Twenty stimulated studies on standardized mean difference and one
continuous study characteristic reported by Hox (2002).
Usage
data(Hox02)
Details
The variables are:
study
Study number
yi
Effect size (standardized mean difference)
vi
Sampling variance of the effect size
weeks
Duration of the experimental intervention in terms of weeks
Source
Hox, J. J. (2002). Multilevel analysis: Techniques and applications. Mahwah, N.J.: Lawrence Erlbaum Associates.
References
Cheung, M. W.-L. (2008). A model for integrating fixed-, random-, and mixed-effects meta-analyses into structural equation modeling. Psychological Methods, 13, 182-202.
Examples
## Not run:
data(Hox02)
#### ML estimation method
## Random-effects meta-analysis
summary( meta(y=yi, v=vi, data=Hox02, I2=c("I2q", "I2hm"), intervals.type="LB") )
## Fixed-effects meta-analysis
summary( meta(y=yi, v=vi, data=Hox02, RE.constraints=0,
model.name="Fixed effects model") )
## Mixed-effects meta-analysis with "weeks" as a predictor
## Request likelihood-based CI
summary( meta(y=yi, v=vi, x=weeks, data=Hox02, intervals.type="LB",
model.name="Mixed effects meta analysis with LB CI") )
#### REML estimation method
## Random-effects meta-analysis with REML
summary( VarComp <- reml(y=yi, v=vi, data=Hox02) )
## Extract the variance component
VarComp_REML <- matrix( coef(VarComp), ncol=1, nrow=1 )
## Meta-analysis by treating the variance component as fixed
summary( meta(y=yi, v=vi, data=Hox02, RE.constraints=VarComp_REML) )
## Mixed-effects meta-analysis with "weeks" as a predictor
## Request likelihood-based CI
summary( reml(y=yi, v=vi, x=weeks, intervals.type="LB",
data=Hox02, model.name="REML with LB CI") )
## End(Not run)