Reviewer says the work is incremental

“The advance over previous work is incremental” is the hardest reviewer comment to answer because it is a judgement about significance, not a defect. The reviewer weighed what the paper adds against what the field already knew and found the increment too small for this venue. The fix is reframing, appealing, or resubmitting to a journal whose criteria do not weigh advance size.

The incremental verdict is also, more often than authors assume, a comment about framing rather than content. A paper can make a substantial advance and still be read as incremental because the introduction positioned it as a gap-filling exercise.

What a reviewer means by incremental

Incremental means the reviewer accepts your results and disputes the size of the advance they establish. The evidence is not in question; the increment is. eLife publishes an explicit public vocabulary for this judgement, and it splits the two apart deliberately: every reviewed paper receives a significance term — landmark, fundamental, important, valuable or useful — alongside an independent strength-of-evidence term — exceptional, compelling, convincing, solid, incomplete or inadequate.

eLife’s vocabulary has no term for “incremental”, but the comment translates into it as the pairing of a high evidence term with a low significance term. eLife defines “useful” as “findings that have focused importance and scope” and “valuable” as “findings that have theoretical or practical implications for a subfield”, against “important” for “findings that have theoretical or practical implications beyond a single subfield”. A paper rated convincing on evidence and useful on significance has no methodological problem to repair. Reviewers at journals with no published vocabulary make the same two-axis judgement in prose, which is why the comment so often arrives attached to a review that criticises nothing.

Reading your own reviews along those two axes tells you which repair is available. If every methodological comment is minor and the objection lives in the summary paragraph, the increment is the issue and no experiment named in the review will move it. If the significance objection rests on a misreading of what was already established, you are answering a different comment — see reviewer says the work is already known.

Common questions about what reviewers mean by incremental

What does it mean when a reviewer says my paper is incremental?

A reviewer who calls a paper incremental accepts the data but judges the advance too small for the venue — a significance verdict, not a defect finding. eLife’s own review vocabulary keeps the two separate: a paper can be rated convincing or solid on evidence while still rated only useful or valuable on significance, and no experiment fixes the second score.

Why did reviewer 2 call my work incremental when the experiments are fine?

An incremental verdict from a reviewer who finds no methodological fault usually means rigor and significance are being scored on separate scales, and the fault the reviewer can’t name lives on the significance axis. NIH’s simplified framework makes the split explicit: Factor 2 (Rigor and Feasibility) and Factor 1 (Importance) each carry a separate 1–9 score, so a paper scores well on one and poorly on the other.

What’s the difference between “incremental” and “already known”?

“Incremental” concedes your result is new but calls the increment too small; “already known” disputes that the result is new at all. The two need opposite responses — narrowing or enlarging the significance claim for the first, versus demonstrating prior novelty for the second — so identify which one you actually received before drafting a reply. See reviewer says the work is already known.

Why incremental cannot be fixed like a defect

Incremental cannot be fixed like a defect because importance and rigour are scored on separate scales, and raising one does not move the other. NIH made that structure explicit in its simplified peer review framework, effective for receipt deadlines on or after 25 January 2025: Factor 1, Importance of the Research, absorbs the former Significance and Innovation criteria and carries its own 1–9 score; Factor 2, Rigor and Feasibility, absorbs Approach and carries a second 1–9 score; Factor 3, Expertise and Resources, receives no numerical score at all.

An application can earn a 2 on Factor 2 and a 6 on Factor 1, and the second number is what sinks it. Journal review runs the same two scales without naming them, and the confusion this produces is predictable: the author receives a low overall verdict, looks for the methodological fault that must have caused it, and cannot find one, because the fault is on the other scale.

NIH’s own guidance complicates a purely mechanical reading of that split. The framework states that “an application does not need to be strong in all categories to be judged likely to have major scientific impact,” making the overall impact score a holistic judgement each reviewer forms rather than an average of the factor scores. Factor 3, Expertise and Resources, is not scored on the 1–9 scale at all: reviewers instead pick one of two dropdown options, “Appropriate” or “Identify need for additional expertise and/or resources,” a pass/fail judgement sitting beside the two numerical scores. In practice, study sections still tend to let a weak Factor 1 dominate the summary statement anyway, because the assigned reviewer writes that paragraph, and a case for importance is harder to make convincingly than a case against it.

