The EO in OBM

Olson, Laraway, and Austin (2001) propose an increased emphasis on the establishing operation in organizational behavior management. Their proposal raises interesting questions about theory, science, and practice. (1) What should be the role of theory in behavior analysis? (2) Should we try to find problems that match our solutions or vice versa ? (3) What is the relative importance of the establishing operation and the performance-management contingency in managing organizational behavior? (4) Should theory and basic research be more informed by the issues raised in applied settings?

The EO in OBM

Olson, Laraway, and Austin (2001) make a thoughtful, well-informed, and well-argued proposal for the importance of the establishing operation (EO) in organizational behavior management (OBM). Furthermore, their proposal raises interesting issues about the way we behavior analysts do our science:

Theory Phobia

I welcome the proposal of Olson et al that we look to the EO as a way of achieving “a more comprehensive theoretical understanding.” And they have gotten off to a good start. Their support of theory is especially important in behavior analysis, because, most empirical articles in behavior-analytic journals are bereft of theoretical analysis. This exemplifies a theory phobia suffered by many, if not most, behavior analysts, a phobia that may result from a misinterpretation of Skinner’s argument against the use of explanatory fictions (i.e., hypothetical constructs and intervening variables) and against the reductionistic need for behavior analysts to reduce their findings to a lower level of analysis (e.g., physiology).

In defense of theory in behavior analysis, I suggest that Skinner was not arguing against an effort to understand complex behavioral procedures in terms of the underlying behavioral concepts and principles, even if such an understanding might involve speculation. In other words, looking for (or even speculating about) the behavioral concepts and principles underlying complex behavioral procedures is not the same as unneeded reductionism, nor is it the same as speculating about underlying explanatory fictions. Behavior analysts seem to have over generalized Skinner’s concerns about these two forms of speculation to a fear of speculation about underlying behavioral processes, which involve neither a different level of analysis nor explanatory fictions.

This fear of speculation has produced a science that, though it has great methodological sophistication, it also has great theoretical naivety. Its practitioners are skilled at detecting the most subtle of methodological flaws, while overlooking the most gross of conceptual errors. For example, behavior analysts consistently commit such simplistic conceptual errors as classifying complex, delayed-outcome, performance-management contingencies as if they were the simple reinforcement contingencies of the Skinner box. (For discussions of the nature of these contingencies see [Malott, 1993; Malott, Shimamune, & Malott, 1993; Michael, 1993; Olson, Laraway, & Austin, 2001].)

Our theoretical naivety is sometimes said to be justified because our behavioral procedures work (e.g., our performance-management contingencies increase productivity). But, the New Guinea aborigines have productive yam patches, as long as they appease the hungry earth god by feeding it with a piece of each yam they harvest; yet we would not argue that the success of these agricultural practices makes a science of biology unnecessary. And even if we do not value the understanding of our universe that results from the sciences of biology and behavior analysis, the history of science and technology suggests that better science will result in better yam patches and more productive, healthy workers.

So, again, I welcome Olson, Laraway, and Austin’s support of theory in behavior analysis. And I encourage the Journal of Organizational Behavior Management, as well as other behavior-analytic journals, to strongly urge their authors to include theoretical analyses addressing the fundamental behavioral processes underlying their complex behavioral interventions, including the underlying EOs, when relevant.

Solutions in Search of Problems vs. Problems in Search of Solutions

There are two, not necessarily exhaustive, approaches to using science and technology to solve real-world problems. The one Olson et al recommend is the solution-in-search-of-a-problem approach. In other words, analyses of EOs seems to be helpful in reducing dysfunctional behavior of clients who are non-verbal; so Olson, et al, propose that we search problems solvable by analyses of the EOs among those problems involving the increasing of functional behavior of clients who are verbal. 

This solution-in-search-of-a-problem approach has become so prominent that it may now dominate behavior analysis (e.g., behavioral momentum, the probability matching law, functional analysis, the EO, itself, and from an earlier era, schedules of reinforcement). Of course, the solution-in-search-of-a-problem is a reasonable, valuable approach; but it has at least two drawbacks. First, behavior analysts become so enamored with the currently fashionable solution, that they force fit the solution in a nearly rote, unthinking manner, with little concern for the appropriateness of that fit (like the little boy with a new hammer who must hammer everything in sight, whether it is hammerable or not). 

