Making Concurrence-Seeking Visible: Groupthink, Discourse Networks, and the 2003 Iraq War

Making Concurrence-Seeking Visible: Groupthink, Discourse Networks, and the 2003 Iraq War Abstract Groupthink is one of the most popular and extensively discussed approaches in studying small group decision-making. However, the methodological question of how to make concurrence-seeking, as the key element of groupthink, visible has received insufficient attention. To make group decision behind closed doors more visible and methodologically advance groupthink theory, I introduce Discourse Network Analysis and apply it to the US invasion of Iraq in 2003. Results demonstrate that the US decision to go to war was based on a sudden and undisputed threat-perception that exhibited hasty concurrence-seeking. Given this threat-perception, and a fierce struggle between advocates of a diplomatic solution and supporters of a military response, the United States chose a military strategy—one that inevitably resulted in war. Since Snyder, Bruck, and Sapin published their landmark study Foreign Policy Decision-Making, the analysis of groups has become one of the most prominent and vibrant research traditions in Foreign Policy Analysis (FPA) scholarship (Hudson 2007, 17–20). Some experts within this tradition focus on bureaucratic (Allison and Zelikow 1999) or organizational routines and cultures (Halperin, Clapp, and Kanter 2006) to explain foreign-policy outcomes, while others adopt small groups as the decisive unit of analysis (Tetlock et al. 1992; Hart, Stern, and Sundelius 1997a). The path-breaking contribution to this literature of small group decision-making is Irving L. Janis’ (1982) book Groupthink: Psychological Studies of Policy and Fiascoes. It has become a standard reference in FPA (e.g., Hart, Stern, and Sundelius 1997b) and the analytical framework for numerous empirical studies (e.g., Smith 1984; Mulcahy 1995; Metselaar and Verbeek 1997; Yetiv 2003; Badie 2010). While scholars discussed and enhanced several elements of Janis's concept—for example, the role of mindguards (Fuller and Alday 1997), antecedent conditions of (Schafer and Crichlow 1996; Baron 2005) and collective risk-taking in a groupthink context (Vertzberger 1997), and the newgroup syndrome (Stern 1997)—methodological questions of groupthink still remain underexplored. Janis himself passed over this question by noting: “For purpose of hypothesis construction (. . .) we must be willing to make some inferential leaps from whatever historical clues we can pick up” (Janis 1982, ix). Tetlock (1979, 1316) first tried to fill this gap. He suggests that Janis’ case study design might have “given too much weight to evidence that supported groupthink hypothesis and too little weight to contradictory evidence.” Using content analysis, Tetlock demonstrates that public statements of decision-makers in groupthink situations “were characterized by significantly lower levels of integrative complexity” (Tetlock 1979, 1322). Other scholars followed Tetlock by linking signs of groupthink to individual perceptions and characteristics, as well as leadership-styles measured by public statements (e.g., Stern and Sundelius 1997; Schafer and Crichlow 2002, 2010). Although these studies have taught us much about the likely relationship between individual characteristics and groupthink, one central question is still unanswered: How can researchers operationalize deliberations in a group context? Or more specifically, how can researchers make concurrence-seeking in groups (the core concept of groupthink) visible? As George (1997, 39) argues, the lack of a satisfactory answer to this core question is the biggest gap in groupthink theory. Up to now, hasty concurrence-seeking and hence groupthink cannot be observed directly but must either be inferred by the presence of antecedent conditions and/or the final symptoms. This paper introduces Discourse Network Analysis (DNA) as a method for visualizing nonpublic decision-making in group settings and hence making concurrence-seeking visible. DNA rests on the assumption that group decision-making involves actors competing for discursive dominance. DNA suggests that by analyzing the discourse of actors in a decision-making process, researchers can cluster actors into coalitions. One can then trace the phases of these coalition-building-processes and see when and how a winning coalition developed. As I argue in this article, understanding decision-making processes as coalitional struggles for discursive dominance advances the groupthink approach to FPA because it makes the discourses that are the foundation of decision-making processes visible. Moreover, it allows scholars to broaden sources of information, thereby integrating verifiable and replicable data into their research. I demonstrate the usefulness of DNA with the case of the Iraq invasion by the United States in 2003. Since this invasion, a great deal of investigative journalism and scholarship has focused on the questions of why the administration of George W. Bush decided to topple Saddam Hussein and occupy Iraq, who the driving forces behind this decision were, and when this decision was made. The elite of US investigative journalists were the first to address these questions by blaming neoconservative ideology (Suskind 2004; Packer 2007; Bamford 2005; Kaplan 2008), the terrorist attacks of September 11, 2001, and the threat of nuclear-armed terrorists and their state supporters (Woodward 2004; Suskind 2007), or a defective decision-making process within the administration (Hersh 2005; Isikoff and Corn 2007). Moving beyond descriptive (and often speculative) accounts of the invasion, political scientists began to systematically analyze the decision-making process and provide answers to the question of the actual motives underlying the invasion. According to these studies, individual characteristics and leadership styles (Shannon and Keller 2007; Dyson 2009), ideas and ideologies (Flibbert 2006; Burgos 2008; Lind 2012), pressure groups (Mazaar 2007; Mearsheimer and Walt 2007; Vanderbush 2009), or misperceptions and a defective decision-making process (Yetiv 2011; Burke 2005; Fitzgerald and Lebow 2006; Jervis 2006; Houghton 2008; Mitchell and Massoud 2009; Badie 2010) are proposed as explanatory variables for the war. Like investigative journalism, these studies view the hawkish group around Cheney and Rumsfeld as the pivotal actors behind the decision. Nevertheless, both investigative journalists and political scientists disagree as to the specific moment when the decision was made. Arguments range from pre-September 11 or September 11 itself, or from spring or fall of 2002, to the claim that the decision process was so incremental as to lack a clear starting date. Although journalists and scholars have advanced our understanding of the decision-making process, their analyses face the problem of sources as a serious methodological shortcoming. While a few studies on decision-making in course of the Iraq invasion have used verifiable and replicable data (Dyson 2006, 2009; Shannon and Keller 2007), most have not. The journalists’ descriptions of the decision-making dynamics primarily rest on whistleblower accounts and hearsay, sources that are neither verifiable nor replicable, with only the individual journalist's reputation giving weight to the message. Consequently, scholarly analyses that base their conclusions on these sources do not fulfill strict methodological standards and therefore run the risk of building their explanations on shifting sands. I argue that DNA helps counteract this problem and provides researchers with tools for making concurrence-seeking visible, based on verifiable and replicable data. More specifically, this article posits two core arguments concerning the phases of the decision-making-process in the run-up to the 2003 invasion of Iraq. First, speeches, interviews, and autobiographies contribute to a significantly better understanding and deciphering of the decision-making process than investigative journalism and are a necessary corrective to it. Second, and similar to the findings of McCauley (1979) and Stern and Sundelius (1997), the decision-making process must be understood as a two-phased process, first, building a consensus about the nature of the problem and, second, agreeing on what to do about it. In the first phase, President Bush and his inner circle built the decision to invade Iraq on a homogeneous and undisputed threat-perception induced by the terrorist attacks of September 11, 2001. This phase featured three signs of groupthink (Janis 1982): a cohesive decision group, a lack of norms and impartial leadership to guarantee an unbiased discussion, and a provocative situational context. In the second phase, given this perception and a dysfunctional group structure, a military policy discourse incrementally replaced the diplomacy-favoring discourse in 2002, a process that inevitably led to war. My analysis thus adds to the discussion of groupthink as an explanation for the Iraq invasion 2003 and puts the argument of Badie (2010), that the shift in the administration's view of Saddam Hussein as an existential threat happened gradually over periods of months until the invasion, into perspective. The threat perception changed immediately after the terrorist attacks of September 11, 2001. The struggle between Bush and his key advisers in the following months, therefore, was not so much about the question of how dangerous Saddam Hussein might be but about how to tackle this threat effectively. This result has consequences for the analysis of decision-making processes in general and groupthink in particular. By splitting a foreign-policy decision-process into two phases (i.e., framing the threat and answering the threat), the shifts within each phase become more detailed. Hence, and because of the path-dependency of the problem-definition and the solution phase, groupthink may be present in the former but absent in the latter, an insight that can be lost by mixing both stages. This article is organized as follows: First, I outline the role of small groups in foreign-policy decision-making and argue that under certain conditions, these groups are decisive for foreign-policy decision-making. Second, I introduce the theoretical and methodological foundations of DNA and outline the research design. Third, I show how the terrorist attacks of September 11, 2001, led to a sudden, undisputed, and homogeneous threat perception. Fourth, I explain how this homogeneous threat perception incrementally transformed the heterogeneous policy discourse into a homogeneous one in the third quarter of 2002, so that the focus shifted primarily to military instead of nonmilitary policy options. Small Groups in Foreign-Policy Decision-Making Determining the principal unit of analysis for explaining foreign-policy decision-making is one of the core topics in foreign-policy analysis. One tradition of classic FPA scholarship seeks to explain a state's external behavior by examining the internal dynamics of groups and organizations. According to this approach, foreign-policy outcomes are explained best by the social context of groups and institutions in which these individuals are embedded, that is, “the definition of the situation” (Snyder et al. 2002, 76). Following this line of reasoning, foreign-policy outcomes can be explained through an understanding of this context and its distinctive processes and dynamics. Although Hart (1994, 132) is skeptical of the role of small groups in foreign-policy outcomes and ascribes more importance to bureaucracies (Allison and Zelikow 1999) or organizational routines and cultures (Halperin, Clapp, and Kanter 2006), Tetlock et al. (1992) conceive small groups as the decisive unit of analysis. As Hart, Stern, and Sundelius (1997a, 8) argue, small group analysis helps us to explain “the crucial, formative policy decisions that shape significant parts of a country's foreign policy.” Hence, small groups as collective social forms are crucial agents of decision-making. Beside the groups’ function as the central locus of policy making, they have the advantage of being able to correct or minimize individual shortcomings. People are not able to act fully rational. They are rather “bounded” by their inability to gather all the information necessary for decision-making and process it adequately in a short period of time (Simon 1985). Furthermore, people exhibit all sorts of cognitive deficiencies that impair their perceptions of a situation and rational responses to it. In a group context, the shortcomings of one individual in a certain area can be compensated for by others. However, as psychological studies have revealed, “groups [too often] actually amplify those mistakes [and] turn out to be even worse than individuals” (Sunstein and Hastie 2015, 2) when it comes to rational decision-making. This observation is especially true when groups act rapidly, emotionally, and intuitively. There are several reasons why groups fail to compensate for the shortcomings of their members. Actors within groups may fail to speak up because more powerful and more competent actors have a different opinion or because they are socially pressured to self-censorship (Sunstein and Hastie 2015, 22–23). In addition, “unreasonably strong group and team members who might capitalize on the situation to force his or her ideas on the group” (Franz 2012, 3) can undermine team effectiveness. Furthermore, group deliberations “usually reduce . . . variance” (Sunstein and Hastie 2015, 28) and reinforce uniformity. This last insight is particularly striking in the effect of “polarization.” Depending on the group's initial predisposition, group decisions are more extreme (either in a more cautious or more risky direction) after the group discussion (Franz 2012, 163 and Sunstein and Hastie 2015, 78–85). As Kaarbo (2008) and Kaarbo and Beasley (2008) argue, polarization often takes place in cabinet coalitions. Minorities within such coalitions can manipulate the decision-making process and push the group towards a more extreme option. This push towards the extreme can be explained by either the “political hijacking of the junior partner,” the “domestic political weakness” of the coalition itself, the “diffusion of accountability” in such groups, or classical “logrolling” (Beasley and Kaarbo 2014, 730). This insight not only applies to coalition cabinets, where a consensus of the parties involved is essential for policy making, but also to presidential systems, like in the United States, where minority views within the administration can push a president towards an extreme decision, such as going to war (Elman 2000, 97). Vertzberger (1997, 284) attributes this polarization effect in the US presidential system to the self-confidence and assertiveness of certain members of the administration's