Researchers Propose Two-Team Cross-Screening To Assess Replicability
Roy et al. (2025) presented at the Harvard Data Science Initiative propose a two-team cross-screening design that nonrandomly splits observational data and assigns separate discovery and confirmation teams to assess replicability across distinct subgroups. They apply the approach to the Wisconsin Longitudinal Study examining unwanted pregnancy effects and argue nonrandom splitting (e.g., Catholic versus non-Catholic women) strengthens robustness checks against unmeasured confounding and multiplicity.
Key Points
- 1Introduce two-team cross-screening that splits data nonrandomly for discovery and confirmation
- 2Show replicability detects robustness against different unmeasured confounders across populations
- 3Enable practitioners to combine discovery and FWER-controlled confirmation within one observational dataset
Scoring Rationale
Novel, directly applicable methodological advance for observational studies, limited by single-study seminar-level presentation.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
