Abstract

We use a simple probabilistic decision problem as the basis for a three-hour exercise taught jointly by faculty from organizational behavior and management science. The exercise requires students to commit to a decision, and then defend their decision to other students. The problem is counter-intuitive, hence the bulk of students make an incorrect decision. This provides insight regarding irrational decision making behaviors, experiential learning of important organizational behavior topics, and an enhanced perspective on the value of analytical techniques. We use a simple probabilistic decision problem as the basis for a three-hour exercise taught jointly by faculty from organizational behavior and management science. The exercise requires students to commit to a decision, and then defend their decision to other students. The problem is counter-intuitive, hence the bulk of students make an incorrect decision. This provides insight regarding irrational decision making behaviors, experiential learning of important organizational behavior topics, and an enhanced perspective on the value of analytical techniques.

Motivation for the Exercise

One of the key drivers in business school management science education is integration with other functional areas (Winston, 1996, Carraway and Clyman, 1997, Jordan et al., 1997, Powell, 1998). Leading management science survey textbooks such as Ragsdale (2000) and especially Winston and Albright (2001) have many examples from finance, marketing, economics and operations management.

Organizational behavior is a required component of most MBA programs. Organizational behavior systematically studies "individual, group, and organizational processes" (Greenberg and Baron, 1995). The field examines the "attitudes and behaviors of individuals and groups in organizations" and "provides insight about effectively managing and changing them. It also studies how organizations can be structured more effectively and how events in their external environments affect organizations." (Johns and Saks, 2001).

Students sometimes acquire the misleading impression that organizational behavior is a "soft" field because many course elements are taught using discussion-based case methods or experiential exercises, and this approach does not always do justice to the highly empirical basis underlying the development of key concepts. By contrast, management science pedagogy stresses the tool and techniques of problem-solving but does not usually embed such practices in a context in which human behavior might have enormous consequences for decision outcomes and quality. Each field asks different questions, studies particular sets of problems, publishes within its own niche, and speaks its own specialized language.

Collaborations between the organizational behavior and management science fields are rare. The fields occupy silos that rarely meet in the classroom. Despite the importance of organizational behavior to business programs, we are aware of no teaching materials that integrate it with management science. We came to collaborate because we taught in an unusual MBA program that promoted flexibility and exchange among disciplines. This program required frequent team meetings among professors. It was during these meetings that we both saw the potential of this exercise for advancing key ideas about decision-making within our respective disciplines. This exercise has been repeated four times - twice for full-time MBA students, once in an Executive MBA program, and once in an academic conference of organizational behaviour scholars -- and the results have been very similar.

Benefits of the Exercise

Because management science and organizational behavior both address decision making, there should be many opportunities for integration. We present an experiential exercise that integrates organizational behavior and management science in the context of a simple stochastic decision problem. The goals of the exercise are to:

• draw out and identify irrational behaviors that managers exhibit in the face of decisions to help students diagnose the types of conditions, problems, settings, and individual characteristics that might result in poor decision-making.


• give students an experience of organizational behavior topics such as escalation of commitment and judgment under uncertainty


• establish the value of rational analysis as an arbiter of complex decisions

The exercise is superior to our prior lecture-based approaches of teaching organizational behavior because it gives students an actual experience of irrational decision-making, escalation of commitment, and judgment under uncertainty. The unique benefit for the management science course is that this exercise provides a situation where quantitative analysis leads to significant, immediate resolution of a vexing problem about which students come to care passionately. We know of no better way to achieve these benefits.

This exercise uses the well-known Let's Make a Deal problem. (An Annotated Selected Bibliography for this problem is in Appendix E.) We emphasize that we use this particular problem not because of any innate interest it may hold. We use it because it is easy to understand, yet strongly counter-intuitive, thus providing an effective vehicle to induce irrational behaviors. It is these irrational behaviors, and the use of rational analysis to counter them, that are the core of this exercise.

