Re: Chris Evans reply

Mon, 13 Mar 1995 13:09:11 -0600 (CST)

I wouldn't view principal components as hierarchical, since all the components
are on one level. Potential disadvantages are that (a) they are indeed all on
one level of extraction and (b) reflect only linear relationships between or
among constructs. A better alternative for both hierarchy and asymmetric
potentially nonlinear relationships might be clustering procedures, such as
HICLAS by de Boeck.

On the other hand, principal components are not without merit: (a) useful
predictions derived from cognitive complexity indices (number of factors,
variance accounted for in the first one or two factors, etc.), (b) improved
clinical understanding by identifying idiosyncratic groupings of constructs
highly loaded on the same factor, relationships of which would not be
nomothetically predicted in the general population but of high and reliable
significance in a given individual's construct system.


From: IN%"" 13-MAR-1995 00:37:56.34
To: IN%""
Subj: RE: Chris Evans reply

Return-path: <>
Received: from (
by KUHUB.CC.UKANS.EDU (PMDF V4.3-11 #9008)
13 Mar 1995 00:37:53 -0600 (CST)
Received: by id <>
( for; Mon, 13 Mar 1995 06:13:59 GMT
Received: from by id
<> ( for with SMTP; Mon,
13 Mar 1995 06:13:56 GMT
Received: by (4.1/SMI-SVR4) id AA08073; Sun, 12 Mar 95 22:11:06 PST
Date: Sun, 12 Mar 1995 22:11:06 -0800 (PST)
From: "Tim A. Connor" <>
Subject: Re: Chris Evans reply
In-reply-to: <1995Mar13.104255-0500@[]>
Message-id: <>
MIME-version: 1.0
Content-type: TEXT/PLAIN; charset=US-ASCII
Content-transfer-encoding: 7BIT
Precedence: list

A question re principal component analysis of grids (I'm no doubt
revealing my statistical naivete, but what the hell): are the factors
conceptualized as unverbalized superordinate constructs? If not, what
are they? I confess to a certain discomfort with entities that have no
non-statistical correlate. I'll probably get over it one of these days...

Tim Connor