Date:         Wed, 27 May 1998 19:05:21 -0400
Reply-To:     Soumen Mukherjee <SMukherjee@MAPSNET.COM>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         Soumen Mukherjee <SMukherjee@MAPSNET.COM>
Subject:      Re: CFA with one factor
Comments: To: "Pedro J. Cuestas" <pcuestas@FCU.UM.ES>
Content-Type: text/plain

In the model you have specified, it is not clear what "f2" is. It is clearly
not a manifest variable. If it is intended as a latent variable like "f1"
then you need to specify its measurement using other manifest variables
(much like you have specified the measurement for f1). If it is intended to
be a second order factor/construct then the model is not identified. It
might help to know a little more about what you are attempting. Also what
specfic error messages did you receive?
 
Soumen.
 
> -----Original Message-----
> From: Pedro J. Cuestas [SMTP:pcuestas@FCU.UM.ES]
> Sent: Thursday, November 27, 1997 1:28 PM
> To:   SAS-L@UGA.CC.UGA.EDU
> Subject:      CFA with one factor
>
> Hi,
>
> I'm trying to run a confirmatory factor analysis with just one factor and
> three observed variables. I've written the following program (see below),
> but I got an error message. Can anyone help me to run this CFA?
>
> Thanks a lot.
>
> -------------------------------------------------
> proc calis data=sasuser.ecos02 method=ml edf=128 all;
>     lineqs
>         pe03v001=v1 f1+e1,
>         pe03v02r=v2 f1+e2,
>         pe03v003=v3 f1+e3,
>         f1=v4 f2+e4;
> std
>    f2=1.0,
>    e1-e3=u21-u23,
>    e4=1.0;
> bounds
>    0<=u21-u23;
> run;
> ------------------------------------------------------