jeem-tables-figures — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited jeem-tables-figures (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Environmental economics is inherently spatial and dynamic, so JEEM exhibits should exploit geography and time, not just stack coefficients. The reader — and the referee — should be able to see the identifying variation and the welfare magnitude from the exhibit alone. Lead with the figure that shows the design (an event study, a map of treatment, an RD plot, a WTP density); relegate the full regression table to support it.
| Result type | The exhibit that sells it | Common failure to avoid |
|---|---|---|
| Staggered regulation DiD | event-study plot with leads/lags + CIs | a single DiD coefficient with no dynamics |
| Spatial treatment / exposure | a clean map of treatment and the outcome gradient | a table that hides where the variation is |
| RD in a standard/threshold | binned RD plot with fitted polynomials | over-fitted high-order polynomial, no bins shown |
| Hedonic capitalization | coefficient + implied $ WTP per unit, with CI | a raw coefficient with no welfare translation |
| Stated-preference WTP | WTP distribution / CV curve with uncertainty | only utility coefficients, no compensating variation |
| Damage / dose-response | dose-response curve with CI band | a single linear slope hiding nonlinearity |
Because environmental variation is spatial, a map often does identification work that a table cannot: it shows where treatment fell, whether treated and control units are comparable, and whether the outcome gradient lines up with the exposure. Treat maps with the same rigor as regression tables. State the coordinate reference system and the unit of aggregation; choose a perceptually uniform, colorblind-safe scale (not rainbow); show treatment and outcome on comparable geographies and at a scale where the reader can see the variation. A sloppy map that exaggerates a gradient is worse than no map, because referees will distrust the rest of the exhibits.
Environmental damages and valuations are often nonlinear — pollution-mortality concentration-response curves, threshold effects in temperature, diminishing WTP in scope. A single linear slope hides exactly the structure a regulator needs (where marginal damage steepens, where a standard should bind). When the relationship is plausibly nonlinear, show the dose-response curve with a confidence band rather than reporting one coefficient, and let the exhibit reveal the shape. A flexible specification that turns out linear is reassuring; a linear specification that hides a kink is a missed contribution and an easy referee point.
Generate exhibits from the fitted result, not by retyping numbers. Full map: execution-with-mcp. JEEM is environmental economics — policy/regulation designs and non-market valuation; the causal chain serves its program-evaluation lane.
etable (multi-model) or did_summary_to_latex straight from the result_id.plot_from_result / enhanced_event_study_plot / event_study_table —axis units and the SE/clustering note baked in.
See a full fitted-result → exhibit chain in the JF execution walkthrough.
A regulation-DiD paper reports a single coefficient: "treatment lowered PM2.5 by 8% (s.e. 2%)." A referee cannot tell whether pre-trends were flat or whether the effect grew over time. The JEEM fix: replace the lone coefficient with an event-study figure showing four pre-period leads (flat, near zero) and the post-period dynamics (the effect builds over three years to 8%), each with a 95% CI band, clustered spatially. The same table now reports the implied avoided-mortality value in dollars using a stated VSL, so the welfare magnitude is legible without leaving the exhibit. A companion map shows treated vs. control counties, making the identifying variation visible at a glance.
The single most common JEEM exhibit failure is leaving the result as a regression coefficient. A coefficient on log price, a hazard ratio, or a logit utility weight is not a welfare statement. Every headline exhibit should carry the conversion the reader needs: a hedonic coefficient → implied $ WTP per unit of the amenity; a mortality coefficient → avoided deaths and their valuation; a discrete-choice utility weight → compensating variation. Put the conversion and its CI in the table note or a dedicated column so the referee never has to do the arithmetic — and so the welfare claim is auditable on the page.
Elsevier house style permits significance stars, and JEEM does not ban them — but at this journal the persuasive object is the magnitude and its uncertainty, not the asterisk. Lead with the point estimate and its confidence interval in interpretable (welfare) units; let stars, if used, be a secondary cue. A table where the reader's eye lands on a row of asterisks instead of on a dollar damage or a WTP has buried the contribution. The discipline is the same as in the robustness suite: show that the welfare-relevant number is both economically meaningful and precisely estimated, and let the reader judge significance from the interval.
Order the exhibits so a referee can follow the argument from the figures alone: (1) a descriptive map or time series that establishes the environmental setting and where/when the variation occurs; (2) the design exhibit — event study, RD plot, first-stage, or balance table — that earns identification; (3) the headline result with its welfare translation; (4) the robustness exhibit showing the welfare number is stable; (5) any heterogeneity or distributional exhibit that carries the policy implication. This sequence mirrors how the paper should read and pre-empts the referee who skims figures first. Avoid front-loading a dense regression table before the reader has seen the variation.
【Journal】Journal of Environmental Economics and Management
【Skill】jeem-tables-figures
【Lead exhibit】event study / map / RD plot / WTP density — shows identifying variation? [Y/N]
【Welfare units】damage $ / $ WTP / % capitalization present in exhibit? [Y/N]
【Uncertainty】CIs/SEs shown; stars not substituting for magnitude? [Y/N]
【Spatial honesty】projection/unit/color scale stated? [Y/N]
【Self-contained notes】sample/unit/SE/estimand in every note? [Y/N]
【Source status】verified URL / 待核实 / not asserted
【Next skill】jeem-writing-style~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.