Convergence issue

Tiffany

unread,

Sep 11, 2014, 5:48:04 PM9/11/14

to lav...@googlegroups.com

Hi everyone,

I am trying to fit a relatively simple SEM, but I keep getting non-convergence error that I can't seem to understand.

Any comment would be much appreciated. Thank you so much.

CODE:

model<-'

# measurement model

XR_H=~HV205+Sanavg

TR_H=~HV201+HV209+HV237+HV213

H_C=~age+V501+H11

D_H=~H11+V404+age

# regressions

D_H~XR_H+TR_H

# residual correlations

'

fit <- sem(model, data = ind_new, link="logit", ordered=c("HV205,H11,V404"))

summary(fit, fit.measures= TRUE, standardized = TRUE, rsquare=TRUE)

WARNING:

In lavaan::lavaan(model = model, data = ind_new, ordered = c("HV205,H11,V404"), :

lavaan WARNING: model has NOT converged!

SUMMARY:

** WARNING ** lavaan (0.5-16) did NOT converge after 10000 iterations

** WARNING ** Estimates below are most likely unreliable

Used Total

Number of observations 47745 48243

Estimator DWLS

Minimum Function Test Statistic NA

Degrees of freedom NA

P-value (Unknown) NA

Parameter estimates:

Information Expected

Standard Errors Robust.sem

Estimate Std.err Z-value P(>|z|) Std.lv Std.all

Latent variables:

XR_H =~

HV205 1.000 0.342 0.739

Sanavg 0.727 0.249 0.824

TR_H =~

HV201 1.000 0.058 0.155

HV209 4.029 0.235 0.611

HV237 2.820 0.164 0.335

HV213 5.138 0.299 0.601

H_C =~

age 1.000 47.731 30.368

V501 -0.000 -0.002 -0.017

H11 0.000 0.000 0.000

D_H =~

H11 0.000 0.000 0.000

V404 0.039 0.136 0.287

age -14.214 -49.928 -31.766

Regressions:

D_H ~

XR_H 5.289 0.515 0.515

TR_H -84.835 -1.407 -1.407

H11 ~

H_C -0.022 -1.068 -3.697

D_H 0.320 1.123 3.885

Covariances:

XR_H ~~

TR_H 0.018 0.889 0.889

H_C -12.555 -0.769 -0.769

TR_H ~~

H_C -2.760 -0.993 -0.993

Intercepts:

HV205 0.312 0.312 0.673

Sanavg 0.334 0.334 1.105

HV201 0.828 0.828 2.198

HV209 0.180 0.180 0.468

HV237 0.405 0.405 0.825

HV213 0.460 0.460 0.922

age 2.913 2.913 1.853

V501 1.016 1.016 8.087

H11 0.092 0.092 0.318

V404 0.662 0.662 1.398

XR_H 0.000 0.000 0.000

TR_H 0.000 0.000 0.000

H_C 0.000 0.000 0.000

D_H 0.000 0.000 0.000

Variances:

HV205 0.097 0.097 0.454

Sanavg 0.029 0.029 0.321

HV201 0.139 0.139 0.976

HV209 0.092 0.092 0.626

HV237 0.214 0.214 0.888

HV213 0.159 0.159 0.639

age 0.364 0.364 0.148

V501 0.016 0.016 1.000

H11 0.082 0.082 0.981

V404 0.205 0.205 0.918

XR_H 0.117 1.000 1.000

TR_H 0.003 1.000 1.000

H_C 2278.254 1.000 1.000

D_H 0.537 0.044 0.044

R-Square:

HV205 0.546

Sanavg 0.679

HV201 0.024

HV209 0.374

HV237 0.112

HV213 0.361

age 0.852

V501 0.000

H11 0.019

V404 0.082

D_H 0.956

Warning message:

In .local(object, ...) :

lavaan WARNING: fit measures not available if model did not converge

Edward Rigdon

unread,

Sep 11, 2014, 6:11:16 PM9/11/14

to lav...@googlegroups.com

Your factors H_C and D_H share two indicators, having only one unique indicator each. The DH factor is dependent on XRH and TRH. HC does not predict anything, and is not predicted, so it just sits there covarying by default with XRH and TRH. Is that the model you meant to specify? Is it important to have HC in the model? The identification problem comes about because HC and DH share indicators. It is possible that you could achieve identification if you specify starting values for the Loadings for the indicators loading on HC and DH.

-Ed Rigdon

Sent from my iPad

--
You received this message because you are subscribed to the Google Groups "lavaan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lavaan+un...@googlegroups.com.
To post to this group, send email to lav...@googlegroups.com.
Visit this group at http://groups.google.com/group/lavaan.
For more options, visit https://groups.google.com/d/optout.

Yves Rosseel

unread,

Sep 18, 2014, 3:00:15 AM9/18/14

to lav...@googlegroups.com

> fit <- sem(model, data = ind_new, link="logit", ordered=c("HV205,H11,V404"))

Note that the 'link' argument is (at least in 0.5-16, silently) ignored
when using the default estimator DWLS.

Yves.

Hashim Javed

unread,

Sep 20, 2022, 2:25:20 AM9/20/22

to lavaan

Dear Tiffany,

Kindly check if you are using sub-sample for analysis, and if you are using a variable to make sub-sample based on its value,

then don't include that variable as covariates. (The variables included using '+' sign)

And also use aliases with all the variables included using '+' sign.

For Example:

XR_H=~HV205+Sanavg

TR_H=~HV201+HV209+HV237+HV213

use instead:

XR_H=~HV205 + a*Sanavg

TR_H=~HV201 + c1*HV209 + c2*HV237 + c3*HV213

Reply all

Reply to author

Forward