Pre-Submission Reviewer Lens (er-reviewer-lens)
This skill does not speculate about the preferences of any specific reviewer, nor does it simulate any particular referee. It only distills therecurrent, predictable, structuralpoints of criticism in 《经济研究》 review into a "self-audit lens," for authors to run through themselves before submission and plug the holes in advance—moving "criticism heard only after submission" forward to "saying it out loud yourself before submission."
Trigger Timing
- The draft has taken shape and is about to enter er-submission
- You feel it is "about there" but cannot articulate what is still missing
- After rejection, you want to systematically check "which high-frequency landmines were stepped on"
- Coauthors review each other and want to use a unified standard to find fault
High-Frequency Criticism Checklist (by paper stage)
For each category, we give "How the reviewer will ask → How you should prevent it in advance / respond in the text."
1. Theoretical Positioning (highest frequency and most lethal at this journal)
| The reviewer will ask |
How you should prevent it in advance |
| "What exactly is the theoretical contribution? This is just a policy evaluation" |
In paragraphs 2–3 of the introduction, clearly state "the marginal advance this paper makes to X theory," rather than merely writing "evaluated the effect of policy Y." See er-introduction、er-theory-hypotheses |
| "It does not engage in dialogue with articles on the same topic in 《经济研究》 over the past 3 years" |
The literature review should explicitly cite the 2–4 most relevant articles from this journal in the past 3 years and clearly state "how this paper differs from them." See er-literature-review |
| "Several marginal contributions overlap with each other / are overstated" |
Write 2–3 contributions that are orthogonal to each other and each of which can be matched to a corresponding subsection in the main text; delete exaggerated wording such as "first" and "fills a gap" |
2. Identification Strategy
| The reviewer will ask |
How you should prevent it in advance |
| "Why is this policy exogenous to individuals?" |
Provide institutional-background evidence (top-down, unrelated to individual characteristics), rather than a sentence saying "we believe it is exogenous" |
| "TWFE + staggered DID does not address heterogeneous treatment effect bias" |
Switch to robust estimators such as CS / SA / dCDH / BJS + Bacon decomposition. See er-identification |
| "IV exclusion restriction does not hold / first stage is weak" |
Argue for the uniqueness of the channel for the exclusion restriction, report effective F and AR robust CI |
| "Parallel trends are only shown in a figure, not tested" |
Joint test of pre-treatment coefficients + power / Honest DiD sensitivity, not just a figure |
| "Is there an anticipation effect / early response?" |
Test coefficients for several pre-treatment periods, discuss anticipation, and if necessary exclude the "front-running" window |
3. Mechanisms
| The reviewer will ask |
How you should prevent it in advance |
| "The mechanism is only stated verbally, with no evidence" |
At least 1 independent mechanism regression table. See er-mechanism |
| "The three-step mediation method / Sobel is not credible" |
According to 江艇 (2022, 《因果推断经验研究中的中介效应与调节效应》, 《中国工业经济》第 5 期): empirically run only the D→M regression; argue M→Y using theory + literature, and do not directly run Y on M |
| "The mechanism contradicts the direction of the main effect but is not explained" |
If the mechanism variable runs opposite to the main effect, an economic explanation must be provided; otherwise, remove it |
4. Heterogeneity category
| Reviewers will ask |
How should you prevent this in advance |
| "Heterogeneity has no theoretical guidance; it looks like post hoc storytelling by picking significant results" |
Each heterogeneity dimension should have a theoretical expectation before looking at results. See er-heterogeneity |
| "The subsample is too small and the estimates are not credible" |
Report subsample sizes; do not split when too small; merge if necessary |
| "No significance test was done for the difference in coefficients between the two groups" |
Use interaction terms / Chow test to confirm that the difference in coefficients between groups is significant; do not draw conclusions from "one significant, one not significant" |
5. Data and variables category
| Reviewers will ask |
How should you prevent this in advance |
| "The measurement of the core variable is questionable" |
Provide the literature basis for the measurement + at least 1 alternative measurement robustness check. See er-data-sample |
| "Sample screening is not transparent" |
List the screening steps and the number of observations removed at each step (sample flow) |
| "There are outliers / unit errors in the descriptive statistics" |
Winsorize or truncate and explain; check units and extreme values |
6. Robustness category
| Reviewers will ask |
