技能 数据科学 学术统计展示规范

学术统计展示规范

v20260724
jbes-tables-figures
本指南提供了撰写高级统计或计量经济学论文时,设计和呈现表格和图表的最佳实践。它确保了复杂的统计结果(如大小-功效曲线、偏差/RMSE比较和实证收益)对专业统计学家和应用经济学家都是清晰易懂的,从而提升学术展示的严谨性。
获取技能
466 次下载
概览

Tables & Figures (jbes-tables-figures)

When to trigger

  • Simulation tables are dense and a reader cannot read off whether the method controls size
  • Size, power, coverage, bias, and RMSE are scattered instead of side-by-side
  • Figures lack confidence bands, or compare against no baseline
  • Exhibit notes do not state the DGP, sample size, replications, or nominal level

What JBES exhibits must communicate

A JBES paper is judged by method experts who read the tables as the evidence. Exhibits carry two distinct jobs: the Monte Carlo exhibits that establish the method's statistical properties, and the empirical exhibits that establish relevance on real data. Both must be readable by a statistician and an applied economist, since the journal bridges the two communities.

Monte Carlo exhibits

  • Put size and power side-by-side across DGPs and sample sizes; mark the nominal level so over/under-rejection is obvious.
  • For interval methods, report coverage and average length together; for point estimators, bias and RMSE together.
  • Always show the incumbent baseline in the same table under identical DGPs.
  • Figures: size-power curves, coverage-vs-n plots, empirical-vs-nominal QQ plots, sampling distributions with the asymptotic reference overlaid.

Empirical exhibits

  • Report estimates with appropriate (HAC/cluster/dependence-robust) standard errors.
  • Show the substantive payoff: what the new method changes relative to the standard approach.

Self-contained notes (every exhibit)

State the DGP or data source, sample size(s), number of Monte Carlo replications, nominal level, the estimator/test, the inference method, and units. A reader should not need the body to interpret the table.

Execution bridge (StatsPAI / Stata MCP)

Generate exhibits from the fitted result, not by retyping numbers (the usual source of body-vs-appendix drift). Full map: execution-with-mcp.

  • Tables: etable (multi-model columns) or did_summary_to_latex straight from the result_id — one variable definition, one set of numbers, body and appendix in sync.
  • 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 (from the result's diagnostics) and states the magnitude in interpretable units.

See a full fitted-result → exhibit chain in the JF execution walkthrough.

Checklist

  • Size and power readable side-by-side, with nominal level marked
  • Coverage shown with average length; bias shown with RMSE
  • Incumbent baseline in the same exhibit under identical DGPs
  • Figures have confidence bands / reference distributions
  • Every note states DGP/data, n, replications, level, method, units
  • Vector output (PDF/EPS); no chartjunk; legible in print
  • Numbers match the manuscript text and the code output

Anti-patterns

  • A wall of numbers where size control cannot be read off at a glance
  • Reporting power without size (or coverage without length)
  • Omitting the baseline, so "improvement" is unquantified in the exhibit
  • Notes that force the reader back into the body to decode the table
  • Figures with no uncertainty and no reference distribution

Worked vignette: a size-and-power table a referee can read

A hypothetical JBES paper reports a Monte Carlo size-power table for a new break test (numbers illustrative). The good version puts size and power side-by-side across n ∈ {120, 240, 480}, marks the 5% level in the caption, and shows the CUSUM benchmark in the same rows — so a reader sees the new test holds size at 5.2% while CUSUM over-rejects at 8.7%. The note states the DGP, replications, MC standard errors, level, and method.

Exhibit-pushback patterns (venue-specific fixes)

JBES referee objection Fix this skill enforces
"I cannot read size control off this table." Put size and power side-by-side; mark the nominal level in the note
"No baseline, so 'improvement' is unquantified." Place the incumbent in the same exhibit under identical DGPs
"The note does not let me interpret the table alone." State DGP/data, n, replications, level, method, and units in every note

Calibration anchor (hedged): JBES exhibits serve two audiences — a statistician reading Monte Carlo properties and an applied economist reading the payoff — so every table must be legible to both. Exact format specifics are live T&F preflight items.

Exhibit pass for Journal of Business & Economic Statistics

Run this as a concrete capability pass. First lock the statistical estimand, identification/simulation evidence, empirical illustration, and reproducibility path; then test whether the manuscript addresses econometrics/statistics reviewers who expect methodological credibility plus a business or economic use case.

  • Primary move: For every table or figure, state the object, sample/case base, uncertainty display, and one sentence the exhibit proves for this venue.
  • Decision ledger: return claim / evidence / blocker / next edit rows so the next pass can patch the manuscript directly.
  • Sibling comparison: compare against Journal of Econometrics for theory-heavy methods, Econometric Theory for proof-first work, Quantitative Economics for economics-theory methods; if the neighboring outlet has the stronger audience claim, recommend re-routing before polishing.
  • Verification floor: before submission-ready advice, re-open resources/official-source-map.md for volatile rules and name the one unresolved fact that could change the recommendation.

Output format

【Exhibit】Monte Carlo / empirical
【Size+power】side-by-side with level marked? [Y/N]
【Coverage/length or bias/RMSE】paired? [Y/N]
【Baseline】incumbent in same exhibit? [Y/N]
【Notes】DGP/data, n, reps, level, method, units present? [Y/N]
【Next step】jbes-writing-style
信息
Category 数据科学
Name jbes-tables-figures
版本 v20260724
大小 6.33KB
更新时间 2026-07-28
语言