There might be fewer strained fits, if the those behavior analysts followed the recommendation of Olson et al and based their science and technology on a sophisticate interest in underlying behavioral mechanisms. Much, though not all, of the inappropriateness of these forced fits is again due to simplistic extrapolations from Skinner-box research with non-verbal rats and pigeons; these extrapolations are to complex, language-based, human contexts. 

My criticism of these extrapolations is not to suggest that there isn’t a little rat in all of us, not to suggest that basic, Skinner-box research is irrelevant to our understanding of complex, human behavior. However, to find the relevance, we must dig deeply into the underlying behavioral processes, more deeply than most behavior analysts usually go. So the first difficulty with the solution-in-search-of-a-problem is that it may often result in inappropriate, forced solutions.

The second difficulty is that this approach has become so dominant, at least in the research literature, that it may be driving out the problem-in-search-of-a-solution approach. For example, researchers who are successful publishers often explicitly ensure that they are doing research on solutions currently fashionable in their targeted journals. And researchers who are unsuccessful publishers often complain that their manuscripts are being rejected because they are not using fashionable solutions. (Of course, this applies to fashionable problems to be solved, as well as fashionable solutions.) And while it is, no doubt, valuable for individual researchers to carry out thematic research, it may be harmful for an entire discipline to do so. (Incidentally, the Journal of the Experimental Analysis of Behavior’s publication of Michael’s original EO article [1982], itself, illustrates that that fashion can be bucked, as does that article’s gradually but steadily growing number of citations in the Journal of Applied Behavior Analysis [see Iwata, Smith, & Michael, 2000].)

I have some concern about both of these difficulties with regard to the proposal of Olson et al; they propose that we look for OBM problems that can be solved by manipulating the EO. First, in my opinion, the authors had to strain considerably to find OBM examples that EO manipulations might possibly solve. This suggests that, to avoid inappropriate, forced fits, those who follow the EO-solutions strategy in OBM will have to do so with considerable more theoretical sophistication than has been evidenced in most empirical OBM articles. 

The second difficulty is that this EO-solution strategy may have the potential to dominate OBM research to the extent that it suppresses other solutions to OBM problems. Fortunately, the empirical work in OBM seems to be less driven by fashionable solutions than is the case in the general field of applied behavior analysis (ABA). However, the theoretical OBM articles do tend to sail in on the breeze of fashion, as the article by Olson et al would indicate, not that this is always bad, especially if such articles are accompanied by a critical, theoretical intelligence, as is the case with that article. 

There is another potential problem related to this second fashionable-solutions difficulty: The EO-solution strategy may also tend to suppress work on problems not readily solved by EO manipulations, a problem of narrowness that may be occurring in the published literature in the general field of ABA. This could be especially unfortunate, if it turns out that the sort of problems EO manipulations can solve do not account for the majority of the problems in OBM, as, in deed, the examples of Olson et al seem to suggest. 

None of this means it might not be of value for some researchers in OBM to aggressively purse the EO as a fundamental cause of some OBM problems and to purse the manipulation of the EO as a solution to those problems. It is only to suggest that, in modeling after the general field of ABA by exploring the importance of the EO, we should not make the more general mistake of the researchers and editors in general ABA, the mistake of becoming too fashion conscious.

There is another interesting issue related to the strategy of solutions in search of problems. It is the concept in search of examples. Lindsley referred to this as the problem of the conceptual zoo (Lindsley, 1977). We can build a framework or matrix of concepts that might more or less exhaustively describe a domain. Then we send explorers out to find animals to fill each of the cages of our zoo, to find examples to fill each of the cells of our matrix or framework, examples to illustrate each concept. But we may discover that creatures do not exist in nature to fill each of the cages of our zoo, in other words, that examples do not exist in nature to fill each of the cells in our framework, to illustrate each of our concepts. For instance, outside the laboratory or the explicit behavior-modification setting, my explorers and I have failed to discover realistic example of traditional shaping (fixed-outcome shaping) or differential reinforcement of other behavior, though these procedures are common and useful in ABA research and application efforts. 