inner circle. These members “with more radically polarized judgements and preferences invest more resources in attempts to exert influence” than others. In summary, groups do not necessarily make better decisions than individuals. Groups have their own shortcomings, which stem from their structures and the dynamics between members. A path-breaking contribution to the field of small group decision-making is Irving L. Janis’s book Groupthink: Psychological Studies of Policy Decisions and Fiascoes (1972/1982). With its focus on decision-making processes in small groups and on the reasons for suboptimal policy outcomes, this book has become a standard reference in foreign-policy analysis and the springboard for a lively discussion on small groups in foreign-policy decision-making (e.g., Smith 1984; Mulcahy 1995; Stern and Sundelius 1997; Stern 1997; Schafer and Crichlow 2002; Yetiv 2003). According to Janis (1982, 9), small-group decision-making is often characterized by “a deterioration of mental efficiency, reality testing, and moral judgment that results from in-group pressure.” In essence, “groupthink” discusses how groups reach a consensus and whether it is the result of a thoughtful deliberation process. Janis focuses on group processes in which there is an emphasis on concurrence-seeking. This is especially valuable to groups during periods of stress when there is “a heightened need for affiliation with other members” (George 1997, 37). As Hart (1991, 247) points out, groupthink may result in defective decision-making processes in which the consensus is a faulty one because of an “excessive form of concurrence-seeking” where “group members have come to value the group (and their being part of it) higher than anything else.” Groupthink occurs when four antecedent conditions prevail. First, when a group is cohesive. According to Franz (2012, 177, 180–81 and 187), cohesion occurs either when members feel good about each other (social cohesion) or when they are committed to a goal (task cohesion). Cohesion can have a positive effect on decision-making, but too much cohesiveness can also be problematic. Highly cohesive groups tend to become more polarized and hence more prone to making risky decisions. Second, the structure of the group is another possible antecedent condition. Groups that isolate themselves from the outside world and prevent alternative views from being offered, risk a suboptimal decision process. Third, groupthink is more likely when the leader is not impartial and hence too directive and does not promote divergent opinions. When leaders do not feel restricted by group norms (that structure the deliberation process and give divergent opinions the chance to be discussed) other group members will stay silent. Such partial leaders push the group too early into a certain direction and thereby risk ignoring viable alternatives (Yetiv 2003, 423–24 and 431). As George and Stern (2002) stress, groups should, therefore, seek to promote a system of multiple advocacy, “in which competition and disagreement among different participants is structured and managed to achieve the benefits of diverse points of view” and where the deliberation process is organized by an honest broker who has no hidden agenda and guarantees impartiality. Finally, concurrence-seeking tendencies become more likely when the group finds itself in a situation characterized by high stress, caused either by an external threat or a group's lack of self-esteem due to recent failures. This tendency leads to premature consensus-building in which the group neglects to assess various policy options. Groupthink therefore finds its expression in the overestimation of the group, a closed-mindedness among its members, and a pressure towards uniformity through self-censorship, the illusion of unanimity, or silencing dissenters. Given these symptoms, the probability increases that the group exhibits an incomplete survey of objectives and alternatives, fails to reappraise initially rejected alternatives, and conducts a poor and biased information search. Although the presence of some of these behaviors does not automatically induce groupthink, the chances of a defective decision-making process increase the more that these symptoms and consequences appear (Janis 1982, 245). In sum, groupthink-theory places special emphasis on the process of consensus building in groups by highlighting the concurrence-seeking effort of its members. However, as Hart (1994, 7) emphasizes, concurrence-seeking per se is not a problem but rather a necessary precondition for decision-making. Groups that are not able to reach a timely consensus also fail to produce good policy outcomes, such as groups that reach a consensus prematurely, neglecting to assess all its options. The central question, therefore, is how a group reaches a consensus and whether or not the group discussed all possible options extensively in a fair, unbiased, and inclusive deliberative process. One of the biggest challenges for Janis's concept is that of causation. As Baron (2005) correctly notes, the symptoms of groupthink do not necessarily follow from concurrence-seeking, and concurrence-seeking itself may not be the result of antecedent conditions. Hence, the central question for researchers must be how to make concurrence-seeking visible. George (1997, 39) points to this weakness of groupthink theory. He argues that concurrence-seeking and hence groupthink cannot be observed directly and must be inferred by either the presence of antecedent conditions and/or the final symptoms. This problem is not limited to groupthink theory. As Bennett and Checkel (2015, 11) note, it concerns other theories as well: “[w]e do not get to observer causality—we make inferences about it.” Especially in small groups like presidential cabinets, researchers are not able to directly observe group deliberations and hence to trace causal mechanisms. A solution to this problem is the use of process tracing as “the central within-case-method” (Bennett and Checkel 2015, 4) in single or small N studies, to “trace the links between possible causes and observed outcomes” (George and Bennett 2005, 6). Process tracing analyzes interviews and documents and, by doing so, seeks to capture causal mechanisms. Whereas alternative sources like oral history or declassified archival materials are often available for earlier cases, researchers of recent cases are mainly left with investigative journalism. Yet, this alternative lacks scientific reliability and replicability. Thus, in the following section, I will outline DNA and argue that this concept is a promising way to make the concurrence-seeking efforts of groups visible and comprehensible and to trace the causal mechanisms of decision-making in groups. Discourse Networks Analysis and the Role of Advocacy Coalitions Discourse Network Analysis rests upon three theoretical foundations—the Advocacy Coalition Framework (ACF), Discourse Coalition Theory (DCT), and Social Network Analysis (SNA). ACF has its origins in public-policy analysis and seeks to explain policy change and its drivers in the course of time. Following Paul A. Sabatier (1988, 130), ACF primarily focuses on the “interaction of political elites within a policy community/subsystem attempting to respond to changing socioeconomic and political conditions.” ACF assumes that policies only change when a “winning-coalition” of likeminded proponents succeeds over its competitors. Hence, political elites must be successful in allying with like-minded actors, develop common strategies, and convince a majority within the policy community of the advantages of their approach. The more actors a coalition unites and the more homogeneous they are (in terms of similar core beliefs and/or common strategies), the higher the probability that they will succeed and become the winning coalition (Sabatier and Weible 2007, 196). Although Sabatier developed the ACF for analyzing larger policy communities comprising a variety of different organizations and actors and observed them over a longer time span, ACF's assumptions are also applicable to the realm of foreign-policy analysis and smaller communities, as Lee (2014) has shown. Applying ACF to President Bush and his advisors and their decision-making process on the Iraq invasion in 2003 is justifiable for two reasons. First, after the terrorist attacks of 9/11, foreign-policy decision-making became increasingly centralized in the administration and was dominated by a few, very powerful actors (Kassop 2003). Second, although in the US presidential system consensus is neither necessary (because of presidential prerogatives in foreign and security policy) nor desirable (George and Stern 2002), presidents nevertheless draw on group deliberations in their inner circles when they make decisions (Vertzberger 1997, 284; Elman 2000, 97). President Bush as commander-in-chief unambiguously emphasized that although the decision to go to war was his own, he nevertheless based this decision on a “vigorous debate” and listened to the “discussion and divergent views” (Bush 2010, 189–90) of his inner circle (Mitchell and Massoud 2009, 272–73; Rice 2011, 86–87; Rumsfeld 2011, 319). Consequently, ACF is a suitable concept to explain which actors and which coalitions of actors within the inner circle of the administration were the crucial ones and how the struggle between different coalitions evolved over time, leading up to the war decision. The second concept on which DNA builds is Maarten A. Hajer's (1993; 1997) Discourse Coalition Theory. Hajer (1993, 44) believes that political problems are socially constructed by those who are able to influence the policy process. This process of construction, in turn, is revealed by analyzing the public discourse of the actors involved. Renshon (2009) supports this claim when he argues that public statements reveal private beliefs and are not an attempt to distract from the actual agenda. These statements provide insight into the private process of policy building and are, therefore, especially valuable when the private transcripts are unknown or unknowable. This claim is further underlined in the case of the Iraq War by the effort of Vice-President Cheney and his followers to plant supportive evidence of their views in the stories of reporters, which then entered the public record and were again used as an evidence of “independent” verification of these views. Hence, private beliefs are recoverable as a subtext in the public records through discourse analysis. But even if one is skeptical about the transparency of public statements, public statements constrain an actor's behavior in the medium to long-term (Schimmelfennig 2001). So it does not matter whether actors only comply with the official lingo or if they have also internalized the message (according to McCauley (1979), a necessary requirement for groupthink). These individual discourses can then be aggregated to form discourse coalitions, which are defined as groups of actors that use similar discourses towards a policy problem. Discourse Coalition Theory operationalizes these public discourses and can be used to locate actors and their views towards a policy problem within a group. The third and final concept on which DNA rests, is Social Network Analysis (SNA). In the last couple of years, SNA has gained prominence in the social sciences (Borgatti et al. 2009) and especially in political science (Lazer 2011). This is due to the fact that core political science concepts, such as power, agency, and structure, must be understood as relational phenomena. As McClurg and Young (2011, 41) write: Only by focusing clearly on the degree to which relations constrain agency can we begin to fully explain our depictions of what the concept of liberty, freedom, and agency mean in different political systems and contexts. SNA, therefore, enables researchers to put agency in context and to show how the context and its structures influence actors and their behavior (Hafner-Burton, Kahler, and Montgomery 2009). In the case of this article, SNA is particularly valuable in explaining the different phases of group deliberations and the role of certain actors in these phases. In this article, I apply three distinctive network metrics, which help me to explain which actors are responsible for the formation or breakup of coalitions, to understand the degree of cohesiveness within groups and coalitions, and to display coalition building processes as such. These metrics are centrality, density and modularity. First, centrality reveals how important a certain actor is within a network. It does so, by calculating the sum of edges (direct relations) of a certain node (the actor) in comparison to all the other nodes and their sums of edges. A special form of this degree centrality is the betweenness centrality, which ascribes importance to a node if it acts as a bridge between other nodes. Betweenness centrality, therefore, allows me to highlight actors with the most crucial relations and consequently the most influence within a network. These actors are responsible for the evolution of a coalition and can be understood as brokers. Second, density describes the ratio between the actual and the possible number of edges between nodes in a network (Kadushin 2012, 29): the higher the value (between 0 and 1), the higher the homogeneity of a network or a subgroup within a network. As Erisen and Erisen (2012) demonstrate in their experimental study, density and thus homogeneity are indicators for the likelihood of the success or failure of decision processes. The higher the density and homogeneity within a group (or subgroup), the more biased a decision process becomes and the more likely this process will end up in a suboptimal result. Consequently, this network metric perfectly fits to groupthink and its argument that hasty concurrence-seeking at an early stage and a too homogeneous deliberation process can threaten results. Finally, modularity uses density to compare the number of actual edges between nodes in a network with the possible number of edges in comparable random networks. By doing so, modularity allows us to divide networks into subgroups or, in our case, into coalitions (Kadushin 2012, 48; Newman 2012, 224). To sum up, Discourse Network Analysis makes three major contributions to the analysis of group decision-making. First, it conceives the building of a winning-coalition between actors in a policy process as a necessary precondition for the change of policies (ACF part of the approach). Second, it identifies these coalitions by operationalizing the discourse1 of their members (DCT part). And finally, it uses social network analysis to display these actors via their discourses and show how this network evolved and changed over time and who the crucial actors in this process were (SNA part). By doing so, DNA captures the causal mechanisms