[From Rick Marken (950102.2100)]
Welcome back Bob Clark. Nice to have you here again.
Bruce Abbott (950102.1615 EST)--
What got me started along this line was Rick's assertion that the
methods outlined in my text are useless for studying "living control
systems."
Bill Powers (950102.1825 MST) --
I agree that Rick is overgeneralizing.
Perhaps just being unclear. If "IV-DV methodology" is taken to refer
simply to the process of manipulating one variable while monitoring
any resulting change in another, then it is an overgeneralization to say
that IV-DV methodology is useless for studying control systems. In
"the test", for example, we manipulate one variable (disturbance) and
look for an effect on an output variable or lack of effect on a
hypothetical controlled variable.
When I refer to "IV-DV methodology", however, I am referring to the
application of a methodology in the context of a cause-effect model; this
is how IV-DV methodology is taught in psychological research methods
texts. The student is taught that the IV is the presumed cause and the
DV is the presumed effect. Most of the statistical methods used
to test for the effect of IV on DV are based on the "general linear
model". With two IVs the model looks like this:
DV = c + k.1IV.1+k.2IV.2 +k.3IV.1*IV.2+error
The third term is the interaction and, of course, you can include higher
order terms too (IV.1^2, IV.1^3, etc). The general linear model explicitly
views the organism as a "transfer function" that transforms input
variables (the IVs) into output variables (DVs). The assumption is that
the coefficients of the terms in the model tell you something about
how the organism responds to input variables. It is this assumption --
which is the basis of the application of what I call IV-DV methodology
in conventional psychology -- that I say is false when you are dealing
with a negative feedback system (ie. a living organism). This is why I
say that conventional IV-DV methodology is useless in the study of
living control systems.
Best
Rick