Similarly, we should remain sensitive to the possibility that, though the EO is an important, fundamental concept in behavior analysis and in general ABA, we might discover few significant instances where knowledge of the relevant EOs will give us nontrivial insights into the nature of the world of OBM and few instance where it will allow us to manage performance better in that world. Olson et al seem sensitive to the possibility that the EO/OBM cage might remain fairly barren; I hope that other conceptual explorers maintain that sensitivity. But, in any case, the only way to know the fate of the EO/OBM cage is to send out a few exploring parties. (Incidentally, Lindsley also suggested that our conceptual zoos should not blind us to the possibility that we might discover in the real world important behavioral phenomena for which we have not prepared a cage.)

The EO (Motivation) vs. the Contingency

In his behavioral engineering model, Gilbert (1978) presents his own six-cell conceptual matrix, with two cells devoted to “motivation.” One of those two cells represents “incentives” and the other, “motives.” The motive cell corresponds roughly to the EO but lacks the sophisticated treatment provided by Olson et al; it antedates Michael’s (1982) elaboration of the EO concept. However, what Gilbert’s model does not have is a distinct cell for contingencies, though the concept seems to be squeeze into the same cell as incentives. But for most OBM problems in most OBM settings, the lack of incentives and the lack of motivation, in the technical, EO sense, does not seem to be the problem. The problem usually is the lack of an effective performance-management contingency using the incentives (reinforcers or aversive conditions) and the EOs already plentiful in the OBM setting. While we can look for interesting, unusual incentives, the most expedient and ubiquitous incentive is probably money; the only problem is to make the money properly contingent on the desirable performance.

In the OBM setting, logistics and social validity almost demand that the incentive or reinforcer be a learned reinforcer, such as money. And, our applied literature reflects a common misconception as to what is the EO for learned reinforcers, an issue about which Olson et al express concern. That misconception is that the proper EO for a learned reinforcer is deprivation of the learned reinforcer (e.g., deprivation of approval or praise is an appropriate EO; and satiation with approval or praise is an appropriate abolishing operation). 

The error of this view might be clearer, if we look inside the Skinner box. What is the appropriate EO for the learned reinforcer (the water dipper click) that has been paired with the unlearned reinforcer (a drop of water in the dipper)? As Michael (2000) implies, the EO is the deprivation of water (a transitive, conditioned, establishing operation), without which, incidentally, the click would not become a learned reinforcer. But more to the point, if the click had previously been established as a learned reinforcer, it would not continue to reinforce behavior after water satiation. In other words, the EO for a learned reinforcer is not deprivation of that learned reinforcer (e.g., the EO is not deprivation of the dipper click). Aside from formidable methodological problems, it would seem naïve to try to satiate the rat with the dipper click (use repeated presentations of the click as the abolishing operation [satiation]). (Those formidable methodological problems involve confounding either click satiation with water satiation or confounding click satiation with the unpairing of the click and the water.)

So, surely something else must be going on when research shows what looks like an EO effect resulting from “deprivation” and “satiation” of praise for little children or “deprivation” and “satiation” of towels for a psychotic lady who hordes towels until she has dozens. That something else may be very complex, but surely it is not an EO effect.

Now, back to the EO in OBM. Money and other learned reinforcers will probably remain the major reinforcers in OBM. And, as we have just seen, the relevant EO for such reinforcers involves deprivation of the unlearned, backup reinforcers with which the learned reinforcers have been paired. But it is unlikely that future performance managers will intervene at the level of food deprivation as the EO to increase the effectiveness of money as a learned reinforcer. Thus, it is unlikely that the EO will ever play a major role in OBM interventions, which is not to argue that it might not be important in the sort of niche problems suggested by Olson et al. Instead, the major goal in OBM should remain the implementation of effective performance-management contingencies.

Incidentally, there may be some ambiguity, or even disagreement, as to what constitutes an effective deprivational EO for such unlearned reinforcers as water. Most clearly, something like 23 hours of water deprivation is the traditional EO. But what about the momentary unavailability of water, once the rat has licked the dipper dry? Is that momentary unavailability also an EO? If so, surely it is a much different sort of EO that 23 hours of water deprivation, so different that, though the momentary unavailability might fit the operational definition of an EO, it seems to me to be almost trivial to consider 23 hours of water deprivation and the momentary unavailability of water as the same class of phenomena. Yes, both deprivation and unavailability are necessary for the rat to make water-reinforced lever presses; but perhaps that should not make them force us to classify both as EOs. The two phenomena may be analogous but not homologous. 

