R help - Apr 2009 - CFA in R/sem package

Hi,

I am not sure if R-help is the right forum for my question. If not,
please let me know.

I have to do some discriminant validity tests with some constructs. I
am using the method of doing a CFA constraining the correlation of a
pair of the constructs to 1 and comparing the chi-square of this
constrained model to the unconstrained model. If the chi-square
difference is not significant, then I cannot reject the null
hypothesis that the two constructs are equal.

Well, if you are going to test, say, 4 constructs (A, B, C, and D),
you will have to have 2*C(4,2) = 12 models to test, 5 constructs, 20
models, and so forth. A tedious and error prone process...

So far, I have been using AMOS for that shake, given that 1) my
university has the license, 2) my other colleagues use it, and 3) I
know it ;)

I would like to know if any of you use R, namely the sem package, for
that application and if you can share your thoughts/experiences on
using it. I don't thing I would have problems "porting" my models
to
R/sem, but I would like to know if there is an optimized process of
doing that tests, without manually coding all the dozens of models.

Best,

Iuri.

Ivan,

I recently put together the sem.additions package over at R forge in  
part for just such a multiple model problem.  THere are a variety of  
methods that make it easy to add/delete links that could be automated  
with a for loop and something from the combn package, I think.

http://r-forge.r-project.org/projects/sem-additions/

-Jarrett

On Apr 9, 2009, at 6:39 AM, Iuri Gavronski wrote:
> Hi,
>
> I am not sure if R-help is the right forum for my question. If not,
> please let me know.
>
> I have to do some discriminant validity tests with some constructs. I
> am using the method of doing a CFA constraining the correlation of a
> pair of the constructs to 1 and comparing the chi-square of this
> constrained model to the unconstrained model. If the chi-square
> difference is not significant, then I cannot reject the null
> hypothesis that the two constructs are equal.
>
> Well, if you are going to test, say, 4 constructs (A, B, C, and D),
> you will have to have 2*C(4,2) = 12 models to test, 5 constructs, 20
> models, and so forth. A tedious and error prone process...
>
> So far, I have been using AMOS for that shake, given that 1) my
> university has the license, 2) my other colleagues use it, and 3) I
> know it ;)
>
> I would like to know if any of you use R, namely the sem package, for
> that application and if you can share your thoughts/experiences on
> using it. I don't thing I would have problems "porting" my
models to
> R/sem, but I would like to know if there is an optimized process of
> doing that tests, without manually coding all the dozens of models.
>
> Best,
>
> Iuri.
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.