[R-meta] Random-effects for correlational studies in rma.mv()

[Yuhang Hu]

Hello Meta community,

I am meta-analyzing a set of studies correlating a variable (Var1) with a
set of other variables (Var2s). Each study uses a single scale to measure
'Var1' and correlates it with different 'Var2's (see data structure below).

Conceptually, would it make more sense to consider ' Var2' nested as in:

random = ~1| scale/ study/ Var2/ obs

Or crossed with 'scales' and 'studies' as in:

random = list(~1| scale/ study/ obs, ~1| Var2)

Many thanks for your time and guidance,
Yuhang

Data structure is below:

study scale Var2 obs ri
1     1     1       1   .5
1     1     2       2   .1
2     2     1       3   .3
3     2     2       4   .4
3     2     3       5   .6
4     1     1       6   .8
4     1     3       7   .7

	[[alternative HTML version deleted]]