Groupthink Versus High-Quality Decision Making in International Relations

Mark Schafer, Scott Crichlow

• Publisher: Columbia University Press
• Year: 2010
• ISBN:
  9780231520188
• Topic: Politics & International Relations
• Subtopic: American Government
• Index: Politics & International Relations

Are good and bad outcomes significantly affected by the decision-making process itself? Indeed they are, in that certain decision-making techniques and practices limit the ability of policymakers to achieve their goals and advance the national interest.

The success of policy often turns on the quality of the decision-making process. Mark Schafer and Scott Crichlow identify the factors that contribute to good and bad policymaking, such as the personalities of political leaders, the structure of decision-making groups, and the nature of the exchange between participating individuals. Analyzing thirty-nine foreign-policy cases across nine administrations and incorporating both statistical analyses and case studies, including a detailed examination of the decision to invade Iraq in 2003, the authors pinpoint the factors that are likely to lead to successful or failed decision making, and they suggest ways to improve the process. Schafer and Crichlow show how the staffing of key offices and the structure of central decision-making bodies determine the path of an administration even before topics are introduced. Additionally, they link the psychological characteristics of leaders to the quality of their decision processing. There is no greater work available on understanding and improving the dynamics of contemporary decision making.

PART I

Groupthink and the Quality of the Foreign-Policy Decision-Making Process

One

INTRODUCTION

More than thirty-five years ago, Irving Janis published his now well-known book Victims of Groupthink: A Psychological Study of Foreign-Policy Decisions and Fiascoes. Janis’s basic argument was that the quality of the process of foreign-policy making is likely to have an effect on the quality of outcomes that stem from the decision. Sometimes decision makers engage in a careful, deliberative process that seeks deep and varied information, specifies objectives, considers alternatives, asks hard questions, checks against biases in the process, questions assumptions, and reevaluates information and choices when needed. Citizens might generally assume that those in the upper echelons of government always conduct such careful, rational decision processes, especially when making major foreign-policy decisions that affect the blood and treasure of the country. Yet Janis’s important work demonstrated that sometimes the process of foreign-policy making can be deeply flawed—and in such cases the probability of having a poor outcome goes up significantly.

His most famous case, the Bay of Pigs “fiasco,” documents how an administration consisting of exceptionally bright individuals failed to conduct anything remotely close to a careful decision-making process. The Kennedy administration failed to question deeply flawed basic assumptions, kept important information out of the process when it challenged the direction the group was heading, and quickly quieted one individual who dared to try to disagree with the president and others in the administration. The result was one of the most embarrassing foreign-policy setbacks in the history of the modern presidency.

Janis referred to the broad pattern of decision-making problems in this case and others as “groupthink.” Although he offers different definitions of the term throughout his work, he is referring generally to a process where group norms and patterns essentially take over and result in deeply flawed decision making. The “group” takes on a life of its own, one that is bigger than the sum of the individuals in the group. Group cohesion is valued above the quality of information processing; dissent is discouraged, suppressed, or eliminated; shortcuts are taken in the process; assumptions by the leader or key advisors go unquestioned; and biases lead to policy. The result of these flaws in the process is that the group reaches “premature consensus” (Janis 1972). The group feels good about its decision, and group cohesion has been maintained, because, after all, dissent was minimal.

Janis supported his argument with a series of case studies (1972, 1982). Some, such as the Bay of Pigs, demonstrate that when groupthink occurred, the outcome was very poor. Others, such as the Cuban Missile Crisis, demonstrate that high-quality decision making—which avoids groupthink—resulted in a very good outcome. There is great intuitive appeal to Janis’s argument. We hope leaders at the top of governments conduct careful decision-making procedures, and when they do not, it makes sense that the outcomes should be less than optimal. Yet, while Janis’s work remains the locus classicus in this area, it also raises many questions. How often does groupthink occur? Is groupthink an either-or condition, either there or not? Or, on the other hand, is the quality of decision making more variable than that? Might some cases display decision making of very poor quality, some display very good decision making, and some in between? How often does poor decision making result in poor outcomes? What conditions give rise to poor decision making? Are there different stages, components, or factors in decision making that affect quality? What can be done by policymakers to improve the quality of decision making?

