Should Your Statistician Be a Co-Author?

3 min read

Sometimes, but not because they were paid or because they ran the numbers. A statistician belongs on the author list when their work shaped the study's design or interpretation and they can stand behind the paper. If they ran an analysis you specified and never saw the manuscript, an acknowledgment is the honest credit.

The standard most journals use

Most biomedical and social science journals follow some version of the authorship criteria from the International Committee of Medical Journal Editors (ICMJE), and they apply to statisticians exactly as they apply to everyone else. An author has to meet all four:

A substantial contribution to the conception or design of the work, or to the acquisition, analysis, or interpretation of the data.
Drafting or critically revising the manuscript for important intellectual content.
Approving the final version before it is submitted.
Agreeing to be accountable for the work, including answering questions about its accuracy.

Note what is missing: payment. Being paid for the work doesn't disqualify someone from authorship, and it doesn't earn it either. A salaried postdoc gets authorship for the same reason a consultant might: the contribution.

When it's clearly co-authorship

They designed part of the study. Chose the sampling scheme, ran the power analysis, or reshaped the randomization or measurement plan.
They made judgment calls in the analysis. Picked the model, handled the missing data, decided how to deal with clustering, and can explain why each choice holds up.
They wrote the methods and results. Or revised them enough that the paper's claims now depend on their reading of the numbers.
They will answer the reviewers. If you expect to forward the statistical comments to them, they are already doing an author's job.

When an acknowledgment fits better

The work was purely technical. You specified the model and they ran it, cleaned a file, or produced a table to your layout.
They never saw the manuscript. Someone who hasn't read the paper can't approve it or be accountable for it, however much analysis they did.
The engagement was advisory. A one-hour consult that pointed you toward a method is worth thanking someone for, not listing them.

The opposite mistake is common too. Adding a statistician's name as a courtesy, without their involvement in the writing, puts them on the hook for claims they never checked. Many statisticians will decline for exactly that reason.

Settle it at the start, in writing

Authorship disputes almost always come from expectations nobody said out loud. Put a line about it in the scope of work before any analysis begins: whether authorship is expected, what involvement it requires (reviewing drafts, responding to reviewers), and what happens if the role grows or shrinks. A study that starts as "just run the regression" and ends with the statistician rewriting the results section has changed, and the credit should follow.

Two practical notes. Some journals require a contribution statement (often using the CRediT roles, such as "Formal analysis" and "Methodology"), so agree on yours early. And if a statistician is an author, budget for their time on revisions: a reviewer's methods questions often land months after the analysis was finished.

If you're planning a study and want a statistician involved deeply enough to earn a place on the paper, from design through peer review, that's how we work with researchers.