|
|
 |
| Interpreting output from complex surv... |
 
|
| Message/Author |
|
|
|
|
I am trying to run SEM on survey data, accounting for weights, strata and cluster. I use following codes for analysis and model statements.
ANALYSIS:
TYPE = COMPLEX RANDOM;
INTEGRATION = MONTECARLO (5000);
MITERATIONS = 2000;
PROCESSORS=6;
MODEL: f1 by a1 a2 a3;
f3 by b1 b2 b3;
f7 by c1 c2 c3;
f4 by d1 d2 d3;
f6 by e1 e2 e3;
f9 by f1 f3 f7;
y1 ON f4 f6 f9 depres;
In the output for path coefficients, I get something like (I did not include the estimates, standard error and p-values):
y1 ON
F4
F6
F9
y1 ON
depres
F6 WITH
F4
F9 WITH
F4
F6
What do the estimates for F6 WITH F4, F9 WITH F4, and F9 WITH F6 mean? Do I only use the estimates corresponding to the ON command? Thank you. |
|
|
|
|
| The WITH statements are covariances between your factors. They are typically needed since they can't be assumed to be zero. |
|
|
| Back to top |
|
|
