Jru Tables Figures

Tables and Figures (jru-tables-figures)

When to trigger

• Choice data are buried in a wall of cells when a figure could show the pattern (the fourfold pattern, an Allais reversal, an Ellsberg gap)
• A structural table lists parameters with no sense of which moment or task identifies them
• A probability-weighting function or value function is described in prose instead of plotted
• Standard errors, confidence intervals, or the estimation method behind each parameter are missing or inconsistent

What JRU exhibits must do

JRU audiences read for decision-theoretic content: they want to see the shape of the weighting function, the location of the reference point, the size of the ambiguity premium, and how precisely each parameter is pinned down. The exhibit's job is to make the behavioral pattern and the parameter estimate immediately legible, with honest uncertainty.

Figures — the workhorses of this field

• Plot the estimated functions, not just their parameters. A Prelec or Tversky–Kahneman w(p) and a reference-dependent value function are far more informative drawn than tabulated; overlay the EU benchmark (the 45° line for w(p)) so the deviation is visible.
• Show the choice pattern directly where it is the point: certainty-equivalent curves, switching-point distributions in a price list, the fourfold pattern of risk attitudes, the Ellsberg matching-probability gap.
• Confidence bands, not just point curves — show estimation uncertainty around the fitted function.
• For experiments, a CONSORT-style flow (recruited → completed → analyzed) and a balance summary belong up front.

Tables — discipline

• Each structural parameter row should make clear what identifies it (which task/moment) and report a standard error or CI and the estimation method.
• Report standard errors / confidence intervals as the primary inferential object; do not let asterisks substitute for showing precision and magnitude.
• Separate structural parameters (curvature, weighting, ambiguity index, λ) from nuisance / control estimates so the reader's eye lands on the primitives.
• Units and stakes stated (lab currency vs. real money, per-period vs. lifetime) so a VSL or premium is interpretable.

Execution bridge (StatsPAI / Stata MCP)

Generate exhibits from the fitted result, not by retyping numbers. Full map: execution-with-mcp. JRU spans decision experiments and applied risk; randomization inference for experiments, DiD/IV for observational claims.

• Tables: etable (multi-model) 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 magnitude in interpretable units.

See a full fitted-result → exhibit chain in the JF execution walkthrough.

Checklist

• The probability-weighting / value function (or ambiguity premium) is plotted, with the EU benchmark overlaid
• Confidence bands accompany any fitted function or key estimate
• Each structural parameter shows its identifying task/moment, an SE/CI, and the estimation method
• The headline behavioral pattern (fourfold / Allais / Ellsberg) is shown, not just described
• Experiment exhibits include sample flow and balance
• Magnitudes are interpretable: stakes, units, and currency stated
• Precision is shown via SE/CI rather than asterisks alone

Reading the exhibit as a referee would

Before finalizing, look at each exhibit cold and ask the questions a JRU referee asks:

• Can I see the deviation from EU without reading the text? (If the w(p) plot has no 45° line, no.)
• Can I tell how precise each estimate is? (If only stars, no.)
• Can I tell what identifies each parameter? (If the table has no identifying-task column, no.)
• Can I convert the magnitude to something interpretable — dollars, an elasticity, a probability? (If units are "points," no.)

An exhibit that fails any of these sends the referee back to the prose to reconstruct what the figure should have shown — friction that costs goodwill in review.