Despite its intuitive appeal and its popularity, Janis’s original work has some important limitations that simply cannot address these questions and others. Janis’s work was strong on providing qualitative narratives of cases but weak on modern social-science methods. He selected a small number of cases that—one could argue—merely exemplified his working thesis. Related to this concern is that Janis himself chose cases that he judged either as successes or failures, and then he subsequently made judgments about the quality of the processes that went into those cases, a method that runs the risk of researcher bias (or at least opens the door to such criticisms). Janis’s qualitative case studies also do not lend themselves to any type of meaningful comparison. We do not know if there is “more” groupthink in one case compared to another, or how much better the quality of decision making is in some cases, or if there is a systematic correlation between qualitative levels of decision-making processes and the quality of outcomes that stem from those processes. Finally, there is the issue of the timeliness of Janis’s work. Since he published this seminal book on groupthink (1972), there have been six different U.S. presidents (Ford, Carter, Reagan, G. H. W. Bush, Clinton, and G. W. Bush), numerous new foreign-policy events and issues, the end of the cold war, the nonstate-actor challenges of the twenty-first century, and a host of other changes in international politics. In addition, there have been many contributions to the foreign-policy decision-making literature and much advancement in social-science research methods. Our broad objectives in this book are to revisit Janis’s hypothesis, report new case studies, apply new social-science methods that let us investigate his hypothesis more broadly and rigorously, and generally expand the scope of the questions he posited so long ago.

Groupthink is now a famous term in many different disciplines, including political science, international studies, psychology, sociology, and business management and administration. It is a term we will use frequently in this book. However, the term is limiting. Janis himself offered several different definitions of groupthink throughout his career. In addition, since Janis’s book was first published, there have been many different studies that looked not only at the phenomenon of groupthink but also at other components related to the quality of decision making. Thus our work looks more broadly at the quality of the decision-making process, of which groupthink is only one part. In some circles, the term groupthink is synonymous with poor-quality decision making—and indeed, when groupthink is present, it is an indicator of poor-quality decision making. However, a case may have poor-quality decision making and not be emblematic of groupthink. All squares are rectangles, but not all rectangles are squares. Therefore, while we clearly place our work in association with the groupthink research program, we think of our analytical frame as being broader than that. We look at the quality of group decision making in its broadest possible terms, one part of which is groupthink.

Groupthink as Part of a Larger Research Program

It is important to place our work and the groupthink research program into the broader context of research in the field of international relations. In short, this study is one component of the broader research area called foreign-policy decision making (FPDM). More than fifty years ago, Snyder, Bruck, and Sapin (1954) revolutionized foreign-policy analysis by imploring that we get inside the “black box” of the decision-making process to better understand and explain the factors affecting policy. By arguing that “the state is its decision makers” (Snyder, Bruck, and Sapin 1962, 65), these scholars realized the importance of the human element in decision making. Foreign-policy making is not simply a matter of a rationalist calculus, which merely requires realist inputs about power and interests to determine choices and outcomes. Instead, we must think of the decision process as a fundamentally human one that includes diverse personalities, competing—and sometimes unclear—role assignments, misperceptions, biased processing, wishful thinking, problematic group dynamics, ineffective lines of authority and communication, and other common human elements. To understand how a state arrives at a decision, we must carefully examine the human processes behind that decision. Without this, we have an incomplete understanding of the case. Every single decision a state makes must flow through humans and human interactions in one form or another.