of decision-making and provides answers to the questions of whether and when a group came to a consensus and who the driving forces behind this consensus were. Consequently, DNA helps us to understand whether or not this concurrence-seeking process was premature and hasty. Research Design To construct discourse networks and to calculate the network metrics, I compiled a dataset consisting of 227 interviews (spontaneous speech-acts) and speeches (prepared speech-acts) of the inner circle of decision-makers (i.e., Bush, Cheney, Powell, Rumsfeld, Rice, Tenet, Armitage, and Wolfowitz) in the period from January 2001 through December 2004. Whenever possible, one interview and one speech per actor and quarter were chosen to ensure a timely representation of discourses, on the one hand, and to control for impression-management (by combining spontaneous with prepared speech-acts), on the other hand. From the total list of all possible speeches and interviews in the sample period, I chose only those in which the keywords “Iraq” and “Hussein” appeared most often. I coded the data by using the Discourse Network Analyzer, a program developed by Leifeld (2011) to compile discourse networks. Statements were coded in four steps: (1) actor issuing the statement, (2) organization the actor belongs to, (3) category the actor is referring to, and (4) approval (yes) or rejection (no) of the category by the actor. So, for example, I coded the statement of President Bush “Iraq continues to flaunt its hostility and to support terror” in his State of the Union address of 29 January 2002 as follows: (1) Bush, (2) White House, (3) support of terrorism, (4) yes. Paragraphs were seen as complete units of meaning. Therefore, statements were coded only once within a paragraph (as long as they were not contradictory). I prepared the twenty-nine categories of the coding scheme inductively on the basis of a small pilot study and organized these categories into two groups representing two subdiscourses of the overall discourse: (1) US threat perception and (2) US policy options toward Iraq. I used this dataset to construct both affiliation and congruence networks. The former displays actors sharing categories; the latter only displays actors, showing how close these actors are to each other based on the number of shared categories. Thus, affiliation networks are used to display the dominant categories of a discourse in a certain period of time and to show who referred to these categories; whereas congruence networks reveal which actors belong together in certain coalitions. 9/11 and the Undisputed Threat Perception I conceive of the decision-making process in the run-up to the Iraq war not as a linear process but as a more complex one (George and Bennett 2005, 212) of two interrelated decisions (i.e., first, framing the threat and then answering it). I apply Janis’s concept of small group decision-making to analyze this process. This requires me to clarify the antecedent conditions under which President Bush and his inner circle made decisions on Iraq. I will do so by analyzing the cohesiveness of the decision group and its structure and by discussing the situational context in which the group operated. Given these antecedent conditions, I will demonstrate how the terrorist attacks of September 11, 2001, instantly homogenized the threat perception of the actors involved, providing the basis for the subsequent policy discourse. The inner circle of advisers of President Bush was an experienced group of actors when it came to foreign and security policy, and it was highly cohesive. The group members knew each other from the early 1960s on, working in the Nixon and Ford administrations (Cheney and Rumsfeld); with Reagan (Cheney, Rumsfeld, Powell, Armitage, and Wolfowitz); and under G. H. W. Bush (Cheney, Powell, Rice, Armitage, and Wolfowitz). During these years, they were not only able to build intense personal relationships (Rumsfeld and Cheney or Powell and Armitage), but they even had to deal with Iraq during the first Iraq war in 1990–91. President Bush and Tenet (Director of Central Intelligence) were the only ones who did not have a record of cooperation in the past. Nevertheless, these actors built one of the most experienced, well known and tightly knit foreign and security policy teams in U.S. history, when the Bush administration began to work in January 2001 (Mann 2004). In addition to this social cohesiveness, the group also exhibited a very high level of task cohesiveness, induced by the provocative situational context of the terrorist attacks of September 11. Keeping the United States safe and preventing a further attack became the dominating task (e.g., Bush 2010, 181; Cheney and Cheney 2011, 333; Rumsfeld 2011, 362; Powell and Koltz 2012, 36) that bound the members of the administration together. Concerning the structure of this group, it is remarkable how important the concept of multiple advocacy became for its members. President Bush (2010, 65–66), Powell (Powell and Koltz 2012, 10), Rumsfeld (2011, 180184), and Rice (2011, 153) all stressed the importance of controversy and were aware of the danger of a deliberation process that was too homogeneous. The widely accepted assertion that the president was not an impartial leader but predisposed toward war (Yetiv 2011, 238; Feith 2009, 243; Pillar 2011, 25) cannot easily be discerned from the biographies of members of the administration or the discourse of the president himself. Rumsfeld (2011, 319 and 328), for example, recalls that Bush never really expressed his preferences and did not make clear what the results and consequences of their deliberations process were. Nevertheless, the deliberative process under Bush lacked clear norms and was characterized as too informal (Pfiffner 2009, 375). This cohesive group of actors, lacking formal norms structuring the deliberation process, faced a totally new security policy context after the terrorist attacks of September 11. Two schools of thought evaluate the impact of these attacks on the United States and its top-level decision-makers in contrasting ways. The first school sees the members of the Bush administration as the victims of 9/11 and the events thereafter. According to this line of reasoning, the terrorist attacks forced them to alter their approaches towards foreign and security policy, giving them little scope for variation (Flibbert 2006, 325–28; Mitchell and Massoud 2009, 279; Badie 2010, 278). The second school, however, sees Bush and his advisers as the agents of this change; September 11, 2001, became a useful pretext for implementing a long-desired re-orientation of US foreign and security policy (Kaufmann 2004; Mazaar 2007, 13; Burgos 2008, 225). Regardless of whether the Bush administration became a victim of the dynamics following the terrorist attacks or was the agents of them, September 11, 2001, transformed the context in which decisions were made. This transformation was clearly expressed in the autobiographies of the different actors. For Bush (2010, 128, 137, 151 and 229), the terrorist attacks became a perceptual shock, redefining and radicalizing his world view. Cheney and Powell, as well as Rumsfeld, perceived 9/11 as a seminal event, bringing the topics of terrorism and nuclear proliferation to the fore (Cheney and Cheney 2011, 10; Powell and Koltz 2012, 218; Rumsfeld 2011, 352–55). Rice (2011, xviii, 79, 148 and 169) depicts 9/11 as a “crack in time,” dramatically changing the way security issues were seen. Tenet (2008, 305) supports her view by writing: “After 9/11 everything changed. Many foreign-policy issues were now viewed through the prism of smoke rising from the World Trade Center and the Pentagon.” Tenet (2008, 174), furthermore, perceives the attacks as a personal and organizational failure. Henceforth, the prevention of further attacks on US soil became the unanimous conviction of the administration. All other aims and objectives were subordinated to this long-term goal (Bush 2010, 154–55; Cheney and Cheney 2011, 6). It was against the backdrop of the cohesiveness of its members, the lack of norms and an honest broker to structure the discussion, and the dramatic transformation of the context—the three antecedent conditions of groupthink—that the decision group came to its first consensus: a sudden, undisputed, and precedent-setting threat-perception. Figure 1 displays the affiliation network of the “threat-discourse” from January 2001 until September 11, 2001. Although all actors perceived Iraq as a threat, they differed on the question what exactly made up this threat. Bush, Cheney, Powell and Rice saw Iraq as an “aggressive actor,” but only Bush and Rice, as well as Armitage, linked this perception to the potential threat of a nuclear-armed rogue state (“rogue state threat with WMD”—weapons of mass destruction). Bush generally believed the United States to be in a “post-Cold-War security environment” with new rules defining the game and shared this view only with Wolfowitz. The deputy secretary of defense, in turn and together with Tenet and Powell, was convinced that the Iraqi threat had its causes in the “tyranny” of the regime. Powell on the other hand, stressed Hussein's “history of WMD-use” and was convinced that Iraq possessed or at least sought to develop WMDs (“possession/development of WMD”). This conviction was shared only by Tenet, who was the only actor who feared nuclear-armed terrorism (“terrorists seek/will use WMD”). In this threat-discourse, Rumsfeld acted as an outsider, sharing no category with any other member of the administration. According to him, the Iraqi threat was generally underestimated (“underestimation of threat”) and was especially troublesome because of Iraqi's “history of developing WMD.” Figure 1. View largeDownload slide Affiliation network of the “threat-discourse” until 9/11 (agreement “yes”). Figure 1. View largeDownload slide Affiliation network of the “threat-discourse” until 9/11 (agreement “yes”). The small value of density of this affiliation network of 0.062 and the relatively high value for modularity of 0.219 (for a small group) of the congruence network in the same period show that the threat-perception was heterogeneous and no actor dominated the discourse. The heterogeneous view of the Iraqi threat is a broad hint to a lively and healthy discussion with respect to the deliberation process. However, this observation changed dramatically and instantly with September 11, 2001. As Figure 2 of the threat-affiliation network from September 11 until March 20, 2003, shows, all members of the Bush administration were now convinced that Iraq was an aggressive actor, possessing or at least developing WMDs. In this “post-9/11 security environment,” Iraq's “history of WMD-use” and the assumed “support of terrorism” by Saddam Hussein were seen as unacceptable. The density of this network rose to 0.19, whereas the modularity of the corresponding congruence network reached zero. Consequently, the discourse became highly homogeneous and one cannot distinguish between subgroups of actors. This finding is even true when the period of observation is reduced to quarters. From 9/11 until the invasion in March 2003, the threat-perception in every single quarter was highly homogeneous with a modularity value of the congruence networks of zero.2 Figure 2. View largeDownload slide Affiliation network of the “threat-discourse” from 9/11 to Iraq invasion (agreement “yes”). Figure 2. View largeDownload slide Affiliation network of the “threat-discourse” from 9/11 to Iraq invasion (agreement “yes”). These discourse networks confirm the critique of several individuals within the administration. Tenet (2008, 305), for example, complains that there was never a serious discussion about the threat facing the United States. Paul O'Neill, the secretary of treasury, supports Tenet's account and emphasizes that the discussion never centered on the question of why to invade Iraq but only on how to do it (Suskind 2004, 76). This sudden and undisputed threat-perception, shared by all key advisers and the president himself, became the precedent-setting basis on which the administration began to think about policy towards Iraq. It was this premature consensus, which was never reexamined, that determined the future discussion and thus the decision-making process. Although failure to reexamine this perception was the only symptom of groupthink in this first phase of the decision process, it was a vital failure. Without this early consensus, the invasion of Iraq would have been unimaginable (Badie 2010, 285). From a Heterogeneous Nonmilitary to a Homogeneous Military Policy Discourse In contrast to the threat-discourse, the discussion about policy options began as a very homogeneous discourse (see Figure 3). Although in the first months of the new administration, a lively discussion in the principal's committee evolved about the usefulness of sanctions—with Powell a fierce advocate and Rumsfeld an opponent (Woodward 2004, 15)—President Bush and his team agreed on most categories and favored nonmilitary options and a modest regime-change policy in dealing with Iraq. “Sanctions” and the “consensus/cooperation with allies” build the backbone of this strategy, which was supported by Bush, Cheney, Powell, and Rice. Only Rumsfeld did not agree on sanctions as an appropriate policy (see Figure 3). Instead, he focused on regime change (although not militarily but by supporting the opposition) and “containment/deterrence,” sharing the first option with Powell and the latter with the president. Furthermore, Powell and Rice asked for “inspections” and were open to “selected military action” if necessary. Modularity of zero in the respective congruence network shows that the administration had a homogeneous policy approach at this early stage. Although they could not agree on how serious the threat posed by Iraq was, they were sure to tackle it by nonmilitary means, such as sanctions, containment, and deterrence, and in cooperation with allies. Figure 3. View largeDownload slide Affiliation network of the “policy-discourse” pre-9/11 (agreement “yes”). Figure 3. View largeDownload slide Affiliation network of the “policy-discourse” pre-9/11 (agreement “yes”). The terrorist attacks of September 11 caused a first shift in this policy (see Figure 4). As Bush (2010, 230) and Rice (2011, 171–72) emphasize in their autobiographies, the administration now followed the policy of “coercive diplomacy,” which rested on two pillars. On the one hand, the United States sought to convince Iraq by diplomatic means to give up its supposed WMD program. On the other hand, the buildup of a military threatening posture would emphasize US intentions and provide a backup if the diplomatic path failed. However, for the time being, the focus should lie on the diplomatic, nonmilitary path. Figure 4. View largeDownload