LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous messageNext messagePrevious in topicNext in topicPrevious by same authorNext by same authorPrevious page (October 2002)Back to main SPSSX-L pageJoin or leave SPSSX-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:         Sun, 20 Oct 2002 14:49:18 -0500
Reply-To:     R Cole <recole73@insightbb.com>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         R Cole <recole73@insightbb.com>
Subject:      Re: New Variables from Factor Analysis
In-Reply-To:  <3DB2BCC0.A40D59@mail.montclair.edu>
Content-Type: text/plain; charset="us-ascii"

I have had that concern with my sample size, but that is what I have to work with in the study. My intention is to raise the threshold for Significant Factor Loadings in determining the factors. In this case, I was considering using .6. This is based on the recommendations of Hair, et al (1998) (p112) where a factor size of .7 is recommended for a sample size of 60, and then somewhat lowering the factor size based on the large number of variables. I had indication the Factor Analysis was appropriate with an Overall Measure of Sampling Adequacy of .716 and a Bartlett Test of Sphericity: 754.650 Significance: .000.

Would these additional measures be considered adequate for use in a study of managerial perceptions of barriers to effective implementation of a new management system? In this case, where the purpose is to provide managers guidance on what barriers to avoid or make allowances for, I am more concerned with type-2 error.

Thanks very much for your advice.

Being new to this list, would it be more appropriate to conduct these discussions off the list? Richard

-----Original Message----- From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of Gil Klajman Sent: Sunday, October 20, 2002 9:25 AM To: SPSSX-L@LISTSERV.UGA.EDU Subject: Re: New Variables from Factor Analysis

In addition, a factor analysis with barely two cases per variable is in itself very unreliable. See, for example a text such as Tabachnik and Fidell, who recommend a minimum of five cases per variable, for detail.

Gil


Back to: Top of message | Previous page | Main SPSSX-L page