Your file request
Your CIOS file request: Q-METHOD/06179063.304 hotline item

-
Received:  by CIOS Mailer; Wednesday 17 Jun 2009 06:33:04
Date:         Wed, 17 Jun 2009 12:22:00 +0200
From:         Ingunn St=?ISO-8859-1?Q?=F8rksen?= 
Subject: Question on Q-samples and their structure
To:           Q-METHOD@LISTSERV.KENT.EDU
List-Help: ,              
List-Unsubscribe: 
List-Subscribe: 
List-Owner: 
List-Archive: 

This is a multipart message in MIME format.
--=_alternative 0038F264C12575D8_=
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: quoted-printable

Dear Q-friends, I have been reading William Stephenson=B4s chapter (IV) on =

Samples and their structure (in The study of behavior, 1953). Although I=20
think I understand the principles of block designs as they are explained=20
and exemplified here, and although I see some of the advantages of block=20
designs, I also have some questions.
It seems to me that the examples given often have a very large number of=20
cells. For instance on page 77 there is an example with 48 cells. This=20
seems extremely fine-graded/detailed, and if I was to find statements for=20
each cell in such a structure I believe I would have had to construct=20
quite a few of the statements myself, rather than relying on interviews.=20
Each statement in such a structure also need to meet 4 criteria/level=20
descriptions (e.g. acei or acek etc), which seems hard ? at least=20
regarding the themes that we are studying at the moment.=20
For the moment we are trying to dig into children?s subjectivity regarding =

family, divorce etc. If you interview children, the statements that come=20
out often seem simple and straightforward, e.g. ?I really love my dad?,=20
and it seems hard to fit this into such fine-graded systems. In stead, in=20
several occasions, we have used simple cross tables (e.g. 2X2 or 2X3)=20
ending up with 4 or 6 cells from structuring the samples. For instance in=20
a single-case study entitled Family narratives through the eyes of an=20
adult child of divorce we simply crossed the two dimensions ?child vs=20
adult perspective? and ?statements on divorce vs living arrangements? and=20
ended up with 4 cells: ( The journal has kindly given us permission to put =

the article on our web-pages here:=20
http://saf.uis.no/getfile.php/SAF/Bilder/Publikasjoner/St%C3%B8rksen%20et%2=
0al%20JHS%202008.pdf=20
)

Child perspective
A
Adult perspective
B
Experiences related to the divorce
C

A/C

B/C
Experiences related to the living-arrangements
D

A/D

B/D
St=F8rksen, I., Berner, K. L., & Thorsen, A. A. (2008). Family narratives=20
through the eyes of an adult child of divorce. Journal of Human=20
Subjectivity, 6 (2), 27-46.
I notice that Stephenson (page 78) states that ?The sample is thus never=20
selected purely randomly from the universe?. The question is: would it be=20
preferable to use the classical ?effect (or independencies) and levels?=20
design (as described by Stephenson and others) for structuring the=20
Q-sample, or is it OK in some cases to use simpler structures such as=20
simple categories or cross tables? On page 74 in the same chapter=20
Stephenson mentions that the structured sample has some advantages=20
regarding error. Are we missing this advantage or any other advantages?
Thanks for all help and advice that any of you are kind to give!
Best whishes from Norway, Ingunn St=F8rksen
Ingunn St=F8rksen
Associate Professor/ Dr. Psychol.
Centre for behavioural research
University of Stavanger
N-4036 STAVANGER
NORWAY
(+ 47)
5183 2900
5183 2934
918 47 157
ingunn.storksen@uis.no=20
http://www.uis.no/q-konferanse
http://www.uis.no/bambi
--=_alternative 0038F264C12575D8_=
Content-Type: text/html; charset="ISO-8859-1"
Content-Transfer-Encoding: quoted-printable



Dear Q-friends, I have been rea=
ding
William Stephenson=B4s chapter (IV) on Samples and their structure (in The
study of behavior, 1953). Although I think I understand the principles
of block designs as they are explained and exemplified here, and although
I see some of the advantages of block designs, I also have some questions.<=
/font>
It seems to me that the examples
given often have a very large number of cells. For instance on page 77
there is an example with 48 cells. This seems extremely fine-graded/detaile=
d,
and if I was to find statements for each cell in such a structure I believe
I would have had to construct quite a few of the statements myself, rather
than relying on interviews. Each statement in such a structure also need
to meet 4 criteria/level descriptions (e.g. acei or acek etc), which seems
hard – at least regarding the themes that we are studying at the mome=
nt.

For the moment we are trying to
dig into children’s subjectivity regarding family, divorce etc. If you
interview children, the statements that come out often seem simple and
straightforward, e.g. “I really love my dad”, and it seems hard=
 to fit
this into such fine-graded systems. In stead, in several occasions, we
have used simple cross tables (e.g. 2X2 or 2X3) ending up with 4 or 6 cells
from structuring the samples. For instance in a single-case study entitled
Family narratives through the eyes of an adult child of divorce
we simply crossed the two dimensions “child vs adult perspective̶=
1; and
“statements on divorce vs living arrangements” and ended up wit=
h 4 cells:
( The journal has kindly given us permission to put the article on our
web-pages here: http://saf.uis.no/getfile.php/SAF/Bilder/Publikasjoner/St%C=
3%B8rksen%20et%20al%20JHS%202008.pdf
)





Child perspective
A


Adult perspective
B



Experiences related =
to
the divorce
C


A/C


B/C


Experiences related =
to
the living-arrangements
D


A/D


B/D
St=F8rksen, I., Berner, K. L., &=
amp;
Thorsen, A. A. (2008). Family narratives through the eyes of an adult child
of divorce. Journal of Human Subjectivity, 6 (2), 27-46.
I notice that Stephenson (page 7=
8)
states that “The sample is thus never selected purely randomly from t=
he
universe”. The question is: would it be preferable to use the =
classical
“effect (or independencies) and levels” design (as described by=
 Stephenson
and others) for structuring the Q-sample, or is it OK in some cases to
use simpler structures such as simple categories or cross tables? On page
74 in the same chapter Stephenson mentions that the structured sample has
some advantages regarding error. Are we missing this advantage or
any other advantages?
Thanks for all help and advice t=
hat
any of you are kind to give!
Best whishes from Norway, Ingunn
St=F8rksen
Ingunn St=F8rksen

Associate Professor/ Dr. Psychol.

Centre for behavioural research

University of Stavanger

N-4036 STAVANGER

NORWAY

(+ 47)

5183 2900

5183 2934

918 47 157

ingunn.storksen@uis.no  

http://www.uis.no/q-konferanse<=
/font>

http://www.uis.no/bambi
--=_alternative 0038F264C12575D8_=--