From: John Fox (firstname.lastname@example.org)
Date: Fri 14 May 2004 - 04:02:55 EST
If you're willing to assume multinormality but don't have in mind a
specific model to estimate, then why not use factanal? The uniquenesses
are estimates of the unreliability of the items.
I hope this helps,
On Thu, 13 May 2004 17:09:14 +0200 (MEST)
"Jens Oehlschlägel" <email@example.com> wrote:
> Dear John,
> Dear Joseph,
> Thank you for your quick answers and the pointer to semnet.
> I try to clarify on my assumptions:
> - yes, I am willing to assume multivariate normality
> - no, I don't want to assume a single factor model
> - I assume there is an unknown number of factors, and I do not know
> items belong to which factors
> but I still want to estimate single item reliabilities
> Is this impossible?
> How can one specify such unspecified model to sem?
> Which (exploratory?) function would me help doing this?
> Or just: what should I read?
> Thanks for any help
> Jens Oehlschlägel
> Rfirstname.lastname@example.org mailing list
> PLEASE do read the posting guide!
Department of Sociology
Hamilton, Ontario, Canada
Remail@example.com mailing list
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
This archive was generated by hypermail 2.1.3 : Mon 31 May 2004 - 23:05:09 EST