Jop Tables Figures

Tables & Figures (jop-tables-figures)

At JOP, exhibits compete directly for pages: tables and figures count toward the 35-page (Research Article) or 10-page (Short Article) limit. Every exhibit must earn its space, stand on its own, and be regenerated by the deposited code so the replication analyst's re-run matches the printed page.

When to trigger

• Designing or revising the main tables and figures
• A reviewer found an exhibit unclear, overcrowded, or redundant
• The paper is over the page budget and exhibits are the obvious place to cut
• Deciding which exhibits belong in the main text vs the Online Appendix

Principles

• Self-contained. Title, units, N, and notes let the exhibit be read without the text. Define abbreviations and estimators in the note.
• Show the quantity of interest. Prefer plots of marginal effects, predicted probabilities, or coefficients with CIs over dense coefficient dumps a general reader cannot parse.
• Legible in print and grayscale. Colorblind-safe palettes; distinguishable line/point styles; vector output (PDF/EPS).
• One job per exhibit. If a table tries to do three things, split it or move two to the appendix.

Manage the page budget

• Tables and figures are part of the page count — a single sprawling table can cost a full page.
• Keep main-text exhibits to the few that carry the argument; move balance tables, full regression grids, and secondary plots to the Online Appendix (≤ 25 pp), which is excluded from the 35.
• A Short Article rarely needs more than one or two exhibits — choose the decisive one.
• Prefer compact, well-designed tables over wide landscape monsters that eat space and break in print.

Reproducibility of exhibits

• Every table/figure is produced by a script, not hand-edited; the numbers must match the text and the deposited package.
• Name the script that generates each exhibit in the readme so the JOP analyst can map exhibit → code.
• Re-run end-to-end before submission so figure files are current.

Execution bridge (StatsPAI / Stata MCP)

Generate exhibits from the fitted result, not by retyping numbers (the usual source of body-vs-supplement drift). Full map: execution-with-mcp. Journal of Politics spans observational and experimental political science; report the identifying assumption and the magnitude, not just stars.

• Tables: etable (multi-model columns) or did_summary_to_latex straight from the result_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 effect size in interpretable units.

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

Anti-patterns

• A coefficient dump where a marginal-effects plot would communicate the result
• Exhibits illegible in grayscale or at print size
• Hand-tuned final figures the code cannot regenerate (fails the replication check)
• Filling the main text with exhibits and blowing the page budget when the appendix exists

Worked micro-example (illustrative)

The AVR-turnout author's first draft puts a six-column regression table in the main text — full coefficients, two pages of print. A referee cannot read the effect off it and the page cost is steep. The fix replaces it with one marginal-effects figure: predicted turnout by adoption status with 95% CIs, colorblind-safe, vector, self-contained note giving N and the estimator. It reads in grayscale and the coefficient table moves to the appendix.

Referee pushback patterns and the JOP fix

• "I cannot read the effect off this table." Replace the coefficient dump with a marginal-effects or predicted-probability plot showing the quantity of interest with uncertainty.
• "The exhibits push the paper over length." Keep only the few that carry the argument in the main text; move balance tables and full grids to the appendix.

Exhibit pass for Journal of Politics

Treat this skill as an executable review pass, not a prose hint. First lock the political mechanism, evidence design, and scope condition; then judge whether the current manuscript answers the venue's real reader: political-science reviewers who want theory, identification or formal logic, and generalizable political implications in balance.

• Do the pass: For every table or figure, state the estimand or object, sample or case base, uncertainty display, and one sentence the exhibit proves for the venue audience.
• Return a ledger: give claim / evidence / risk / manuscript location rows, so the next agent can edit rather than rediscover the issue.
• Sibling guard: compare against APSR for field-wide political science, AJPS for design-heavy empirical work, World Politics for comparative/international politics; if a sibling owns the contribution, recommend re-routing before polishing format.
• Stop condition: do not give submission-ready advice until the pack's resources/official-source-map.md has been checked for volatile rules and the manuscript has one concrete fix for the largest venue-specific risk.

Output format

【Exhibit list】main-text vs appendix split
【Each exhibit】self-contained? shows quantity of interest? [Y/N]
【Legibility】grayscale/colorblind-safe, vector output? [Y/N]
【Reproducible】generated by named script, matches text? [Y/N]
【Page cost】main-text exhibits within budget? [Y/N]
【Next】jop-writing-style

Supplementary resources

• ../../resources/external_tools.md — plotting and exhibit-export tooling (coefplot, ggplot2, marginaleffects, sf/tmap)
• ../../resources/official-source-map.md — JOP page-limit rules (exhibits count toward the limit)

Source: brycewang-stanford/Awesome-Journal-Skills → Journal-of-Politics-Skills/skills/jop-tables-figures/SKILL.md