The practical consequence is that the standard revision instinct is aimed at the wrong number. Adding controls, adding replicates and tightening the statistics all raise the rigour score and leave the importance score exactly where it stood. A reviewer who asks for more controls is telling you the evidence is thin. A reviewer who calls the work incremental is telling you the evidence is fine and the claim it supports is small. Those two comments require opposite responses, and answering the second as though it were the first wastes a revision cycle.

Three ways good work reads as incremental

The introduction argues from a gap rather than from a question. “X has not been studied in Y” tells the reader nobody has done it. It does not tell them why it matters that nobody has. A gap is a fact about the literature; a reason is a fact about the world. Reviewers who read a gap-shaped introduction reach for “incremental” because the paper never told them what turns on the answer.

The diagnostic is mechanical. Read the final paragraph of your introduction and mark each sentence as a statement about the literature or a statement about the world. A gap-shaped introduction ends on the literature — who has studied what, and where the coverage stops. A question-shaped introduction ends on a consequence: a treatment that cannot be chosen between two options, a mechanism that two competing models explain differently, a measurement that current practice assumes and has never verified. The literature sentences do not have to go; they move up to become the setup for the consequence rather than the conclusion of the argument.

The advance is buried. If the most interesting result is figure 5 and the abstract leads with figure 1, reviewers weigh the paper by what they were shown first. Reordering requires no new experiments and is sometimes decisive.

Reviewers typically reach the title and abstract first and the figures next, turning to the methods when something looks wrong, so the increment judgement is usually formed before the methods are read. If figure 5 carries the advance, it belongs in the second sentence of the abstract and it belongs earlier in the figure sequence, unless the logical order genuinely forbids it. Test the abstract by deleting everything after its second sentence and asking whether what remains still states an advance. If it states only a topic and a method, the paper will be read as a method applied to a dataset.

The comparison is to the wrong prior work. Papers positioned against the nearest neighbour look small. Positioned against the standing assumption they overturn, the same result looks larger. Both framings can be honest; only one is accurate if the standing assumption is what actually changed.

The nearest neighbour is the natural comparison because it is the paper you read most closely while designing the study, and it is almost always the wrong denominator. Against the closest prior paper, every result is a delta. Against the belief the field held before either paper existed, the same result may be the thing that makes the belief untenable. Naming the right comparison is the substance of most novelty objections.

A worked example: one result, two framings

Consider a constructed but ordinary case: a manuscript reports that a kinase inhibitor reduces tumour volume in a mouse model, after an earlier paper reported the same effect in a different model.

Framed against the nearest neighbour, the contribution is that the earlier finding extends to a second model. A reviewer reads one more model, one more effect size in the same direction, and writes “incremental” — correctly, given the framing offered.

Framed against the standing assumption, the same experiment says something else. If the earlier model carried the activating mutation that the field assumed was the reason the inhibitor worked, and the second model lacks that mutation and shows the effect anyway, the paper does not extend a finding. It removes a requirement. The presumed mechanism is not necessary for the response, which changes which patients are plausible candidates and which biomarker is worth measuring. Same animals, same data, same figures.

The second framing is available only if the second model genuinely lacks the mutation and the manuscript measured and reported it. If the mutation status was never characterised, the framing is an assertion the design cannot support, and a reviewer will convert it into a request for the genotyping experiment — see mouse model relevance. The reframing is not a writing exercise. It is a claim about what the data establish, and it is subject to the same evidentiary test as any other claim in the paper.

The reframing also assumes the two models are comparable on every axis but the mutation, and that assumption is the one a specialist reviewer checks first. If the first study used an immunodeficient xenograft and the second a syngeneic or genetically engineered model with an intact immune system, the reviewer’s next question is which host factor changed, not only which mutation did — an intact immune system is itself a plausible alternative explanation for a drug response that an immunodeficient host cannot show. The manuscript has to rule that confound out explicitly, in the text, rather than let the reframing imply it was controlled.

