I have a SEM built with lavaan and I would like to assess spatial autocorrelation. I found this code written by Jared Byrnes that does exactly that:
However in this page there is warning that as of 2016/10/1 lavaan is not producing correct errors. I browsed lavaan's development page and it seems the bug has not been fixed. Jared Byrnes recommends that users calculate the errors by hand.
Does anyone know of an alternative way of assessing spatial autocorrelation? Alternatively, how can I calculate the errors by hand?