brycewang-stanford Referee Strategy (eer-referee-strategy)
When to trigger
- The paper is near-complete and you want to pre-empt predictable referee objections
- You are unsure who the likely referees are and what they will attack
- An editor desk-screen risk needs reducing (breadth/novelty/fit)
- Before submission, to convert anticipated objections into in-paper defenses
The EER referee reality
EER review is single-anonymized: the referee sees who the authors are. The referee is typically an expert in your exact design or field and frequently someone you cite. Two gates precede them: an editor desk-screen (out-of-scope or low-novelty papers are rejected without review, and the submission fee is non-refundable), then ≥2 external referees for an independent assessment (检索于 2026-06；以官网为准). The first decision is usually fast; the strategic move is to pre-empt the predictable objections in the paper itself so a referee has less to grab.
Map objection → pre-emption
| Likely objection | Pre-emption (where it lives) |
|---|---|
| "Not general interest / too narrow" | breadth hook on page one (eer-topic-selection, eer-writing-style) |
| "Contribution is incremental / already known" | named frontier delta + concurrent-work distinction (eer-literature-positioning) |
| "Identification is not credible" | design-appropriate estimator + key-assumption defense (eer-identification) |
| "TWFE is biased here" | heterogeneity-robust DiD + flat leads (eer-identification) |
| "Exclusion restriction is assumed, not defended" | theory + institutions + falsification (eer-identification) |
| "Result is fragile / spec-dependent" | coefficient-stability battery (eer-robustness) |
| "Inference ignores clustering/dependence" | correct clustering + wild bootstrap / dependence-robust SEs (eer-robustness) |
| "Model is decorative" | model delivers the tested comparative static (eer-theory-model) |
| "Can't reproduce this" | replication package ready + data statement (eer-replication-package) |
| "Exhibits unreadable / stars instead of SEs" | one headline table with SEs (eer-tables-figures) |
Strategy moves
- Pre-empt the top three objections in the paper, not just the response letter — a referee who finds the answer already there has less to demand.
- Write for the cited referee: be accurate and fair to the literature; do not dismiss the person likely refereeing you.
- Reduce desk-screen risk: make breadth and novelty unmissable in the abstract and first paragraph.
- Pre-build the robustness/identification appendix the obvious referee will ask for, and signal it in-text.
- Keep claims calibrated — overclaiming is the fastest route to a hostile report.
Checklist
- Top 3 likely objections listed, each with where it is pre-empted in the paper
- Breadth + novelty visible enough to survive the editor desk-screen
- Identification/robustness defenses already in the draft (not deferred to rebuttal)
- Literature treatment fair and accurate (referee may be a cited author)
- Claims calibrated to evidence; no overselling
- Replication materials ready to disarm a reproducibility objection
Anti-patterns
- Assuming anonymity and writing dismissively about the likely referee's own work
- Saving every defense for the rebuttal instead of pre-empting in the paper
- Submitting a marginal-fit paper and losing the non-refundable fee to a desk reject
- Overclaiming and handing the referee an easy "the evidence does not support this"
- Ignoring an obvious robustness test the field referee will certainly request
Worked vignette (illustrative)
A finance paper anticipates three objections: (1) "endogenous treatment" → adds an IV with a defended exclusion and a placebo; (2) "driven by the crisis years" → leave-out-period robustness in-text; (3) "incremental vs. paper X" → a sentence naming exactly how it relaxes X's key assumption. All three are answered in the submitted paper. When the report arrives, the referee's main asks are already addressed, shortening the R&R.
Output format
【Likely referee profile】field/design expert; possibly cited author
【Top 3 objections】each + where pre-empted in the paper
【Desk-screen risk】breadth/novelty visible up front? [Y/N]
【Pre-built defenses】identification/robustness/replication ready? [Y/N]
【Claim calibration】matched to evidence? [Y/N]
【Next step】eer-submission (preflight) → eer-rebuttal (after decision)