Confa for confirmatory factor analysis

[Marie Sofie]

I have tried using confa (as adviced in previous posts) for a confirmatory factor analysis, so I did something like this:

confa (energy: ener_sec1 ener_sec2 ener_sec3 ener_sec4 )

and it has given me this message. Why? What does it mean?

initial: log likelihood = -36015.811
rescale: log likelihood = -36015.811
rescale eq: log likelihood = -1516.252
could not calculate numerical derivatives
flat or discontinuous region encountered
convergence not achieved
r(430);


I need to conduct a simple confirmatory factor analysis, one laten variable and 4 observed variables Thank you

[Phil Bromiley]

Why not use the Stata supported sem package? Stata supports it which means it is more validated and may work better with other parts of Stata than user written procedures. Many more of us use that routine than confa which is user written.

This looks like a very simple cfa which confa looks like it should do easily. Why this doesn't work is impossible to determine without data. Are your indicators highly correlated?