Agent Skills
› brycewang-stanford/Awesome-Journal-Skills
› mksc-rebuttal
mksc-rebuttal
GitHub用于撰写Marketing Science修改重投（R&R）的逐点回复信。在修订完成后，针对编辑和审稿人关于识别、假设、反事实及复现等核心问题，结构化地陈述已完成的修改、模型推理及代码更新位置，以有效回应并解决冲突意见。
Trigger Scenarios
收到MKSC R&R决定并完成修订后
需逐点回应审稿人关于识别或假设的质疑
需协调不同审稿人的冲突意见并向AE说明
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill mksc-rebuttal -g -y
SKILL.md
Frontmatter
{
"name": "mksc-rebuttal",
"description": "Use when drafting the revision and response letter after a Marketing Science revise-and-resubmit — structuring point-by-point responses to the Senior\/Associate Editor and reviewers across a multi-round process, especially on assumptions, identification, counterfactuals, computation, and replication. Drafts the response after revising; it does not interpret the decision (mksc-review-process)."
}
R&R Response & Rebuttal (mksc-rebuttal)
When to trigger
- You received an MKSC R&R and have planned the revision (via
mksc-review-process) - You have actually made (or scoped) the manuscript and code changes and need to write the response
- You must reconcile conflicting reviewer demands for the AE
- A later-round decision asks you to defend or extend prior changes
Write the response after revising the manuscript and re-running the analysis — it documents changes, not promises. The process is developmental and typically multi-round; resubmit the revised manuscript, response document, and updated replication materials through ScholarOne (mc.manuscriptcentral.com/mksc), keeping the same handling team where possible and the manuscript blinded.
Response-document structure
- Opening letter to the Senior/Associate Editor. Thank the team; summarize the most important changes in 2–4 sentences; explicitly address the AE's prioritized concerns. Note where reviewers conflicted and how you resolved it. (The cover letter/response is seen by the Editor/AE; keep the manuscript anonymized.)
- Per-reviewer sections. Restate every comment (numbered/faithfully), then respond.
- Point-by-point format: Comment → Response (what you did and the modeling/econometric reasoning) → Location (section/table). Quote new manuscript text where helpful.
Responding to MKSC's signature demands
- "Identification is not credible." Add the explicit identification argument, stronger/alternative instruments, a sensitivity-of-estimates analysis, or a Monte Carlo recovering known parameters — and explain why it resolves the concern (see
mksc-data-analysis). - "The model is too restrictive / an assumption drives the result." Relax the assumption and re-estimate, or show the result is robust to it; report the new comparative static/counterfactual.
- "The counterfactual is not valid." Re-solve allowing firms to re-optimize, report magnitudes with uncertainty, and bound the scope of validity.
- "Computation / multiple equilibria." Report solver settings, convergence, multiple starting values, and how multiplicity is handled.
- "Replication." Update the data-and-code package; confirm the master script reproduces every revised table/figure/counterfactual.
- Adding analysis vs. length. Heavy new derivations or robustness usually live in the online appendix; for a Frontiers paper, the 6,000-word total cap still binds, so move detail online.
Before/after: answering the signature "identification" objection
The modal MKSC first-round comment is some form of "I am not convinced the model identifies the effect." Compare responses:
- Weak (citation-padding, will draw a second round): "We thank the reviewer. We have added references to prior structural work establishing that this class of models is identified (Section 3)."
- Strong (re-analysis, closes the point): "We agree the argument was implicit. We now (i) state the exclusion restriction explicitly — cost shifter Z shifts supply but not demand conditional on our controls (Section 3.2); (ii) add a Monte Carlo that recovers the known parameters from simulated data within 3% (Appendix C.1); and (iii) report a sensitivity analysis showing the counterfactual moves less than 5% as we perturb the restriction (Table C4). Together these address the concern that [assumption] drove the estimate."
The strong version does real work — an explicit restriction, a recovery experiment, and a sensitivity bound — and tells the AE exactly where each lives.
Reconciling conflicting reviewers for the AE
When reviewers pull in opposite directions (R1: "the model is too stylized"; R2: "the model is over-parameterized"), do not silently satisfy one. State the trade-off to the AE, choose a resolution, and justify it: e.g., "We keep the parsimonious specification R2 favored and add R1's richer variant as a robustness check in Appendix D, showing the main counterfactual is unchanged." Surfacing the conflict with a defended choice reads as judgment; picking a side silently reads as evasion.
Tone and tactics
- Be respectful and substantive; concede gracefully and make the change.
- Disagree with evidence (an alternative estimation, a robustness result), not assertion.
- Address every point; silent omissions read as evasion.
- Surface reviewer conflicts to the AE with your chosen resolution.
Checklist
- Manuscript and code actually revised before the letter was written
- AE's prioritized concerns addressed first and explicitly
- Every reviewer comment restated and answered, with locations
- Identification / assumption / counterfactual demands met with real re-analysis
- Computation and replication-package updates documented
- Declined requests justified with evidence and an alternative test
- Manuscript stays blinded; length/Frontiers cap respected
Anti-patterns
- Writing the response before re-running the analysis.
- Answering "deepen identification" by adding citations instead of re-estimating.
- A counterfactual left unchanged when firm re-optimization was the objection.
- Skipping an inconvenient comment or de-blinding the manuscript.
Output format
【Round】1st R&R / 2nd / ...
【AE priorities addressed】1... 2... 3...
【Per-reviewer coverage】R1 x/x, R2 x/x, R3 x/x — all answered?
【Major changes】model/identification: ... counterfactual: ... computation: ...
【Replication】updated package reproduces revised exhibits?
【Declined + justification】...
【Status】drafted / final → resubmit via ScholarOne
【Next step】on next decision → mksc-review-process
Version History
- 1839142 Current 2026-07-05 14:04