Essentially, the FPDM research program takes us inside the state. This is different from other schools of thought in international relations exemplified by the “billiard-ball” metaphor. This metaphor suggests that state interaction in international politics is similar to balls interacting on a billiard-ball table: the solid balls periodically come into external contact with one another, causing collisions and caroms. The part of this metaphor most relevant to our discussion here is the idea that states are solid entities and that whatever exists inside the state is irrelevant. One school of thought related to this notion is the rational-actor school. It assumes that states are rational calculators trying to achieve their interests. In this conception, think of a state as a computer program: it takes the inputs from the system and simply calculates the output, with the decision rule specified as maximizing state interests. It weights the same factors the same way every time; it does so without error, biases, or emotions; and it does not allow for variance in interaction patterns because, conceptually speaking, there are none. In short, there is no human element in the billiard-ball model.

The FPDM school strongly diverges from this conception, and indeed, Snyder, Bruck, and Sapin so long ago wrote quite the opposite—that the decision process within a state is quintessentially a human process, fraught with those things that are human: biases, errors, emotions, interpretations, projections, personalities, and many others. This does not mean that actors within states are irrational. Nor does it mean that actors are not trying to maximize interests. Of course they are. But they are human beings, and they have human limitations. In fact, human beings are involved every step of the way in decision making—from trying to discern what “interests” are, to interpreting ambiguous information, to weighing possible alternatives designed to accomplish those “interests,” to conducting the implementation of the policy. And since each human being is unique, who the individuals are inside the apparatus of the state matters. If someone else had been president at the time, would the United States have employed a blockade in the Cuban Missile Crisis? Invaded Iraq in 2003? Opened relations with China? Who governs can matter, and states are not solid entities.

Once we grasp this, then many avenues of research are opened to us. Research coming out of the FPDM school includes such areas as the psychological study of leaders; analysis of public opinion; consideration of bureaucracies and their role in the process; and the analysis of the media, special-interest groups, particular governmental structures, the political culture and norms, and many other areas. The groupthink research program is one of the most well-known programs to emerge out of FPDM. It fits well with the notion that decision making is a human process and that in order to understand what a state does and how it arrived at that decision, we must get inside the state and look at the processes within.

The particular contribution of the groupthink research program is its thesis that the quality of the process in decision making is likely to be connected to the quality of the outcome associated with the decision; poor process is likely to increase the probability of a poor outcome, and vice versa. Sometimes human beings interact more effectively than other times; sometimes group norms overwhelm group members who might otherwise oppose a policy; sometimes power plays in a group can override good analytical processes. When decision making breaks down at the top of the state apparatus, problems such as poor information search and biased information processing are much more likely to emerge. Under such circumstances, human-interaction patterns result in something that does not look like the precise functioning of a software program.

The quality of the decision-making process certainly is not the only explanation for poor outcomes. They may come about because of a variety of factors associated with the case: the other side may take a totally unexpected action, public opinion may decline quickly, or the international community may react very negatively. Janis’s thesis suggests that the quality of the decision-making process is one key variable—though not the only one—affecting the quality of the outcome. Importantly, the quality of the process is something over which decision makers can have great influence. While it is not always possible to influence the other side’s action, public opinion, or the international community, it is possible to set up better structures and procedures for decision processing within the administration. For that reason, a better understanding of these matters can have powerful prescriptive benefits. By altering the way a government is designed and decision making is carried out, political leaders have the ability to affect the likelihood that their endeavors will be successful, and the countries they lead may see major benefits as a result.

Going back to two of Janis’s cases, the Bay of Pigs and the Cuban Missile Crisis, we have an example of a decision-making process he describes as terribly problematic and one he describes as excellent within the same administration. In our research, this kind of major transformation is rare, but it exemplifies the possibility that an administration can implement better structures and procedures. And implementing such structures and procedures can have a tangible, positive effect on policymaking. Kennedy quickly and publicly took personal blame for the Bay of Pigs disaster and then went to his administration and told them that new and better decision-making procedures were needed. He introduced a system where some advisors were assigned the role of challenging assumptions and asking hard questions. He included a larger number of advisors in the process and brought in outsiders who had different kinds of expertise and could challenge intra-administration perspectives. Concerned that his own presence might hinder free discussion at meetings, he pointedly stepped out of some meetings, allowing his advisors to speak their minds more freely and reducing the potential appearance of leader bias (Janis 1982). Each of these acts created structures and processes that allowed for more critical thinking and better information processing.