A game-show contestant (the student) is offered the choice of three doors. Behind one is a car; behind the other two doors are goats. After the contestant chooses a door, the host reveals a goat behind a door the contestant did not select. The host asks the contestant whether he or she will stay with the unopened original choice of door or switch to the remaining unopened door. Almost all students prefer to stay with their original choice of door. Not only do they prefer this alternative, but they strongly argue that they are either equally likely, or even more likely (because they believe in intuition) to win by staying with their original choice. These students are wrong. The mathematically correct answer is that switching is the winning strategy two out of three times. (See the Annotated Bibliography for references on the game and its ability to fool even professional statisticians.) Because the exercise hinges on the ability to explain the problem, it is effective with students who have had prior exposure to it. Although they might recognize the right answer, they still struggle to understand it.

One of the intriguing benefits of running this simple exercise in a classroom setting is that many students become more and more deeply entrenched in their commitment to staying with their original choice of door. Rather than seeking to solve the problem using any of the management science techniques they might have learned earlier in the degree program, they become deeply convinced that they are correct and that no verification is required. At this point, which usually takes less than an hour, we have simulated a genuine escalation of commitment to a losing course of action. As Ross and Staw put it, "Much of organizational theory can be reduced to two fundamental questions -- how do we get organizations moving, and how do we get them stopped once they are moving in a particular direction?" This second question is particularly relevant, given that a series of empirical studies has documented the tendency of individuals to become "locked into losing situations" (1992). Ross and Staw measured the cost of this persistence in one mega project alone, the ill-fated Shoreham nuclear power plant, as being over $5 billion. The determinants of escalation behavior may be found in the nature of the project, the psychology of the decision-makers, the social context, and the organizational factors surrounding the project (Brockner, 1992). Within this literature are such concepts as sunk costs or opportunity costs, calculations of risk and reward, the political nature of relationships and obligations, and the various decision-making traps into which humans frequently though inadvertently fall (the latter described in the landmark article by Tversky, and Kahneman, 1974).

To be able to create the conditions leading to escalation of commitment, to have students actually feel it in a very personal way, and then to be able to guide a class out of a trapped situation through the use of techniques they already know -- this is the essence of the exercise.

After the experiential portion of the exercise is over, students are given the opportunity to examine the rigorous empirical and theoretical literature in both organizational behavior and management science. Instead of simply debriefing the exercise and closing it down, we encourage students to explore some of the classic articles in two fields. Although the classroom portion of the exercise takes less than one day, the learning continues through the guided assignments we created for use over the next weeks.

Because we used this exercise with both full-time MBA and Executive MBA classes and tested it in a major university teaching conference, we anticipate it could be successfully used in any post-graduate business program. We hypothesize this exercise could be useful with undergraduates.

Structure of Exercise

The exercise is done in a 3-hour session taught jointly by an organizational behavior professor and a management science professor. To prepare for the exercise, the organizational behavior professor gives the students a short assignment a few days before the joint class. (We purposely distribute the assignment in the organizational behavior class. We want students to see for themselves the limitations of intuitive, non-quantitative analysis, and we want them to experience their own tendency to think inside the silo of a given course. An interesting extension to the research on defective group decision-making would be to contrast the outcomes of the exercise had it been assigned in a management science class versus an organizational behaviour class. To what extent does the context influence the decision processes employed?) Students hand in their written solution to the problem just prior to the joint exercise. The assignment is in Appendix A. Examples of written feedback from the organizational behavior professor on the written assignment is in Appendix B.

At the conclusion of the voting, each student has publicly selected one of three groups:

• STAY: It is best to retain one's original choice of door.


• INDIFFERENT: It doesn't matter if one stays or switches


• SWITCH: It is best to choose the other unopened door.

It is important for planning this exercise to appreciate that about 70 percent of students will hand in assignments that argue for staying, 15 percent are indifferent, and a maximum of 15 percent will opt to switch. Because at least 85 percent of students initially get the answer wrong, the exercise itself generates enormous controversy.

We inform students that the class needs to reach agreement on the correct answer. Students are formed into groups of three to five like-minded voters. We provide each group with blank transparencies and markers. We instruct the groups to meet for 20 minutes. Their task is to create a presentation to persuade other students that their answer is correct.

Each group gives a 2-5 minute presentation using an overhead projector. Presentations are followed by extensive and highly energetic discussion. Eventually, analysis is brought to bear on the problem and most students are persuaded of the correct answer. Thoughtful facilitation is required.

Important Information About the Game

There is one correct answer. Other answers are wrong.

