NEWS.md

In SamCH93/ReplicationSuccess: Design and Analysis of Replication Studies

ReplicationSuccess 1.3.3

• new functions pEdgington, powerEdgington, and sampleSizeEdgington to design and analyze replication studies based on sum of p-values (Edgington's) method (see https://doi.org/10.1098/rsos.240149 for details)
• updated some references

ReplicationSuccess 1.3.2

• new maintainer Samuel Pawel
• Rewrote conversion functions ci2se, ci2estimate, ci2z, ci2p, z2p, and p2z.
• Functions now use vapply() instead of Vectorize()
• Leads to a substantial speedup
• Might lead to changes in output format of these functions
• Fixes behaviour in computation of z <- estimate / se in case where se == 0

ReplicationSuccess 1.3.1.1

• Rewrote conversion functions ci2se, ci2estimate, ci2z, ci2p, z2p, and p2z.
• Functions now use vapply() instead of Vectorize()
• Leads to a substantial speedup
• Might lead to changes in output format of these functions
• Fixes behaviour in computation of z <- estimate / se in case where se == 0

ReplicationSuccess 1.3.1

• Removed functions from exports:
• .PPpSceptical_
• .T1EpSceptical_
• .effectSizeReplicationSuccess_
• .effectSizeSignificance_
• .levelSceptical_
• .p2z_
• .pSceptical_
• .powerReplicationSuccess_
• .powerSignificance_
• .predictionInterval_
• .sampleSizeReplicationSuccess_
• .z2p_

ReplicationSuccess 1.3

• Change of maintainer.
• Transfer ownership of development repository to CRS
• Update links throughout the entire repository following maintainer and ownership change
• Implemented "controlled" level
• Added one-sided p-values in `RProjects'
• Removed alternatives "greater" and "less" in functions T1EpSceptical and PPpSceptical. Alternative "one.sided" now returns the T1E rate (respectively project power) assuming both alternatives are one-sided in the same direction.
• Changed arguments and output of function hMeanChiSqCI:
• Renamed argument level to conf.level in the new version
• Removed argument wGamma entirely
• When called with alternative = "none", the returned list does not contain elements gammaMean and gammaHMean anymore.
• Removed argument type = "liberal" in all functions that had this option
• Removed functions levelEquivalent and unirootAll
• Added systematic tests to all functions in the package
• Updated function documentation
• Updated vignette

ReplicationSuccess 1.2

• New data set protzko2020
• Fixed a bug in the error message of effectSizeReplicationSuccess
• Added license information and improved documentation of data set RProjects

ReplicationSuccess 1.1.1

• Fixed a bug that prevented vignette building on R devel with updated texlive

ReplicationSuccess 1.1.0

• New maintainer Samuel Pawel
• Bugfixes hMeanChiSqMu
• Replace rootSolve::uniroot.all with custom function unirootAll

ReplicationSuccess 1.0

• CRAN release

ReplicationSuccess 0.2

• sampleSizeSignificance, powerReplicationSuccess, sampleSizeReplicationSuccess now use analytic instead of numerical implementation

• Development version of the package migrated to GitHub

• Vectorization in functions now with Vectorize instead of mapply

• Unit tests migrated to testthat

• Documentation migrated to roxygen2

ReplicationSuccess 0.1-4

• New function T1EpSceptical

• New function PPpSceptical

• NEWS file added

• New function Qtest

ReplicationSuccess 0.1-3

• Golden threshold implemented (see https://arxiv.org/abs/2009.07782)

• Recalibration of pSceptical now implemented via type argument

• powerReplicationSuccess and sampleSizeReplicationSuccess now also take type argument

• pSceptical returns per default recalibrated sceptical p-value (golden recalibration)

• sampleSizeReplicationSuccess and sampleSizeSignificance can now also be computed based on relative effect size (before only based on power)

• New function effectSizeReplicationSuccess

• New function effectSizeSignificance

• Name of thresholdSceptical changed to levelSceptical

SamCH93/ReplicationSuccess documentation built on Oct. 23, 2024, 9:17 p.m.