Psci tables figures

Tables & Figures (psci-tables-figures)

In Psychological Science, exhibits carry a lot of the argument because the prose is so tight. They use APA 7th-edition conventions and — unlike many journals — are embedded in the main text near where they are discussed, not collected at the end. Good figures also help show effect sizes and uncertainty that the format leaves little room to describe.

When to trigger

• Designing the main results figure/table
• Deciding what goes in the article vs. supplemental online material
• A reviewer found an exhibit unclear or non-APA
• Visualizing distributions, individual data, and uncertainty (not just means)

Principles

1. Show the data and the uncertainty. Prefer figures that display the distribution or individual points with means and confidence intervals (e.g., violin/dot + interval) over bar-of-means plots that hide variability. Magnitude and uncertainty should be visible.
2. Self-contained + APA 7th. Titles, notes, axis labels, Ns, and units make each exhibit intelligible alone; follow APA table/figure formatting.
3. Embedded placement. Place tables and figures within the text near their first discussion, per the journal's instruction (an exception to the APA end-placement default).
4. Earn the space. The tight format means every exhibit must add evidence; push secondary exhibits (full models, manipulation checks, extra studies) to the supplemental online material.
5. Reproducible + accessible. Generated by the shared analysis script so values match; colorblind- safe and grayscale-legible.

Worked micro-example — the main results figure (illustrative)

For the two-study attention package, the primary figure must carry the argument the prose cannot.

Figure 1. Attentional-capture cost by condition, Studies 1–2.
  Geometry:  per-participant points + group mean + 95% CI (dot/interval
             or violin), NOT a bar of means.
  Panels:    Study 1 (N = 240) | Study 2 (N = 300, direct replication).
  Annotation: d = 0.34, 95% CI [0.08, 0.59] (S1); d = 0.29 [0.06, 0.51] (S2).
  Note:      defines the capture-cost metric, Ns, exclusion count, and that
             intervals are 95% CIs — readable without the main text.
  Placement: embedded at first discussion (APA 7th override of end-placement).
  Source:    rendered by the deposited R script so values match exactly.

Exhibit triage — article vs. supplemental online material

Exhibit	Home	Reason
Primary effect with CI across studies	main text	this is the contribution
Full regression / model table	supplement	costs words, secondary to the estimate
Manipulation / attention-check results	supplement	needed for credibility, not the headline
Robustness / specification grid	supplement	summarized in one main-text sentence
Extra studies beyond the core two	supplement	keep the article focused

Exhibit-stage reviewer pushback and the venue fix

• "Bar chart hides the spread" → switch to dot/violin + 95% CI; show individual data where N allows.
• "I can't read the figure without the text" → make the note self-contained (metric, Ns, what the error bars are).
• "Table reports only stars" → add effect-size and CI columns; this is a post-reform expectation here.
• "Figure values don't match Table 2" → regenerate both from the single deposited script.

Exhibit calibration anchors

• Because the prose budget is so tight, the figure is often where magnitude and uncertainty actually live; design it to stand alone if a hurried editor reads only the exhibits.
• Show the data generating process where feasible: raw points behind the mean, paired lines for within-subject designs, and the interval that the claim depends on — not a decorative error bar.
• Embedding near first discussion is an explicit instruction here and a frequent reason manuscripts are returned; verify the current placement rule against the journal's submission guidelines, but never default to APA end-of-document placement.
• Accessibility is part of credibility: colorblind-safe palettes and grayscale-legible encodings so the 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. Psychological Science is short-format experimental psychology with strong open-science norms; preregister, run randomization inference, and report effect sizes with family-wise corrections.

• 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

• Bar plots of means that hide the distribution and uncertainty
• Exhibits requiring the prose to be intelligible (not self-contained)
• Tables/figures dumped at the end rather than embedded near discussion
• Reporting only stars/p-values in tables with no effect size or CI
• Exhibit values that don't match the shared analysis script

Output format

【Main exhibit】what it shows + why a figure/table
【Shows distribution + uncertainty?】(CIs / individual data) [Y/N]
【APA 7th + self-contained + embedded near text?】[Y/N]
【Article vs supplement】split decided
【Reproducible + accessible?】matches script, grayscale/colorblind-safe? [Y/N]
【Next】psci-writing-style

Supplementary resources

• ../../resources/external_tools.md — plotting tools and papaja for APA exhibits
• ../../resources/official-source-map.md — APA 7th style and embedded-exhibit instruction