I am attempting to examine factor invariance (as per Phil Ender's Hierarchy of Factor Invariance presentation) with a 2nd order factor. I am confused because my output for the first step (separate CFAs for each group) is not giving me a likelihood ratio test/chi-square against the saturated model, which usually given at the end of the factor loadings and variance estimates. There is no error message. My command is (from SEM builder):
sem (F4@1 -> L1) (F4 -> L2 L3) (L1 <- _cons@0) (L1@1 -> s5pfs1) (L1 -> s5pfs2 s5pfs5 s5pfs10 s5pfs11 s5pfs13) (s5pfs1 <- _cons@0) (L2@1 -> s5pfs3) (L2 -> s5pfs4 s5pfs6 s5pfs7) (s5pfs3 <- _cons@0) (L3@1 -> s5pfs8) (L3 -> s5pfs9 s5pfs12 s5pfs14 s5pfs15) (s5pfs8 <- _cons@0) if female==0 [pweight = wgt_student_5], latent(F4 L1 L2 L3 ) mean(F4) var( e.L1@0 e.s5pfs1@0 e.s5pfs3@0 e.s5pfs8@0) nocapslatent
Thanks so much for your assistance
sem (F4@1 -> L1) (F4 -> L2 L3) (L1 <- _cons@0) (L1@1 -> s5pfs1) (L1 -> s5pfs2 s5pfs5 s5pfs10 s5pfs11 s5pfs13) (s5pfs1 <- _cons@0) (L2@1 -> s5pfs3) (L2 -> s5pfs4 s5pfs6 s5pfs7) (s5pfs3 <- _cons@0) (L3@1 -> s5pfs8) (L3 -> s5pfs9 s5pfs12 s5pfs14 s5pfs15) (s5pfs8 <- _cons@0) if female==0 [pweight = wgt_student_5], latent(F4 L1 L2 L3 ) mean(F4) var( e.L1@0 e.s5pfs1@0 e.s5pfs3@0 e.s5pfs8@0) nocapslatent
Thanks so much for your assistance