Manuscript:   [title]
Version:      [commit, checksum or date of the main file]
Files read:   [main file, included files, bibliography, supplement]
Not provided: [figures as images only, data, code, appendix ...]
Request:      [referee report | pre-submission check | editor assessment | revision check]
Venue, field: [if known]
Audit:        [n errors, n warnings, n notes; includes not read: none | list]

# Referee report

Manuscript: [title]
Version reviewed: [version]
Date: [date]

## Summary of the submission

[Three to five sentences the authors would accept as fair: the question, the data and
method, the main result, the claimed contribution.]

## Overall assessment

[One paragraph: what is convincing, what is not, and what the recommendation turns on.]

## Strengths

1. [Specific, with location.]

## Major comments

1. **[Short title.]** [What is wrong or not shown, quoting the manuscript.]
   Where: [section, exhibit, file:line]
   Why it matters: [which conclusion depends on it]
   What would resolve it: [analysis, clarification or narrower claim]
   Status: [confirmed | not shown]

## Minor comments

[Grouped by section or exhibit. One or two lines each, with location.]

## Exhibits

[Table by table and figure by figure, only where there is something to say.]

## Citations and numbers

[Keys that do not resolve, attributions that could not be verified (say "not verified"),
numbers that disagree, with both locations.]

## Replication

[What a replicator is missing. Whether provided code or data were run, inspected or neither.]

## Questions for the authors

1. [A question whose answer would change the assessment.]

## Not reviewed

[Figures available only as images, supplementary files, data, code, anything outside the
reviewer's competence.]

Dear [corresponding author],

Thank you for submitting "[title]". [It has been read by n referees | I have read it
closely.] The decision is: [accept | minor revision | major revision | reject].

[One paragraph: what the paper does and what is valuable in it.]

[One paragraph: the reasons for the decision, in order of importance.]

A revision must address the following points. I will check each one against the revised
manuscript:

1. [Essential point, with the referee comment it comes from.]
2. [Essential point.]

The following are recommended but are left to your judgment:

- [Optional point.]

Please return, with the revised manuscript, a response letter that quotes each comment,
states what was changed and where (section, page or line), and marks the changes in the
text. [Deadline, format requirements.]

[For a rejection: the reasons that could not be addressed by a revision, stated plainly,
and what a different paper built on this work would need.]

Sincerely,
[editor]

### Data
- [ ] Each data source is named, with version, vintage or access date
- [ ] Access route is stated (public archive with identifier, application, purchase, restricted)
- [ ] Sample construction is listed step by step, with the count remaining after each filter
- [ ] The final sample size matches the N reported in the tables
- [ ] Every variable used is defined (construction, unit, transformation)
- [ ] Treatment of missing values and outliers is stated

### Method
- [ ] The estimating equation or algorithm is written out
- [ ] Standard errors: type and clustering level are stated
- [ ] All tuning choices are reported (bandwidths, lags, priors, hyperparameters) and how they were chosen
- [ ] Every stochastic step has a stated seed
- [ ] Analyses that were run but not reported are acknowledged

### Code and environment
- [ ] Code is available, with a persistent location
- [ ] Software and package versions are listed
- [ ] There is a map from each table and figure to the script that produces it
- [ ] Run time and hardware are given when they are substantial
- [ ] A licence or terms of use are stated

### Verification done in this review
- [ ] Code inspected (yes / no)
- [ ] Code run (yes / no; which exhibits were reproduced, to what precision)
- [ ] Discrepancies found: [list with locations]

| ID | Role | Type | Severity | Where | Quote | Problem | Status | Reason for status |
|---|---|---|---|---|---|---|---|---|
| A-01 | audit | wrong | minor | paper.tex:27 | "412 of 1,250 respondents (38.0%)" | 412/1250 is 33.0% | confirmed | recomputed |
| R-01 | referee | not shown | major | Section 5 | "[exact words]" | [problem] | not shown | searched sections 5 to 7 and the appendix |

| # | Comment (previous report) | Authors' response | Change found in new version | Verdict |
|---|---|---|---|---|
| M1 | [quote] | [summary of the response] | [section, file:line, or "none found"] | resolved / partly / not resolved / resolved differently / new problem |

## Changes not mentioned in the response letter
[From the version comparison: numbers that changed, sections added or removed, citations
dropped. Each with old and new location.]

## New findings in the revised version
[From rerunning the audit and rereading changed sections.]

## Assessment of the revision
[Which essential points are resolved; what remains; recommendation.]