How to Choose Keywords for a Journal Submission
A practical workflow for selecting manuscript keywords that represent the study accurately, fit a journal's current instructions, and avoid empty repetition of the title.
Readiness scan
Find out if this manuscript is ready to submit.
Run the Free Readiness Scan before you submit. Catch the issues editors reject on first read.
How to use this page well
These pages work best when they behave like tools, not essays. Use the quick structure first, then apply it to the exact journal and manuscript situation.
Question | What to do |
|---|---|
Use this page for | Getting the structure, tone, and decision logic right before you send anything out. |
Most important move | Make the reviewer-facing or editor-facing ask obvious early rather than burying it in prose. |
Common mistake | Turning a practical page into a long explanation instead of a working template or checklist. |
Next step | Use the page as a tool, then adjust it to the exact manuscript and journal situation. |
Quick answer: To choose keywords for a journal submission, first identify the study's specific population or system, intervention or exposure, method, and outcome. Then check the target journal's live author instructions for its required number, format, and controlled vocabulary. Keep terms that add information beyond the title, remove duplicates and ambiguous abbreviations, and test whether each remaining term would help a researcher find this exact paper.
Use this alongside the pre-submission checklist, journal-selection guide, and manuscript quality check.
From our manuscript review practice
The best manuscript keywords do not try to market the paper. They make the study's population, intervention or exposure, method, and outcome findable without repeating the title.
Who this is for
This workflow is for an author who has chosen a target journal and needs to complete the title-page or submission-system metadata accurately. It is not an SEO exercise, a substitute for a clear title and abstract, or a reason to change the study's claims. Keywords improve how a paper is classified and discovered only when they describe the work truthfully.
The first decision is whether the journal asks for free-text keywords, terms from a controlled list, subject classifications, or a combination. Current publisher guidance varies. For example, Oxford Academic journals may specify an article-type range and a pre-defined list with optional free text; the Journal of Navigation asks authors to choose up to four terms through the online system rather than place them in the manuscript; and other journals advise authors not to repeat title words. The current journal instruction is the authority.
The keyword-selection workflow
Build a candidate list before narrowing it. Start from the study, not from broad field labels or terms you hope will attract attention.
Step | What to write down | Decision test | Common mistake |
|---|---|---|---|
1. Name the research object | Population, organism, material, disease, dataset, or setting | Could a reader distinguish this study from a general field review? | Using only a broad discipline such as "oncology" or "nanotechnology". |
2. Name the central action or relation | Intervention, exposure, mechanism, comparison, or analytic task | Does the term describe what the study actually tests? | Using a causal term for an observational association. |
3. Name the method or design | Trial, cohort, systematic review, imaging, assay, model, or dataset type | Would this term help the right methods reader find the paper? | Listing every instrument rather than the method that changes interpretation. |
4. Name the outcome or application | Primary endpoint, phenotype, performance measure, or use context | Is it supported by the results, not merely discussed? | Choosing a future application the data do not validate. |
5. Check the journal rule | Number, format, controlled vocabulary, and submission location | Is each term allowed by the target journal today? | Applying a generic "five keyword" rule to every journal. |
6. Remove overlap | Terms already fully conveyed by the title or duplicates | Does each surviving term add a distinct retrieval path? | Repeating the title with word-order changes. |
What makes a useful keyword?
A useful keyword is specific enough to distinguish the work, stable enough that a field reader would recognize it, and honest about the study design. The keyword should point to a real component of the manuscript: the population or system, the exposure or intervention, the method, the outcome, or the setting.
Consider a cohort study assessing a post-operative infection outcome after a specific intervention. A weak list might be surgery, infection, outcomes, and clinical research. Those are labels for a library shelf, not a route to the article. A stronger candidate set might name the procedure class, patient population, intervention, study design, and precisely defined outcome, subject to the journal's allowed number and taxonomy.
The same principle applies outside clinical research. A materials study may need the material composition, processing route, measurement technique, and performance context. A computational paper may need the task, data modality, model family, and evaluation setting. Do not include a term merely because it is fashionable or because it appears often in another journal's articles.
Keyword decision table
Use this decision table to evaluate candidates before submitting. The point is not to maximize a total; it is to make each term defensible in the paper and identify the next action when it is not.