The game show host always offers you the opportunity to switch. This is different than the old "Let's Make a Deal" television game show, where the host (Monty Hall) would choose to offer the switching opportunity conditional on the door originally chosen. Such a situation hinges on a subjective assessment of the host's behavior and lacks a mathematically correct answer.

This is one of those unusual situations where thinking outside the box is undesirable-there is sufficient challenge within the box. For example, one can't detect the second goat by observing the behavior of the first goat. Nor can one entice the goats with pheromones or imitation goat sounds, etc. (creative answers that we actually received from our students).

Details of the Exercise

Duration: A minimum of three hours is necessary to allow full debate. We usually allocate three hours and find that students continue arguing and discussing the simulation long after the class is over.

Setting: We use a case-style classroom that is suited for class discussions. In the front of the room there is an overhead projector. A computer attached to a projection device is located discreetly near the front of the room. The computer is connected to the web, and is running in a web browser a computer simulation of the game. (See the "Let's Make a Deal Applet" in the Resources section below.) The students are not allowed to see the computer screen until the appropriate time. The professors have props hidden for later use: 4 toy goats and 2 toy cars.

Commitment: We want the students to commit strongly to their answer-preferably the wrong answer-in order to give them a personal experience of escalation to commitment. To achieve this, we force the students to commit to an answer in writing, then vote publicly, and then group like-minded students together with the task of persuading others. Because one of the goals of the exercise is for students to actually experience entrenchment and escalation of commitment, we purposefully avoid grouping students of different opinions. (There might well be different outcomes if the groups were organized to reflect a diversity of opinion rather than unanimity. Organization behaviour researchers could easily adapt this exercise to test for the effects of diversity of opinion on group dynamics and decision outcomes.)

Vote: Since the assignment is to decide whether it is best to "stay" or "switch", we initially vote on these two choices. We let the option of "indifferent" emerge during the voting process. (We don't present this option because we want students to discover the richness of the problem on their own, and we want them to observe their switch/stay choice when faced with indifferent alternatives.)

Presentations: We provide transparencies and pens for the students. The students huddle in break-out rooms, study carrels, corners of the classroom, or the hall to discuss and prepare their transparencies. Upon the students' return, we have each group present briefly. We have the "stay" groups present first, followed by "indifferent", followed by "switch". This helps entrench the wrong ("stay") answer, and minimizes the risk of a uniquely persuasive "switch" presentation persuading the class before they have had a chance to engage in unstructured debate. (This debate is where the strongest personal learning seems to occur.)

Debate: The debate is intense and unstructured. Students will go the front to present new material, break into ad-hoc groups, wander the room looking for others to talk to, or even give up and read the newspaper. (These personal reactions are all points to bring out in the debriefing.) We have filmed the exercise, and encourage the use of visual records for later debriefing. The debate is emotional and intense. Each individual will likely exhibit the personal behaviors they engage in under stress-withdraw, act aggressive, talk without listening, etc. (This is pointed out during the debrief, and we suggest to them that they reflect on their reaction after the exercise.)

Generally some students change their mind from "stay" to "switch" but many do not. Eventually when the debate gets completely stuck we point out that they have become a classic "deadlocked board of directors" and work with them to resolve the issue by encouraging some sort of analysis.

Analysis: Students can easily play the game using scraps of paper. Students can easily analyze the game using enumeration. However, in our experience, the class is resistant to performing analysis or doing experimentation. We generally have to push the class to do experiments or analysis.

One way to get a discussion of analysis started is to use the props and start quietly simulating the game in a corner of the classroom. Those students not engaged in the arguments will notice something is going on. This sets up the idea of doing experimentation/simulation.

Computer: The computer is running a Java applet that simulates the game. We discretely reposition the computer so that it is visible to students who have chosen to walk up to the projector during the Debate phase. Eventually a student will notice the game and start clicking on it. The provenance of the simulation is unknown to the students-it could be a plant, or it could be programmed wrong. However, the simulation can be very persuasive. This leads into debriefing points about not believing "black boxes", and the power of simulation and data for resolving conflict.

Facilitation: Initially facilitation is very light. We want the students to commit emotionally to their answers. We want them to experience "escalation of commitment" and become entrenched in their answers. We want them, not the faculty, to have responsibility for the debate. We resist their calls to referee the debate. If they suggest that expert opinion should be called in, we decline, and remind them that in the real world experts can disagree; they need to sort it out themselves.

