EFA rotation: lavaan support for oblimin gamma and geomin epsilon

[zhao...@gmail.com]

Hi all,

I just found out that Lavaan supports the oblimin gamma rotation by passing rotation.args = list(oblimin.gamma = 0.2) to the function cfa(). Similar for the geomin epsilon.

I am wondering what the consensus or recommendation is among social science/psychology researchers in:

1. oblimin or geomin?

2. whether to explore different values of gamma and epsilon?

My dataset seems to provide "neater" rotation (interpretable factors despite more complex) with epsilon = 0.5 in geomin (as recommended in the literature) but I would like to understand better how it is different from oblimin with a different gamma value? 

Any suggestions would be much appreciated.

Thanks!