NEWS

ClusterBootstrap 2.0.0 (2025-08-21)

Maintenance update: changed maintainer email address
Added generalized (i.e., applicable to any function) clusterBootstrap function for cluster resampling at multiple levels, allowing for replace specification at each level.
Added check for return value of detectCores() in clusbootglm. detectCores() returns NA in specific situations (e.g., when using webR), breaking the function (issue #18)
Added error handling for clusbootglm (issue # 19)
Added proper exit handling for RNGkind() and makeCluster() when using multiple cores in clusbootglm and for plot functions
All functions up-to-date with R >= 4.0.0, examples require R >= 4.1.0 (base pipe)
Fixed warning "cannot xtfrm data frames" for emm
Added general predict function for clusbootglm
Update package title, as the package is no longer solely for glms
Rewrite internal function calls to accomodate non-invasive export in NAMESPACE using exportPattern
All non-helper functions are explicitly exported for better experience of tab completion and GUI (e.g., RStudio) info popups

A paper about the package has been published in Behavior Research Methods (see citation("ClusterBootstrap"))
Function clusbootglm can handle lm/glm class objects as model input as an alternative to a formula

Added functionality to calculate and plot estimated marginal means
Added permutation test and a plot function to assess observed data under null distribution(s)

Added parameter estimate.type to summary(), removing inconsistency with output from coefs() by defaulting to bootstrap estimates

Renamed some function arguments for consistency, clarity and aesthetic reasons
Added medication dataset from the accompanied website of Singer & Willett (2003)
Removed redundant bootstrap estimates from summary.clusbootglm()
Added and prettified some error handlings
Changed confint() interval.type argument default to match the one from summary() (BCa interval)