s-news
[Top] [All Lists]

Differences in DF for factanal

To: <s-news@lists.biostat.wustl.edu>
Subject: Differences in DF for factanal
From: "Monroe, Dean" <dean.monroe@okstate.edu>
Date: Wed, 22 Sep 2004 11:14:57 -0500
Thread-index: AcSgv02MDqccGk4tSRag4A2vOMPwqQ==
Thread-topic: Differences in DF for factanal
Hello:
 
Recently I was working on a Factor analysis project using the factanal command in Splus with mle estimation.  I entered 9 variables and wanted 5 factors, then ran the algorithm.  The significance test showed 5 factors tested with 1 df.  I then ran 4 factors and got 4 factors tested with 6 df.  
 
Why, by changing the number of factors 5 to 4 did the df change by 5?  I would have expected maybe a 2 df change.  Has anyone researched this problem out to see exactly how Splus calculates df in factanal?
 
Dean Monroe 
<Prev in Thread] Current Thread [Next in Thread>
  • Differences in DF for factanal, Monroe, Dean <=