Use when defending the research design of an American Sociological Review (ASR) manuscript — causal/observational inference for quantitative and demographic work, case selection and comparison for comparative-historical work, site and informant logic for ethnography, and network/computational designs. ASR judges each tradition on its own terms. Strengthens the design; it does not write code.
SkillScore23Maintained
Automatic
◆Research pack
Last verified 8/15/2026
What it changes
Passed 5 of 21 test cases with the skill, and 3 of 21 without it.
Without skill14%
With skill23%
Checks it passed that the model alone missed
Includes the exact header tag 【What it identifies】
Includes the exact routing tag 【Next】asr-data-analysis
Includes the exact header tag 【Key assumption(s) or selection logic】
and 3 more in the full report
3 fixed2 passed either way16 still failing1 broke
Pass-rate lift
+9 pp
Tokens
+29%
Agent turns
0%
Response time
-28%
Model gemini-3.6-flashMeasured 8/15/2026v15 of 22 cases passed with skill · 21 comparable
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 2 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 2
Prompt
My historical sociology study explores revolutionary outcomes across Latin American agrarian regimes. I plan to present my narrative as showing necessary conditions without discussing what evidence would disprove my thesis. How should I structure my research design?
What was checked, without and with the skill
Checked
Without skill
With skill
Includes the exact header tag 【Robustness / negative cases / sensitivity】
Missed
Passed
Requires specifying disconfirming evidence or counterfactual conditions that would falsify the argument
Missed
Passed
Includes the exact routing tag 【Next】asr-data-analysis
Missed
Passed
case-07 · gemini-3.6-flash · 8/15/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md129 linessingle file
Research Design (asr-research-design)
ASR welcomes many methods but is exacting about each. The design must credibly link the argument
(asr-theory-building) to evidence and rule out the leading alternative. Pick the section matching
your method.
When to trigger
Specifying identification, case selection, sampling, or site/informant logic
A reviewer questioned causal claims, generalization, selection, or a confound
Justifying why your design adjudicates the rival account from asr-literature-positioning
Quantitative / demographic
Be honest about what the design identifies. Much of sociology is observational; distinguish
description, association, and causation. If causal, state the assumptions (ignorability, parallel
trends, exclusion) and defend them.
Designs: panel/fixed-effects, DID/event study (modern staggered estimators, not naive TWFE),
IV, RDD, matching/weighting with balance and sensitivity; decomposition for inequality; event
history for timing; age-period-cohort for demographic change.
Inference & sampling: respect complex survey design (weights, clustering, strata); cluster at
the right level; report uncertainty.
Sensitivity: how strong must an unobserved confounder be to overturn the result?
Comparative-historical
Case selection by design logic (most/least-likely, paired comparison, deviant) — say what each
case is a case of; justify the comparison and the counterfactual.
Specify the causal form (necessary/sufficient conditions, conjuncture, path dependence; QCA where
apt) and what evidence would have disconfirmed the argument.
Ethnographic / interview
Site and informant selection justified theoretically, not by access alone; state positionality
and access conditions.
Plan how depth, saturation, and negative cases are handled; how claims map to observed evidence.
Ethics/IRB, consent, and confidentiality planned from the start (see asr-data-and-transparency).
Network / computational
Define boundary specification, missing-ties, and the generative process (e.g., ERGM/SAOM logic).
For computational text/ML: validate against human-labeled samples; report stability.
Use when defending the research design of an American Journal of Political Science (AJPS) manuscript — causal identification for observational work, experimental and survey-experimental design, formal-empirical linkage, or case-based inference. AJPS reviewers are quantitatively demanding, so identification must license the claim being made. Strengthens the design; it does not write code.