slide Affiliation network of the “policy-discourse” from 9/11 until 31 December 2001 (agreement “yes”). Figure 4. View largeDownload slide Affiliation network of the “policy-discourse” from 9/11 until 31 December 2001 (agreement “yes”). At first, this policy realignment found its expression in the shift of focus from “sanctions” (only Powell supported this policy) to “inspections” (supported by Powell, Wolfowitz, Bush, and Cheney). Whereas Cheney, Rumsfeld, and Armitage envisioned “containment/deterrence,” President Bush cautioned against the “risk of inaction.” Although the actors were again quite homogeneous in their approach (modularity of zero in the respective congruence network) and favoring mostly nonmilitary measures, the focus of discussion shifted from a more passive (sanctions) to a more active approach (inspections). In the first six months of 2002, this homogeneous discourse coalition fell apart but without being replaced immediately by a new consensus. In the first quarter of 2002 (see Figure 5), only Powell and Rice urged for “inspections” and “sanctions.” All the other members of the decision group did not support this strategy any more. Instead, they referred to categories like “risk of inaction” (Bush, Rumsfeld, and Wolfowitz) or “all options are on the table” (Bush, Wolfowitz, and Powell). Especially the latter category became an indicator that an open war was now imaginable for an increasing number of decision-makers. Furthermore, Rumsfeld, supported by Rice, favored a “regime change.” Figure 5. View largeDownload slide Affiliation network of the “policy-discourse” in the first quarter 2002 (agreement “yes”). Figure 5. View largeDownload slide Affiliation network of the “policy-discourse” in the first quarter 2002 (agreement “yes”). This development climaxed in the second quarter of 2002 (see Figure 6). Powell and Rice were the only actors calling for “sanctions,” but Powell stood alone in urging for “inspections.” Rumsfeld (supported by Powell) again pressed for a “regime change.” Cheney and Wolfowitz stressed the “risk of inaction,” whereas the President reaffirmed that “all options are on the table.” Figure 6. View largeDownload slide Affiliation network of the “policy-discourse” in the second quarter 2002 (agreement “yes”). Figure 6. View largeDownload slide Affiliation network of the “policy-discourse” in the second quarter 2002 (agreement “yes”). As the congruence networks and the values for modularity (0.189 in the first and 0.375 in the second quarter of 2002) illustrate, the initial consensus and the homogeneity of the group broke completely apart. In those six months, discussion within the administration shifted incrementally from a nonmilitary discourse (i.e., “sanction” and “inspections”) toward a military discourse (i.e., “risk of inaction,” “regime change,” and “all options are on the table”). This finding supports Woodward's (2004, 108 and 119) argument that the military pillar of the “coercive diplomacy” policy gained the upper hand in the first half of 2002. According to Powell, the “[m]omentum within the administration was building toward military action” (Powell and Koltz 2012, 209) in August 2002. Rice (2011, 158159) agrees with Powell and sees Cheney and Rumsfeld as proponents of this realignment. She is convinced that the president became increasingly sympathetic to this shift. From the third quarter of 2002 until the invasion in March 2003 (see Figure 7) the “risk of inaction” and “war” became the dominant categories, shared by all. Sanctions were not an issue any more. The decision group suddenly became homogeneous again (modularity of zero), and the policy discourse was solely characterized by military options. Diplomacy as the first pillar of the coercive-diplomacy strategy was abandoned in favor of military pressure. This shift unleashed a dynamic that gathered momentum and inevitably led to the war in March, as Tenet (2008, 333) argues. By turning to the military pillar of the strategy, according to Bush (2010, 250252), the United States could not back down when Iraq refused to fulfill US claims. The United States was forced to go to war in order not to lose its credibility. The crucial period of time, thus, was the first six months of 2002, as a new consensus centering around military options replaced the initial consensus on nonmilitary policy options. The dominant actors in this phase of the decision-making process were Rumsfeld and Powell, who had the highest values for betweenness centrality, 4.33 and 2.0, respectively. These two actors served as brokers between all the others. This, again, indicates that the decision process in this crucial phase was shaped by an intense controversy between Powell, who favored nonmilitary options and thus the diplomatic pillar of the coercive-diplomacy strategy, and Rumsfeld, who demanded a tougher stance towards Iraq and the pursuit of the military pillar. Figure 7. View largeDownload slide Affiliation network of the “policy-discourse” from 1 October 2002 until 31 March 2003 (agreement “yes”). Figure 7. View largeDownload slide Affiliation network of the “policy-discourse” from 1 October 2002 until 31 March 2003 (agreement “yes”). This controversy was furthered by an increasingly dysfunctional group with personal rivalries, missing norms, and the lack of an impartial leader who could structure the decision process. Bush (2010, 65–66) acknowledges in his biography that he knew of the importance of an effective and heterogeneous decision group as a precondition for good advice. He, therefore, appointed Rumsfeld as secretary of defense to balance Powell, the highly admired secretary of state and potential dominant foreign and security policy figure in the administration. This decision was the starting point for a bureaucratic fight between members of the State Department and the Department of Defense, as well as the vice president and his office, over the dominance of the foreign and security policy process (Mann 2004, 271–75). Bush (2010, 89) concedes that the group became more and more ineffective as time passed: “What started as creative tension, had turned destructive.” The other group members shared this view of an increasingly dysfunctional National Security Council (NSC). Rumsfeld (2011, 319 and 328) laments that the whole decision process lacked structuring norms. Furthermore, Bush never made clear where he stood at the end of an intensive discussion and what the results had been. Rumsfeld (2011, 325–26) blames Rice and the president for this failure. Rice not only failed in structuring the process, according to Rumsfeld, but also failed to act as an impartial leader. She tried to forge a consensus that offered President Bush watered down compromises instead of distinctive policy choices. In contrast to Rumsfeld, Powell complains that his advice was no longer in demand and that crucial decisions were often made in his absence. Consequently, Bush did not get the full spectrum of opinions (Powell and Koltz 2012, 36 and 55). Rice confirms both Rumsfeld's as well as Powell's view. She tried to be the honest broker that Rumsfeld and Powell objected to for very different reasons (Rice 2011, 18–19). Tenet (2008, 138) also blames Rice for the dysfunction of the decision body. She was neither willing nor able to put up a fight against such dominant actors as Cheney and Rumsfeld, who began to marginalize her and take over the decision-making process. This finding is in line with other analyses of the decision-making process in the run-up to the 2003 Iraq War, which also found: the increase of influence of actors like Cheney and Rumsfeld and their allies within the inner circle of advisers; the creation of alternative centers for intelligence analysis to bolster their preferred policy choices and strategies; the pressure put on dissenters; and the neglect of a comprehensive and realistic postwar and contingency planning (Pfiffner 2004; Woodward 2004; Hersh 2005; Isikoff and Corn 2007; Badie 2010). These consequences of the groupthink syndrome also underline how the actors struggled for discourse dominance and why Cheney, Rumsfeld, and others succeeded in doing so and convinced President Bush, the crucial actor in this process, of their approaches. What these networks further demonstrate is that Secretary of State Powell, who is often regarded as an opponent of the war, also shifted to the dominant military discourse beginning in the third period of 2002. This finding confirms Rumsfeld's (2011, 457) claim that the decision to go to war was unanimous, with no one verbally dissenting. Although Bush (2010, 251) is convinced that Powell had serious doubts about the war, the secretary of state did not question the decision in public (an indicator of self-censorship). Conclusion The literature on the decision-making process in the run-up to the 2003 Iraq War faces a serious dilemma. It builds its argument largely on investigative journalism, sources for which are neither scientifically verifiable nor replicable. Primary documents, that is, speeches, interviews, and autobiographies, are often neglected in these analyses because they are either seen as useless for deciphering private decision-making processes or too biased to be reliable. Consequently, scholars do not take advantage of significant sources. In contrast, this article relies upon them, as well as on other primary sources to explain foreign-policy decision-making. They are corrective to accounts that rely upon investigative journalism and, beyond that, are verifiable and replicable. The article also shows the value of integrating Discourse Network Analysis into analyses of foreign-policy decision-making (i.e., groupthink in the case of this study). First, DNA permits political scientists to broaden their sources of information by bringing speeches and interviews into their analyses. Second, DNA contributes to the understanding of decision-making processes in groups. It reveals struggles between different coalitions of actors over discursive hegemony. It does so by relating the individual discourses of actors to the group discourse and thereby showing how coalitions are constructed and deconstructed. Network metrics help us to comprehend these dynamics by displaying the properties of groups and coalitions, such as density and modularity. DNA, therefore, provides a methodological means to make the consensus building efforts (in form of concurrence-seeking) visible and to display the phases of this process in more detail. In particular, DNA provided substantial empirical evidence for examining the intermediate steps in the decision-making process and for answering why the United States invaded Iraq, who the driving forces behind this decision were, and when the decision was made. First, the decision-making process in the run-up to the war occurred in two phases. In the first phase, decision-makers agreed on a common threat perception. Although there was no consensus on the dangers Iraq posed at the beginning of the administration, the terrorist attacks of 9/11 changed that. This consensus building process featured the three antecedent conditions of groupthink: a cohesive group of actors; a dysfunctional group lacking norms and impartial leadership that could structure the decision process; and a provocative situational context. Given the homogeneous threat-perception that was an outcome of the first phase, the second phase required the decision group to find a consensus concerning the right policy towards Iraq. This process was characterized by the incremental collapse of the originally dominant discourse, the focus on diplomatic means and nonmilitary policies, and the rise of a new, more hawkish military approach. After redirecting US policy towards Iraq on to the military path, the administration could not back down. It had to walk this path until the end in order not to lose face. Second, the driving forces behind this policy shift were a group of actors around Vice President Cheney and Secretary Rumsfeld. It was especially the latter who, as a broker, had the most influence on the deliberation process according to the discourse networks. Rumsfeld's discourse was more compatible with that of the other actors. Consequently, the discourse coalition surrounding him was more successful in convincing President Bush than his principal opponent, Secretary of State Powell. This finding supports those accounts that see the Iraq War decision as the result of a fierce fight between “hawks” (Cheney, Rumsfeld, and Wolfowitz) and “doves” (especially Powell). Finally, the replacement of the former diplomatic discourse with a more military one took place in the first half of 2002. In the third quarter of 2002, Bush and his inner circle (including Powell) once again shared a common approach towards Iraq. This finding suggests that the decision to go to war or at least the trend-setting decision to focus solely on the military path of the policy of coercive diplomacy, which led to the war, was taken in the last weeks of the second or early weeks of the third quarter of 2002. The use of DNA reveals why actors and groups follow certain policy options instead of others. Future research should not just apply this concept to case studies with small decision groups but to larger communities, including actors like bureaucracies and organizations, parliaments, pressure groups and media. Although DNA provides valuable insights into small group dynamics, it is much more powerful when the number of actors involved increases. 1 Following Holzscheiter (2014, 144) I define “discourse“ and “discourse analysis” as follows: “[D]iscourse analysis is engagement with meaning and the linguistic and communicative processes through which social reality is constructed. Discourse can therefore be defined as, basically, the space where intersubjective meaning is created, sustained, transformed and, accordingly, becomes constitutive of social reality.” Furthermore, I use a micro-interaction approach, focusing on the communicative and deliberative exchange of meanings between agents in the struggle for the better argument (Holzscheiter 2014, 146–150). 2 The only exception was the second quarter of 2002. However, in this case modularity reached 0.065, which is almost zero and thus not significant. Supplemental Information Supplemental information is available at the Foreign Policy Analysis data archive. References Allison Graham T., Zelikow Philip. 1999. Essence of Decision: Explaining the Cuban Missile Crisis . 2nd ed. New York, NY: Longman. Badie Dina. 2010. “Groupthink, Iraq, and the War on Terror: Explaining US Policy Shifts toward Iraq.” Foreign Policy Analysis  6: 277– 96. Google Scholar CrossRef Search ADS   Bamford James. 2005. A Pretext for War: 9/11, Iraq, and the Abuse of American Intelligence Agencies . New York, NY: Anchor Books. Baron Robert S. 2005. “So Right It's Wrong: Groupthink and the Ubiquitous Nature of Polarized Group Decision Making.” Advances in Experimental Social Psychology  37: 219– 53. Google Scholar CrossRef Search ADS   Beasley Ryan K., Kaarbo Juliet. 