Your CIOS file request: Q-METHOD/06099070.606 hotline item

-
Received:  by CIOS Mailer; Tuesday 9 Jun 2009 07:06:06
Date:         Tue, 9 Jun 2009 12:03:33 +0100
From:         Louise Bryant 
Subject: Variance explained and bipolar factors
Comments: cc: Emma Van der Gucht 
To:           Q-METHOD@LISTSERV.KENT.EDU
List-Help: ,              
List-Unsubscribe: 
List-Subscribe: 
List-Owner: 
List-Archive: 

Dear all,

Another query - which links the earlier discussion related to Sam Hopper's =
variance issue and my own about separating bipolar factors.

In our study, the pre-rotated four factors we selected accounted for 47% of=
 the variance (26, 8, 7, 6).

Because two of the four exemplars for Factor 2 were negative ones, we split=
 F2 out into 2 factors - resulting in a five factor solution. The .lis file=
 now shows the variance accounted for as 17, 8, 8, 10, 12 - which adds up t=
o  55, i.e. 47% + another 8%. I would have expected the variance of the ori=
ginal F2 to have been split in some way across the new Factors F2 & F3 - wh=
y hasn't this happened? This contravenes the laws described previously abou=
t the same amount of variance being accounted for before and after rotation=
.

Thanks

Louise

Dr Louise Bryant
Lecturer in the Psychology of Healthcare
University of Leeds
Leeds Institute of Health Sciences
Room 1.12, Charles Thackrah Building
101 Clarendon Road,=20
Woodhouse
Leeds LS2 9LJ
=20
tel: 0113 343 1882
fax: 0113 343 0850