PET_PEESE: PET_PEESE

In JConigrave/msemtools: Routines, tables, and figures for metaSEM analyses

PET_PEESE

Usage

PET_PEESE(
  m,
  transf = NULL,
  digits = 2,
  alpha = 0.1,
  type = c("PET-PEESE", "PET", "PEESE"),
  pattern = "{est} [95% CI {lower}, {upper}]"
)

Arguments

transf	function with which to transform results
digits	number of digits for rounding
alpha	the alpha threshold used in PET-PEESE. Defaults to 0.1 (10 \item typeShould PET-PEESE, PET or PEESE be used? \item xa metaSEM or meta_ninja model

Calculates a PET-PEESEE adjusted estimate PET-PEESE uses meta-regression in order to adjust estimates for publication bias. When there is publication bias the sampling variance becomes correlated with effect size. By using the standard error (PET) or the sampling variance (PEESE) as moderators in a meta-regression, estimates (the intercept) can be made which partial out the correlation between effect size and variance. PET-PEESE first tests whether the variance component is a significant (p < 0.05) predictor of effect size. If it is, PEESE is used. Otherwise, PET is used.

JConigrave/msemtools documentation built on Aug. 10, 2022, 12:09 p.m.