This commentary is critical of Goltz and Hietapelto’s operant
model of power (2002) but critical with great respect because the
model reflects analytical thoughtfulness and a commendable desire
for conceptual parsimony and because the model is presented in a clear,
well-written exemplary manner, with many ingenious applications. In
other words, the model and its presentation are both well thought-out.
According to Goltz and Hietapelto’s model, “power”
is the ability to control the operant behavior of others through the
manipulation of behavioral contingencies; thus, to be powerful, one
must have access to effective behavioral consequences and the skills
to use them in an appropriately contingent manner. And according to
this model, people resist organizational change, to the extent that
it reduces their power. However, I think this operant model too greatly
restricts both the concept of “power” and the sources
of resistance to change.
What
The Powerful Do
Administrators may have power to implement new programs, a new product
line in manufacturing, a new college in a university. But the administrator
will not necessarily be responsible for assigning control over the
program to a particular person or group. And if the administrator
does make that assignment, the assignment might be made as a result
of finding the best fit from a nation-wide job search, rather than
as an effort to “reinforce” or control the operant behavior
of the recipient of that job assignment through the manipulation of
that behavioral consequence.
Powerful people distribute resources and conditions, that are often
reinforcers and aversive conditions; but for Goltz and Hietapelto
to call those resources “rewards” and “punishers”
or “consequences” suggests the powerful use them contingently.
However, often, perhaps most often, powerful people do not distribute
those resources and conditions contingent on worker behavior. For
example, our university administration controls the thermostat settings
in all the buildings on campus; so the secretaries sweat during the
summer to save air-conditioning electricity, regardless of the quantity,
quality, and nature of their work. Our administration is not using
the air-conditioning as a behavioral consequence, though they do have
and use the power to control it. Therefore, I think reference to the
quasi-behavioral terms, “rewards,” “ punishers,”
and “consequences,” is unwarranted.
Why do managers want resources? I suspect it is so they can do their
jobs better, not so that they can use those resources as contingent
reinforcers to control their workers’ behavior, in other words,
not so they can exert “social power” over their subordinates.
They want to buy more and better equipment for all their workers,
so those workers can produce more and better-quality products. And
they want to hire more workers, so their unit can produce more and
better-quality products. That does not mean that some managers do
not distribute a small amount of their resources contingently, either
intentionally or unintentionally. So I think the social-power model
may lead us away from the relevant controlling variables, rather than
point us to them, the relevant variables often being the opportunity
to do a better job.
Concerning power potential, I wonder if worker ratings of managers
reflects perceived potential for determining the distribution of resources
and conditions more than the actual, contingent distribution. In fact,
I suspect a worker’s immediate supervisor and fellow workers
exert more differential control over resources and conditions than
does the CEO.
Networking
Goltz and Hietapelto say, “individuals . . . may seek to develop
horizontal and vertical [social networks],” with, what I take
to be the implication that such seeking is common practice. But, while
I think people do end up in horizontal and vertical networks, I doubt
if it is usually a result of seeking to do so; and to suggest that
the result is intentional may be almost teleological. Promotions up
the vertical network usually result in better pay, better working
conditions, and more prestige, as well as more social power. It is
not clear that social power is usually the most powerful of those
potential incentives. Similarly, husbands have social power over their
wives in that wives typically adopt the voting pattern of their husbands
(at least they used to); but I doubt if most husbands seek the horizontal
social network of marriage in order to maximize their contribution
to the political party of their choice.
Although some people clearly seek to develop social networks to achieve
social power over others, that may not be typical of most. And, in
fact, it may be that much network seeking that actually occurs is
to gain or maintain control over ones own destiny not over the destiny
of others. In other words, I doubt that “when a person’s
position is low in [social power], he or she is likely to seek to
increase [that social power] to enhance his or her influence over
consequences, and as a result, over behavior [of others].” I
think he or she is more likely to seek better pay, etc. If I am wrong,
I would love to see the data.
Concerning the importance of being in a social network, my informal
observation would be that people use their network position much more
to ensure that they themselves rather than someone else get the desired
resources, with the result that people less powerfully placed do not
get those resources. But those well placed are not using their influence
over resources as reinforcers to control the behavior of those further
removed from the decision making.
