Re: To Lois on Construct Synthesis

Lois Shawver (
Wed, 6 Mar 1996 18:20:01 -0800 (PST)


At my request you were setting up an example to help me understand
what you mean by "construct weaving". Let me ask a few more questions to
see if I understand your example.

To help illustrate your ideas, you presented the following grid:

50 persons (elements) along the 5 constructs.
Grid Elements (N=50)
Mary Willie Joseph ........ Mike
Intelligent 1 5 4 3
Competent 3 7 1 1
Careful 7 6 2 5
Wise 5 5 1 4
Educated 1 4 7 7

And you said:
Note that the configuration does not require that
Competent be uncorrelated with Wise. It may or
may not be. Both Wise and Competent may
share a common source in a fourth variable;

I want to ask (I'm just trying to understand this language):

Isn't the correlation of competent with wise an empirical matter?
If the subjects had not responded:
Competent 3 7 1 1
Wise 5 5 1 4

but had responded instead with:
Competent 7 3 1 1
Wise 5 5 1 2

wouldn't we have had a higher correlation between these two variables?
And, when you are talking about the "common source in a fourth variable"
are you talking about both competent and wise being more correlated with
"education" than with each other? As might be the case in the following

Competent 3 7 1 1
Wise 5 5 1 4
Education 4 6 1 2

Is this what you mean?

If so, then going to your next point, you say:

These numbers are just random and not meant at this
time to simulate real formal causation.

You mean by "causation" the higher correlation with the "Education" than
between "comepetent and wise"? Is your imagery that the
variable"Education" is working behind the scenes, so to speak, to affect
the variation in "competent and wise" such that it is? Putting it another
way, are you thinking that "competent" and "wise" are being treated like
euphemisms for the _real_ variable "education"?

And is something like the account I just gave what you see as involved when
you say:
In the real world we would be interested in discovering
if the client is implicity synthesizing any of the
constructs from the other constructs.

..Lois Shawver