From jle-skills
Builds a targeted robustness suite for JLE manuscripts: specification, sample, jurisdiction, and inference checks to address referee concerns about stability and researcher degrees of freedom.
How this skill is triggered — by the user, by Claude, or both
Slash command
/jle-skills:jle-robustnessThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
- The main estimate of a legal/regulatory effect is in hand and you must show it is not an artifact of one specification
JLE referees — economists who know the institution — probe whether the estimated effect of the rule is stable, honestly inferred, and not the product of researcher degrees of freedom, with special attention to legal-design choices: how you dated the rule, which jurisdictions you treated as controls, whether enforcement was uniform. Robustness here is not a wall of regressions; it is a targeted set of checks each tied to a specific threat to the legal identification. Map every plausible objection to the one check that answers it, and show the point estimate barely moves.
| Threat to the result | The check that answers it |
|---|---|
| Omitted confounders | Oster δ / coefficient-stability bounds; controls added in steps |
| Wrong treatment date | re-date to signing vs. effective vs. enforcement onset; donut around the date |
| Contaminated control jurisdictions | drop jurisdictions with contemporaneous reforms; alternative donor pools; placebo on uncovered legal areas |
| Specification search | a specification curve; declare the primary spec up front |
| Functional form | levels vs. logs, alternative outcome/penalty definitions, nonparametric version |
| Sample / jurisdiction selection | leave-one-state-out, balanced vs. unbalanced panel, drop the largest jurisdiction |
| Inference too narrow (few jurisdictions) | cluster at the legal-variation level; wild-cluster bootstrap; randomization/permutation inference |
| Design-specific fragility | DiD: honest-DID bounds; RD: bandwidth/donut; IV: weak-IV-robust set |
Keep the robustness section about whether the estimate of the legal effect is stable, and do not let it drift into re-arguing the rule's normative merits. A referee wants to know the number survives re-dating, control swaps, and correct inference — not your view on whether the rule is good policy. Park the welfare and policy discussion in its own section (see jle-theory-model) so the robustness evidence reads as clean, mechanical stress-testing of the identified effect.
A DiD estimate that an entry-licensing law raised consumer prices is 6% (s.e. 2). The robustness suite: (i) re-dating from the statute's signing to its effective date shifts the estimate trivially (6.1%); (ii) dropping the three states with simultaneous occupational-licensing reforms leaves it at 5.7%; (iii) a leave-one-state-out sweep stays within [5.2%, 6.4%]; (iv) Oster δ implies selection on unobservables would need to be 2.1× selection on observables to nullify it; (v) with 11 treated states a wild-cluster bootstrap keeps the 95% interval away from zero, whereas naive clustering over-rejects; (vi) a placebo on an unlicensed adjacent service is null. The point estimate barely moves — the JLE target.
Most JLE empirical designs exploit variation across a small number of legal units — 50 states, a dozen circuits, a handful of countries, one agency's enforcement regions. With few clusters, conventional clustered standard errors over-reject, so a result that looks significant may not survive correct inference. Treat this as the baseline expectation, not a corner case:
【Primary spec】declared? [Y/N] — estimate: ___ (s.e. ___)
【Threat → check map】confounders: ___ | date: ___ | controls: ___ | form: ___ | sample: ___ | inference: ___ | design: ___
【Inference】clustering level: ___; few-cluster method: ___
【Design sensitivity】honest-DID / RD bandwidth / weak-IV set: ___
【Estimate stability】range across checks: [___, ___]; checks that move it: ___
【Next step】jle-tables-figures
npx claudepluginhub brycewang-stanford/awesome-journal-skills --plugin jle-skillsOrganizes robustness and sensitivity checks for JLEO manuscripts by targeting institutional threats (selection, endogeneity, measurement, alternative mechanisms) with specific defusal tests.
Organizes robustness checks for JEEA manuscripts around threats a general-interest referee would raise, ensuring headline results are stable to specification, sample, inference, and assumption perturbations.
Builds robustness batteries and falsification logic for JPE manuscripts whose main result rests on a single specification. Runs specification checks, mechanism discrimination tests, and structural sensitivity analysis.