First release of the standalone zcurve3 package. Keeps the z-curve 2.0
power estimates (EDR/ERR/FDR) and adds:
- Effect-size estimation & heterogeneity from a distribution-free ncp
mixture (yi/sei), with a floor correction for SE-driven apparent tau. - Prediction interval for a new study — the range of true effects a new
study from the same literature could show, folding in sampling error in
both the mean and tau. - Directional & clustered analyses — one-sided local power; cluster bootstrap.
- Rewritten forest plot — adjusted per-study effects with bootstrap
intervals, auto-sizing to a file.
Quality: R CMD check --as-cran clean; 89 testthat tests; CI on Windows,
macOS, and Linux (R release, devel, oldrel-1).
Install: remotes::install_github("UlrichSchimmack/zcurve3@v3.0.0")
Authors: Ulrich Schimmack, František Bartoš, Jerry Brunner.