How to tell a reframing from a flattering one

A reframing is accurate when you can name the sentence it makes false. Three tests separate an accurate reframing from a flattering one, and a framing that fails all three should not be submitted.

Name the sentence. Point to a specific claim, in a specific prior paper or review, that your result contradicts, narrows or makes conditional. Write it down as a quotation. If the strongest sentence you can produce is “our findings extend our understanding of”, there is no standing assumption being revised and the reframing is decoration.

Name the decision that changes. State who does something differently as a result, and what they did before. A clinician choosing between two agents, an experimentalist including a control they previously omitted, a modeller dropping a parameter. “The field will be interested” is not a decision.

Check that the framing survives the methods section. A framing the design cannot support is caught at review, and the failure mode is worse than the original comment: it converts a novelty objection into a rigour objection. A novelty objection is a judgement you can appeal, argue against or route to a better-matched journal. An overclaim is a defect you cannot deny, and it puts the honesty of the rest of the paper in question — the substance of an overclaim check. Reframe up to what the data establish and stop there.

When the verdict is a real judgement

An incremental verdict is a real threshold judgement, rather than a framing failure, when the reviewers converge on the increment while agreeing on nothing else. Sometimes the work genuinely is a modest extension, and no reframing changes that. The reviewer is applying the venue’s threshold, which is a scope decision rather than a rigour one.

Three signals mark a genuine threshold judgement rather than a framing failure. Both reviewers reach the same verdict on the increment while disagreeing about everything else, which means they converged on the size of the advance independently rather than on any specific flaw. The editor’s letter uses scope language — priority, fit, readership — rather than reporting a scientific concern. And the reviews are short. A reviewer who found the work sound and the increment small has little to write, so a two-paragraph review with no methodological requests is a stronger signal of a threshold judgement than a long one. The same distinction governs an out of scope decision made before review.

The options are then: add the experiment that would raise the advance, or move to a journal whose scope fits. The second is legitimate and usually faster. A well-matched venue publishes in months; chasing a scope threshold with new experiments can take a year and often fails again.

Authors underuse this. A solid paper in a good specialist journal contributes more than a year spent trying to clear a bar that was set for a different kind of result.

Choosing a venue that matches the increment

Choosing a venue that matches the increment means reading the journal’s stated evaluation criteria before submitting, not its impact factor, because journals differ in whether they assess the size of an advance at all — and the difference is stated in published policy rather than inferable from reputation. PLOS ONE puts its position in one sentence: “We evaluate research on the basis of scientific validity, strong methodology, and high ethical standards—not perceived significance.” Its seven criteria for publication require original research, absence of prior publication, analyses “performed to a high technical standard”, conclusions “supported by the data”, intelligible standard English, ethical compliance and adherence to reporting guidelines and data-availability standards. None of the seven names novelty, significance or impact.

Selective journals sit at the opposite end, applying an explicit test of conceptual advance or breadth of interest that the reviews may never mention by name. eLife occupies a third position between the two. Under its reviewed-preprint model, every paper it sends for review is published as a Reviewed Preprint carrying the reviews and an eLife assessment, whatever the assessment says. A paper judged useful on significance is published and labelled rather than rejected, which means the increment judgement becomes a public annotation instead of a gate. Combinations such as “Valuable/Incomplete” and “Useful/Solid” appear routinely among eLife’s published Reviewed Preprints — concrete evidence that a low significance rating does not by itself block publication under this model.

The operational advice is to read the criteria page before the impact factor. A journal that publishes an explicit soundness-only standard cannot reject you for an increment, and a journal whose criteria page asks for a conceptual advance or broad interest will apply that test whatever the reviews say about the methods. Matching the increment to the stated standard is the whole of journal selection for this objection, and it is the step most often skipped after a rejection, when authors resubmit to a journal of similar selectivity and receive a similar verdict.

