Tables & Figures (humrel-tables-figures)
When to trigger
- Your exhibits exist but a reader cannot trace the inference from data to theory
- A qualitative paper has quotes scattered in prose but no data-structure or evidence table
- A process study has no figure showing the temporal/phase structure
- Quantitative tables are dense, asterisk-driven, or report numbers without interpretation
- You are checking exhibits for anonymization before a double-anonymous submission
Exhibits must make the argument visible
At HR, exhibits are part of the theoretical apparatus, not an appendix of proof. The job differs by tradition, but the test is constant: each exhibit should let a reader see the inference that the prose claims.
Qualitative & critical exhibits
-
Data-structure figure (the spine of a qualitative paper). Show first-order codes → second-order themes → aggregate dimensions (Gioia) or the equivalent. This is often the single most scrutinized exhibit; build it from
humrel-data-analysis.
-
Evidence / data-to-theory table. Representative quotations mapped to constructs, so the data-to-theory link is auditable. Quotes here are proof quotes; reserve a few vivid power quotes for the body.
-
Process / phase model. For process theory, a figure showing the sequence, events, and feedback over time — the visual argument of how the phenomenon unfolds.
-
Informant / site table. Roles, counts, durations — establishing the evidentiary base without revealing identifying detail (anonymize site and organization names).
Quantitative exhibits
-
Descriptives + correlations with reliabilities on the diagonal where applicable.
-
Main models in a clean regression table; report standard errors / confidence intervals, and do not let significance asterisks be the result — interpret magnitudes in the text.
-
Interactions plotted, not left as raw coefficients; show simple slopes for moderation.
-
Multilevel results labelled by level; variance components reported where theory is cross-level.
House-style and submission constraints
-
SAGE Harvard (author-date) governs in-text references and notes within exhibits (检索于 2026-06；以官网为准).
- The 13,000-word total cap includes everything; bloated tables and a long appendix compete with the main text for that budget — every exhibit must earn its space.
-
Anonymize all exhibits for double-anonymous review: no organization names, no identifying figures, no author-revealing acknowledgements in notes.
- Figures should be legible in greyscale and at the journal's column width.
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. Human Relations blends critical/qualitative and quantitative work; apply the chain below to its survey / experimental quantitative papers.
-
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.
Checklist
Anti-patterns
- A qualitative paper with no data-structure or evidence table (inference invisible)
- Quote-dumping in tables with no mapping to constructs
- Tables where significance asterisks substitute for substantive interpretation
- Process claims with no figure showing the temporal structure
- Exhibits that reveal the field site or author institution
- An appendix so large it crowds out the main argument under the word cap
Output format
【Journal】Human Relations
【Skill】humrel-tables-figures
【Qual exhibits】data structure / evidence table / process model status
【Quant exhibits】SEs-CIs reported; magnitudes interpreted (yes/no)
【Inference visible】can a reader trace data → theory? (yes/no)
【Anonymization】exhibits clean of site/author tells (yes/no)
【Budget】exhibits justified against 13k cap (yes/no)
【Next skill】humrel-writing-style
