Skip to content

revision

Category: drafting
Field: economics
License: private (curator-owned)
Updated: 2026-05-20
Stages: paper-drafting

Curator-private skill — copy text from 100xOS/shared/skills/writing/revision.md.

Paper Revision Workflow

Purpose

Guide the systematic revision of a manuscript in response to reviewer reports and an editor's decision letter. This skill covers the full workflow from reading reports to producing a revised manuscript with point-by-point responses.

For formatting the response letter itself, see writing/referee-response.

Revision Phases

Phase 1: Triage the Reports

Read all reviewer reports and the editor's letter. Classify every comment:

Category Description Priority
Fundamental Challenges identification, causality, core contribution Must address first
Methodological Estimation issues, robustness, alternative specs High
Data/measurement Variable definitions, sample selection, data quality High
Exposition Clarity, structure, writing, notation Medium
Bibliography Missing citations, DOI, reference formatting Medium
Minor Typos, formatting, cosmetic Low

Key rule: Do not start writing until you have read ALL reports and mapped every comment to a paper section. Some comments interact — a causality concern from Referee 1 may be addressed by the robustness check requested by Referee 2.

Phase 2: Address Causality and Identification First

Reviewers' concerns about causal identification are the single most common reason for rejection in empirical papers. Address these before anything else.

Checklist for estimation / causality concerns:

  • Is the identification strategy clearly stated? (source of variation, exclusion restriction, parallel trends assumption)
  • Are threats to identification listed and addressed?
  • Are standard errors correct? (clustering level, heteroskedasticity)
  • Is the estimator appropriate? (OLS when you need IV? DiD without pre-trends?)
  • Are the key coefficients causally interpretable or merely correlational? If correlational, is this stated clearly?
  • Do robustness checks address the specific concerns raised, not generic ones?
  • Are placebo tests, falsification tests, or pre-trend analyses included where applicable?

When the reviewer says "this is not causal": Do not just add caveats. Either (a) strengthen the identification strategy with additional analysis, or (b) reframe the contribution around the descriptive/correlational finding and explain why it is valuable on its own terms. Half-measures ("we acknowledge this is correlational" while still making causal claims) will be rejected.

Phase 3: Systematic Section-by-Section Revision

Map each reviewer comment to the specific section(s) it affects. Then revise section by section, addressing all mapped comments for that section in one pass.

Paper section Common reviewer concerns
Introduction Contribution unclear, overpromising, missing related work
Literature review Missing key references, shallow engagement, no DOI
Model/Theory Assumptions not justified, notation inconsistent
Data Sample selection unclear, variable definitions missing
Estimation/Results Causality, robustness, alternative specifications
Discussion Over-interpreting, ignoring limitations
Conclusion Repeating results without insight
References Missing DOI, inconsistent formatting, orphan citations

Phase 4: Bibliography and DOI Enforcement

Every reference MUST have a DOI. This is a hard requirement, not a suggestion.

  1. Run a DOI completeness check on the .bib file.
  2. For each entry missing a DOI, look it up via CrossRef Simple Text Query.
  3. Use \usepackage{doi} in the preamble so DOIs render as clickable links.
  4. Use plainnat bibliography style (or another DOI-displaying style).
  5. Only exceptions: pre-DOI publications, unpublished manuscripts, datasets.

Phase 5: Draft the Response Letter

Use the writing/referee-response skill for formatting. Key principles:

  • Quote every comment verbatim.
  • Address every comment — never skip one.
  • Provide page/line references for all changes.
  • For causality concerns, explain the specific analysis added (not just "we revised the section").
  • Cross-reference when multiple referees raise the same issue.

Phase 6: Verification

Before submitting:

  • Every reviewer comment has a response
  • All new tables/figures are numbered and referenced
  • DOI present for all references (run reference checker)
  • \usepackage{doi} in preamble, plainnat or DOI-enabled .bst
  • Track changes or marked-up manuscript prepared
  • No orphan citations or orphan references
  • Causality/identification section has been strengthened (not just caveated)
  • Response letter has page/line references for all changes

Integration with Workers

Task Worker Key skills
Read & triage reviewer reports virtual_coauthor This skill + reasoning/argument-audit
Revise estimation/causality econometrics_worker, causal_inference econometrics/*, causal-inference/*
Revise prose sections paper_drafter writing/paper-structure, this skill
Draft LaTeX response letter latex_drafter (section: referee-response) writing/referee-response
Check bibliography DOIs reference_checker review/reference-check, latex/bibtex
Full quality review of revision deep_review review/*

Anti-Patterns

  • Defensive revisions. Adding caveats ("we acknowledge this limitation") without substantive changes. Reviewers see through this.
  • Cosmetic robustness. Adding 10 robustness tables that all use the same flawed identification strategy. Fix the strategy, then show robustness.
  • Ignoring the editor. The editor's letter prioritizes concerns. Address the editor's specific requests first.
  • Scope creep. Adding new analyses not requested by reviewers. Stick to what was asked unless a new analysis directly addresses a concern.
  • Missing DOI in the revision. Reviewers specifically check for this. It signals carelessness.