From crim-skills
Defends research designs for criminology manuscripts: causal identification, longitudinal/life-course methods, criminal-career trajectories, place-based experiments, and qualitative case selection. Mode-aware and tailored to journal standards.
How this skill is triggered — by the user, by Claude, or both
Slash command
/crim-skills:crim-research-designThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
*Criminology* accepts many methodologies but is demanding about each. The design must credibly connect
Criminology accepts many methodologies but is demanding about each. The design must credibly connect
the mechanism (crim-theory-building) to crime evidence. This skill is mode-aware: pick the section
that matches your work and defend it against the strongest rival explanation.
crim-literature-positioningcrim-data-and-transparency).For the single strongest rival theory, write one sentence: "If the rival mechanism were operating instead of mine, the crime data would look like ___; instead they look like ___." If you cannot, the design does not yet identify the contribution.
A defensible Criminology design names the threat reviewers are trained to raise and the move that neutralizes it. Selection into offending and into treatment is the recurring worry.
| Design | Identifying assumption | Threat a referee names | Defensive move |
|---|---|---|---|
| Hot-spot / policing RCT | randomization, no spillover | displacement contaminates controls | measure diffusion vs. displacement |
| Staggered deterrence-policy DID | parallel trends across adopters | bad-comparison TWFE | staggered estimator + pre-trends |
| Life-course turning point | within-person change isolates effect | selection into marriage/work | fixed-effects/hybrid + sensitivity |
| RDD at a sentencing threshold | continuity at the cutoff | manipulation at the line | McCrary density + bandwidth robustness |
A state raises a sentencing penalty in some counties before others. A naive TWFE gives a 9% drop (illustrative); a referee flags invalid comparisons among staggered adopters. Refit with a heterogeneity-robust staggered estimator: flat pre-trends and a credibly identified 4% first-year drop. Cluster at the county (assignment) level; with 14 treated counties add a wild-cluster bootstrap, and note a NIBRS transition could inflate pre-period UCR counts.
【Mode】quant-causal / longitudinal-life-course / place-experiment / qualitative
【Estimand or claim】what is being identified/shown (and within- vs. between-person)
【Crime measure】reported / victimization / self-report + dark-figure note
【Key assumption(s)】and how each is defended
【Rival ruled out】the adjudication sentence
【Robustness/sensitivity】planned checks
【Next】crim-data-analysis
../../resources/external_tools.md — trajectory/survival/spatial packages and longitudinal datasets../../resources/official-source-map.md — Criminology scope and methodological breadthnpx claudepluginhub brycewang-stanford/awesome-journal-skills --plugin crim-skillsGuides analysis and reporting for Criminology manuscripts: count models, longitudinal panels, trajectory models, survival/recidivism, robustness checks, and reproducibility.
Defends research design for APSR manuscripts: causal identification, case selection, process tracing, experimental design, and formal-empirical linkage.
Defends a Social Forces manuscript's research design—causal identification, demographic methods, case selection/process tracing, or network/computational pipelines—to meet reviewer standards.