A six-trial, source-verified investigation: stroke counts look straightforward, until person-time and prevention populations change the interpretation. Make seven methodological decisions before the answer key appears.
Evidence boundaries: Based on historical metadat::dat.hart1999 rows, not invented patient outcomes. This open-book self-check is not an independently blinded, psychometrically validated competence assessment, a complete systematic review or current clinical guidance. No personal/clinical data are uploaded.
Language status: English-only provisional activity. The existing twelve Synthesis course-language routes remain available; this English-only new content has not been independently reviewed or translated.
Source dossier: six archived trials
The dataset distinguishes participants (N) from exposure time (person-years). Warfarin is listed first, then control. Competing baseline risks and intervention comparability cannot be adjudicated from this table alone.
Metadat Hart 1999 archive counts and exposure times
Choose one response per question. No feedback is revealed until all seven answers are submitted. Because the source page is public, this is an open-book new-data task, not a controlled unseen examination.
Worked source-based feedback
Statistical scope: the browser reproduces a common-effect inverse-variance log-incidence-rate-ratio calculation. The official metafor six-trial REML example happens to estimate τ² = 0; this is not a claim that common-effect and random-effects methods are generally interchangeable. A different effect measure (incidence rate difference) yields a different I².
Continue in the Research Studio or visit AllMeta to explore available analysis tools. Do not paste person-years into binary-event participant denominator fields; check that an app supports the required incidence-rate estimand before importing.