Once they are entrenched, we need to help the class dig out. The tools for digging out of an entrenched position are experimentation and analysis. We do not allow the easy option of simply taking a vote and letting the majority decide. The challenge is to get students to start analyzing and experimenting. The props (goats and cars) can be used to start some students playing the game. Side conversations can be used to get other students enumerating the game. The computer applet can be used to get students to use a computer simulation of the game.

Once the experimentation/analysis gets underway, the class will start to converge on the correct answer. We don't need for them all to agree because our discussion goals relate to the process of individual and group decision-making, not the analysis required to get the right answer.

Facilitators need to be prepared for a fractious class. Students have verbally attacked the game and verbally attacked the professors for working "against the class".

Debrief: Because of the emotional nature of the exercise, it is important to do at least some debriefing before the end of class. We take a 10 minute break before the debrief to make it clear that the Debate phase is over, and give emotionally charged students a chance to calm down. We start with a few opening remarks, and then let the students start the debrief. We make a point of touching on the following issues during the debriefing discussion:

• acknowledge that this is not what they expected
• thank the students for their efforts in a very difficult exercise
• make a point of appreciating the contribution to the class of individual students who did things they might be regretting-this is a valuable teaching to the class
• discuss the readings on escalation of commitment in light of this experience
• explore their reaction to the computer simulation, whose provenance and accuracy is unknown
• in light of the previous point discuss the power of simulation and data for persuasion
• remind them of the simplicity of the exercise and the richness of the reaction to it
• discuss their emotional responses throughout the exercise

In addition, we guide students through structured homework-based debreifings. Appendix C shows example homework questions for the organizational behavior class, and Appendix D contains example homework questions for the management science class.

Students with Prior Knowledge of the Game

Because this game is well known, some students may have some prior knowledge of the game. Such students can be segregated and act as observers, or allowed to play the game. We ask that these students disclose their prior knowledge in the assignment that is handed in prior to the start of the in-class exercise.

Prior knowledge of the game seems to pose less difficulty than one might expect. First, some students with prior knowledge recall the wrong answer. They receive a powerful lesson about the limitations of received knowledge, and fall into the trap with the bulk of the class. Second, most students with prior knowledge do not understand why the answer is correct. They still struggle to understand and explain why it is correct. Finally, any students with prior knowledge who are technically astute enough to understand why the answer is correct will learn that being right is nice, but is not sufficient for a manager. One must be able to persuade others. For some technically skilled students, this can be a vital learning experience. In a particularly fractious class of 30, one student arrived at the correct answer and presented it well in front of his peers. He was articulate, logical, and accurate. Students heckled him, shouting "loser, loser" despite his best efforts.

We have tried it both ways, having students with prior experience serve as observers, and alternatively take part as participants. We prefer to include all students in the game. The key learnings of the game come from the experience of personal reactions and group dynamics, not from the articulation of the correct answer. We believe that any students with prior knowledge learn more if they experience the game rather than observe it. Having students with prior knowledge of the game does not seem to harm the learning of the other students.

Analytical Skills That Can Be Used to Find the Correct Answer

Students have multiple techniques they can use to solve this problem. The exact mix of available techniques depends on prior preparation and the content of the management science course.

• Decision Trees
• Exhaustive Enumeration (equivalent to a decision tree)
• Monte Carlo Simulation in a computer
• Repeatedly Play the Game (Experimentation)
• Argument by extreme case (You have 1,000,000 doors, the host opens 999,998 doors with goats)

Why is This a Powerful Exercise?

After learning via the case method, experiential exercises, and other hands-on approaches, many students (particularly MBA students) persist in making the following excuses for poor performance:

• if I had more time…
• if I had more information…
• if the instructions were clearer…
• there was so much information that I couldn't sort it out…
• if I could have worked alone rather than in a group
• if I could have worked in a group rather than alone
• the statistics were so complicated

The delightful aspect of this exercise is that it is so simple: so easily tested through experimentation, with a large variety of methods that could use logic and reasoning, any one of which would result in the correct answer.

The frustration of this exercise is that students prematurely develop an intuitive answer which is wrong, and refuse to employ any analytical tools to test their intuition and instead devote their energies to justifying their wrong answer. The problem is significant because of the light it sheds on human nature, the limitations of rational models of decision-making, and the resistance to information that contradicts our intuitions.

