Dear all,
I am running a sem on the following data. My command is
sem trust->std_retweet std_favorite std_quote,latent(trust) nocapslatent.
The model fails to converge.
I learned from the manual that the problem is possible due to the starting value for some variables. As all three variables have been standardized, I tried an initial value 20 percent large than is expected.
sem trust->std_retweet std_favorite std_quote,latent(trust) nocapslatent var(e.std_retweet, init(1.2)). It still does not solve the problem of failing to converge. Could anyone offer some tips on troubleshooting?
0 Response to Problems of setting intial values in sem
Post a Comment