Leadership
Goltz and Hietapelto say, “In the operant model of power, leadership
is defined as being an individual’s skill in using the consequences
under his/her direct or indirect control to influence behavior toward
goals that will obtain rewards for the unit.” And I think being
a good contingency manager is an important attribute of a good leader,
though I am still not convinced that most good leaders are good contingency
managers; but I would be a happy man, if I could see clear, solid,
data to the contrary, data that do not require too great an inferential
leap of faith.
Also, I think their operant model neglects some of the most important
characteristics of leadership--the power to select the goals toward
which the leader will lead, and hopefully to select those goals wisely.
Of only slightly less importance is the ability to use rational and
moral persuasion to convince the subordinates to work toward those
goals. For example, political leaders often attempt and sometimes
succeed in persuading the electorate to vote for political candidates
on rational and moral grounds, without the promise of some resources
being swung their way (e.g., abortion and gun control). This is not
to deny the advice to follow the money trail when analyzing political
decisions; but even on the money trail, the money does not necessarily
function as a contingent reinforcer controlling the behavior of subordinates.
The Goltz and Hietapelto model predict that people will exert most
control over the behavior of others if they have the most control
over resources and dispense those resources with the most contingency-management
skill. Maybe, but I am not convinced. The CEO of GM, exerted tremendous
control over the entire city of Flint, simply by closing the GM plant
and moving it to Mexico. And he also exerted tremendous control over
workers in Mexico simply by giving them the jobs. But I think it would
strain the credulity of an operant model to say the CEO was strategically
punishing the behavior of the citizens of Flint, because the workers
couldn’t work for Mexican wages and that he was reinforcing
the behavior of the Mexican workers, because they could. The CEO’s
actions are more analogous to simply taking the rat out of the Skinner
box and destroying the box, not a technique studied extensively in
JEAB.
This is not to say that contingency management could not
be an important, effective performance management tool, in organizations,
as well as everywhere else in our lives. But I doubt if it actually is. Furthermore, I am not convinced it could ever be the most important
performance management tool in organizations, when we look at performance
management from a power perspective and include activities that the
powerful often do, like closing and opening factories, divisions,
departments, product lines, colleges, as well as one of their most
prominent contemporary activities—down sizing and also knowledgably
dumping their stock before anyone else gets a chance.
Now, I would be happy to say these activities of closing, opening,
firing, and hiring are not really what we mean by performance management;
but when we start looking at the prerogatives of power, I think we
overstate the importance, both current and potential, of behavior
analysis to the effectiveness of the powerful. Furthermore, I suspect
the powerful understand this and are more receptive to behavior analysis,
when we do not oversell.
Stretching
the Skinner-box Metaphor
It has become common in applied behavior analysis, including OBM,
to couch analyses in terms of those animal-laboratory concepts currently
fashionable, prominent among such concepts being the matching law.
However, often such usage is conceptually erroneous or gratuitous.
In the present case, reference to the matching law seems gratuitous.
Although studies of the matching law address multiple dimensions of
behavioral consequences, I suspect that which specific dimensions
and combinations of dimensions the powerful control is not crucial,
as long as the combination is significant. In addition, it does not
take the matching law, to predict that both rats and workers will
press the lever that produces the most reinforcing outcome; and we
are a long way from worrying about hyperbolic functions.
Furthermore, I think the behavior analyzed in this power model is
controlled by indirect-acting, rule-governed analogs to behavioral
contingencies (e.g., analogs to reinforcement, punishment, escape
and avoidance contingencies), not controlled by the direct-acting
contingencies studied in the Skinner box with non-verbal animals (Malott
& Trojan, in press). But animal research involving direct-acting
contingencies forms the empirical basis for the matching law and another
currently fashionable concept, behavioral momentum. So, I think it
is an error to implicate these concepts in analysis of rule-governed
behavior (Malott, 1993; Malott, Malott, & Shimamune, 1993; Malott,
Shimamune, & Malott, 1993).
Resistance
to Change
Power
Concerning downsizing, Goltz and Hietapelto say, “… managers
will not be happy with this decreased potential for influencing behavior.”
But, I think managers will be equally or more unhappy with the decreased
potential for their unit to get its job done without remaining employees
having to work much harder than before and with the certainty of the
suffering the downsizing will cause those employees being fired.
