.

Bibliographic databases are the base layer. OpenAlex is open and free with a documented API, Semantic Scholar and Europe PMC are open in their own domains, and Scopus, Web of Science and Dimensions are licensed products your library may already provide. They differ in coverage, especially for books, non-English work and conference proceedings, so a name absent from one is not absent from the field.

Author identifiers disambiguate the results. ORCID links a person to their outputs across databases and is the fastest way to separate two researchers with the same name or follow one who has moved institution three times.

Funding databases show who is doing the work now rather than who published it three years ago, and they name the institution that holds the money. CORDIS for EU projects, NIH RePORTER for NIH awards, Gateway to Research for UKRI.

Learned societies, their special interest groups, their committees and their prize lists are curated by the field itself, and they cover the people whose contribution is institutional rather than bibliometric.

Journal editorial boards and conference programme committees are lists of people the field has already vetted for exactly the judgement you may be looking for.

Domain registries answer questions publication records cannot: ClinicalTrials.gov names principal investigators and sites, patent databases name inventors, and repositories and code platforms name the people who built the tools everyone else uses.

Trace the field through its citations

Databases return lists. Tracing returns a map, which is what you actually need when you do not yet know the field well enough to judge the list. Three moves, done in order, cover most topics in an afternoon.

Start from one paper you trust and read its reference list for the works it treats as foundational. Those are the field's fixed points and the names attached to them recur.
Then go forward: find everything that has cited that paper since, which is where the current work is. Forward citation is the single most productive move available and it is free in the open databases.; Read the recent citing papers for their authors rather than their conclusions. A group that has cited the same anchor paper four times in three years is working on it continuously.; Look for the review articles, and then look at who wrote them. A recent authoritative review is both a map of the field and a statement by its authors that they can survey it.; Watch for the boundary. When forward citations start arriving from a discipline you did not expect, you have found the neighbouring community that works on the same problem under another name, and it is usually where the least obvious experts are.; Track people, not only papers. Once a name recurs, read their recent output as a whole: what they are doing now is often not what the paper you started from was about.

Tell expertise apart from visibility

Every source above is biased towards people who are already visible, and visibility and expertise correlate imperfectly. Some deliberate checks are what keep a map honest.

Citation counts measure attention over time, which advantages older work and older researchers. A first author with three recent papers on the precise question may know more about it than a much-cited figure whose contribution was a decade ago.; Journal prestige is a property of a journal, not of a person or a paper. Assessment initiatives such as DORA exist because using it as a proxy for individual quality is a known error.; Look at what someone has done recently, not at their total. The most useful single check is the last three years of output against the specific question.; Check for retractions and corrections when the stakes justify it. Retraction Watch data is now part of the open Crossref record, as announced by Crossref, so it can be checked without a subscription.
Check interests when independence matters: funding sources, competing interest declarations on recent papers, advisory or board roles, patents. This is public information and it is the step most often skipped.; Distinguish someone who studies a thing from someone who does it. For a practical question about a technique or a clinical service, the person with the hands is frequently a better source than the person with the citations.

What the map is for, and what it cannot do

A list of names is an input to something, and the something determines how much verification is worth doing. Four uses cover most of it, and all four have the same blind spots.

Literature work: starting from people rather than citations tells you which groups produced a body of work, which of them disagree, and where a review would be superficial. It is also how you notice a field dominated by one laboratory.
Horizon scanning and due diligence: establishing whether anyone independent has examined a claim or a technology, and who could assess it. Here independence matters more than eminence, and the check on interests is the substance of the exercise.
Journalism and public communication: finding someone who can speak to a topic, with the added requirement that they are not a party to it. Editorial boards, learned societies and press offices are faster routes than a database for this, and the interest check is not optional.
Panels, advisory boards and evaluation: building a candidate list with the coverage and the independence the exercise demands. For peer review specifically, the requirements are stricter and our guide to choosing reviewers covers them.; The blind spots are consistent across every source on this page: work published in languages other than English is under-covered, books and reports are under-covered, industry work that never leaves the company is invisible, and people who have changed field are easy to mis-classify. Absence from your map is not evidence of absence from the field.; Our research expert finder automates one step of this, and only one. Upload a paper, a report or a policy brief and it returns researchers from the public literature whose work is closest to that document, with a reason for each. Because the query is a whole document rather than a keyword, it reaches groups using a different vocabulary for the same problem. It is not a ranking of authority, it does not know who is right, it does not know who is available, and it gives you no way to contact anyone. Everything in this section is still yours to do.

Frequently asked questions

Decide first which kind of expertise you need: topic, method, context, or independence, because they come from different sources. Then work the published record with a bibliographic database such as OpenAlex, Semantic Scholar or a database your library licenses, trace the field forwards from one trusted paper through its citations, and cross-check against funding databases, learned societies, editorial boards and domain registries. Finish by verifying recency and interests on the names you keep.