[R] Is SEM package of R suitable for sem analysis

[Yihui Xie]

Well, I guess it will be extremely difficult to persuade your teacher
to use the sem package in R. AFAIK, AMOS and LISREL have become the
golden standard for SEM in China, and SEM has become a golden model
for analysis in social science and psychology. I have been a cynic to
SEM for a long time, because I strongly believe this model has been
completely abused, and perhaps the "easy-to-use" AMOS has been helping
people abuse SEM. A common phenomenon I see is, people do not care
about the model at all - all they care about is whether AMOS can
successfully compute the coefficients; if AMOS failed, they will try
to (randomly) set constraints to the model, i.e. add/remove this
arrow, drag some variables into/out of the model, set this
coefficient/variance to 1.

Anyway, I encourage you to ask your teacher why AMOS is "more
suitable". I have a couple of very ridiculous stories about the SEM
software, in which I can only feel the blind faith in software.
Sigh...

The good thing is, R is open source, so is the sem package. You can
see everything in it, and you can extend it as you wish (in case your
teacher think it is less suitable).

Regards,
Yihui
--
Yihui Xie <xieyihui at gmail.com>
Phone: 515-294-6609 Web: http://yihui.name
Department of Statistics, Iowa State University
3211 Snedecor Hall, Ames, IA



On Sat, Dec 26, 2009 at 8:26 PM, Reeyarn_李智洋_10928113 <reeyarn at gmail.com> wrote:
> Dear Bruno and Joe,
>
> Thanks for advising!
>
>
> Reeyarn
>
> On Sat, Dec 26, 2009 at 9:32 PM, Bruno Falissard <falissard_b at wanadoo.fr> wrote:
>
>> A few years ago it could have been true, but now the package has improved
>> (especially with the bootstrap procedure).
>> At the moment there is no argument to recommend AMOS.
>
> On Sun, Dec 27, 2009 at 12:30 AM, Joe King <jp at joepking.com> wrote:
>
>> I am going to take SEM this next quarter in my doctoral program. My
>> suggestion is to use the program your professor suggests and try to
>> re-create your models in R using the SEM package.
>
> ______________________________________________
> 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.
>