lavaan and AMOS differences with Wheaton classic data

[golo...@gmail.com]

Hi,

I have compared the results from lavaan and AMOS. I have used the Wheaton classic data covariance matrix. A model with factor loadings of the first of the indicators fixed to 1.0, can produce exact results of chi-square and other coefficients, but some minimal differences in variances, when compare lavaan and AMOS results. But a model with the variance of latent variables fixed to 1.0, can produce exact results of coefficients as chi-square, CFI, TLI, and some minimal differences in variances, but some awesome sign differences in some results. To fix to 1.0 variances, I have used the next sintaxis

fit <- sem(wheaton.model, std.lv=TRUE,sample.cov = wheaton.cov,sample.nobs = 932,likelihood = "wishart")

This is the model I have used in lavaan:

wheaton.model <- '

ses     =~ education sei

alien67 =~ anomia67 powerless67

alien71 =~ anomia71 powerless71

alien71 ~ alien67 ses

alien67 ~ ses'

The graphic model in Amos is in the file attached.

lavaan results

Regressions:

  alien71 ~

        ses   -0.233    0.070   -3.316    0.001   -0.151   -0.151

  alien67 ~

    ses       -0.688    0.062  -11.147    0.000   -0.567   -0.567

Amos results

Regression Weights: (Group number 1 - Default model)

			Estimate	S.E.	C.R.	P	Label
Alien67	<---	SES	.688	.062	11.147	***
Alien71	<---	SES	.233	.070	3.316	***

lavaan results

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

Latent variables:

  ses =~

    education         2.581    0.124   20.797    0.000    2.581    0.833

    sei              13.761    0.793   17.364    0.000   13.761    0.649

Amos results

EDUCATIO	<---	SES	-2.580	.124	-20.797	***
SEI	<---	SES	-13.754	.792	-17.364	***

I repeat this analysis several times, but I have the same result. 

Best,

Al

[Jeremy Miles]

The direction of the signs is arbitrary. The two models are exactly the same, because:

a * -b is equal to -a * b.

You can set starting values in either Amos or Lavaan that will start the model in a similar place,and give you the same result.

Jeremy

--
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.

[golo...@gmail.com]

Hi Jeremy,

I will try your solution.

Thank you very much!

Al