However, if my concerns are unfounded, then it may be that the momentary withholding of the dipper click, the momentary withholding of praise, and the momentary withholding of money are also EOs. In that case, the EO might play a major role in mainstream, performance-management, OBM analyses and interventions. But, also should that appear to be that case, we may want to be sure we are wearing watertight goulashes, should we fall through the thin ice of operational definitionism into the water of trivial categorizing and end up all wet.

From Theory to Application to Theory

OBM consultants often argue that our behavior-analytically informed consulting is superior to the competition’s because our behavioral consulting is based on basic, SCIENTIFIC, laboratory research, where the competition’s is based on mysticism. And while the criticism of the competition may be well placed, I think our behavioral consulting practices have much shallower roots in basic behavior-analytic science than we usually acknowledge, mainly because those roots are fertilized with simplistic analogies that do not reflect how our laboratory-based, behavioral principles and concepts actually underlie or fail to underlie our consulting practices. 

And, applied behavior analysis researchers have been chastised for failing to make adequate reference to basic, laboratory research and perhaps to behavior-analytic theory. Instead they operate from a pragmatic bag of common-sense tricks disguised as science. That disguise is especially deceptive because a few of those tricks may have been empirically verified, though much as the feeding-the-earth-god trick has been empirically verified in the Journal of Aboriginal Yam Growers; complete with multiple-baseline designs it has both have been shown to work. 

So, we OBM practitioners yearn for the mantel of respect basic science gives us, and we OBM researchers yearn for the relief from guilt basic-science references give us. But, deny it though they often do, the basic-science, experimental analysts also have a yearning; and that yearning is for relevance. So these yearnings have combined to send a torrential flow (or at least a torrential trickle) of basic science and advanced theory into the laps of OBM practitioners and researchers. 

But now, it may be time to start the flow in the other direction. Now it may be time for basic science and advanced theory to more aggressively reflect and address the issues encountered in applied settings, including OBM. Now it may be time for applied practitioners and researchers to stop forcing ill-fitting concepts and principles onto applied phenomena; now it may be time applied workers to help the basic researchers and theorists refine their efforts so as to reflect the phenomena of the world outside the experimentalist’s laboratory and outside the theorist’s office. 

Furthermore, the EO might be a good conceptual area to start, if it should need refining in order to increase its applied utility in OBM. Here is why the EO might be an especially opportune area in which to reverse the science/theory =>applied flow: Jack Michael, the developer of the EO, has proven exceptional in his efforts at continuous, quality improvement of that concept (see Michael, 2000). Therefore he might be especially adept at working with the OBM researchers and practitioners in refining the concept of the EO to meet these new challenges. Furthermore, Olson, Laraway, and Austin have demonstrated such a high level of mastery of the complex intricacies of the EO concept that they would be ideal to play an active role on the OBM side.

At the end of a rant such as this, it is only proper that I add: Of course, these are just my opinions; and I may be wrong.

References

Gilbert, T. F. (1978). Human competence: Engineering worthy performance. New York: McGraw-Hill.

Iwata, B. A., Smith, R. G., & Michael, J. (2000). Current research on the influence of establishing operations on behavior in applied settings. Journal of Applied Behavior Analysis, 33, 411-418.

Lindsley, O. R. (1977). What we know that ain’t so. Invited address, Third Convention Midwestern Association for Behavior Analysis, Chicago, IL.

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., 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.

Michael, J. M. (1982). Distinguishing between discriminative and motivational functions of stimuli. Journal of the Experimental Analysis of Behavior, 37, 149-155.

Michael, J. M. (1993). Concepts and principles of behavior analysis. Kalamazoo, MI: Society for the Advancement of Behavior Analysis. 

Michael, J. M. (2000). Implications and refinements of the establishing operation concept. Journal of Applied Behavior Analysis, 33, 401-410.

Olson, R., Laraway, S., & Austin, J. (2001) Unconditioned and conditioned establishing operations in organizational behavior management. Journal of Organizational Behavior Management. 21

Comments

[Shawn S. Jacksob]

The article written by you very good, I like it very much. I -Coach key holderswill keep your new article.