What to do in the response

A response to an incremental verdict does exactly one of three things — reframe, appeal or concede — and responses fail most often by attempting two at once. Decide which before drafting a sentence.

If you are reframing, be specific about what changed and why it is more accurate — not merely more flattering. Point to the standing assumption the work revises, or to the question the field has been unable to answer.

If you are appealing, address the increment directly: state what was known, what is known now, and what becomes possible that was not. Vague assertions of importance make it worse.

If you are conceding, say so and withdraw rather than arguing. A short, gracious withdrawal takes an afternoon; a losing argument with an editor who has already decided takes weeks, and the paper is usually better served elsewhere in that time.

Two structural points about where the argument goes. A novelty objection is answered to the editor, not to the reviewer, because the editor owns the scope decision and the reviewer owns the science; put it in the covering paragraph of the response rather than in the numbered reply to comment 14, where it reads as a rebuttal of a colleague. And answer it in three sentences of the form “the field held X; this work establishes Y; Z is now possible or decidable”, with the middle sentence pointing at a figure. The general structure of the letter is covered in how to write a response to reviewers, and the decision of whether to write one at all in when to appeal a rejection.

What reviewers actually write

Reviewers rarely write the single word “incremental”. They write one of a small set of sentences, and recognising them tells you which objection you have before you start drafting.

“The advance over previous work is incremental and does not meet the bar for this journal.” “The study is well executed but largely confirms what is already established.” “The novelty is limited to the application of an existing method to a new dataset.” “The authors do not articulate what is conceptually new here.” “This is a useful contribution but would be better suited to a more specialist journal.” “The introduction establishes a gap but not its importance.” “The findings are expected given the prior literature.” “The work confirms rather than extends.” “I would be considerably more enthusiastic if the mechanism were established.”

Notice what these sentences have in common: not one of them names an experiment. That absence is the diagnostic. The single exception is the last, which is a mechanism request wearing novelty clothing and has a different remedy — see reviewer says the mechanism is not shown. “Well executed but” and “useful contribution but” are the clearest openers to watch for, because both concede the rigour axis explicitly in the clause before the objection.

Questions about responding to an incremental verdict

How do I respond when a reviewer says the advance is incremental?

Address an incremental verdict to the editor, not the reviewer, in three sentences: what the field held before, what this work establishes now, and what becomes decidable that was not. Point the middle sentence at a specific figure. Then choose exactly one path — reframe against the standing assumption, appeal directly, or withdraw and resubmit to a journal that does not weigh advance size.

Should I add more experiments if reviewer 2 says my paper is incremental?

Only if the experiment changes what the paper claims, not just how well the existing claim is supported. Adding controls or replicates raises the rigor score, which an incremental verdict usually already concedes, and leaves the significance score untouched. An experiment that lets you name a decision that changes — a clinician choosing between treatments, a modeller dropping a parameter — is the kind that moves the second axis.

My reviewer said the paper is incremental but didn’t ask for a single new experiment — what should I actually do next?

A reviewer who names no experiment in an incremental verdict is scoring significance, not rigor, so no experiment you add will move that score. Read the review’s summary paragraph for scope language — priority, fit, readership — then pick one response: reframe the introduction against the standing assumption the data revise, appeal to the editor directly, or resubmit to a better-matched journal.

Can I appeal a rejection for lack of novelty?

Yes: most journals permit an appeal against a novelty verdict, and appeals occasionally succeed — usually when a reviewer misread what was already established or missed where the paper’s advance actually sat. An appeal that just restates the paper’s importance rarely works, because importance is the exact thing under dispute; a well-matched resubmission elsewhere is normally faster and less likely to draw the same reviewers again.

What the decision letter tells you that the reviews do not

The editor’s letter carries the scope decision and the reviews carry the science, so read the letter first and read it for which of the two it is exercising. Four phrasings recur and each implies a different next step.

“Not a priority for publication in this journal” and “of more interest to a specialist audience” are scope statements. No revision addresses them, because the editor has not asked a question. Submit elsewhere.

“The reviewers raise substantive concerns” paired with an invitation to resubmit means the increment is negotiable and the editor is willing to see the argument again. That invitation is the clearest signal that a reframing is worth the effort.

