jeea-tables-figures
GitHub用于构建或修订JEEA论文表格与图表，确保符合期刊规范并面向广泛读者。遵循三线表、无显著性星号、报告标准误等原则，通过StatsPAI/Stata工具生成自包含、清晰的展示内容，突出核心结论。
Trigger Scenarios
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill jeea-tables-figures -g -y
SKILL.md
Frontmatter
{
"name": "jeea-tables-figures",
"description": "Use when building or revising tables and figures for a Journal of the European Economic Association (JEEA) manuscript so exhibits read clean for a general-interest audience and respect JEEA house norms. Shapes exhibits; it does not change the underlying estimates."
}
Tables & Figures (jeea-tables-figures)
When to trigger
- Tables are dense, over-decorated, or still carry significance asterisks
- A figure is the better vehicle for the headline result but you are leading with a table
- Notes are missing the sample, units, inference, or estimator
- Exhibits are dumped at the end rather than supporting the argument
The JEEA exhibits bar
JEEA is general interest, so exhibits must be legible to an economist outside the subfield: the headline result should be readable from one table or one figure with its note alone. JEEA follows the standard modern-economics presentation discipline — report standard errors / confidence sets, not significance asterisks or boldface for significance — and the final accepted version is copyedited to OUP house style, so do not over-format the submission. Lead with the exhibit that carries the claim; everything else supports it.
Table craft
- Three-line tables (
booktabs): top/mid/bottom rules only; no vertical rules, no shading for emphasis. - No significance stars. Report point estimate and standard error (or confidence set) beneath it; never
***/**/*or boldface to denote significance. - Self-contained notes: sample and period, unit of observation, dependent variable and units, estimator, inference (clustering level), and the number of observations/clusters.
- One claim per table. The main table should let a reader recover the headline magnitude and its uncertainty without the text.
- Structural fit tables: show targeted vs. untargeted moments (data vs. model) so identification is visible.
Figure craft
- Event-study plots with confidence bands and a visible pre-period (flat leads = the parallel-trends evidence).
- Model-fit overlays (data vs. simulated moments) and counterfactual curves with uncertainty bands for structural work.
- Vector output (
.eps/.pdf) preferred for final files; raster at ≥300 dpi; readable axis labels and units; colorblind-safe palettes; no chartjunk. - Prefer a figure over a table when the shape of the result (dynamics, heterogeneity, distribution) is the point.
Execution bridge (StatsPAI / Stata MCP)
Generate exhibits from the fitted result, not by retyping numbers. Full map:
execution-with-mcp. JEEA is a general-interest European economics flagship; credible identification across applied fields.
- Tables:
etable(multi-model) ordid_summary_to_latexstraight from theresult_id. - Figures:
plot_from_result/enhanced_event_study_plot/event_study_table— axis units and the SE/clustering note baked in. - Every note names the estimator + clustering and states the magnitude in interpretable units.
See a full fitted-result → exhibit chain in the JF execution walkthrough.
Checklist
- No significance asterisks or boldface-for-significance anywhere
- Standard errors / confidence sets reported beneath estimates
- Each table/figure note is self-contained (sample, units, estimator, inference, N)
- Headline result recoverable from one exhibit + its note
- Event-study / fit / counterfactual figures carry uncertainty bands
- Three-line tables; vector figures at publication quality
- Exhibits support the argument in-text, not dumped at the end
Exhibits by paper type
- Applied micro / development: lead with the event-study figure (dynamics + parallel-trends evidence) or a single main-effect table; relegate the specification grid to the online appendix.
- Theory: a figure of the key comparative static or a numerical example can make an abstract proposition concrete; do not over-tabulate.
- Structural / quantitative: a model-fit table (targeted vs. untargeted moments) plus a counterfactual figure with uncertainty bands; a parameter-estimate table with standard errors.
- Macro / finance: impulse-response or factor figures with confidence bands; avoid factor-zoo tables no generalist can parse.
Worked vignette (illustrative)
A draft reports the main result in a 9-column regression table with three rows of asterisks and bolded coefficients. A general-interest referee cannot find the headline number. The JEEA-grade fix: move seven columns to the online appendix, keep the two that carry the claim, replace asterisks with the point estimate and its standard error in parentheses, and add a self-contained note (sample, period, unit, dependent variable in levels, clustering at the state level, N and number of clusters). The headline magnitude is now readable from the table alone — and the event-study figure, promoted to the main text, shows the dynamics the table cannot.
Anti-patterns
- Significance stars or bolded "significant" coefficients (against JEEA house norms)
- A table no one can read without hunting through three paragraphs of text
- Notes missing the clustering level, units, or sample definition
- Leading with a regression table when an event-study figure tells the story better
- Over-formatting to mimic the typeset journal style the copyeditor will apply anyway
- Counterfactual or fit figures with no uncertainty representation
- A factor-zoo or specification-grid table dumped in the main text with no headline exhibit
Output format
【Headline exhibit】[table N / figure N] — recoverable from its note? [Y/N]
【Significance display】SEs/CIs only, no asterisks? [Y/N]
【Notes complete】sample/units/estimator/inference/N? [Y/N]
【Figures】event-study / fit / counterfactual with bands? [Y/N]
【Next step】jeea-writing-style
Version History
- 1839142 Current 2026-07-05 14:01
