Can Package SEM do mean structural analysis?
Dear John: Thank you very much for your prompt explanation on the present possibilities of SEM. I am not a statistician so that I can not figure out the raw sums-of-squares-and-products matrix. And I will do with SEM in the area without mean structural equation. I appreciate your good teaching at every time. --------========---------- Mitsuo Igarashi mitsu5 at ruby.famille.ne.jp ----------------
John Fox <jfox at mcmaster.ca> wrote:
Dear Mitsuo, My intention in writing the sem package was to provide a basic structural-equation facility for R. I haven't made explicit provision for models with means, but it might be possible to fit such models by using the raw sums-of-squares-and-products matrix among the observed variables (perhaps divided by n) as input. It might be necessary to make small modifications to degrees of freedom, etc. I'm afraid that I don't have time now to figure out whether this will work and, if so, precisely how to do it, but you're welcome to try. If you get it to work, I'd be interested in the answer. If this approach doesn't work, then incorporating means would likely require major modifications to the program. There are, by the way, several other things that dedicated structural-equation software does that the sem package does not -- e.g., analysis by groups and for ordinal observed variables. I may, at some point, extend the package in these directions, but I don't have plans at the moment to do so. Good luck, John At 10:53 PM 5/24/2003 +0900, you wrote:
I am wondering whether Package SEM can do with intercepts and means in its structural analysis. If it can not calculate, how can I make a supplemental function in R?