“We are unable to offer publication, but would consider a new submission that establishes X” is a conditional offer with the condition named. Treat X as the specification it is, and decide whether you can meet it before deciding whether to appeal.

Rejection with no scope language at all, over two mixed reviews, is the case where an appeal has a genuine chance, because the editor deferred to reviewers who did not agree. What to do with each of these is set out in rejected after review, what next.

Frequently asked questions

Should I submit elsewhere when a reviewer calls the work incremental?

Submitting elsewhere is usually the right call when both reviewers converged on the increment independently and the editor’s letter uses scope language such as “priority” or “readership” rather than naming a scientific concern. PLOS ONE states plainly that it evaluates “scientific validity, strong methodology, and high ethical standards—not perceived significance,” so a sound paper cannot be rejected there for the size of its advance.

Can rewriting the introduction fix an incremental verdict?

Rewriting the introduction often does answer an incremental verdict, because the verdict is frequently a reaction to a gap-shaped introduction rather than to the data. Replace “X has not been studied in Y” with what turns on the answer — a treatment that cannot be chosen, a mechanism two models explain differently. The rewrite is only honest if the data support the larger claim; otherwise it becomes an overclaim.

I got rejected for being “incremental” even though both reviewers said the methods were sound — is it even worth appealing, or should I just resubmit somewhere else?

Read the editor’s letter before deciding: “not a priority for this journal” is a scope statement no appeal answers, while “reviewers raise substantive concerns” plus an invitation to resubmit means the increment is genuinely negotiable. Two mixed reviews with no scope language from the editor is the pattern where an appeal has a real chance, because the editor deferred to reviewers who disagreed with each other.

When the obvious remedy is unavailable

The remedy a novelty objection implies — one more experiment that enlarges the claim — is frequently unavailable, and the alternatives are more specific than “submit elsewhere”.

The experiment needs animals, cohorts or instrument time you do not have. Narrow the claim instead of enlarging the evidence. “Incremental” is a ratio between what is claimed and what was already known, and both terms can be moved. A title and abstract that claim a bounded, exactly supported result frequently draw a better reception than the same paper claiming the general case on the same data.

A competing group published the obvious follow-up while you were under review. The increment shrank after submission, through no decision of yours. Cite the new paper, state plainly what your work establishes that theirs does not, and do not omit it hoping the reviewer missed it. Independent confirmation under different conditions is a genuine contribution, and the way to make it count is to name precisely which conditions differed — see reviewer says the experiment is not replicated.

The study cannot be repeated. A closed trial, a completed field season, a registry extract under an expired agreement or an exhausted sample set means the dataset is the only one there will be. Say so explicitly in the discussion. Irreplaceability is a significance argument, and it counts only if you make it — which is also the honest way to handle registry and database limitations.

The reviewer wants a mechanism the system cannot deliver. Some systems do not permit the clean perturbation, and the right response states the constraint and reports the strongest available indirect evidence rather than promising an experiment that will not work.

Before submission

An incremental verdict is the objection the study section lens is built to surface before a reviewer reaches it. That lens reads the manuscript as a grant reviewer would rather than as a referee: whether the significance argument is demonstrated or asserted, whether the innovation claim is genuine novelty or merely unreported work, whether the introduction gives a reason or only a gap, and what the first sceptical objection will be. It is the one agent that will sometimes tell you the science is fine and the framing is the problem — which is exactly the finding this comment represents.

Run it while the introduction is still editable. The framing defects above are inexpensive to repair before submission and expensive afterwards, because after review a reframing has to be argued against a reviewer who has already formed a view.

Related

Novelty objections · Reviewer says it is already known · Reviewer says the mechanism is not shown, where the same scope-versus-rigour distinction applies · Why papers get rejected

PerfectPaper reads the introduction, the abstract and the figure order together, and reports where the paper’s stated advance is smaller than the advance its data support. It names the standing assumption the results bear on, so the significance argument is written before a reviewer has to guess at it.

Review my manuscript