How should you prevent this in advance |
| "Key robustness checks are missing" |
According to the identification strategy, complete the standard robustness checks for that design. See er-robustness |
| "The conclusions change when the measurement / sample is changed" |
Run these perturbations yourself first; if the conclusions are fragile, figure out why before submitting rather than avoiding it |
7. Writing and standards category
| Reviewers will ask |
How should you prevent this in advance |
| "Empty talk and clichés; policy recommendations are grandiose and vague" |
Delete template sentences; every sentence must be informative. See er-style |
| "The abstract has no quantitative results" |
Write the direction + magnitude of the core effect in the abstract. See er-abstract |
| "The reference style does not conform" |
This journal usesauthor–date system: in the text use (author, year), and at the end arrange by author–year,do not use [1][2] / [J][M] |
| "Too many tables and figures" |
Keep only the core tables and figures in the main text; put the rest in the appendix. See er-tables-figures |
| "Grammar / language problems" |
Read through once to eliminate typos and faulty sentences—this journal explicitly treats grammar / language as one of the review criteria |
8. Policy implications category
| Reviewers will ask |
How should you prevent this in advance |
| "The policy implications are too thin" |
Elevate the implications to the "meaning level": explain what the conclusions imply for understanding some type of economic mechanism / institution, rather than merely listing recommendations |
| "The policy recommendations are too operational, like a memo written to the government" |
Avoid items like "relevant departments should…"; derive implications naturally from the research findings. See er-policy-implication |
10-question self-check before submission (the 10 issues most likely to lead to rejection)
Ask yourself item by item, "Can I answer this on the spot?":
- State in one sentence this paper'stheoretical contribution(not the policy effect)—can you?
- List which 2–3 articles in this journal and in Economic Research Journal over the past 3 years this paper dialogues with, and where the differences lie—can you?
- Why is the treatment / policyexogenousto individuals—is there evidence?
- Did the staggered DID use a heterogeneity-robust estimator and perform a Bacon decomposition—was it done?
- Parallel trends / weak instruments / smoothness, weretestsconducted (rather than only looking at plots)—were they done?
- Does the mechanism have an independent regression table and avoid the unreliable three-step mediation method—is this satisfied?
- Were the heterogeneity dimensionstheoretically motivated before examining results, and were between-group differences subjected to significance tests—is that so?
- Do the core variables have alternative measures, and are the conclusions stable after changing the measure—has this been verified?
- Does the abstract containquantifiedcore results—does it?
- Are the references in author–year style, and are there any typos or grammatical errors in the full text—has this been checked?
If you cannot answer any one question, it is a high-risk item; fix it before submitting.
Must-check list
- The theoretical contribution can be stated in one sentence and is not equivalent to the policy effect
- Explicit dialogue with articles on the same topic in this journal over the past 3 years
- 2–3 marginal contributions, non-overlapping, without exaggerated wording
- There is evidence for identification exogeneity; staggered DID uses a robust estimator
- Parallel trends / weak instruments etc. were tested rather than just shown in plots
- The mechanism has an independent table and mediation is handled according to the Jiang Ting (2022) standard
- Heterogeneity is theoretically guided, and between-group differences are tested for significance
- Core variables have robustness checks with alternative measures
- The abstract contains quantified results; policy implications fall at the "meaning level"
- References are in author–year style; grammar and wording have been read through without errors
Anti-patterns
- Treating this checklist as something to look at only after a submission is rejected—it is abefore submissionmirror
- Only doing superficial work to get by: adding a sentence, pasting a figure, without truly responding to the criticism
- Ignoring theory-positioning criticisms and only focusing on adding technical details—no matter how solid the technique, it will still be rejected if the theory does not hold
- Self-check reveals fragile conclusions but choosing to avoid them and betting the reviewers will not notice
Output format
【自检命中高风险项】
- [环节] 质疑：…… | 严重度：高 / 中 / 低 | 修补指向：er-xxx
- [理论定位] 贡献=政策效果，无理论推进 | 高 | er-introduction / er-theory-hypotheses
- [识别] 交叠 DID 裸 TWFE | 高 | er-identification
- [机制] 仅口头说明，无回归表 | 中 | er-mechanism
- ……
【10 问自检】答不上来的题号：[ … ]
【整体判断】可投 / 需大修 / 需小修后投
【下一步】er-submission
