Open research data · Launched August 6, 2026
Download real, openly licensed datasets prepared to run immediately in ERN Meta-Analysis Studio. Every curated file identifies its original source, associated publication, license, and the changes ERN made for compatibility.
Ready for analysis
Curated launch collection
These files were curated from the openly distributed metadat collection. ERN Institute reformatted them for compatibility but does not claim ownership of the original observations.
Thirteen studies comparing tuberculosis-positive cases in vaccinated and non-vaccinated groups.
Source: Colditz et al. (1994), distributed as dat.colditz1994 in metadat. ERN calculated group totals from the original event and non-event counts.
Nine studies comparing hospital stays under specialized stroke-unit care and routine care.
Source: Normand (1999), distributed as dat.normand1999. Original means, standard deviations, and sample sizes are retained; negative effects favor specialized care.
Twenty correlations from multisection courses examining whether instructor ratings track student achievement.
Source: Cohen (1981), distributed as dat.cohen1981. Correlations and section-level sample sizes are retained; independent samples sharing a citation are distinguished.
Sixteen studies examining associations between conscientiousness and adherence to prescribed medication.
Source: Molloy et al. (2014), distributed as dat.molloy2014. Correlations and sample sizes are retained, with measurement and quality details added to the notes.
One real dataset containing continuous and binary outcomes. Use Run all different effect-size metrics to fit both models and display both complete result sets.
Source: Gibson et al. (2002), distributed as dat.gibson2002. ERN mapped the continuous and event-count outcomes into one Studio-compatible file. One zero event cell appropriately triggers the Studio’s disclosed continuity-correction warning.
Author-submitted datasets are reviewed before publication and will be identified with the badge above. Submission is optional and entirely separate from files loaded privately in the Studio.
Transparent provenance
ERN-ready files retain the source observations. Preparation is limited to schema mapping, stable effect identifiers, compatible group totals where required, and explanatory notes. The Studio may produce different results from a publication when the original analysis used another effect-size measure, model, estimator, continuity correction, or inference method.
The launch collection is redistributed under GPL-2.0-or-later, matching the metadat source package. Cite both metadat and the original publication listed on each source page.
