I want to use exploratory factor analysis to understand my test structure with correlated factors. With so many rotations available how does one pick the rotation that best describes the true relationship between each item and factor? Would it be useful to simulate data with a known factor structure and then compare the rotation methods? Has this been done and/or would this be useful?
We choose the default rotation based on what we think will work best in most situations. You can read about the different rotations in the references we provide in the user's guide to see if one of the rotations other than the default seems better for your data. You can also simulate data according to your data and see which rotation is best in your case.