It seems I have raised a storm of activity, when that was not my
intention. I thank you all for the many valuable emails you have sent
me. However, I must apologize for having raised a tempest in a
teapot. I consider myself internally FLAMED because my ignorance has
cost some of you time which needn't have been spent.
I DO need additional information regarding the best way to reflect
constructs/contrasts in the current study. Several responses from you
have been fruitful, but often off-target because I had not been
suficiently specific regarding the nature of this study.
Following is a description of the current study. Let me emphasize that
elements (columns) of the grid, as well as the rows
(constructs/contrasts) are dealt with via hierarchichal cluster
analysis. These analyses rely on distance measures, and not on correlations.
The problem which has arisen has to do with two facts: (1) Many Ss
produce the same constructs/contrasts under different labels; such
constructs/contrasts should be organized into a single construct/contrast
and (2) Many Ss organize two constructs/contrasts as direct opposites of
each other -- as an hypothetical example, one construct may display
GOOD/BAD while the other may display BAD/GOOD. These two
constructs/contrasts may be identical, excepting only that the S wrote
them in reverse scores. This is the situation in which reflection of a
construct is necessary, and this is the situation in which I request the
help of the PCP network.
Following is a more complete description of the current study. I hope it
is helpful.
METHODS AND TECHNIQUES IN THE CURRENT STUDY
The study uses an 18 element x 18 construct/contrast grid. Triadic
comparisons involve 3 direct comparisons in each row, and each column, of
the grid.
The elements (role stereotypes) are as follows:
A female friend
Me as I am
A male friend
Someone I disappointed
Someone who is dependable
Someone who is knowledgeable
My father
A female I dislike
Someone who is unhappy
Me as my spouse sees me
My mother
Someone I respect
A male I dislike
Me as I would like to be
Someone who is angry
My spouse
Someone who is funny
Someone who disappointed me
A copy of the grid is available via snail-mail if I have a dirt address
to send it. This displays the exact nature of comparisons/contrasts.
Each subject is instructed to use a plus sign (+) to indicate those who
are "alike", a minus sign (-) to indicate those who are "different", and
to leave blank the rest. These responses are subsequently numerically
coded as +1, -1, and 0, respectively. Distances between pairs of columns,
and pairs of rows, of the grid are calculated using the Euclidean metric.
The numerical data are then subjected to an hierarchical cluster analysis.
Clustering uses the method provided by Johnson (Psychometrika, V. 32, No.
3, 1967, pp. 241-254) as modified by Hubert (Psychometrika, V37, No. 3,
1972, pp. 261-274). The clustering pattern selected for inclusion into
the study consists of that pattern "containing the maximum number of
clusters provided that (Hubert's) goodness-of-fit criterion is .95 or
greater."
Sample composition is very restricted in order to enhance replicability.
Subjects must be legally married; it must be their first marriage; they
must have been married for 2 or more years; they must currently be
cohabiting; they must be between 20 and 60 years of age; they must be
Caucasian (and non-Hispanic); their parents must have been married and
cohabiting up and until the time they (the S's) were emancipated;
English must be the primary language spoken by the Ss; Ss must never
have been in psychotherapy apart from brief premarital pastoral
counseling; and neither spouses' parents could currently be living with S.
The sample is intended to be large so that any "effects" NOT found would
be without meaningful significance (of cource ignoring validity of
instrumentation). The targeted sample size is 80 couples (160 Ss);
this might be increased to a total of 95 couples -- the future will tell.
Instrumentation includes the Locke-Wallace Marital Adjustment Scale, the
Areas Of Change Questionnaire (a measure of wished-for changes in the
marriage), the Marlowe-Crown Social Desirability Scale, and a demographic
questionnaire (available via snail mail).
Two distinct studies are being conducted on this database. STUDY 1:
Marital satisfaction is higher when spouse and opposite sexed parent are
seen as similar (but not too similar because of incest guilt). This
Freudian tenent has been held for over 70 years, but never demonstrated.
There are a couple of lesser hypotheses being tested in Study 1. STUDY
2: Marital satisfaction is greater to the extent that cognitive
complexity (the complexity of constructs/contrasts) is similar between
husband and wife. Marital satisfaction is greater to the extent that
relational complexity (the complexity of grid elements) is similar
between husband and wife. Marital satisfaction is better predicted by
the similarity (between husband and wife) of cogitive complexity and
relational complexity together, than by either measure alone.
I still ask for your help, and hopefully have been more clear in
explaining the problem so that I no longer waste your time.
Sincerely,
F. Reid Creech
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%