How to Find a Methods Reviewer

The cheapest routes cost you nothing socially and are consistently the most underused. Exhaust them before writing to a stranger.

Statistical or methodological support units. Many universities, medical schools, and large research institutes run one, sometimes under names such as biostatistics core, research design and analysis service, methods hub or data science unit. Access is often free to internal researchers and always requires booking further ahead than people expect.

Clinical trials units and research design services, in health-facing institutions. They exist to be consulted, they carry deep expertise in design and analysis, and consulting them at the design stage is what they most want to be used for.

Core facilities and platform teams. Bioinformatics, imaging, survey methodology and qualitative research platforms all employ people whose entire job is the method, and who are frequently willing to read a paper that uses their platform.

The graduate school. Doctoral programmes in methodology, biostatistics or quantitative social science contain senior students and postdocs who are current with the technique, are often glad of the visibility, and are frequently better matched to a recent method than a senior professor would be.

When the institution cannot cover it, the search moves to the wider community. The organising principle is to search by method rather than by topic, which is the opposite of how most people search for reviewers.

Methods departments and their staff pages

statistics, biostatistics, epidemiology and public health, econometrics, quantitative psychology, sociology methods, computational and data science. Read the staff pages, since methodologists describe their techniques there in a way their papers do not.

The methodological literature itself. Find the paper that introduced or evaluated the technique you used, and the papers that critique it, and look at who wrote them. Someone who has published on the pitfalls of your method is the ideal reader.

Learned societies of methods, and their special interest groups: the Royal Statistical Society, the American Statistical Association and the International Biometric Society are the general ones, and most fields also have a methods section within their disciplinary society. Section membership lists, committees and conference programmes are directories of exactly the right people.

The reporting guideline communities. The authors and contributors listed for the checklist that applies to your study type, collected on the EQUATOR Network, are by definition people who think carefully about how your kind of study should be reported.

Software and package authors. Whoever wrote and maintains the implementation you used knows its failure modes better than anyone, and package communities and issue trackers are public.

Your own network, one step out. Ask two colleagues who they would ask. A named referral from someone the person knows converts far better than any cold email, and it is the single most effective move in this whole section.

Matching on the word statistician is not matching. Fifteen minutes of checking prevents the two bad outcomes: a decline that could have been predicted, and an acceptance from someone who does not actually know the technique.

Check for the specific method, not the general field. Statistics is as wide as biology, and a survival analyst, a survey methodologist and a machine learning researcher share very little practical overlap.

Look at the last three to five years of their output, not at career totals. Methods move, and currency matters more here than eminence.

Check whether they have worked with data like yours. A method applied to clinical trial data and the same method applied to administrative records or to text raise different problems, and someone who has met your data type will find more.

Check the conflicts before you ask, in both directions

a competitor, a co-author on the work, or someone whose own paper your result contradicts is the wrong reader for this. Reading a paper before it is published carries an obligation, and it should not create an awkward position for either of you.

The failure mode of these requests is not rudeness, it is size. An unbounded ask from a stranger is declined by default, and a bounded one is frequently accepted.

Lead with the specific question, in the subject line and the first sentence. Whether a difference in differences design holds when treatment timing is staggered is an answerable question; please review my paper is not.

Bound the time explicitly and honestly. Thirty minutes on a call, or an hour on the analysis section, is a commitment someone can evaluate. If you think it is two days, say two days.

Say what you will send and confirm it is ready. Code, data or a synthetic version, the analysis plan and the reporting checklist should exist before you ask, not after they agree.

State the recognition you are offering and let them choose

acknowledgement by default, co-authorship if their contribution becomes substantive, or payment through a consultancy arrangement where your institution allows it. Do not promise co-authorship as an inducement before knowing what they will contribute.

Make the no easy. One sentence saying that a decline is completely fine and that a suggestion of someone else would be welcome converts more declines into referrals than anything else you can write.

This is the part most pages on this subject leave out. Knowing where the process breaks tells you when to stop searching and pay for it instead.

There is no register. Nothing lists methodologists who are available to review, and every route on this page is an approximation built on databases, staff pages and personal networks.

We do not have a tool for this. Our finders match published work to a document you upload, and none of them is built for methods matching, so we are not going to pretend otherwise. If enough people read this page, building one becomes a reasonable thing to do, and that is a decision we would rather make from evidence than from optimism.

Searching by topic finds the wrong people. This is worth repeating because it is the default behaviour of every literature search tool: topic searches surface the field's own researchers, who are precisely the ones who share your blind spot about the method.

Availability is invisible from outside. Nothing in any public source tells you whether a person has time, and the only way to find out is to ask, which is why the reserve list matters.

Seniority is often the wrong axis. For a technique that entered the field in the last five years, a postdoc or a late-stage doctoral student in a methods department is frequently the better reader, and much more likely to say yes.

It costs someone real time. If you would not act on a negative answer, do not spend a colleague's day getting one, and consider whether a reporting checklist would have told you the same thing.

Where goodwill is not the right currency, methods expertise can be bought or contracted. These routes are legitimate, are common in some fields, and remove the awkwardness entirely.

Institutional consultancy. Many statistics departments and support units take internal or external work formally, at a costed rate, and the arrangement is documented rather than personal.

Named collaboration written into a grant. If a project will produce several papers with a demanding analysis, a costed methodologist as a project partner is far better than repeated favours, and funders expect to see it.

Independent statistical consultants, including those who advertise through learned society directories. Rates and quality vary, so ask for two references from academic clients and check that the person publishes.

Journal-side statistical review, where it exists. Some journals run a formal statistical review track, so choosing a venue that does is itself a way of getting the analysis checked, and our journal finder can help you compare candidate venues on their stated policies.

Whatever the route, declare it properly. Paid methodological input is disclosed in the acknowledgements or the contributions statement, and substantive intellectual contribution to design or analysis meets most authorship criteria, including those of the ICMJE.