## Calculate Effect Sizes

Consider the meta-analytic researcher who is interested in calculating effect sizes (ES) for a new project. She ends up with several studies providing relevant data to compute standardized mean differences ES (Cohen's δ). However, a few studies do not provide the necessary data to compute the ES (means & standard deviations), although they provide alternate data that can potentially be used to calculate an ES. In this case, one of several functions from the compute.es, RcmdrPlugin.MA, or MAd package may be useful. For example, suppose the primary study reported the output of a independent t-test and the sample size for treatment and comparison groups but not the means of each group. The ES can be computed easily using one of these freely available R packages. For example, at the R command prompt, type:

install.packages('compute.es') # INSTALL THE compute.es PACKAGE (IF NOT ALREADY IN LIBRARY)

library(compute.es) # LOADS THE PACKAGE

tes(t=2.5, n.1=30, n.2=30)

where the the t-value = 2.5 and the treatment sample size= 30 and comparison = 30. The output will generate several statistics but to obtain only the standardized mean differences and its variance, one can do the following:

tes(t=2.5, n.1=30, n.2=30, verbose=FALSE)[ , 12:13]

There are several additional options for calculating ESs. Click here or here for additional options and here for more info on compute.es.

install.packages('compute.es') # INSTALL THE compute.es PACKAGE (IF NOT ALREADY IN LIBRARY)

library(compute.es) # LOADS THE PACKAGE

tes(t=2.5, n.1=30, n.2=30)

where the the t-value = 2.5 and the treatment sample size= 30 and comparison = 30. The output will generate several statistics but to obtain only the standardized mean differences and its variance, one can do the following:

tes(t=2.5, n.1=30, n.2=30, verbose=FALSE)[ , 12:13]

There are several additional options for calculating ESs. Click here or here for additional options and here for more info on compute.es.