[R-meta] New major version of R package meta available on CRAN

[Guido Schwarzer]

Hi all,

I would like to announce that version 5.0-0 of meta is available on CRAN.

The most significant changes are

- use of the restricted maximum likelihood (REML) estimator for the 
between-study variance instead of the DerSimonian-Laird estimator 
(argument 'method.tau'),

- switch of the behaviour of print() and print.summary(), i.e., print() 
provides the meta-analysis results, while print.summary() provides also 
individual study results,

- renaming some argument names (including 'fixed' and 'random' instead 
of 'comb.fixed' and 'comb.random'),

- print "Common effect model" instead of "Fixed effect model".

The DerSimonian-Laird estimator can be used as default in all 
meta-analyses with the command settings.meta(method.tau = "DL"). The 
command settings.meta("meta4") resets more defaults from meta, version 
4, including the text "fixed effect model". The behaviour of print and 
summary cannot be changed.

The old argument names still work but will result in a warning.

More details on the changes are available at 
https://www.researchgate.net/project/Implementation-of-statistical-methods-for-meta-analysis-in-R 
.

New readers of "Meta-Analysis with R (Use R!)" should install a legacy 
version of meta: https://tinyurl.com/dt4y5drs

Best, Guido