2014. “Explaining Extremity in the Foreign Policies of Parliamentary Democracies.” International Studies Quarterly  58: 729– 40. Google Scholar CrossRef Search ADS   Bennett Andrew, Checkel Jeffery T.. 2015. “Process Tracing: From Philosophical Roots to Best Practices.” In Process Tracing: From Metaphor to Analytic Tool , edited by Bennett Andrew, Checkel Jeffery T.. Cambridge: Cambridge University Press. Borgatti Stephen P., Mehra Ajay, Brass Daniel J., Labianca Giuseppe. 2009. “Network Analysis in the Social Sciences.” Science  323: 892– 95. Google Scholar CrossRef Search ADS PubMed  Burgos Russell A. 2008. “Origins of Regime Change: ‘Ideapolitik’ on the Long Road to Baghdad, 1993–2000.” Security Studies  17: 221– 56. Google Scholar CrossRef Search ADS   Burke John P. 2005. “Condoleezza Rice as NSC Advisor: A Case of the Honest Broker Role.” Presidential Studies Quarterly  35: 554– 75. Google Scholar CrossRef Search ADS   Bush George W. 2010. Decision Points . New York, NY: Crown. Cheney Dick, Cheney Liz. 2011. In My Time: A Personal and Political Memoir . New York, NY: Simon & Schuster. Dyson Stephen Benedict. 2006. “Personality and Foreign Policy: Tony Blair's Iraq Decision.” Foreign Policy Analysis  2: 289– 306. Google Scholar CrossRef Search ADS   Dyson Stephen Benedict. 2009. “‘Stuff Happens’: Donald Rumsfeld and the Iraq War.” Foreign Policy Analysis  5: 327– 47. Google Scholar CrossRef Search ADS   Elman Miriam Fendius. 2000. “Unpacking Democracy: Presidentialism, Parliamentarism, and Theories of Democratic Peace.” Security Studies  9: 91– 126. Google Scholar CrossRef Search ADS   Erisen Elif, Erisen Cengiz. 2012. “The Effect of Social Networks on the Quality of Political Thinking.” Political Psychology  xx: 1– 27. Feith Douglas J. 2009. War and Decision: Inside the Pentagon at the Dawn of the War on Terrorism . New York: Harper. Fitzgerald Michael, Lebow Richard Ned. 2006. “Iraq: The Mother of all Intelligence Failure.” Intelligence and National Security  21: 884– 909. Google Scholar CrossRef Search ADS   Flibbert Andrew. 2006. “The Road to Baghdad: Ideas and Intellectuals in Explanations of the Iraq War.” Security Studies  15: 310– 52. Google Scholar CrossRef Search ADS   Franz Timothy M. 2012. Group Dynamics and Team Interventions: Understanding and Improving Team Performance . Malden, MA: Wiley-Blackwell. Fuller Sally Riggs, Alday Ramon J.. 1997. “Challenging the Mindguards: Moving Small Group Analysis Beyond Groupthink.” In Beyond Groupthink: Political Group Dynamics and Foreign Policy-making , edited by Hart Paul ’t, Stern Eric K., Sundelius Bengt. Ann Arbor, MI: The University of Michigan Press. George Alexander L. 1997. “From Groupthink to Contextual Analysis of Policy-Making Groups.” In Beyond Groupthink: Political Group Dynamics and Foreign Policy-making , edited by ’t Hart Paul, Stern Eric K., Sundelius Bengt. Ann Arbor, MI: The University of Michigan Press. George Alexander L., Bennett Andrew. 2005. Case Studies and Theory Development in the Social Sciences . Cambridge, MA: MIT Press. George Alexander L., Stern Eric K.. 2002. “Harnessing Conflict in Foreign Policy Making: From Devil's to Multiple Advocacy.” Presidential Studies Quarterly  32: 484– 508. Google Scholar CrossRef Search ADS   Hafner-Burton Emilie M., Kahler Miles, Montgomery Alexander H.. 2009. “Network Analysis for International Relations.” International Organization  63: 559– 92. Google Scholar CrossRef Search ADS   Hajer Maarten A. 1993. “Discourse Coalitions and the Institutionalization of Practice: The Case of Acid Rain in Britain.” In The Argumentative Turn in Policy Analysis and Planning , edited by Fisher Frank, Forester John. Durham, NC and London: Duke University Press. Google Scholar CrossRef Search ADS   Hajer Maarten A. 1997. The Politics of Environmental Discourse: Ecological Modernization and the Policy Process . Oxford: Oxford University Press. Google Scholar CrossRef Search ADS   Halperin Morton H., Clapp Priscilla A., Kanter Arnold. 2006. Bureaucratic Politics and Foreign Policy . 2nd ed. Washington, DC: Brookings Institution Press. Hart Paul ’t. 1991. “Irving L. Janis’ Victims of Groupthink.” Political Psychology  12: 247– 78. Google Scholar CrossRef Search ADS   Hart Paul ’t. 1994. Groupthink in Government: A Study of Small Groups and Policy Failure . Baltimore, MD: The Johns Hopkins University Press. Hart Paul ’t, Stern Eric K., Sundelius Bengt. 1997a. “Foreign Policymaking at the Top: Political Group Dynamics.” In Beyond Groupthink: Political Group Dynamics and Foreign Policy-Making , edited by Hart Paul ’t, Stern Eric K., Sundelius Bengt. Ann Arbor, MI: The University of Michigan Press. Hart Paul ’t, Stern Eric K., Sundelius Bengt, eds. 1997b. Beyond Groupthink: Political Group Dynamics and Foreign Policy-making . Ann Arbor, MI: The University of Michigan Press. Hersh Seymour M. 2005. Chain of Command: The Road from 9/11 to Abu Ghraib . New York, NY: Harper. Holzscheiter Anna. 2014. “Between Communicative Interaction and Structures of Signification: Discourse Theory and Analysis in International Relations.” International Studies Perspectives  15: 142– 62. Google Scholar CrossRef Search ADS   Houghton David Patrick. 2008. “Invading and Occupying Iraq: Some Insights From Political Psychology.” Peace and Conflict  14: 169– 92. Google Scholar CrossRef Search ADS   Hudson Valerie M. 2007. Foreign Policy Analysis: Classic and Contemporary Theory . Lanham: Rowman & Littlefield. Isikoff Michael, Corn David. 2007. Hubris: The Inside Story of Spin, Scandal, and the Selling of the Iraq War . New York, NY: Three Rivers Press. Janis Irving L. 1982. Groupthink: Psychological Studies of Policy Decisions and Fiascoes . 2nd ed. Boston, MA: Wadsworth. Jervis Robert. 2006. “Reports, Politics, and Intelligence Failures: The Case of Iraq.” The Journal of Strategic Studies  29: 3– 52. Google Scholar CrossRef Search ADS   Kaarbo Juliet. 2008. “Coalition Cabinet Decision Making: Institutional and Psychological Factors.” International Studies Review  10: 57– 86. Google Scholar CrossRef Search ADS   Kaarbo Juliet, Beasley Ryan K.. 2008. “Taking It to the Extreme: The Effect of Coalition Cabinets on Foreign Policy.” Foreign Policy Analysis  4: 67– 81. Google Scholar CrossRef Search ADS   Kadushin Charles. 2012. Understanding Social Networks: Theories, Concepts, and Findings . Oxford - New York, NY: Oxford University Press. Kaplan Fred. 2008. Daydream Believers: How a Few Grand Ideas Wrecked American Power . Hoboken, NJ: John Wiley & Sons. Kassop Nancy. 2003. “The War Power and Its Limits.” Presidential Studies Quarterly  33: 509– 29. Google Scholar CrossRef Search ADS   Kaufmann Chaim. 2004. “Threat Inflation and the Failure of the Marketplace of Ideas: The Selling of the Iraq War.” International Security  29: 5– 48. Google Scholar CrossRef Search ADS   Lazer David. 2011. “Networks in Political Science: Back to the Future.” PS: Political Science & Politics  44: 61– 68. Google Scholar CrossRef Search ADS   Lee Su-Mi. 2014. “Understanding the Yalta Axioms and Riga Axioms through the Belief Systems of the Advocacy Coalition Framework.” Foreign Policy Analysis  DOI: 10.1111/fpa.12040:1-21. Leifeld Philip. 2011. Discourse Network Analyzer Manual . Technical report EAWAG Konstanz: URL: http://www.philipleifeld.de/cms/upload/Downloads/dna-manual1.30.pdf. Lind Michael. 2012. “Neoconservativism and American Hegemony.” In Why Did the United States Invade Iraq?  edited by Cramer Jane K., Thrall A. Trevor. Routledge Global Security Studies London - New York, NY: Routledge. Mann James. 2004. Rise of the Vulcans: The History of the Bush's War Cabinet . New York, NY: Penguin Books. Mazaar Michael J. 2007. “The Iraq War and Agenda Setting.” Foreign Policy Analysis  3: 1– 23. Google Scholar CrossRef Search ADS   McCauley Clark. 1979. “The Nature of Social Influence in Groupthink: Compliance and Internalization.” Journal of Personality and Social Psychology  57: 250– 60. Google Scholar CrossRef Search ADS   McClurg Scott D., Young Joseph K.. 2011. “Political Networks: Editor's Indroduction - A Relational Political Science.” PS: Political Science & Politics  44: 39– 43. Google Scholar CrossRef Search ADS   Mearsheimer John J., Walt Stephen M.. 2007. The Isreal Lobby and U.S. Foreign Policy . New York, NY: Farrar, Straus and Giroux. Metselaar Max V., Verbeek Bertjan. 1997. “Beyond Decision Making in Formal and Informal Groups: The Dutch Cabinet and the West New Guinea Conflict.” In Beyond Groupthink: Political Group Dynamics and Foreign Policy-making , edited by Hart Paul ’t, Stern Eric K., Sundelius Bengt. Ann Arbor, MI: The University of Michigan Press. Mitchell David, Massoud Tansa George. 2009. “Anatomy of Failure: Bush's Decision-Making Process and the Iraq War.” Foreign Policy Analysis  5: 265– 86. Google Scholar CrossRef Search ADS   Mulcahy Kevin V. 1995. “Rethinking Groupthink: Walt Rostow and the National Security Advisory Process in the Johnson Administration.” Presidential Studies Quarterly  25: 237– 50. Newman M. E. J. 2012. Networks: An Introduction . Oxford - New York, NY: Oxford University Press. Packer George. 2007. The Assassin's Gate: America in Iraq . London: Faber & Faber. Pfiffner James P. 2004. “Did President Bush Mislead the Country in His Arguments for War with Iraq?” Presidential Studies Quarterly  34: 25– 46. Google Scholar CrossRef Search ADS   Pfiffner James P. 2009. “Decision Making in the Bush White House.” Presidential Studies Quarterly  39: 363– 84. Google Scholar CrossRef Search ADS   Pillar Paul R. 2011. Intelligence and U.S. Foreign Policy: Iraq, 9/11, and Misguided Reform . New York, NY: Columbia University Press. Powell Colin, Koltz Tony. 2012. It Worked for Me: In Life and Leadership . New York, NY:   Harper. Renshon Jonathan. 2009. “When Public Statements Reveal Private Beliefs: Assessing Operational Code at a Distance.” Political Psychology  30: 649– 61. Google Scholar CrossRef Search ADS   Rice Condoleezza. 2011. No Higher Honor: A Memoir of My Years in Washington . New York, NY: Crown. Rumsfeld Donald. 2011. Known and Unknown: A Memoir . London: Sentinel. Sabatier Paul A. 1988. “An Advocacy Coalition Framework of Policy Change and the Role of Policy-Oriented Learning Therein.” Policy Sciences  21: 129– 68. Google Scholar CrossRef Search ADS   Sabatier Paul A., Weible Christopher M.. 2007. “The Advocacy Coalition Framework.” In Theories of the Policy Process , edited by Sabatier Paul A.. Cambridge, MA: Westview Press. Schafer Mark, Crichlow Scott. 1996. “Antecedents of Groupthink: A Quantitative Study.” Journal of Conflict Resolution  40: 415– 35. Google Scholar CrossRef Search ADS   Schafer Mark, Crichlow Scott. 2002. “The Process-Outcome Connection in Foreign Policy Decision Making: A Quantitative Study Building on Groupthink.” International Studies Quarterly  46: 45– 68. Google Scholar CrossRef Search ADS   Schafer Mark, Crichlow Scott. 2010. Groupthink Versus High-Quality Decision Making in International Relations . New York, NY: Columbia University Press. Schimmelfennig Frank. 2001. “The Community Trap: Liberal Norms, Rhetorical Action, and the Eastern Enlargement of the European Union.” International Organization  55: 47– 80. Google Scholar CrossRef Search ADS   Shannon Vaughn P., Keller Jonathan W.. 2007. “Leadership Style and International Norm Violation: The Case of the Iraq War.” Foreign Policy Analysis  3: 79– 104. Google Scholar CrossRef Search ADS   Simon Herbert A. 1985. “Human Nature in Politics: The Dialogue of Psychology with Political Science.” The American Political Science Review  79: 293– 304. Google Scholar CrossRef Search ADS   Smith Steve. 1984. “Groupthink and the Hostage Rescue Mission.” British Journal of Political Science  15: 117– 26. Google Scholar CrossRef Search ADS   Snyder Richard C., Bruck H. W., Sapin Burton M., Hudson Valerie M., Chollet Derek H., Goldgeier James M.. 2002. Foreign Policy Decision-Making (Revisited) . New York, NY: Palgrave Macmillan. Google Scholar CrossRef Search ADS   Stern Eric K. 1997. “Probing the Plausibility of Newgroup Syndrome: Kennedy and the Bay of Pigs.” In Beyond Groupthink: Political Group Dynamics and Foreign Policy-making , edited by Hart Paul ’t, Stern Eric K., Sundelius Bengt. Ann Arbor, MI: The University of Michigan Press. Stern Eric K., Sundelius Bengt. 1997. “Understanding Small Group Decisions in Foreign Policy: Process Diagnosis and Research Procedure.” In Beyond Groupthink: Political Group Dynamics and Foreign Policymaking , edited by Hart Paul ’t, Stern Eric K., Sundelius Bengt. Ann Arbor, MI: The University of Michigan Press. Sunstein Cass R., Hastie Reid. 2015. Wiser: Getting Beyond Groupthink to Make Groups Smarter . Boston, MA: Harvard Business Review Press. Suskind Ron. 2004. The Price of Loyalty: George W. Bush, the White House, and the Education of Paul O'Neill.  New York, NY: Simon & Schuster Paperbacks. Suskind Ron. 2007. The One Percent Doctrine: Deep Inside America's Pursuit of Its Enemies Since 9/11 . New York, NY: Simon & Schuster Paperbacks. Tenet George. 2008. At the Center of the Storm: The CIA during America's Time of Crises . New York, NY: Harper. Tetlock Philip E. 1979. “Identifying Victims of Groupthink From Public Statements to Decision Makers.” Journal of Personality and Social Psychology  37: 1314– 24. Google Scholar CrossRef Search ADS   Tetlock Philip E., Peterson Randall S., McGuire Charles, Chang Shi-jie, Feld Peter. 1992. “Assessing Political Group Dynamics: A Test of the Groupthink Model.” Journal of Personality and Social Psychology  63: 403– 52. Google Scholar CrossRef Search ADS   Vanderbush Walt. 2009. “Exiles and the Marketing of U.S. Policy toward Cuba and Iraq.” Foreign Policy Analysis  5: 287– 306. Google Scholar CrossRef Search ADS   Vertzberger Yaacov Y. I. 1997. “Collective Risk Taking: The Decision-making Group.” In Beyond Groupthink: Political Group Dynamics and Foreign Policy-making , edited by Hart Paul ’t, Stern Eric K., Sundelius Bengt. Ann Arbor, MI: The University of Michigan Press. Woodward Bob. 2004. Plan of Attack . New York, NY: Simon & Schuster. Yetiv Steven A. 2003. “Groupthink and the Gulf Crisis.” British Journal of Political Science  33: 419– 42. Google Scholar CrossRef Search ADS   Yetiv Steven A. 2011. Explaining Foreign Policy: U.S. Decision-Making and the Persian Gulf Wars . 2nd ed. Baltimore, MD: The Johns Hopkins University Press. © The Author (2017). Published by Oxford University Press on behalf of the International Studies Association. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Foreign Policy Analysis Oxford University Press

Making Concurrence-Seeking Visible: Groupthink, Discourse Networks, and the 2003 Iraq War

Foreign Policy Analysis , Volume Advance Article – Oct 11, 2017

Loading next page...
