Building Paper Screening Rubrics

Automated paper screening uses defined relevance criteria to filter papers at scale. Build a screening rubric by collaboratively designing clear, testable criteria with AI guidance, then apply it to hundreds of papers consistently and accurately, eliminating manual review bottlenecks.

Relevance depends on your research context and criteria. This skill helps you brainstorm and articulate what "relevant" means for your topic—capturing nuances like methodology, scope, and subject matter—then documents those rules in a machine-readable format for reproducible screening.

Test-driven refinement validates your rubric on a small sample of real papers before bulk processing. Run your criteria against sample papers, identify misclassifications, refine the rules iteratively, and verify accuracy until the rubric consistently identifies papers you need.

Yes. Screening rubrics export as structured JSON files that encode your relevance criteria in automation-ready format, enabling integration into screening pipelines and batch processing tools while maintaining reproducibility and versioning of your decision rules.

Use a screening rubric for literature searches of 50+ papers where consistency and speed matter. Manual review becomes error-prone and time-intensive at scale; a rubric captures your criteria once, applies them uniformly, and scales to hundreds of papers without fatigue-driven inconsistency.

This skill generates a screening-criteria.json file that documents your rubric rules, scoring logic, and decision thresholds in structured format. Versioning this file alongside your research ensures reproducibility, enables team collaboration, and lets you track how criteria evolved across rescreening cycles.