[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Re: [renamed] Factor Analysis and Missing Data

From   Maarten buis <>
Subject   Re: st: Re: [renamed] Factor Analysis and Missing Data
Date   Tue, 2 Feb 2010 16:04:12 +0000 (GMT)

--- On Tue, 2/2/10, Joseph Coveney wrote:
> <snip> I wonder how badly misled Jet would be by submitting
> a matrix of pairwise correlation coefficients to 
> -factormat , forcepsd-. Is anyone aware of anything on this,
> pro or con?  It could be worth a simulation study if there's
> nothing already out there on this.

There is a paragraph on using pairwise correlation matrices in
linear regression in Paul Allison (2002) Missing Data. Thousand
Oaks: Sage, pp 8-9, with a number of references to articles from
the '60s and '70s. Main message seems to be that you are not
doing too much dammage if your data is MCAR, but things can go
wrong when the data is MAR. Moreover, there isn't that much 
advantage to using pairwise correlaitons when the data is MCAR,
so Paul's conclusion is that the advantages don't outweight the

-- Maarten

Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen


*   For searches and help try:

© Copyright 1996–2020 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index