From jde-skills
Stress-tests causal identification strategies (RCT, DID, IV, RDD) for Journal of Development Economics manuscripts in low- and middle-income settings.
How this skill is triggered — by the user, by Claude, or both
Slash command
/jde-skills:jde-identification-strategyThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
- The empirical core is OLS + controls with an undefended causal claim
Development economics has been a leader in the credibility revolution, and JDE referees apply a demanding identification standard while respecting that field settings are messy. The implicit credibility ranking (strong → weaker):
A theoretical paper is judged on the development relevance and rigor of its mechanism, not on a research design. A paper with novel, hard-to-assemble data answering a first-order development question can carry reduced-form evidence — but the question and the data discipline must be exceptional.
JDE also runs a permanent pre-results review / Registered Reports track: a prospective design (hypotheses, procedures, statistical analysis plan, power analysis, pilot data if applicable) can be reviewed and accepted in principle before results exist. If your design is prospective, build it to that standard — it both strengthens identification and unlocks that route (see jde-review-process).
Hypothetical: a national school-construction program rolled out district-by-district over six years — a quasi-experiment exploiting staggered policy timing to estimate effects on years of schooling.
| Referee concern | What clears the JDE bar |
|---|---|
| "Staggered TWFE is biased here" | Modern estimator + Bacon decomposition + event-study plot |
| "Spillovers leak treatment into controls" | Spillover/SUTVA test; bound it; defend the controls |
| "IV exclusion is asserted, not argued" | Theory + institutions + falsification, all three |
| "RDD cutoff may be manipulated" | Density test (McCrary / CJM) + covariate smoothness |
| "One site — why generalize?" | Pin the LATE/population; mechanism evidence that travels |
【Design】RCT / RDD / DID / IV / theory / descriptive
【Identifying variation】one sentence
【Diagnostics done】[PAP, balance, attrition/Lee bounds, density, first-stage F, pre-trends, ...]
【Diagnostics missing】[...]
【Inference】clustering level + few-cluster handling + MHT
【Interpretation】LATE / ATE / external validity note
【Pre-results route?】[Y/N]
【Next step】jde-data-analysis
npx claudepluginhub brycewang-stanford/awesome-journal-skills --plugin jde-skillsStress-tests causal identification strategies (RCT, DID, IV, RDD, event study) against the QJE general-interest bar before tables are drafted.
Stress-tests causal identification arguments (RCT, DiD, RD, IV, shift-share) for AEJ: Applied manuscripts, ensuring designs meet the journal's credibility bar before exhibits are finalized.
Stress-tests causal identification designs (DiD, IV, RDD, experiment) for EER manuscripts, ensuring credibility before finalizing exhibits.