Advancing the Groupthink Research Program

While Janis’s work was groundbreaking and important, it also has some limitations and leads to many new questions. We do not approach this project as doubters of Janis’s basic thesis or as critics of his findings—indeed, quite the opposite. Janis’s cases provide intriguing evidence, and reading histories of foreign-policy cases over the years shows abundant anecdotal support for the hypothesis that there is an important connection between the quality of the decision-making process and the quality of the outcomes associated with that process. Nonetheless, the limitations in Janis’s research and the new questions that emerge out of it must be addressed. Janis’s cases included not only the two from the Kennedy administration but such others as the level of naval preparedness before Pearl Harbor, U.S. escalation in Vietnam, and the development of the Marshall Plan. These are major matters that affected life and death, and they fundamentally altered international politics. If the quality of the process mattered in them—and many other cases since—then indeed the stakes are very high in this research, and it is incumbent upon us to be as rigorous as possible.

There are several important methodological innovations and advancements in the research presented in this book. While Janis himself judged both the outcome and the quality of the case, we separate the evaluation of these two parts of the puzzle. The dependent variable—the outcome of the case—is judged by a survey of foreign-policy experts from around the country (who were kept blind to the research hypotheses). The quality of decision making—our primary independent variable—was separately assessed by us, using case-study techniques. With these case studies our work looks similar to Janis’s. We gather high-quality materials for each case. We include such things as memoirs, autobiographies, accounts by journalists and historians, and other works that, as much as possible, let us get inside the decision-making process. We read these materials, identify key actors, cross-check the facts, develop timelines, and generally come to know what went on inside each decision process as much as possible.

However, our methods also make improvements over Janis’s approach. Before beginning the research, we specified a set of operationalized variables pertaining to the situational conditions and variables relating to the quality of decision making. These allow us to consistently and systematically evaluate these factors across the set of cases in our research. These operationalizations are derived from earlier research on the quality of group decision making. For example, Janis argues that one symptom of groupthink is poor information search in the case. Based upon other existing research, we specified a definition of this concept that could be applied the same way across every case in our study:

Poor Information Search: The group fails to pursue information that may be available and necessary for critically evaluating policy options. This can be the result of a variety of problems, including arrogance, ignorance, misguided deference, or even a conscious decision to not contact someone who has relevant information and is available to talk.¹

These operationalizations add a higher level of objectivity, consistency, and structure to our reading and evaluation of the case studies. In addition, these operationalized variables lend themselves to quantification of the underlying concepts. For example, based on the above definition, we are able to code every case for the presence or absence of poor information search. When we combine this variable with others (such as stereotyping, biased processing, and suppressing dissent) in the decision-processing stage of a case, we end up with a numerical value—essentially a measurement—for the number of decision-processing faults in that case. Though Janis never took his research to this level, from his overview of the cases we can be c...

Table of contents

• Cover 
• Half title
• Title
• Copyright
• Dedication
• Contents 
• List of Tables and Figures
• Acknowledgments
• Part I: Groupthink and the Quality of the Foreign-Policy Decision-Making Process
• Part II: Case Studies in American Foreign-Policy Decision Making
• Part III: Statistical Analyses
• Part IV: Conclusions
• Appendix A: Cases Included in the Analysis
• Appendix B: Operational Definitions of Situational-Context Variables
• Appendix C: Operational Definitions of Group-Structural Variables
• Appendix D: Operational Definitions of Decision-Processing Variables
• Notes
• References
• Index