Citing PyMetaAnalysis

PyMetaAnalysis does not yet have a DOI or software paper. Until one is available, cite the exact software version and repository used in the analysis.

The repository includes machine-readable CITATION.cff metadata. GitHub's Cite this repository control can export it to common bibliographic formats. Release maintainers keep its version synchronized with the package metadata.

Retrieve the installed version with:

import meta_analyze as ma

print(ma.__version__)

A provisional citation can use:

PyMetaAnalysis contributors. PyMetaAnalysis (version <version>):
a pandas-first meta-analysis library for Python.
https://github.com/ZhaoboDing/PyMetaAnalysis

Include the access or release date required by the selected citation style. When citing a tagged release, link to that release rather than an unpinned branch.

Report the analysis, not only the package

A reproducible publication should also state:

  • effect measure and treatment/control direction;
  • common-effect or random-effects model;
  • pooling and tau-squared estimator;
  • confidence-interval and prediction-interval methods;
  • continuity-correction and RD boundary policies where applicable;
  • included/excluded studies and reasons;
  • PyMetaAnalysis, Python, NumPy, pandas, and SciPy versions;
  • code and data availability.

result.method_details() provides a starting point for method prose, and result.report() captures resolved methods, diagnostics, provenance, warnings, and optionally the row-level table. Review generated text against the protocol and journal requirements before publication.