 
/lp/ou_press/making-concurrence-seeking-visible-groupthink-discourse-networks-and-w0AKk5Xnpf
Publisher
Wiley-Blackwell
Copyright
© The Author (2017). Published by Oxford University Press on behalf of the International Studies Association. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1743-8586
eISSN
1743-8594
D.O.I.
10.1093/fpa/orx009
Publisher site
See Article on Publisher Site

Abstract

Abstract Groupthink is one of the most popular and extensively discussed approaches in studying small group decision-making. However, the methodological question of how to make concurrence-seeking, as the key element of groupthink, visible has received insufficient attention. To make group decision behind closed doors more visible and methodologically advance groupthink theory, I introduce Discourse Network Analysis and apply it to the US invasion of Iraq in 2003. Results demonstrate that the US decision to go to war was based on a sudden and undisputed threat-perception that exhibited hasty concurrence-seeking. Given this threat-perception, and a fierce struggle between advocates of a diplomatic solution and supporters of a military response, the United States chose a military strategy—one that inevitably resulted in war. Since Snyder, Bruck, and Sapin published their landmark study Foreign Policy Decision-Making, the analysis of groups has become one of the most prominent and vibrant research traditions in Foreign Policy Analysis (FPA) scholarship (Hudson 2007, 17–20). Some experts within this tradition focus on bureaucratic (Allison and Zelikow 1999) or organizational routines and cultures (Halperin, Clapp, and Kanter 2006) to explain foreign-policy outcomes, while others adopt small groups as the decisive unit of analysis (Tetlock et al. 1992; Hart, Stern, and Sundelius 1997a). The path-breaking contribution to this literature of small group decision-making is Irving L. Janis’ (1982) book Groupthink: Psychological Studies of Policy and Fiascoes. It has become a standard reference in FPA (e.g., Hart, Stern, and Sundelius 1997b) and the analytical framework for numerous empirical studies (e.g., Smith 1984; Mulcahy 1995; Metselaar and Verbeek 1997; Yetiv 2003; Badie 2010). While scholars discussed and enhanced several elements of Janis's concept—for example, the role of mindguards (Fuller and Alday 1997), antecedent conditions of (Schafer and Crichlow 1996; Baron 2005) and collective risk-taking in a groupthink context (Vertzberger 1997), and the newgroup syndrome (Stern 1997)—methodological questions of groupthink still remain underexplored. Janis himself passed over this question by noting: “For purpose of hypothesis construction (. . .) we must be willing to make some inferential leaps from whatever historical clues we can pick up” (Janis 1982, ix). Tetlock (1979, 1316) first tried to fill this gap. He suggests that Janis’ case study design might have “given too much weight to evidence that supported groupthink hypothesis and too little weight to contradictory evidence.” Using content analysis, Tetlock demonstrates that public statements of decision-makers in groupthink situations “were characterized by significantly lower levels of integrative complexity” (Tetlock 1979, 1322). Other scholars followed Tetlock by linking signs of groupthink to individual perceptions and characteristics, as well as leadership-styles measured by public statements (e.g., Stern and Sundelius 1997; Schafer and Crichlow 2002, 2010). Although these studies have taught us much about the likely relationship between individual characteristics and groupthink, one central question is still unanswered: How can researchers operationalize deliberations in a group context? Or more specifically, how can researchers make concurrence-seeking in groups (the core concept of groupthink) visible? As George (1997, 39) argues, the lack of a satisfactory answer to this core question is the biggest gap in groupthink theory. Up to now, hasty concurrence-seeking and hence groupthink cannot be observed directly but must either be inferred by the presence of antecedent conditions and/or the final symptoms. This paper introduces Discourse Network Analysis (DNA) as a method for visualizing nonpublic decision-making in group settings and hence making concurrence-seeking visible. DNA rests on the assumption that group decision-making involves actors competing for discursive dominance. DNA suggests that by analyzing the discourse of actors in a decision-making process, researchers can cluster actors into coalitions. One can then trace the phases of these coalition-building-processes and see when and how a winning coalition developed. As I argue in this article, understanding decision-making processes as coalitional struggles for discursive dominance advances the groupthink approach to FPA because it makes the discourses that are the foundation of decision-making processes visible. Moreover, it allows scholars to broaden sources of information, thereby integrating verifiable and replicable data into their research. I demonstrate the usefulness of DNA with the case of the Iraq invasion by the United States in 2003. Since this invasion, a great deal of investigative journalism and scholarship has focused on the questions of why the administration of George W. Bush decided to topple Saddam Hussein and occupy Iraq, who the driving forces behind this decision were, and when this decision was made. The elite of US investigative journalists were the first to address these questions by blaming neoconservative ideology (Suskind 2004; Packer 2007; Bamford 2005; Kaplan 2008), the terrorist attacks of September 11, 2001, and the threat of nuclear-armed terrorists and their state supporters (Woodward 2004; Suskind 2007), or a defective decision-making process within the administration (Hersh 2005; Isikoff and Corn 2007). Moving beyond descriptive (and often speculative) accounts of the invasion, political scientists began to systematically analyze the decision-making process and provide answers to the question of the actual motives underlying the invasion. According to these studies, individual characteristics and leadership styles (Shannon and Keller 2007; Dyson 2009), ideas and ideologies (Flibbert 2006; Burgos 2008; Lind 2012), pressure groups (Mazaar 2007; Mearsheimer and Walt 2007; Vanderbush 2009), or misperceptions and a defective decision-making process (Yetiv 2011; Burke 2005; Fitzgerald and Lebow 2006; Jervis 2006; Houghton 2008; Mitchell and Massoud 2009; Badie 2010) are proposed as explanatory variables for the war. Like investigative journalism, these studies view the hawkish group around Cheney and Rumsfeld as the pivotal actors behind the decision. Nevertheless, both investigative journalists and political scientists disagree as to the specific moment when the decision was made. Arguments range from pre-September 11 or September 11 itself, or from spring or fall of 2002, to the claim that the decision process was so incremental as to lack a clear starting date. Although journalists and scholars have advanced our understanding of the decision-making process, their analyses face the problem of sources as a serious methodological shortcoming. While a few studies on decision-making in course of the Iraq invasion have used verifiable and replicable data (Dyson 2006, 2009; Shannon and Keller 2007), most have not. The journalists’ descriptions of the decision-making dynamics primarily rest on whistleblower accounts and hearsay, sources that are neither verifiable nor replicable, with only the individual journalist's reputation giving weight to the message. Consequently, scholarly analyses that base their conclusions on these sources do not fulfill strict methodological standards and therefore run the risk of building their explanations on shifting sands. I argue that DNA helps counteract this problem and provides researchers with tools for making concurrence-seeking visible, based on verifiable and replicable data. More specifically, this article posits two core arguments concerning the phases of the decision-making-process in the run-up to the 2003 invasion of Iraq. First, speeches, interviews, and autobiographies contribute to a significantly better understanding and deciphering of the decision-making process than investigative journalism and are a necessary corrective to it. Second, and similar to the findings of McCauley (1979) and Stern and Sundelius (1997), the decision-making process must be understood as a two-phased process, first, building a consensus about the nature of the problem and, second, agreeing on what to do about it. In the first phase, President Bush and his inner circle built the decision to invade Iraq on a homogeneous and undisputed threat-perception induced by the terrorist attacks of September 11, 2001. This phase featured three signs of groupthink (Janis 1982): a cohesive decision group, a lack of norms and impartial leadership to guarantee an unbiased discussion, and a provocative situational context. In the second phase, given this perception and a dysfunctional group structure, a military policy discourse incrementally replaced the diplomacy-favoring discourse in 2002, a process that inevitably led to war. My analysis thus adds to the discussion of groupthink as an explanation for the Iraq invasion 2003 and puts the argument of Badie (2010), that the shift in the administration's view of Saddam Hussein as an existential threat happened gradually over periods of months until the invasion, into perspective. The threat perception changed immediately after the terrorist attacks of September 11, 2001. The struggle between Bush and his key advisers in the following months, therefore, was not so much about the question of how dangerous Saddam Hussein might be but about how to tackle this threat effectively. This result has consequences for the analysis of decision-making processes in general and groupthink in particular. By splitting a foreign-policy decision-process into two phases (i.e., framing the threat and answering the threat), the shifts within each phase become more detailed. Hence, and because of the path-dependency of the problem-definition and the solution phase, groupthink may be present in the former but absent in the latter, an insight that can be lost by mixing both stages. This article is organized as follows: First, I outline the role of small groups in foreign-policy decision-making and argue that under certain conditions, these groups are decisive for foreign-policy decision-making. Second, I introduce the theoretical and methodological foundations of DNA and outline the research design. Third, I show how the terrorist attacks of September 11, 2001, led to a sudden, undisputed, and homogeneous threat perception. Fourth, I explain how this homogeneous threat perception incrementally transformed the heterogeneous policy discourse into a homogeneous one in the third quarter of 2002, so that the focus shifted primarily to military instead of nonmilitary policy options. Small Groups in Foreign-Policy Decision-Making Determining the principal unit of analysis for explaining foreign-policy decision-making is one of the core topics in foreign-policy analysis. One tradition of classic FPA scholarship seeks to explain a state's external behavior by examining the internal dynamics of groups and organizations. According to this approach, foreign-policy outcomes are explained best by the social context of groups and institutions in which these individuals are embedded, that is, “the definition of the situation” (Snyder et al. 2002, 76). Following this line of reasoning, foreign-policy outcomes can be explained through an understanding of this context and its distinctive processes and dynamics. Although Hart (1994, 132) is skeptical of the role of small groups in foreign-policy outcomes and ascribes more importance to bureaucracies (Allison and Zelikow 1999) or organizational routines and cultures (Halperin, Clapp, and Kanter 2006), Tetlock et al. (1992) conceive small groups as the decisive unit of analysis. As Hart, Stern, and Sundelius (1997a, 8) argue, small group analysis helps us to explain “the crucial, formative policy decisions that shape significant parts of a country's foreign policy.” Hence, small groups as collective social forms are crucial agents of decision-making. Beside the groups’ function as the central locus of policy making, they have the advantage of being able to correct or minimize individual shortcomings. People are not able to act fully rational. They are rather “bounded” by their inability to gather all the information necessary for decision-making and process it adequately in a short period of time (Simon 1985). Furthermore, people exhibit all sorts of cognitive deficiencies that impair their perceptions of a situation and rational responses to it. In a group context, the shortcomings of one individual in a certain area can be compensated for by others. However, as psychological studies have revealed, “groups [too often] actually amplify those mistakes [and] turn out to be even worse than individuals” (Sunstein and Hastie 2015, 2) when it comes to rational decision-making. This observation is especially true when groups act rapidly, emotionally, and intuitively. There are several reasons why groups fail to compensate for the shortcomings of their members. Actors within groups may fail to speak up because more powerful and more competent actors have a different opinion or because they are socially pressured to self-censorship (Sunstein and Hastie 2015, 22–23). In addition, “unreasonably strong group and team members who might capitalize on the situation to force his or her ideas on the group” (Franz 2012, 3) can undermine team effectiveness. Furthermore, group deliberations “usually reduce . . . variance” (Sunstein and Hastie 2015, 28) and reinforce uniformity. This last insight is particularly striking in the effect of “polarization.” Depending on the group's initial predisposition, group decisions are more extreme (either in a more cautious or more risky direction) after the group discussion (Franz 2012, 163 and Sunstein and Hastie 2015, 78–85). As Kaarbo (2008) and Kaarbo and Beasley (2008) argue, polarization often takes place in cabinet coalitions. Minorities within such coalitions can manipulate the decision-making process and push the group towards a more extreme option. This push towards the extreme can be explained by either the “political hijacking of the junior partner,” the “domestic political weakness” of the coalition itself, the “diffusion of accountability” in such groups, or classical “logrolling” (Beasley and Kaarbo 2014, 730). This insight not only applies to coalition cabinets, where a consensus of the parties involved is essential for policy making, but also to presidential systems, like in the United States, where minority views within the administration can push a president towards an extreme decision, such as going to war (Elman 2000, 97). Vertzberger (1997, 284) attributes this polarization effect in the US presidential system to the self-confidence and assertiveness of certain members of the administration's inner circle. These members “with more radically polarized judgements and preferences invest more resources in attempts to exert influence” than others. In summary, groups do not necessarily make better decisions than individuals. Groups have their own shortcomings, which stem from their structures and the dynamics between members. A path-breaking contribution to the field of small group decision-making is Irving L. Janis’s book Groupthink: Psychological Studies of Policy Decisions