Yi-fu posted on Thursday, March 02, 2006 - 10:13 pm
Hi, Dr. Muthen,
I know this may be covered in other discussion topic, but I can't find it right a way. So, please forgive me to ask it again. I have a second-order factor model with two latent variables loaded on a second-order factor. One latent variable contains three indicators and the other has 5 indicators. When I tried to run this model, the standard error of one of the latent loadings (related to second-order factor) is ****** in the printout. Could I know what this mean? There was no error message in my printout.
bmuthen posted on Friday, March 03, 2006 - 6:32 am
*** means huge numbers. For SEs, this happens when the model is not identified. Your model is not identified because it takes at least 3 first-order factors to identify a second-order factor model.
I am trying to do a CFA to a 5 first-order factors model by using data observed on 188 objects, and there are 12 indicators in total for these 5 factors. But I cannot get the model indices (CFI/TLI, RMSEA, etc).
M-plus returns the following information:
THE MODEL ESTIMATION TERMINATED NORMALLY THE STANDARD ERRORS OF THE MODEL PARAMETER ESTIMATES COULD NOT BE COMPUTED. THE MODEL MAY NOT BE IDENTIFIED. CHECK YOUR MODEL. PROBLEM INVOLVING PARAMETER 46.
What does this mean? How to solve this problem? Thank you very much.
p.s. Here are the codes I am using
TITLE: RBANS_CFA on full Data Set DATA: FILE IS "1.dat"; VARIABLE: NAMES ARE LISTIMM STORYIMM FIGCOPY LINE NAMING FLUENCY DIGSPAN CODING LISTREC RECOG STORYREC FIGREC; missing are *; USEVARIABLES ARE LISTIMM STORYIMM FIGCOPY LINE NAMING FLUENCY DIGSPAN CODING LISTREC RECOG STORYREC FIGREC; MODEL: IM BY LISTIMM* STORYIMM; VC BY FIGCOPY* LINE; La BY NAMING* FLUENCY; Att BY DIGSPAN* CODING; DM BY LISTREC* RECOG STORYREC FIGREC; IM@1VC@1La@1Att@1DM@1; ANALYSIS: TYPE = MISSING H1; OUTPUT: standardized ;