Tables & Figures (eursr-tables-figures)
Exhibits are where a double-blind reviewer checks whether the comparative or longitudinal result is
real. ESR excludes tables and figures from the ~8,000-word count (endnotes and references count) —
so a clear exhibit costs nothing against length, but it must earn its place and stand on its own.
When to trigger
- Designing the main results table/figure or a key descriptive/cross-national exhibit
- Deciding what belongs in the article vs. the (supplementary) appendix
- A reviewer found an exhibit unclear, mislabeled, or not self-contained
- Presenting a cross-level interaction or country-level pattern visually
Principles
-
Self-contained. Title, row/column labels, and a complete note make each exhibit intelligible
alone. State the data source (ESS/EU-SILC/SOEP/EVS or register), sample, N (and number of countries),
units, weighting, estimator, and what the estimate is.
-
Show magnitude and uncertainty. Effect sizes and intervals — not stars alone. Predicted-
probability and marginal-effects plots usually beat dense coefficient walls for a comparative
reader, and coefficient/forest plots convey cross-country variation better than tables.
-
Comparative & longitudinal exhibits. Country forest plots / caterpillar plots of random effects;
interaction plots showing the cross-level moderation; mobility/transition tables; survival curves and
cumulative-incidence plots for event history; growth-trajectory plots for panels.
-
Accessible. Colorblind-safe palettes; legible in grayscale; no chartjunk or 3D.
-
Reproducible. Generated by the master script so numbers match the deposited replication package
(see
eursr-transparency-and-data).
Format
- Follow OUP/ESR table and figure conventions; concise notes carrying all interpretive detail.
- Keep identifying information out of exhibits (double-blind review).
- Tables and figures are excluded from the ~8,000-word limit but must still be necessary and clear.
Exhibit conventions a double-blind ESR reviewer expects
| Result type |
Workhorse exhibit |
The note must state |
| Cross-national effect |
country forest/caterpillar plot |
data, N, # countries, estimator |
| Cross-level interaction |
predicted-margins interaction plot |
what is held constant, CI basis |
| Mobility / attainment |
transition / mobility table |
origin–destination coding, N |
| Event history |
survival / cumulative-incidence curve |
risk set, time scale, censoring |
| Panel / growth |
trajectory plot with CIs |
within vs. between, attrition handling |
Worked micro-example (illustrative)
A main results table for a 24-country scarring study is redesigned for ESR.
Before: 6 columns of multilevel coefficients, stars only, no note → reviewer can't tell magnitude,
sample, or number of countries
After: (1) a marginal-effects interaction plot — scar shrinks as activation spending rises (CI band);
(2) a country caterpillar plot of random slopes showing where the effect is strong/weak
Self-contained note (illustrative): "Predicted within-person wage penalty from a two-level model,
N = 86,000 in 24 countries, survey-weighted; bands are 95% CIs (wild cluster bootstrap);
reproduces from master.R, seed = 2026."
The redesign stands alone, leads with the substantive magnitude and the cross-national variation, and
ties the numbers to the deposited script.
Referee pushback → ESR-specific fix
-
"I can't read this table without the text." → Add a complete note (source, N, # countries, units,
weighting, estimator) so it is intelligible alone.
-
"Stars don't tell me if this matters." → Replace with marginal effects and intervals; for a
comparative claim, show the country-level spread with a forest/caterpillar plot.
-
"Where is the cross-level interaction?" → Plot predicted margins across the macro variable rather
than reporting a bare interaction coefficient.
-
"Your figure dies in grayscale." → Re-encode with colorblind-safe, grayscale-legible channels.
Calibration anchors
-
Exhibits are excluded from the word cap — use the room. ESR counts endnotes and references but not
tables/figures, so a clear, self-contained exhibit costs nothing against length.
-
Show the comparison. A country forest/caterpillar plot communicates cross-national variation a
generalist can grasp far better than a coefficient column.
-
Numbers must match the replication package. Exhibit values that disagree with the deposited code
read as a credibility failure under ESR's replication mandate.
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. ESR is comparative quantitative sociology; cross-country panels with confounded institutions — foreground fixed effects and clustering.
-
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
- Tables that require the prose to be intelligible (not self-contained)
- Significance stars with no effect size or interval
- Reporting a bare cross-level interaction coefficient with no predicted-margins plot
- Color-only encoding that fails in grayscale or for colorblind readers
- Exhibit values that don't match the analysis script / replication package
Output format
【Main exhibit】what it shows + why a table/figure
【Self-contained?】title + labels + note + source/N/# countries/weights present? [Y/N]
【Magnitude + uncertainty shown?】[Y/N]
【Comparative variation shown?】(forest/interaction plot where relevant) [Y/N]
【Reproducible?】matches master script / replication package? [Y/N]
【Next】eursr-writing-style
Supplementary resources
