## 7.1 Subgroup Analyses using the Mixed-Effects-Model

To conduct subgroup analyses using the Mixed-Effects-Model (random-effects-model within subgroups, fixed-effects-model between subgroups), you can use the subgroup.analysis.mixed.effects function we prepared for you. To use the function, meta and metafor need to be installed and loaded in your library.

As the code for the function is pretty long, we don’t display it here. To access the function, use this link. Again, R doesn’t know this function yet, so we have to let R learn it by copying and pasting the code from the website in its entirety into the console on the bottom left pane of RStudio, and then hit Enter ⏎.

For the subgroup.analysis.mixed.effects function, the following parameters have to be set:

Code Description
m The output of you meta-analysis. In my case, this is ‘m.hksj’
subgroups The variable in our dataset in which we coded which study belongs to which subgroup. Note that we also tell the function in which dataset this variable is stored. In my case, this was the ‘madata’ dataset i used to get the meta-analysis output ‘m.hksj’. The dataset and the subgroup variable have to be connected with $(e.g. madata$Control). It is important that the subgroup variable is defined for each effect size we used to calculate the initial meta-analysis
exclude This controls if one or more subgroups should be excluded from the subgroup analysis. To exclude one or more subgroups, we have to provide the function with the name of the subgroup as it appears in our dataframe (in quotation marks; e.g., exclude=‘psychology’). One can also exclude more than one subgroup at the same time. To do this, one has to concatenate the names of the subgroups we want to exclude (e.g. exclude = c(‘psychology’,‘general’)). If you do not want to exclude subgroups (as is advised), you can leave out this parameter.

In my madata dataset, which i used previously to generate my meta-analysis output m.hksj, i stored the subgroup variable Control. This variable specifies which control group type was employed in which study. There are three subgroups: WLC (waitlist control), no intervention and information only.

The function to do a subgroup analysis using the mixed-effects-model with these parameters looks like this.

subgroup.analysis.mixed.effects(m=m.hksj,
subgroups=madata\$Control)
## Subgroup Results:
## --------------
##           Subgroup k        TE       seTE  LLCI  ULCI            p
## 1 information only 3 0.4048358 0.07875304 0.250 0.559 2.739011e-07
## 2  no intervention 8 0.5107720 0.11945084 0.277 0.745 1.902797e-05
## 3              WLC 7 0.7836032 0.20993337 0.372 1.195 1.894924e-04
##           Q   I2 I2.lower I2.upper
## 1  1.144426 0.00     0.00     0.82
## 2 16.704467 0.58     0.08     0.81
## 3 22.167163 0.73     0.42     0.87
##
## Test for subgroup differences (mixed/fixed-effects (plural) model):
## --------------
##                       Q df         p
## Between groups 3.031562  2 0.2196365
##
## - Total number of studies included in subgroup analysis:  18
## - Tau estimator used for within-group pooling:  SJ

The results of the subgroup analysis are displayed under Subgroup Results. We also see that, while the pooled effects of the subgroups differ quite substantially (g = 0.41-0.78), this difference is not statistically significant.

This can be seen under Test for subgroup differences (mixed/fixed-effects (plural) model) in the Between groups row. We can see that $$Q=3.03$$ and $$p=0.2196$$. This information can be reported in our meta-analysis paper.