Situation | Action | Risk if you keep the wrong term | Evidence to check |
|---|---|---|---|
Candidate names the specific population, material, or system | Keep it when it identifies what was actually studied | Readers retrieve a broader literature than the sample or material supports | Title, Methods, eligibility or material description |
Candidate names an intervention, exposure, or mechanism | Keep it only when it is the central tested relation | The metadata converts a hypothesis into an apparent result | Research question, Methods, Results |
Candidate names a study design or method | Keep it when the method changes how readers interpret evidence | A routine instrument crowds out the term that makes the result findable | Methods and reporting checklist |
Candidate names a primary outcome or application | Keep it when the Results measure or validate it | A hoped-for implication becomes an unsupported retrieval claim | Results table, figure, or endpoint definition |
Candidate is a controlled vocabulary term | Use it when the journal requires or recommends it and it matches the paper | A forced classification distorts the scientific meaning | Current author instructions and taxonomy |
How we assessed this workflow
We reviewed current publisher author guidance from Elsevier, Taylor & Francis, Oxford Academic, Cambridge, and the Royal Society. They agree on the durable principle that keywords should help the right audience find the paper, but their counts, format rules, and controlled-vocabulary requirements differ. This page therefore provides a decision method, not a universal keyword count or a promise of search performance.
In our pre-submission review work, we find three specific named failure patterns more often than an outright lack of keywords: title repetition that adds no retrieval path, a broad field label that hides the actual study, and an outcome term that overstates the evidence. The checks below make those patterns visible before upload.
A title copied into the keyword field
If every keyword already appears in the title, the metadata may add little. Keep the title's central concept where it belongs, then use the keyword field to name a meaningful population, method, setting, controlled term, or outcome that the title cannot carry without becoming unreadable. This is especially important when the target journal explicitly asks authors to avoid title repetition.
Check whether the title, abstract, and keywords make the same accurate claim.
A broad term masking the real study
"Machine learning," "cancer," or "biomaterials" can be appropriate only if the paper is genuinely broad. In practice, a reader looking for a particular use case needs the data type, disease, material, population, intervention, or analytic setting that makes the study distinguishable. Replace one generic label with a term a relevant researcher would plausibly use to locate this specific evidence.
Check whether the manuscript names its population, method, and outcome precisely.
A result word that the methods cannot support
Keywords can overclaim just as an abstract can. An observational study should not use an unqualified causal term; an exploratory model should not be tagged as a validated clinical tool; and a laboratory result should not be described as a patient outcome. Match the term to the study design and the endpoint actually reported.
A five-minute final check
- Open the target journal's current author instructions and record the exact number, format, and controlled-vocabulary rule.
- Write a candidate list from the title, abstract, Methods, and primary results without looking at a keyword generator.
- Label each candidate as population or system, intervention or relation, method or design, outcome, or classification term.
- Remove a candidate if it duplicates the title, names an unsupported implication, or is too ambiguous for the intended field.
- Search each remaining phrase in a field database or the journal's recent articles to see whether it retrieves the relevant literature rather than an unrelated meaning.
- Keep a record of any controlled term selected in the submission system, especially if it differs from the manuscript's free-text keywords.
Readiness check
Run the scan to see how your manuscript scores on these criteria.
See score, top issues, and what to fix before you submit.
Think twice before finalizing when
Use the workflow when: the manuscript has a settled primary question and the author instructions are available. It is particularly useful when several closely related terms compete for the limited metadata fields.
Think twice when: the title has not yet named the study population or primary outcome; a keyword describes an application that is not measured in the Results; or a controlled term changes the scientific meaning of the work. Fix the title, abstract, methods, or result framing first. Keywords cannot repair a vague research question or an unsupported conclusion.
A pre-submission readiness check can test the alignment among the title, abstract, methods, results, and submission metadata before the final upload.
Frequently asked questions
Use the number and format specified by the target journal. Requirements vary: some journals set a fixed range, some require terms from a controlled list, and others collect keywords only in the submission system. Do not assume a universal number.
Usually no. Several current journal author guides explicitly tell authors to avoid repeating title words when selecting free-text keywords. Use the title for the central finding and the keywords for the population, method, material, setting, or controlled term that the title leaves out.
Use an abbreviation only when it is firmly established in the field and allowed by the target journal. Otherwise include the full term, especially when a reader or indexer could interpret the abbreviation in more than one way.
Not always. A submission system may require keywords, a controlled vocabulary, subject classifications, or all three. Preserve the distinction and follow the target journal's current author instructions.
Final step
Find out if this manuscript is ready to submit.
Run the Free Readiness Scan. See score, top issues, and journal-fit signals before you submit.
Anthropic Privacy Partner. Your manuscript is never used to train any model.