Tables & Figures (joap-tables-figures)

In JAP, exhibits do a lot of the evidentiary work: the means-SD-correlation-reliability table, the path/CFA diagram, the multilevel result table, and (for syntheses) the forest/funnel plot. They follow APA 7th-edition conventions and must be self-contained and consistent with the reported analysis. Good exhibits show effect sizes, confidence intervals, and reliabilities, not just stars.

When to trigger

• Designing the descriptives/correlation table or the main results exhibit
• Drawing a CFA/SEM path diagram or a multilevel result table
• Building a meta-analytic forest/funnel plot
• A reviewer found an exhibit unclear, non-APA, or inconsistent with the text

Principles

1. The Table 1 standard. The first table is almost always means, SDs, and the full correlation matrix with reliabilities on the diagonal and level-of-analysis noted. Reviewers scan it for construct distinctness, multicollinearity, and red flags before reading the model.
2. Show estimates with uncertainty. Path tables and figures report standardized/unstandardized coefficients with confidence intervals (and fit indices for SEM); report indirect effects with their bootstrap/Monte Carlo CIs.
3. Self-contained + APA 7th. Titles, notes, Ns at each level, units, what the error bars/CIs are, and significance conventions make each exhibit intelligible alone; follow APA table/figure format.
4. Make the model legible. A path diagram should show the hypothesized structure, levels (L1/L2), standardized estimates, and which paths are significant — not a decorative box-and-arrow sketch.
5. Earn the space; defer the rest. Push secondary exhibits (full invariance tables, alternative models, manipulation checks) to online supplemental materials; summarize in one sentence.
6. Reproducible + accessible. Generated by the shared analysis script so values match; colorblind- safe and grayscale-legible.

Worked micro-example — the core exhibits (illustrative)

For the servant-leadership package, two exhibits carry the argument.

Table 1. Means, SDs, correlations, reliabilities (team level, k = 74).
  Diagonal: ω reliabilities (.91, .88, .87); note ICC(1)/ICC(2) for aggregated vars,
  N at L1 and L2, and the significance convention. Self-contained.

Figure 1. Multilevel mediation path model (2-2-2).
  Geometry:  leadership (L2) → safety (L2) → performance (L2), with the L1
             voice mediator shown; standardized estimates + 95% CIs on each path.
  Annotate:  indirect = .13, 95% CI [.05, .23]; direct path ns; fit indices in note.
  Note:      defines levels, estimator, Ns at L1/L2, and that intervals are 95% CIs —
             readable without the main text.
  Source:    rendered from the deposited Mplus/R script so values match exactly.

Exhibit triage — article vs. supplemental materials

Exhibit-stage reviewer pushback and the venue fix

• "No reliabilities on the diagonal" → add ω/α to Table 1; reviewers check construct distinctness there.
• "Path table reports only betas and stars" → add CIs and fit indices; report the indirect-effect CI.
• "Figure doesn't show levels" → label L1/L2 and the cross-level paths in the diagram.
• "Table 3 doesn't match the text" → regenerate both from the single deposited script.
• "Forest plot has no weights / CI scale" → add study weights, CIs, and the pooled estimate with its CI.

Exhibit calibration anchors

• Table 1 is a credibility signal: a clean, complete descriptives/correlation table with reliabilities reassures reviewers before they reach the model.
• The path figure should let a hurried editor reconstruct the hypotheses, the levels, and the result from the exhibit alone.
• Accessibility is part of rigor: colorblind-safe palettes and grayscale-legible encodings so evidence survives printing and reuse.

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. JAP is organizational psychology — multilevel survey/field data and experiments; cluster at the right level and apply mediation/moderation discipline.

• 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 correlation table without reliabilities or level-of-analysis notes
• Path tables/figures reporting only coefficients and stars (no CIs, no fit)
• Path diagrams that hide level of analysis or which paths are significant
• Exhibits requiring the prose to be intelligible (not self-contained)
• Exhibit values that don't match the shared analysis script

Output format

【Table 1】means/SD/correlations + reliabilities + level notes? [Y/N]
【Main exhibit】path model/table with estimates + CIs + fit indices? [Y/N]
【APA 7th + self-contained?】[Y/N]
【Article vs supplement】split decided
【Reproducible + accessible?】matches script, grayscale/colorblind-safe? [Y/N]
【Next】joap-writing-style

Supplementary resources

• ../../resources/external_tools.md — papaja, semPlot/tidySEM, metafor plots, APA table tooling
• ../../resources/official-source-map.md — APA 7th style and exhibit expectations