Predict in Higher-order CFA

[Maggie SY Fok]

Hello everyone. I try to follow the official example in estimating higher-order confirmatory factor analysis models.

use http://www.stata-press.com/data/r13/sem_hcfa1, clear

sem (Phys -> phyab1 phyab2 phyab3 phyab4) ///
    (Appear -> appear1 appear2 appear3 appear4) ///
    (Peer -> peerrel1 peerrel2 peerrel3 peerrel4) ///
    (Par -> parrel1 parrel2 parrel3 parrel4) ///
    (Nonacad -> Phys Appear Peer Par)

I would like to predict the factor scores of Nonacad and use the following codes:

predict double fa_Nonacad, latent(Nonacad)

but it gives the error message

predict not possible with summary statistics data
r(198);

How should I get the factor scores?

Thanks,
Maggie

[Erik Ruzek]

I think the problem is that this file is the summary statistics data (ssd) from Marsh and Hocevar (1985).

use http://www.stata-press.com/data/r13/sem_hcfa1, clear
describe
notes

See sem Intro 11. With ssd, you don't have the 251 individual values for each of these variables. And presumably, your idea with predicting factor scores is to have such scores for all the individuals in the data. Accordingly, predicting factor scores is incompatible with summary statistics data.