I agree that power and control over our social and nonsocial environments
are strong learned reinforcers, but I think we should distinguish
between power or control as a hedonic learned reinforcer (one that
has become of value in its own right, like the smile of a passing
stranger and comes to exert almost irrational control over our behavior)
and power or control as an instrumental reinforcer (a learned reinforcer
that is of only transient value because it is needed to get some more
important reinforcer). As an example of power-induced resistance to
organizational change, Goltz and Hietapelto describe faculty and student
refusal to stop using their old PCs in place of a new information
appliance called a “Sunray.” But that refusal may have
been as much or more based on the possibility that the Sunray would
be more difficult to learn to use, more difficult to use, and less
effective as a tool, rather than that the resistance was based merely
or primarily on the reinforcing value of power.
From an applied perspective, what I see as an over emphasis on the
operant power model might well cause practitioners to neglect more
readily accessible variables influencing resistance to change. For
example the change agent might greatly reduce resistance by providing
proper training in the use of the new technology, job aids to facilitate
its use, and examples of its value. In addition, a more careful systems
analysis might sometimes show that worker resistance is appropriate
and not just an irrational urge of the workers to control their own
activities.
Goltz and Hietapelto suggest that loss of power to control the dispensation
of resources explains why managers resist implementing new systems
of performance appraisal requiring them to grade on a curve. However,
my impression is that managers resist most forms of performance appraisal
because performance appraisal is often confrontational and because
the managers have usually not collected the data they need to do an
adequate performance appraisal, with the result that managers frequently
ignore the mandate that they do face-to-face performance appraisals.
Leadership
Behavior analysts often impose a simple Skinner-box reinforcement
model on organizational behavior, calling complex, sustained units
of behavior “responses,” and talking about reinforcing
those “responses,” even though the delay between the terminal
component of those complex units of behavior and their consequences
may be days, weeks, or months, far beyond the delay ever involved
in a Skinner-box demonstration of reinforcement. In recent years,
JOBM authors often substitute the everyday term “rewarded”
for “reinforced,” allowing themselves the freedom to imply
a reinforcement mechanism, without the responsibility of providing
a technical analysis (this terminological vagueness also makes it
easier to inappropriately talk about “rewarding” entire
organizations).
Here is an example of what I think is too metaphorical an extension
of the simple, Skinner-box reinforcement model. Goltz and Hietapelto
address ambiguity in task responsibility resulting from organizational
restructuring. In part, the analysis is that “it is not always
clear whether the new responses will be rewarded . . . .” But,
the very term “responsibility” implies what I think are
the contingencies operating in most organizations: If you fail to
meet your responsibilities, you will get in trouble; if you meet your
responsibilities, you will be left alone. In other words, a rule-governed
analog to avoidance of disaster and reprimand controls the fulfillment
of most of our responsibilities, both personal and organizational.
Goltz and Hietapelto’s example of aligning “the firm’s
incentive and compensation system with” organizational priorities
is far from representative of the real organizational world; and because
of their implicit or explicit deadlines, even such priority-aligned
incentive systems are rule-governed analogs to avoidance of the loss
of the opportunity to receive the incentives, rather than simple,
direct-acting, Skinner-box reinforcement contingencies (Malott and
Trojan, in press).
I think power and leadership are far from being reinforcers of overwhelming
influence. For example, in universities, most faculty members attempt
to decrease rather than increase the size of their classes; and they
also greatly restrict the number of graduate students under their
supervision. While there may be good reasons for this small-is-beautiful
orientation, it would seem to contradict an inherent lust for power
and leadership. Furthermore, it is rare, even among behavior systems
analysts, that faculty members attempt to devise instructional systems
that will maximize the number of students they can teach (i.e., lead
and exert power over).
Resistance
to Pay-for Performance
Goltz and Hietapelto also analyze the common occurrence of workers
resisting the implementation of a pay-for-performance incentive system
that involves a reduction in base pay, presumably to be more than
made up for by the incentive pay. The analysis is that pay is a consequence
and the workers resist because of their loss of control over that
consequence. But I think, most often, pay is not a consequence, at
least not for work done; in most work settings, the worker has little
control over those consequences. For example, typically a worker might
double productivity with little if any increase in pay as a result.
We need not invoke behavior analysis to understand that workers resist
pay cuts that might or might not be made up by a locally untried compensation
system. In general, I think we should not give in to the temptation
to increase the face validity of behavior analysis by labeling all
resources as “consequences;” instead, we should wait until
those resources are explicitly used as consequences for behavior.
Practical Implications
Goltz and Hietapelto list three practical implications, expressed
in terms of “consequences” and “power.” However,
I think those three guidelines for understanding resistance to change
may be more generally and more usefully stated without reference to
“consequences” and “power.”
When planning change:
-
Clarify the changes in responsibilities.
-
Clarify the changes in resource allocation and try
not to greatly redistribute those resources.
-
Select change agents from the most influential at
all levels of the organization.
The
Sermon
Skinner criticized classical learning theorists’ hypothetico-deductive
method of theorizing, suggesting that those theorists were more interested
in their theory and their hypothetical constructs than in the data
itself. I am afraid this same criticism applies to much, perhaps most,
behavior-analytic theorizing as well, especially in theoretical extrapolations
from behavior-analytic laboratory research to applied and everyday
settings.
While I think it is good that at least some behavior analysts are
willing to risk more than the simple (but not easy) demonstration
of individual, empirical, functional relations, which constitutes
the bulk of behavior-analytic work, I also think we should be more
concerned about understanding the applied phenomena under consideration
than demonstrating the explanatory power of behavior analysis in general,
or the explanatory power of particular theories or concepts such as
the matching law, behavioral momentum, or the power model. My concern
is that our most brilliant behavior-analytic theorists are too eager
to demonstrate the universality of their theories and, as a result,
ignore crucial differences between the Skinner box and the applied
or everyday settings to which they extrapolate, thereby producing
uncomfortable forced fits.
Questionable extrapolations from basic to applied behavior analysis
may be fuelled by both sides—experimental behavior analysts
overly eager to justify their basic research by demonstrating its
relevance to applied settings and applied behavior analysts overly
eager to sell the adoption of their procedures by appealing to basic
research. I am not arguing against a close relation between basic
and applied research but rather against superficial extrapolation
from one area to the other.
A related problem is our tendency to gratuitously behavioralize English
and common sense. For example, most often, though not always, we add
little of value by suffixing verbs with the nouns “behavior”
and “response” (e.g., “I was emitting some verbal
behavior” vs. “I was talking”). A mere behavior-analytic
translation of common sense, everyday phenomena, or mentalistic terms
is of little value, if it does not provide new insights.
All too often, the overly used metaphor of the eager little boy with
the new hammer is all too apt. This is not to suggest that the boy
should not seek targets to which he can apply his new hammer and thereby
contribute to the well being of humanity. And this is not to suggest
that behavior analysts should not seek targets to which they can apply
their new theoretical behavior analysis and thereby contribute to
the well being of humanity, both by increasing our understanding of
the world and by improving the condition of that world. Both the hammer
and behavior-analytic theorizing are great tools, but we must take
more careful aim at more carefully selected targets. In developing
and applying our behavior-analytic theory, we must be as critical
of our own theorizing as we are of the theorizing of the mentalists—not
an easy task.
A
Humble Caveat
This commentary is based on an earlier draft of Goltz and Hietapelto’s
article; therefore, it does not reflect subsequent changes in that
article. And, of course, this entire diatribe is only my opinion;
so I could be wrong, though not likely.
Read
replies to this article
References
Goltz, S. M. & Hietapelto, A. M. (2002) Using the operant and
strategic contingencies models of power to understand resistance to
change. Journal of Organizational Behavior Management. 22, 3-22.
Malott, R. W. (1993). A theory of rule-governed behavior and organizational
behavior management. Journal of Organizational Behavior Management,
12, 45-65.
Malott, R. W., Malott, M. E., & Shimamune, S. (1993). Comments
on rule-governed behavior. Journal of Organizational Behavior Management,
12, 91-101.
Malott, R. W., Shimamune, S., & Malott, M. E. (1993). Rule-governed
behavior and organizational behavior management: An analysis of interventions.
Journal of Organizational Behavior Management, 12, 103-116.
Malott, R. W. & Trojan, E. W. (In press.) Elementary principles
of behavior (fifth edition). Upper Saddle River, NJ: Prentice Hall.
Read
replies to this article