I am an amateur in OpenMx. I am trying, for a start, to fit a one-factor model with OpenMx. I am willing to use the path version.
I use type=RAM, and mxData(...type="raw")
I have 1064 observations and 8 items.
I compare the summary() with that of factanal(). I get the same loadings and uniquenesses estimations with one weird problem: When I use "raw" and not "cov", I have a total number of observed statistics = 8512 (=1064*8) instead of the correct = 44. Several statistics are NA (chi-square, RMSA) but the estimates of the loadings and uniquenesses are the same (and means) !!!
I use the code found in the documentation as an example for: Factor Analysis, Path Specification.