Binary outcome regressed on a cfa PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 daniel adkins posted on Thursday, November 30, 2006 - 3:56 pm
are there any special concerns with this model. it seems to be using some sort of quadrature in the estimation and takes forever to converge. the code i used is below. any advice would be great. thanks.

=================================
TITLE: Binary outcome on CFA
DATA: FILE IS L:\data_mil_1_3a.dat;
FORMAT IS free;
TYPE IS individual;
OUTPUT: STAND;
VARIABLE: NAMES ARE mil hhinc medu fedu hhincx unempx;
USEVARIABLES ARE mil hhinc medu fedu hhincx unempx;
CATEGORICAL IS mil;
MISSING ARE .;
ANALYSIS: TYPE IS missing meanstructure;
ESTIMATOR IS ML;
MODEL: mil ON ses;
ses BY hhinc medu fedu hhincx unempx;
 Linda K. Muthen posted on Friday, December 01, 2006 - 10:02 am
This model requires numerical integration. I think it would be only one dimension so I don't know why it would be particularly slow. If you send the input, data, output, and your license number to support@statmodel.com, I can look into it.
 Belal Fattah posted on Wednesday, July 10, 2013 - 8:06 pm
Muthen and Muthen,

Quick question regarding a binary DV SEM model. I am running a logit SEM model with a binary DV and I noticed that there is no r-squared in the output. Is there any way of obtaining an r-squared or do I interrupt this using the odds ratios?

Below is my mplus code:

DATA: FILE IS SouthKorea.csv;

VARIABLE: NAMES ARE V1....V253 PCuse2 d_PL d_JP d_SK d_PL1 ed5 ed6 lowinc highinc;

USEVARIABLE ARE V1....V25 PCuse2 d_PL d_JP d_SK d_PL1 ed5 ed6 lowinc highinc;

MISSING = ALL(-9999);
CATEGORICAL = PCuse2;

ANALYSIS:
estimator = ml;
INTEGRATION = MONTECARLO(500);

MODEL:
gnT by V4 V6 V7 V8;
PCuse2 on V23 V23 V24
ed5 ed6 lowinc highinc gnT;
 Linda K. Muthen posted on Thursday, July 11, 2013 - 6:58 am
The STANDARDIZED option of the OUTPUT command produces R-square.
 Belal Fattah posted on Thursday, July 11, 2013 - 5:20 pm
Thank you! I used STDYX in the model. I received this output:

R-Square

Observed Estimate S.E. Est./S.E. P-Val
Variable

PCUSE2 .640 .053 12.232 0.000
etc.
etc.


Questions: Is the R-sqr the coefficient listed under 'Estimate'? Or how would I interpret this section of the output?
 Linda K. Muthen posted on Friday, July 12, 2013 - 6:02 am
Yes. The first column is R-square. The second is its standard error. The third is the ratio of the estimate to its standard error which is a z-test in large samples. The p-value for this z-test is given in the last column.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: