Power in Organizations

This critique of Goltz and Hietapelto’s operant model of power suggests:

• The definition of power and leadership are too narrow.
• Powerful leaders rarely manage performance through operant contingencies.
• The opportunity to manage the behavior of others is rarely the reinforcer controlling the behavior of the powerful.
• The aversiveness of control by the powerful is rarely the basis for resistance to organizational change.
• Much behavior-analytic extrapolation from the Skinner box is unwarranted.
• Much behavior-analytic theorizing is uncomfortably close to the hypothetico-deductive theorizing about which Skinner warned us.

---

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.

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.

Comments on this article:

Although I have not read their article, your review gave me some idea of their position and, of course, your concerns about the application of behavioral analysis in explaining individual behavior in an organizational setting. Your observations struck a chord (harmonic, of course) with me. Certainly the use of the contingency model has been of value in my experience both clinically and organizationally, but only after I learned through monumental failures and frustrations to use the model when an analysis of the relationship of known variables/facts affecting a condition or situation has been performed. Assuming relationships which are correlational to be functional and 'assigning' reinforcing values is a recipe for disaster. In addition, knowledge of a behavioral history individually and organizationally is required but often ignored. The most successful and powerful leaders I've encountered show no evidence of imposing their will or influence on those they control. In fact, they seem more than willing to give away the control by attributing their own success to others. Others feel better for having known and served under them. Their profiles are low - they are not the Donald Trumps of the world displaying power on The Apprentice. The Trump-types are the second generation entrepreneurs following World War II. An interesting book describing the first generation is called Kindred Spirits by Callahan - post-WWII Harvard Business School group. Remember - The counter-control capabilities of subordinates is often overlooked and underestimated. The rat can bite the bar as well as the experimenter. In short, I enjoyed the thought provoking topic. But, of course, I, too, could be wrong, but I don't think so! - Bob Taylor