------=_NextPart_000_000A_01BD8EDB.3FE387E0
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
My Doctorate was in Architecture in the early 70's and I definitely
considered grid methods.
My understanding is that you have done the method of triads WITHOUT having
your architects apply the elicited construct/contrast pair to all the
streetscapes.
As to quantitative modelling or Mapping of the grids you have a problem.
What you have done is collect a very limited and biased number of triadic
comparisons.
For each pair of streetscapes you could count the number of times they have
been put in the same pair as a proportion of the number of times they were
in the same triad. If you had enough you could then generate a similarity
matrix between the streetscapes and produce a perceptual map on that basis.
This would not include the elicited constructs in the map. However you
probably can't even do that. That is where the bias comes in. You let your
architects select the triads. To be able to extract anything useful for
quantiative you have to generate the triads. IF you aren't going to have the
architects rate all streetscapes on all constructs.
There might be a possiblility of generating a similarity matrix for all the
architects triadic judgements.
However I'd advise you to forget the quantitative and just do content
analysis of the words used - you should be able to do something with that
without any suggestions!
Dr Stephen K Tagg - s.k.tagg@strath.ac.uk
vCard Attachment for further details (needs Outlook)
-----Original Message-----
From: pcp-request@mailbase.ac.uk [mailto:pcp-request@mailbase.ac.uk]On
Behalf Of R.Rezazadeh
Sent: Tuesday, June 02, 1998 6:09 PM
To: pcp@mailbase.ac.uk
Subject: Technical advice needed
Dear pcp group
I am conducting my PhD research in urban design. Part of my study
involves the perceptions of the architects of the streetscapes. I
have conducted a test using 16 streetscenes as stimuli and 30
architects as subjects. While showing them the 16 streetscenes laid
on a table in random order, I have asked each respondent to find two
similar and one different streetscene and state the aspect of
similarity and difference. Each respondent has continued the task,
until he/she could not identify any new aspects for
similarity/difference. Therefore for each respondent I have a grid in
which two similar stimuli and a different one is identified and the
aspect of similarity/difference is stated. Can I use the PCP approach
for analysing my data and finding the constructs within the perception
of each person, or the group as a whole? If yes, how? If no, what
other method do you suggest. Thank you.
------=_NextPart_000_000A_01BD8EDB.3FE387E0
Content-Type: application/octet-stream;
name="Stephen K Tagg (E-mail 2).vcf"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: attachment;
filename="Stephen K Tagg (E-mail 2).vcf"
BEGIN:VCARD
VERSION:2.1
N:Tagg;Stephen;;Dr.
FN:Stephen K Tagg (E-mail 2)
ORG:University of Strathclyde
TITLE:Lecturer in Marketing
TEL;WORK;VOICE:+44 (0141) 548 2409
TEL;HOME;VOICE:+44 (01360) 770971
TEL;WORK;FAX:+44 (0141) 552 2802
ADR;WORK;ENCODING=3DQUOTED-PRINTABLE:;;Stenhouse Building=3D0D=3D0A173 =
Cathedral Street;GLASGOW;Scotland;G4 0RQ;Unite=3D
d Kingdom
LABEL;WORK;ENCODING=3DQUOTED-PRINTABLE:Stenhouse Building=3D0D=3D0A173 =
Cathedral Street=3D0D=3D0AGLASGOW, Scotland G4 0RQ=3D
=3D0D=3D0AUnited Kingdom
ADR;HOME;ENCODING=3DQUOTED-PRINTABLE:;;Lower Lodge=3D0D=3D0ABlane =
Avenue;Blanefield;Stirlingshire;G63 9HU;United King=3D
dom
LABEL;HOME;ENCODING=3DQUOTED-PRINTABLE:Lower Lodge=3D0D=3D0ABlane =
Avenue=3D0D=3D0ABlanefield, Stirlingshire G63 9HU=3D0D=3D0AUn=3D
ited Kingdom
EMAIL;PREF;INTERNET:s.k.tagg@strath.ac.uk
REV:19980501T100130Z
END:VCARD
------=_NextPart_000_000A_01BD8EDB.3FE387E0--
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%