This page provides tips on how to carry out meta-analysis in R.
Under the fixed effects model, we assume that all studies in the meta-analysis share the same true effect size. The variation among the studies are assumed to result only from sampling error (i.e., chance).
To estimate the mean effect size, we weight each effect size in the sample according to the magnitude of its sampling variance, which takes into account the different sample sizes of the studies. Calculate each weight as the inverse of the squared standard error,
w = 1/SEz2,
where z is the effect size of interest and SE is its standard error.
weighted.mean command to calculate a weighted mean of the effect sizes using the weights (see the help page for the command).
Calculate the standard error of the weighted mean as the square root of the inverse of the sum of sample weights. This standard error measures uncertainty of the estimate of true effect size.
SEz = √(1/∑w).
To calculate an approximate 95% confidence interval for the weighted mean, use the normal approximation. The critical value (1.96) is obtained as
qnorm(1 - 0.05/2)
The limits of the 95% confidence interval are then calculated as the estimated weighted mean – 1.96 SEz and the mean + 1.96 SEz.
Under the random effects model we assume that each study estimates a system-specific effect size. There is no “one true” effect size under this model, only a mean effect size. This is more realistic than the fixed effects model for most meta-analyses in ecology and evolution.
To fit the random effects model, estimate the variance among the system-specific effect sizes, Τ2 (“tau squared”). One way to estimate it involves calculating the heterogeneity among the observed effect sizes (Q), and then “correcting” by subtracting the within-study sampling variance. The correction is needed because the variance among the observed effect sizes among studies is inflated by within-study sampling error.
First, calculate Q, the weighted heterogeneity among the observed effect sizes z, as
Q = ∑w(z – z)2
Then calculate Τ2 by subtraction, being careful not to allow a negative value (since Τ2 is a variance, which can’t be negative). In the following, N refers to the number of studies in the meta-analysis.
If Q < N – 1, then set Τ2 = 0.
Τ2 = (Q – (N – 1))/(∑w – ∑w2/∑w).
Τ2 is then used to calculate new weights for the effect sizes of each study under the random effects model, taking account of both within-study sampling variance and between-study variance in effect sizes. SEz2 is the squared standard error of each effect size z (same as we computed in the fixed effects section above).
w’ = 1/(SEz2 + Τ2)
To calculate the weighted mean effect size z under the random effects model, use the same procedure as that used before for the fixed effects model except use the new weights w’ calculated here.
The formula for the standard error (SE) of the mean effect size z is the same as that in the fixed-effects model, except use the new weights, w’.
The 95% confidence interval for the mean effect size under the random effects model is calculated in the same way as under the fixed effects model, except that you would use the standard error for the random effects mean.
Begin by loading the
meta package in R.
metagen command is used as follows,
z <- metagen(TE, seTE, data = mydata) summary(z)
TE is the vector of effect sizes (TE stands for "treatment effect"), and seTE is the corresponding vector of standard errors.
Other variations of the command are as follows. In the following example,
mygroupvar is the arbitrary name of a category variable in the data frame
mydata that identifies different types of studies (for example, experimental or observation, etc). In this case, the meta-analysis results are produced separately for each group in
mygroupvar. The arguments
comb.fixed = FALSE, comb.random = TRUE indicates that you only want the random (mixed) effects meta-analysis results.
z <- metagen(TE, seTE, data = mydata, byvar = mygroupvar, comb.fixed = FALSE, comb.random = TRUE) summary(z)
To obtain a funnel plot or forest plot with the
meta package, use one of the following commands (warning: if you have many studies in your meta-analysis, the
forest command will run slowly and the results won't fit on the graphics window). The arguments
comb.fixed = FALSE, comb.random = TRUE will add a line to the funnel plot for the mean effect size under the random effects model.
funnel(z, comb.random = TRUE, comb.fixed = FALSE, xlab = "Effect size") forest(z, comb.random = TRUE, comb.fixed = FALSE)