More on models

CSGnet Archive CSG1997
1

The dialogue between Tracy and Hans on PCT and models of human behavior
is an interesting strand which I'd like to follow up on. Going back to
Wiener's work on cybernetics, the main focus on any of these models of
human behavior or living systems is feedback and information.
Mechanistic models can be created which seemingly capture the essence of
human systems along these linesl, as Herbert Simon and folks like Bill
Powers have argued.

But information and feedback are not sufficient concepts for modeling
human behavior. As I argued in a previous post, human systems are
unique because culture provides meaning through symbols that go beyond
the simple information paradigm or model. A good example of the failure
of cybernetics to take this into account is illustrated in a paper by
Overman and Loraine entitled "Information for Control: Another
Management Proverb?", in Public Administration Review 54 (2):193-196
(1994). The cybernetic assumption about management and organizations is
that information increases management control. The idea is that if
feedback of information is properly understood or codified, it can lead
to greater control of the environment and lead to more productive,
efficient organizations.

The authors looked at a random sample of 99 Air Force projects
undertaken between 1968 and 1989. Each project was required to have a
management information system in place to provide feedback on how the
program was proceeding. Some projects required elaborate work breakdown
structures and reports every month, while others required relatively
little detail, with reports occurring perhaps every quarter.

The cybernetic or engineering hypothesis is known as the signal
hypothesis, namely, that the process of information increasing control
in organizations is presumed to be a natural process of information
feedback and deviation correction. Cybernetic theory and the signal
hypothesis state that a direct, positive linear relationship exists
between the information being collected and various aspects of project
control. So the hypotheses are: information decreases cost variance,
information decreases schedule variance, and information decreases
quality variance.

Without getting into operationalization of concepts, no significant
correlations were found between levels of information and "control,"
that is, certain positive outcomes regarding schedule variance, cost,
and quality. What is more appropriate than a simple
information/feedback model (the signal hypothesis) is a symbol
hypothesis. The symbol hypothesis, developed primarily out of sociology
and anthropology, suggests that information is embedded in social and
organizational norms making it highly symbolic. So the symbol
hypothesis would suggest that the relationship between information and
control is much less direct and governed by social norms and managerial
behaviors regarding both information and control. Collecting data
might, in fact, have negative consequences on the performance of the
system in terms of lower employee morale, distrust, etc.

To what extent does PCT model or take into account these realities?

Dr. James J. Chriss
Kansas Newman College
Sociology Department

2

[From Bill Powers (970314.1515 MST)]

James H. Chriss (970314) --

Your post confirms what I said before: you need to learn a lot more about
PCT before attempting to use it in your work. Consider this:

The cybernetic assumption about management and organizations is
that information increases management control. The idea is that if
feedback of information is properly understood or codified, it can lead
to greater control of the environment and lead to more productive,
efficient organizations.

The authors looked at a random sample of 99 Air Force projects
undertaken between 1968 and 1989. Each project was required to have a
management information system in place to provide feedback on how the
program was proceeding.

That concept stems from a half-hearted attempt to interpret the word
"feedback," and it completely misses the main phenomenon studied in PCT.
Feedback isn't something that one person can give to or withhold from
another (by any ordinary means). It's the immediately perceived consequence
of a person's own actions. To prevent it you would have to destroy all of a
person's sensory equipment. Without feedback it is impossible to perform any
kind of regular behavior, even so simple as standing erect or getting food
into your mouth. Feedback, to a painter, doesn't mean having an art critic
standing by giving gratuitous opinions about the progress of a picture; it
means being able to see how the canvas looks when you add paint or scrape
some off, so you can tell whether it is approaching the state that YOU THE
PAINTER, not the art critic, want to see. Try blindfolding the painter,
cutting off her feedback, and see how well she paints, assuming you can
persuade her even to try!

Some projects required elaborate work breakdown
structures and reports every month, while others required relatively
little detail, with reports occurring perhaps every quarter.

The cybernetic or engineering hypothesis is known as the signal
hypothesis, namely, that the process of information increasing control
in organizations is presumed to be a natural process of information
feedback and deviation correction. Cybernetic theory and the signal
hypothesis state that a direct, positive linear relationship exists
between the information being collected and various aspects of project
control. So the hypotheses are: information decreases cost variance,
information decreases schedule variance, and information decreases
quality variance.

Without getting into operationalization of concepts, no significant
correlations were found between levels of information and "control,"
that is, certain positive outcomes regarding schedule variance, cost,
and quality.

Of course not; whoever put these studies together didn't have the faintest
idea of the role of feedback in human behavior. Generating "elaborate work
breakdown structures and reports" is a silly bureaucratic conceptualization
of what "feedback" means; I'll bet that whoever wrote those elaborate
reports never checked to see if anyone cared about them or was even reading
them (as opposed to checking the box showing that it went across one's
desk), which would be a true example of feedback relative to the
report-writer (not the recipient of the reports). Such reports aren't
feedback unless the person receiving them demands them and unless the
information in the reports is directly relevant to the person's own goals,
to what the person is trying to accomplish. Feedback in management isn't
receiving mountains of unwanted opinions from a bystander; it's finding out
whether that crucial letter you gave your secretary was ever actually
mailed, or exactly when, or if, the supplier shipped the crucial missing parts.

To what extent does PCT model or take into account these realities?

You have the question backward. To what extent do the so-called "realities"
you mention (actually, theories) take the PCT model into account?

Best,

Bill P.