and Fiascoes (1972/1982). With its focus on decision-making processes in small groups and on the reasons for suboptimal policy outcomes, this book has become a standard reference in foreign-policy analysis and the springboard for a lively discussion on small groups in foreign-policy decision-making (e.g., Smith 1984; Mulcahy 1995; Stern and Sundelius 1997; Stern 1997; Schafer and Crichlow 2002; Yetiv 2003). According to Janis (1982, 9), small-group decision-making is often characterized by “a deterioration of mental efficiency, reality testing, and moral judgment that results from in-group pressure.” In essence, “groupthink” discusses how groups reach a consensus and whether it is the result of a thoughtful deliberation process. Janis focuses on group processes in which there is an emphasis on concurrence-seeking. This is especially valuable to groups during periods of stress when there is “a heightened need for affiliation with other members” (George 1997, 37). As Hart (1991, 247) points out, groupthink may result in defective decision-making processes in which the consensus is a faulty one because of an “excessive form of concurrence-seeking” where “group members have come to value the group (and their being part of it) higher than anything else.” Groupthink occurs when four antecedent conditions prevail. First, when a group is cohesive. According to Franz (2012, 177, 180–81 and 187), cohesion occurs either when members feel good about each other (social cohesion) or when they are committed to a goal (task cohesion). Cohesion can have a positive effect on decision-making, but too much cohesiveness can also be problematic. Highly cohesive groups tend to become more polarized and hence more prone to making risky decisions. Second, the structure of the group is another possible antecedent condition. Groups that isolate themselves from the outside world and prevent alternative views from being offered, risk a suboptimal decision process. Third, groupthink is more likely when the leader is not impartial and hence too directive and does not promote divergent opinions. When leaders do not feel restricted by group norms (that structure the deliberation process and give divergent opinions the chance to be discussed) other group members will stay silent. Such partial leaders push the group too early into a certain direction and thereby risk ignoring viable alternatives (Yetiv 2003, 423–24 and 431). As George and Stern (2002) stress, groups should, therefore, seek to promote a system of multiple advocacy, “in which competition and disagreement among different participants is structured and managed to achieve the benefits of diverse points of view” and where the deliberation process is organized by an honest broker who has no hidden agenda and guarantees impartiality. Finally, concurrence-seeking tendencies become more likely when the group finds itself in a situation characterized by high stress, caused either by an external threat or a group's lack of self-esteem due to recent failures. This tendency leads to premature consensus-building in which the group neglects to assess various policy options. Groupthink therefore finds its expression in the overestimation of the group, a closed-mindedness among its members, and a pressure towards uniformity through self-censorship, the illusion of unanimity, or silencing dissenters. Given these symptoms, the probability increases that the group exhibits an incomplete survey of objectives and alternatives, fails to reappraise initially rejected alternatives, and conducts a poor and biased information search. Although the presence of some of these behaviors does not automatically induce groupthink, the chances of a defective decision-making process increase the more that these symptoms and consequences appear (Janis 1982, 245). In sum, groupthink-theory places special emphasis on the process of consensus building in groups by highlighting the concurrence-seeking effort of its members. However, as Hart (1994, 7) emphasizes, concurrence-seeking per se is not a problem but rather a necessary precondition for decision-making. Groups that are not able to reach a timely consensus also fail to produce good policy outcomes, such as groups that reach a consensus prematurely, neglecting to assess all its options. The central question, therefore, is how a group reaches a consensus and whether or not the group discussed all possible options extensively in a fair, unbiased, and inclusive deliberative process. One of the biggest challenges for Janis's concept is that of causation. As Baron (2005) correctly notes, the symptoms of groupthink do not necessarily follow from concurrence-seeking, and concurrence-seeking itself may not be the result of antecedent conditions. Hence, the central question for researchers must be how to make concurrence-seeking visible. George (1997, 39) points to this weakness of groupthink theory. He argues that concurrence-seeking and hence groupthink cannot be observed directly and must be inferred by either the presence of antecedent conditions and/or the final symptoms. This problem is not limited to groupthink theory. As Bennett and Checkel (2015, 11) note, it concerns other theories as well: “[w]e do not get to observer causality—we make inferences about it.” Especially in small groups like presidential cabinets, researchers are not able to directly observe group deliberations and hence to trace causal mechanisms. A solution to this problem is the use of process tracing as “the central within-case-method” (Bennett and Checkel 2015, 4) in single or small N studies, to “trace the links between possible causes and observed outcomes” (George and Bennett 2005, 6). Process tracing analyzes interviews and documents and, by doing so, seeks to capture causal mechanisms. Whereas alternative sources like oral history or declassified archival materials are often available for earlier cases, researchers of recent cases are mainly left with investigative journalism. Yet, this alternative lacks scientific reliability and replicability. Thus, in the following section, I will outline DNA and argue that this concept is a promising way to make the concurrence-seeking efforts of groups visible and comprehensible and to trace the causal mechanisms of decision-making in groups. Discourse Networks Analysis and the Role of Advocacy Coalitions Discourse Network Analysis rests upon three theoretical foundations—the Advocacy Coalition Framework (ACF), Discourse Coalition Theory (DCT), and Social Network Analysis (SNA). ACF has its origins in public-policy analysis and seeks to explain policy change and its drivers in the course of time. Following Paul A. Sabatier (1988, 130), ACF primarily focuses on the “interaction of political elites within a policy community/subsystem attempting to respond to changing socioeconomic and political conditions.” ACF assumes that policies only change when a “winning-coalition” of likeminded proponents succeeds over its competitors. Hence, political elites must be successful in allying with like-minded actors, develop common strategies, and convince a majority within the policy community of the advantages of their approach. The more actors a coalition unites and the more homogeneous they are (in terms of similar core beliefs and/or common strategies), the higher the probability that they will succeed and become the winning coalition (Sabatier and Weible 2007, 196). Although Sabatier developed the ACF for analyzing larger policy communities comprising a variety of different organizations and actors and observed them over a longer time span, ACF's assumptions are also applicable to the realm of foreign-policy analysis and smaller communities, as Lee (2014) has shown. Applying ACF to President Bush and his advisors and their decision-making process on the Iraq invasion in 2003 is justifiable for two reasons. First, after the terrorist attacks of 9/11, foreign-policy decision-making became increasingly centralized in the administration and was dominated by a few, very powerful actors (Kassop 2003). Second, although in the US presidential system consensus is neither necessary (because of presidential prerogatives in foreign and security policy) nor desirable (George and Stern 2002), presidents nevertheless draw on group deliberations in their inner circles when they make decisions (Vertzberger 1997, 284; Elman 2000, 97). President Bush as commander-in-chief unambiguously emphasized that although the decision to go to war was his own, he nevertheless based this decision on a “vigorous debate” and listened to the “discussion and divergent views” (Bush 2010, 189–90) of his inner circle (Mitchell and Massoud 2009, 272–73; Rice 2011, 86–87; Rumsfeld 2011, 319). Consequently, ACF is a suitable concept to explain which actors and which coalitions of actors within the inner circle of the administration were the crucial ones and how the struggle between different coalitions evolved over time, leading up to the war decision. The second concept on which DNA builds is Maarten A. Hajer's (1993; 1997) Discourse Coalition Theory. Hajer (1993, 44) believes that political problems are socially constructed by those who are able to influence the policy process. This process of construction, in turn, is revealed by analyzing the public discourse of the actors involved. Renshon (2009) supports this claim when he argues that public statements reveal private beliefs and are not an attempt to distract from the actual agenda. These statements provide insight into the private process of policy building and are, therefore, especially valuable when the private transcripts are unknown or unknowable. This claim is further underlined in the case of the Iraq War by the effort of Vice-President Cheney and his followers to plant supportive evidence of their views in the stories of reporters, which then entered the public record and were again used as an evidence of “independent” verification of these views. Hence, private beliefs are recoverable as a subtext in the public records through discourse analysis. But even if one is skeptical about the transparency of public statements, public statements constrain an actor's behavior in the medium to long-term (Schimmelfennig 2001). So it does not matter whether actors only comply with the official lingo or if they have also internalized the message (according to McCauley (1979), a necessary requirement for groupthink). These individual discourses can then be aggregated to form discourse coalitions, which are defined as groups of actors that use similar discourses towards a policy problem. Discourse Coalition Theory operationalizes these public discourses and can be used to locate actors and their views towards a policy problem within a group. The third and final concept on which DNA rests, is Social Network Analysis (SNA). In the last couple of years, SNA has gained prominence in the social sciences (Borgatti et al. 2009) and especially in political science (Lazer 2011). This is due to the fact that core political science concepts, such as power, agency, and structure, must be understood as relational phenomena. As McClurg and Young (2011, 41) write: Only by focusing clearly on the degree to which relations constrain agency can we begin to fully explain our depictions of what the concept of liberty, freedom, and agency mean in different political systems and contexts. SNA, therefore, enables researchers to put agency in context and to show how the context and its structures influence actors and their behavior (Hafner-Burton, Kahler, and Montgomery 2009). In the case of this article, SNA is particularly valuable in explaining the different phases of group deliberations and the role of certain actors in these phases. In this article, I apply three distinctive network metrics, which help me to explain which actors are responsible for the formation or breakup of coalitions, to understand the degree of cohesiveness within groups and coalitions, and to display coalition building processes as such. These metrics are centrality, density and modularity. First, centrality reveals how important a certain actor is within a network. It does so, by calculating the sum of edges (direct relations) of a certain node (the actor) in comparison to all the other nodes and their sums of edges. A special form of this degree centrality is the betweenness centrality, which ascribes importance to a node if it acts as a bridge between other nodes. Betweenness centrality, therefore, allows me to highlight actors with the most crucial relations and consequently the most influence within a network. These actors are responsible for the evolution of a coalition and can be understood as brokers. Second, density describes the ratio between the actual and the possible number of edges between nodes in a network (Kadushin 2012, 29): the higher the value (between 0 and 1), the higher the homogeneity of a network or a subgroup within a network. As Erisen and Erisen (2012) demonstrate in their experimental study, density and thus homogeneity are indicators for the likelihood of the success or failure of decision processes. The higher the density and homogeneity within a group (or subgroup), the more biased a decision process becomes and the more likely this process will end up in a suboptimal result. Consequently, this network metric perfectly fits to groupthink and its argument that hasty concurrence-seeking at an early stage and a too homogeneous deliberation process can threaten results. Finally, modularity uses density to compare the number of actual edges between nodes in a network with the possible number of edges in comparable random networks. By doing so, modularity allows us to divide networks into subgroups or, in our case, into coalitions (Kadushin 2012, 48; Newman 2012, 224). To sum up, Discourse Network Analysis makes three major contributions to the analysis of group decision-making. First, it conceives the building of a winning-coalition between actors in a policy process as a necessary precondition for the change of policies (ACF part of the approach). Second, it identifies these coalitions by operationalizing the discourse1 of their members (DCT part). And finally, it uses social network analysis to display these actors via their discourses and show how this network evolved and changed over time and who the crucial actors in this process were (SNA part). By doing so, DNA captures the causal mechanisms of decision-making and provides answers to the questions of whether and when a group came to a consensus and who the driving forces behind this consensus were. Consequently, DNA helps us to understand whether or not this concurrence-seeking process was premature and hasty. Research Design To construct discourse networks and to calculate the network metrics, I compiled a dataset consisting of 227 interviews (spontaneous speech-acts) and speeches (prepared speech-acts) of the inner circle of decision-makers (i.e., Bush, Cheney, Powell, Rumsfeld, Rice, Tenet, Armitage, and Wolfowitz) in the period from January 2001 through December 2004. Whenever possible, one interview and one speech per actor and quarter were chosen to ensure a timely representation of discourses, on the one hand, and to control for impression-management (by combining spontaneous with prepared speech-acts), on the other hand. From the total list of all possible speeches and interviews in the sample period, I chose only those in which the keywords “Iraq” and “Hussein” appeared most often. I coded the data by using the Discourse Network Analyzer, a program developed by Leifeld (2011) to compile discourse networks. Statements were coded in four steps: (1) actor issuing the statement, (2) organization the actor belongs to, (3) category the actor is referring to, and (4) approval (yes) or rejection (no) of the category by the actor. So, for example, I coded the statement of President Bush “Iraq continues to flaunt its hostility and to support terror” in his State of the Union address of 29 January 2002 as follows: (1) Bush, (2) White House, (3) support of terrorism, (4) yes. Paragraphs were seen as complete units of meaning. Therefore, statements were coded only once within a paragraph (as long as they were not contradictory). I prepared the twenty-nine categories of the coding scheme inductively on the basis of a small pilot study and organized these categories into two groups representing two subdiscourses of the overall discourse: (1) US threat perception and (2) US policy options toward Iraq. I used this dataset to construct both affiliation and congruence networks. The former displays actors sharing categories; the latter only displays actors, showing how close these actors are to each other based on the number of shared categories. Thus, affiliation networks are used to display the dominant categories of a discourse in a certain period of time and to show who referred to these categories; whereas congruence networks reveal which actors belong together in certain coalitions. 