From jleo-skills
Organizes robustness and sensitivity checks for JLEO manuscripts by targeting institutional threats (selection, endogeneity, measurement, alternative mechanisms) with specific defusal tests.
How this skill is triggered — by the user, by Claude, or both
Slash command
/jleo-skills:jleo-robustnessThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
- The main result holds but a referee will ask "is this an artifact of the sample / specification / measurement?"
JLEO referees — institutional and organizational economists — do not reward a wall of robustness tables. They reward a robustness program organized by the specific threats to the institutional claim. For each headline result, name the threats, then show the check that defuses each. The institution-specific threats cluster as follows:
| Threat to the institutional claim | The check that defuses it |
|---|---|
| Selection into the governance form / institution | IV / selection model / matching robustness; bound the selection (Oster δ) |
| Endogenous reform timing or anticipation | event-study leads; placebo timing; controls for pre-trends; alternative timing windows |
| The institutional construct is mismeasured | alternative proxies for the construct; measurement-error bounds; cross-validate the measure |
| An alternative mechanism explains the pattern | a test that separates your mechanism from the rival (heterogeneity prediction, falsification) |
| Inference is fragile (few clusters, serial correlation) | wild-cluster bootstrap; alternative clustering levels; randomization inference |
| Functional form / spec mining | a small, principled spec curve; pre-specified controls; not 40 arbitrary specs |
A paper claims judicial independence reduces expropriation of firms. Threats: (1) independence is endogenous to economic conditions; (2) the independence index is noisy; (3) general rule-of-law improvement, not independence specifically, drives the result. The threat-organized program: an event study around a constitutional reform isolating timing (threat 1); two alternative independence indices that agree (threat 2); and a placebo on a rule-of-law dimension that the mechanism says should not matter for this outcome (threat 3). Each table is captioned by the threat it kills, not by its position in the appendix.
【Headline result】[...]
【Threats to the institutional interpretation】1) ___ 2) ___ 3) ___
【Check per threat】threat 1 → ___ ; threat 2 → ___ ; threat 3 → ___
【Construct validity】alternative measure(s): ___
【Mechanism vs. rival test】[...]
【Inference robustness】wild-cluster / randomization / alt-clustering: ___
【Selection bound】Oster δ or sensitivity: ___
【Next skill】jleo-tables-figures
npx claudepluginhub brycewang-stanford/awesome-journal-skills --plugin jleo-skillsBuilds a targeted robustness suite for JLE manuscripts: specification, sample, jurisdiction, and inference checks to address referee concerns about stability and researcher degrees of freedom.
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.
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.