Even for those students who recall from previous experience what the correct answer is, they in general are unable to explain why the answer is correct. Thus the exercise remains suitable for long-term use.

There is an additional benefit for the Organizational Behavior class. Students sometimes complain that the OB class lacks rigor because there is often no "correct" answer to an organizational issue. This exercise has one mathematically correct answer, and the students cannot rationalize that a wrong answer was somehow valid in an appropriate context.

Student Feedback

Student feedback is mixed. Students are most positive in classes in which the original (incorrect) answer does not count greatly towards grades. The first time we offered the exercise, we erred in assigning ten percent of the final grade in the organizational behaviour course towards the initial answer (stay or switch, and the justification). We wanted students to take their initial assignment seriously. This didn't work well. Given that most people arrive at the wrong answer, students had a legitimate grievance when this counted too strongly against them. In subsequent courses, we made students commit to a specified checklist of certain courses of action if they are wrong. We asked "How sure are you of your answer [to stay, to switch]? If you are wrong, would you donate money to a specified charity? Would you participate in the Terry Fox run for cancer research? Would you bake cookies for the class before the end of the semester?" and so on. This is a much better morale-building approach. We raised hundreds of dollars for charity, had over a dozen runners, and ate cookies for the remainder of the semester. This is the approach we advocate. It adds a more light-hearted touch without reducing the power of the exercise once students come to the class.

Course evaluations of the instructors do not offer the possibility of scoring the power of the exercise on a Likert-type scale. In open-ended sections, however, students volunteered that the exercise was "powerful", "frustrating", "unnerving" and so on. The distinctive feature of the feedback we receive relative to other teaching experiences is that it comes to us in unusual ways and from non-student sources. Students took the exercise home with them and tested their parents, spouses, children and siblings. Their work colleagues were treated to the exercise. Sometimes we would be accosted in the hallways, at parties, and at our places of worship by students' family members who wanted to argue about the Let's Make a Deal answer. There is no doubt that the exercise is engaging, and that it captures attention well beyond any conventional teaching techniques.

Origin of this exercise

This exercise originated when the authors were teaching in the Enterprise MBA program at the University of Calgary. This program was very tightly integrated with the explicit goal of breaking down disciplinary boundaries in teaching. Both instructors were aware that the other "taught something about decision-making" but we had little mutual understanding.

Although the structure and incentives provided by the Enterprise MBA program were important to creating this exercise, we see no reason that they are necessary. The management science instructor has successfully used this exercise in the Executive MBA program, partnering with a different organizational behavior instructor.

We encourage management science faculty to approach their colleagues in organizational behavior to collaborate in delivering this exercise, or seek other forms of educational collaboration.

Better understanding of organizational behavior will help the management science instructor enliven his own teaching. In addition, the more that other departments understand about modern business school management science, the more likely it is that management science will flourish at that school.

References

Carraway, R. L., and D. R. Clyman. (1997), "Managerial Relevance: The Key to Survival for OR/MS". Interfaces Vol. 27, No. 6, pp. 115-130.

Greenberg, J. and R. A. Baron. (1995), Behavior in Organizations,  Prentice Hall, p. 6., Englewood Cliffs, NJ.

Johns, G. and A. M. Saks. (2001), Organizational Behaviour: Understanding and Managing Life at Work., p. 8. Addison Wesley Longman, Toronto.

Jordan, E., L. Lasdon, M. Lenard, J. Moore, S. Powell, and T. Willemain. (1997), "Report of the Operating Subcommittee of the INFORMS Business School Education Task Force," OR/MS Today Vol. 24, No. 1, pp. 36-41.
Available at http://education.forum.informs.org/magnanti.html

Powell, S. G. (1998), "Requiem for the Management Science Course?," Interfaces, Vol. 28, No. 2, pp. 111-117.

Ragsdale, C. T. (2000), Spreadsheet Modeling and Decision Analysis, Southwestern, Cincinnatti, OH.

Winston, W. L. (1996), "Management Science with Spreadsheets for MBAs at Indiana University," Interfaces Vol. 26, No. 2, pp. 105-111.

Winston, W. and S. C. Albright (2001), Practical Management Science., Duxbury, Pacific Grove, CA.

^ã INFORMS