wber-rebuttal

Rebuttal Strategy (wber-rebuttal)

When to trigger

• A WBER decision letter arrived (R&R, major/minor revision, or reject-with-encouragement)
• You need to triage which referee asks are deal-breakers vs. optional
• The two referees disagree, or one is identification-focused and the other policy-focused
• You are deciding what new analysis is worth running before resubmitting
• You need to draft the point-by-point response letter

Read the letter through the WBER lens

WBER revisions almost always hit the journal's two-part bar: the identification referee wants the design hardened; the policy referee wants relevance, external validity, and cost. The editor's letter is the binding document — the editor's synthesis outranks any single referee, and the editor will signal which concerns are essential. Read the editor's framing first, then map each referee point to a referee archetype so your tone and evidence fit the reader. Single-anonymized review means the referees know who you are; engage them as informed colleagues, not adversaries.

Triage every point

| Tier | What it is | How to handle |

|------|-----------|----------------|

| Deal-breaker | The editor or a referee flags it as essential (e.g., "the design is not credible as is") | Do the work fully; lead the response with it; show the new result in the paper |

| Substantive | A real concern short of fatal (a robustness gap, an external-validity question) | Address with new analysis or a clear argument; revise the text |

| Framing | "Reads like a project report," "clarify the contribution" | Reframe and rewrite; usually cheap, high-return |

| Optional / misread | A referee misunderstood, or asks for out-of-scope work | Clarify respectfully; push back with evidence, do not capitulate to scope creep |

Resolve referee disagreements by appealing to the editor's synthesis and to evidence — if Referee 1 wants a structural model and Referee 2 wants none, state which serves the paper's question and why, and let the editor's priorities adjudicate.

Writing the response letter

• Open with a short summary of the main changes and the headline that survived.
• Point-by-point, quote each comment, then give: what you did, where in the revised paper, and the result.
• Show numbers, not promises. "We re-estimated with Callaway–Sant'Anna; the ATT is 4.1pp (s.e. 1.0), up from 3.8pp" beats "we addressed this."
• Match tone to archetype. For the identification referee, lead with the diagnostic and the robust estimate. For the policy referee, lead with external validity, scale-up, and cost.
• Update the data/code package to match the revised exhibits — WBER's release is a condition of publication, so the deposit must track the final results.
• Be honest about trade-offs. If a requested check weakens precision, report it and explain; do not hide it.
• Respect the 40-page cap when adding material — move secondary additions to the supplementary appendix.

Sequencing the revision work

Order the work so the deal-breakers cannot sink the resubmission:

1. Do the essential analysis first — the design re-estimation or new evidence the editor flagged as binding. If it changes the headline, everything downstream (exhibits, abstract, model) must follow.
2. Lock the revised result, then rebuild exhibits and the data/code package against it.
3. Rewrite the framing — intro, contribution, external-validity — to match the new result.
4. Draft the response letter last, once you can report concrete numbers and locations.

Never write the response letter promising work you have not yet done; WBER editors and the same referees will re-read, and an unfulfilled promise reads worse than an honest "this check weakened precision, and here is why we still believe the result."

Checklist

• [ ] Editor's letter read first; essential concerns identified and led with
• [ ] Every referee point triaged (deal-breaker / substantive / framing / optional) and answered
• [ ] Each response quotes the comment, states the change, the location, and the new number
• [ ] Identification asks answered with diagnostics + robust estimates
• [ ] Policy asks answered with external validity / scale-up / cost
• [ ] Referee disagreements resolved by appeal to the editor's synthesis + evidence
• [ ] Data/code package and DOI-linked DAS updated to the revised results
• [ ] Added material respects the 40-page cap; secondary content in the appendix

Anti-patterns

• Responding to referees before parsing the editor's synthesis (the editor decides)
• "We thank the referee and have addressed this" with no specific change or number
• Capitulating to scope creep that would dilute the paper's question
• Answering the identification referee with policy prose, or vice versa
• Letting the revised paper drift over 40 pages with reactive additions
• Leaving the replication package stale relative to the new exhibits
• Treating single-anonymized referees as anonymous adversaries rather than informed colleagues

Tone and the single-anonymized referee

WBER's single-anonymized review means the referees know who you are, but you do not know them — so write as if to a respected colleague, never as if scoring points. Thank substantive criticism genuinely, concede what is right, and push back only with evidence and a clear reason. When you disagree, frame it as "we considered this and here is why the alternative would not serve the question," not as a dismissal. Editors weigh the tone of a response: a defensive or evasive letter can sink a revision that the analysis would otherwise have saved.

Worked vignette (illustrative)

A reject-with-encouragement letter says the editor likes the question but Referee 1 calls the TWFE design biased and Referee 2 says the result "won't generalize beyond three districts." Triage: both are deal-breakers. The team re-estimates with a heterogeneity-robust estimator (ATT now 4.1pp, s.e. 1.0, with flat pre-trend leads) and adds an external-validity section comparing the three districts to the national distribution plus a fiscal-cost-at-scale figure. The response letter leads with the editor's two priorities, quotes each referee, shows the new numbers and their locations, and notes the openICPSR-style package and DOI-linked DAS now match the revised tables. The paper stays at 39 pages by moving older specifications to the appendix.

Output format

【Decision】R&R / major / minor / reject-with-encouragement
【Editor's essential concerns】led-with items
【Triage】deal-breakers / substantive / framing / optional
【Identification responses】diagnostic + robust estimate + location
【Policy responses】external validity / scale-up / cost + location
【Disagreement resolution】how editor synthesis adjudicates
【Package updated】data/code + DOI DAS match revised exhibits? [Y/N]
【Page cap】revision ≤40 pages? [Y/N]
【Next step】resubmit via ScholarOne