The metafor Package

A Meta-Analysis Package for R

User Tools

Site Tools

analyses:gleser2009

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
Next revision		Previous revision
analyses:gleser2009 [2022/08/03 11:17] – Wolfgang Viechtbaueranalyses:gleser2009 [2024/06/09 16:19] (current) – Wolfgang Viechtbauer
Line 79: Line 79:
 Now the model described in equation (19.5) can be fitted with: Now the model described in equation (19.5) can be fitted with:
 <code rsplus> <code rsplus>
-res <- rma.mv(yi, V, mods = ~ factor(trt) - 1, data=dat)+res <- rma.mv(yi, V, mods = ~ 0 + factor(trt), data=dat)
 res res
 </code> </code>
Line 127: Line 127:
 V <- bldiag(lapply(split(dat, dat$study), calc.v)) V <- bldiag(lapply(split(dat, dat$study), calc.v))
  
-res <- rma.mv(yi, V, mods = ~ factor(trt) - 1, data=dat)+res <- rma.mv(yi, V, mods = ~ 0 + factor(trt), data=dat)
 res res
 </code> </code>
Line 165: Line 165:
 V <- bldiag(lapply(split(dat, dat$study), calc.v)) V <- bldiag(lapply(split(dat, dat$study), calc.v))
  
-res <- rma.mv(yi, V, mods = ~ factor(trt) - 1, data=dat)+res <- rma.mv(yi, V, mods = ~ 0 + factor(trt), data=dat)
 res res
 </code> </code>
Line 202: Line 202:
 V <- bldiag(lapply(split(dat, dat$study), calc.v)) V <- bldiag(lapply(split(dat, dat$study), calc.v))
  
-res <- rma.mv(yi, V, mods = ~ factor(trt) - 1, data=dat)+res <- rma.mv(yi, V, mods = ~ 0 + factor(trt), data=dat)
 res res
 </code> </code>
Line 283: Line 283:
 The analysis can now be carried out with: The analysis can now be carried out with:
 <code rsplus> <code rsplus>
-res <- rma.mv(yi, V, mods = ~ factor(trt) - 1, data=dat)+res <- rma.mv(yi, V, mods = ~ 0 + factor(trt), data=dat)
 print(res, digits=3) print(res, digits=3)
 </code> </code>
Line 336: Line 336:
 dat$yi <- round(with(dat, (m2i-m1i)/sdpi), 3) dat$yi <- round(with(dat, (m2i-m1i)/sdpi), 3)
 dat$vi <- round(with(dat, 1/n1i + 1/n2i + yi^2/(2*(n1i+n2i))), 4) dat$vi <- round(with(dat, 1/n1i + 1/n2i + yi^2/(2*(n1i+n2i))), 4)
-dat$covi <- round(with(dat, (1/n1i + 1/n2i) * ri + (rep(sapply(split(dat$yi, dat$school), prod), each=2) / (2*(n1i+n2i))) * ri^2), 4)+dat$covi <- round(with(dat, (1/n1i + 1/n2i) * ri +  
 +                            (rep(sapply(split(dat$yi, dat$school), prod), each=2) /  
 +                            (2*(n1i+n2i))) * ri^2), 4)
 </code> </code>
 The contents of the resulting dataset are: The contents of the resulting dataset are:
Line 359: Line 361:
 The variance-covariance matrix for the entire dataset can be constructed with: The variance-covariance matrix for the entire dataset can be constructed with:
 <code rsplus> <code rsplus>
-V <- bldiag(lapply(split(dat, dat$school), function(x) matrix(c(x$vi[1], x$covi[1], x$covi[2], x$vi[2]), nrow=2)))+V <- bldiag(lapply(split(dat, dat$school),  
 +            function(x) matrix(c(x$vi[1], x$covi[1], x$covi[2], x$vi[2]), nrow=2)))
 V V
 </code> </code>
Line 383: Line 386:
 Finally, we can fit a model allowing for a different treatment effect depending on the outcome with: Finally, we can fit a model allowing for a different treatment effect depending on the outcome with:
 <code rsplus> <code rsplus>
-res <- rma.mv(yi, V, mods = ~ outcome - 1, data=dat)+res <- rma.mv(yi, V, mods = ~ 0 + outcome, data=dat)
 print(res, digits=3) print(res, digits=3)
 </code> </code>
analyses/gleser2009.txt · Last modified: 2024/06/09 16:19 by Wolfgang Viechtbauer