<- file stat 97sem.html -> Two topics: Confirmatory factor analysis; N for Lisrel.

Confirmatory factor analysis

=======================Jeremy Miles, 12 Feb 1997==========spss Message-ID: <3.0.1.32.19970212101711.010c4634@unix1> From: Jeremy Miles <J.N.V.Miles@DERBY.AC.UK> Subject: Re: Factor Analysis: Confirmatory & Exploratory At 22:00 11/02/97 EST, Steven E. Barkan wrote: >Could someone please inform me how to have SPSS do confirmatory >factor analysis. I've used factor analysis a lot, but have recently >been told I've been doing exploratory factor analysis but for a particular >project need to do confirmatory factor analysis. Thanks in advance for >any help. > Here we go again. :-) You can do a kind of confirmatory factor analysis (CFA), using ML extraction in SPSS Factor, examining the chi-square statistic, and selecting the smallest number of factors which are not significant. As far as I know (and I may be wrong about this) the chi-square statistic is prone to a number of problems, which havbe been documented in CFA literature, but which I have not seen in EFA literature (but I have never looked hard). What is usually referred to as CFA is done using a structural equation modelling package, such as (anyoen guess what I am going to say?) AMOS, which SPSS now distribute on some platforms, or LISREL, which they distribute on others. If you have a choice, go for AMOS. A good introduction to this sort of thing is Loehlin, J. Latent variable models, which describes CFA and relates it to EFA and regression (in which it is, at least in part, rooted). Its published by Erlbaum. For more detailed stuff see Bollen, KA - structural equations with latent variables, Wiley) ong, JS - Confirmatpry factor analysis (Sage), or Bollen and Long - Testing structural equation models (Sage again). Also check out the SEMNET FAQ: http://www.gsu.edu/~mkteer/semfaq.html

LISREL sample sizes

=======================TorNeilands, 17 Mar 1997==========spss Message-ID: <v01510107af5300ab72c5@[146.6.54.17]> From: Tor Neilands <IceT@MAIL.UTEXAS.EDU> Subject: Re: sample size for LISREL At 12:00 AM 3/17/97, Automatic digest processor wrote: > >Who can tell me how to calculate the sample size for using LISREL >analysis? Thanks very much! My e-mail is chiou@dolphin.upenn.edu > You might try one (or both) of the following two sources: Loehlin, J.C. (1992) Latent variable models. Hillsdale, NJ: Lawrence Erlbaum Publishers Kaplan, D. (1995) Statistical power in structural equation modeling. In R. Hoyle (Ed.) Structural Equation Modeling: Concepts, issues, and applications. Newbury Park, CA: Sage Publications Good luck, * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Document by Rich Ulrich. E-mail to wpilib+@pitt.edu

FAQ top.

Ulrich home page.

Ulrich FAQ. http://www.pitt.edu/~wpilib/stats99.html