9/11 and the Undisputed Threat Perception I conceive of the decision-making process in the run-up to the Iraq war not as a linear process but as a more complex one (George and Bennett 2005, 212) of two interrelated decisions (i.e., first, framing the threat and then answering it). I apply Janis’s concept of small group decision-making to analyze this process. This requires me to clarify the antecedent conditions under which President Bush and his inner circle made decisions on Iraq. I will do so by analyzing the cohesiveness of the decision group and its structure and by discussing the situational context in which the group operated. Given these antecedent conditions, I will demonstrate how the terrorist attacks of September 11, 2001, instantly homogenized the threat perception of the actors involved, providing the basis for the subsequent policy discourse. The inner circle of advisers of President Bush was an experienced group of actors when it came to foreign and security policy, and it was highly cohesive. The group members knew each other from the early 1960s on, working in the Nixon and Ford administrations (Cheney and Rumsfeld); with Reagan (Cheney, Rumsfeld, Powell, Armitage, and Wolfowitz); and under G. H. W. Bush (Cheney, Powell, Rice, Armitage, and Wolfowitz). During these years, they were not only able to build intense personal relationships (Rumsfeld and Cheney or Powell and Armitage), but they even had to deal with Iraq during the first Iraq war in 1990–91. President Bush and Tenet (Director of Central Intelligence) were the only ones who did not have a record of cooperation in the past. Nevertheless, these actors built one of the most experienced, well known and tightly knit foreign and security policy teams in U.S. history, when the Bush administration began to work in January 2001 (Mann 2004). In addition to this social cohesiveness, the group also exhibited a very high level of task cohesiveness, induced by the provocative situational context of the terrorist attacks of September 11. Keeping the United States safe and preventing a further attack became the dominating task (e.g., Bush 2010, 181; Cheney and Cheney 2011, 333; Rumsfeld 2011, 362; Powell and Koltz 2012, 36) that bound the members of the administration together. Concerning the structure of this group, it is remarkable how important the concept of multiple advocacy became for its members. President Bush (2010, 65–66), Powell (Powell and Koltz 2012, 10), Rumsfeld (2011, 180184), and Rice (2011, 153) all stressed the importance of controversy and were aware of the danger of a deliberation process that was too homogeneous. The widely accepted assertion that the president was not an impartial leader but predisposed toward war (Yetiv 2011, 238; Feith 2009, 243; Pillar 2011, 25) cannot easily be discerned from the biographies of members of the administration or the discourse of the president himself. Rumsfeld (2011, 319 and 328), for example, recalls that Bush never really expressed his preferences and did not make clear what the results and consequences of their deliberations process were. Nevertheless, the deliberative process under Bush lacked clear norms and was characterized as too informal (Pfiffner 2009, 375). This cohesive group of actors, lacking formal norms structuring the deliberation process, faced a totally new security policy context after the terrorist attacks of September 11. Two schools of thought evaluate the impact of these attacks on the United States and its top-level decision-makers in contrasting ways. The first school sees the members of the Bush administration as the victims of 9/11 and the events thereafter. According to this line of reasoning, the terrorist attacks forced them to alter their approaches towards foreign and security policy, giving them little scope for variation (Flibbert 2006, 325–28; Mitchell and Massoud 2009, 279; Badie 2010, 278). The second school, however, sees Bush and his advisers as the agents of this change; September 11, 2001, became a useful pretext for implementing a long-desired re-orientation of US foreign and security policy (Kaufmann 2004; Mazaar 2007, 13; Burgos 2008, 225). Regardless of whether the Bush administration became a victim of the dynamics following the terrorist attacks or was the agents of them, September 11, 2001, transformed the context in which decisions were made. This transformation was clearly expressed in the autobiographies of the different actors. For Bush (2010, 128, 137, 151 and 229), the terrorist attacks became a perceptual shock, redefining and radicalizing his world view. Cheney and Powell, as well as Rumsfeld, perceived 9/11 as a seminal event, bringing the topics of terrorism and nuclear proliferation to the fore (Cheney and Cheney 2011, 10; Powell and Koltz 2012, 218; Rumsfeld 2011, 352–55). Rice (2011, xviii, 79, 148 and 169) depicts 9/11 as a “crack in time,” dramatically changing the way security issues were seen. Tenet (2008, 305) supports her view by writing: “After 9/11 everything changed. Many foreign-policy issues were now viewed through the prism of smoke rising from the World Trade Center and the Pentagon.” Tenet (2008, 174), furthermore, perceives the attacks as a personal and organizational failure. Henceforth, the prevention of further attacks on US soil became the unanimous conviction of the administration. All other aims and objectives were subordinated to this long-term goal (Bush 2010, 154–55; Cheney and Cheney 2011, 6). It was against the backdrop of the cohesiveness of its members, the lack of norms and an honest broker to structure the discussion, and the dramatic transformation of the context—the three antecedent conditions of groupthink—that the decision group came to its first consensus: a sudden, undisputed, and precedent-setting threat-perception. Figure 1 displays the affiliation network of the “threat-discourse” from January 2001 until September 11, 2001. Although all actors perceived Iraq as a threat, they differed on the question what exactly made up this threat. Bush, Cheney, Powell and Rice saw Iraq as an “aggressive actor,” but only Bush and Rice, as well as Armitage, linked this perception to the potential threat of a nuclear-armed rogue state (“rogue state threat with WMD”—weapons of mass destruction). Bush generally believed the United States to be in a “post-Cold-War security environment” with new rules defining the game and shared this view only with Wolfowitz. The deputy secretary of defense, in turn and together with Tenet and Powell, was convinced that the Iraqi threat had its causes in the “tyranny” of the regime. Powell on the other hand, stressed Hussein's “history of WMD-use” and was convinced that Iraq possessed or at least sought to develop WMDs (“possession/development of WMD”). This conviction was shared only by Tenet, who was the only actor who feared nuclear-armed terrorism (“terrorists seek/will use WMD”). In this threat-discourse, Rumsfeld acted as an outsider, sharing no category with any other member of the administration. According to him, the Iraqi threat was generally underestimated (“underestimation of threat”) and was especially troublesome because of Iraqi's “history of developing WMD.” Figure 1. View largeDownload slide Affiliation network of the “threat-discourse” until 9/11 (agreement “yes”). Figure 1. View largeDownload slide Affiliation network of the “threat-discourse” until 9/11 (agreement “yes”). The small value of density of this affiliation network of 0.062 and the relatively high value for modularity of 0.219 (for a small group) of the congruence network in the same period show that the threat-perception was heterogeneous and no actor dominated the discourse. The heterogeneous view of the Iraqi threat is a broad hint to a lively and healthy discussion with respect to the deliberation process. However, this observation changed dramatically and instantly with September 11, 2001. As Figure 2 of the threat-affiliation network from September 11 until March 20, 2003, shows, all members of the Bush administration were now convinced that Iraq was an aggressive actor, possessing or at least developing WMDs. In this “post-9/11 security environment,” Iraq's “history of WMD-use” and the assumed “support of terrorism” by Saddam Hussein were seen as unacceptable. The density of this network rose to 0.19, whereas the modularity of the corresponding congruence network reached zero. Consequently, the discourse became highly homogeneous and one cannot distinguish between subgroups of actors. This finding is even true when the period of observation is reduced to quarters. From 9/11 until the invasion in March 2003, the threat-perception in every single quarter was highly homogeneous with a modularity value of the congruence networks of zero.2 Figure 2. View largeDownload slide Affiliation network of the “threat-discourse” from 9/11 to Iraq invasion (agreement “yes”). Figure 2. View largeDownload slide Affiliation network of the “threat-discourse” from 9/11 to Iraq invasion (agreement “yes”). These discourse networks confirm the critique of several individuals within the administration. Tenet (2008, 305), for example, complains that there was never a serious discussion about the threat facing the United States. Paul O'Neill, the secretary of treasury, supports Tenet's account and emphasizes that the discussion never centered on the question of why to invade Iraq but only on how to do it (Suskind 2004, 76). This sudden and undisputed threat-perception, shared by all key advisers and the president himself, became the precedent-setting basis on which the administration began to think about policy towards Iraq. It was this premature consensus, which was never reexamined, that determined the future discussion and thus the decision-making process. Although failure to reexamine this perception was the only symptom of groupthink in this first phase of the decision process, it was a vital failure. Without this early consensus, the invasion of Iraq would have been unimaginable (Badie 2010, 285). From a Heterogeneous Nonmilitary to a Homogeneous Military Policy Discourse In contrast to the threat-discourse, the discussion about policy options began as a very homogeneous discourse (see Figure 3). Although in the first months of the new administration, a lively discussion in the principal's committee evolved about the usefulness of sanctions—with Powell a fierce advocate and Rumsfeld an opponent (Woodward 2004, 15)—President Bush and his team agreed on most categories and favored nonmilitary options and a modest regime-change policy in dealing with Iraq. “Sanctions” and the “consensus/cooperation with allies” build the backbone of this strategy, which was supported by Bush, Cheney, Powell, and Rice. Only Rumsfeld did not agree on sanctions as an appropriate policy (see Figure 3). Instead, he focused on regime change (although not militarily but by supporting the opposition) and “containment/deterrence,” sharing the first option with Powell and the latter with the president. Furthermore, Powell and Rice asked for “inspections” and were open to “selected military action” if necessary. Modularity of zero in the respective congruence network shows that the administration had a homogeneous policy approach at this early stage. Although they could not agree on how serious the threat posed by Iraq was, they were sure to tackle it by nonmilitary means, such as sanctions, containment, and deterrence, and in cooperation with allies. Figure 3. View largeDownload slide Affiliation network of the “policy-discourse” pre-9/11 (agreement “yes”). Figure 3. View largeDownload slide Affiliation network of the “policy-discourse” pre-9/11 (agreement “yes”). The terrorist attacks of September 11 caused a first shift in this policy (see Figure 4). As Bush (2010, 230) and Rice (2011, 171–72) emphasize in their autobiographies, the administration now followed the policy of “coercive diplomacy,” which rested on two pillars. On the one hand, the United States sought to convince Iraq by diplomatic means to give up its supposed WMD program. On the other hand, the buildup of a military threatening posture would emphasize US intentions and provide a backup if the diplomatic path failed. However, for the time being, the focus should lie on the diplomatic, nonmilitary path. Figure 4. View largeDownload slide Affiliation network of the “policy-discourse” from 9/11 until 31 December 2001 (agreement “yes”). Figure 4. View largeDownload slide Affiliation network of the “policy-discourse” from 9/11 until 31 December 2001 (agreement “yes”). At first, this policy realignment found its expression in the shift of focus from “sanctions” (only Powell supported this policy) to “inspections” (supported by Powell, Wolfowitz, Bush, and Cheney). Whereas Cheney, Rumsfeld, and Armitage envisioned “containment/deterrence,” President Bush cautioned against the “risk of inaction.” Although the actors were again quite homogeneous in their approach (modularity of zero in the respective congruence network) and favoring mostly nonmilitary measures, the focus of discussion shifted from a more passive (sanctions) to a more active approach (inspections). In the first six months of 2002, this homogeneous discourse coalition fell apart but without being replaced immediately by a new consensus. In the first quarter of 2002 (see Figure 5), only Powell and Rice urged for “inspections” and “sanctions.” All the other members of the decision group did not support this strategy any more. Instead, they referred to categories like “risk of inaction” (Bush, Rumsfeld, and Wolfowitz) or “all options are on the table” (Bush, Wolfowitz, and Powell). Especially the latter category became an indicator that an open war was now imaginable for an increasing number of decision-makers. Furthermore, Rumsfeld, supported by Rice, favored a “regime change.” Figure 5. View largeDownload slide Affiliation network of the “policy-discourse” in the first quarter 2002 (agreement “yes”). Figure 5. View largeDownload slide Affiliation network of the “policy-discourse” in the first quarter 2002 (agreement “yes”). This development climaxed in the second quarter of 2002 (see Figure 6). Powell and Rice were the only actors calling for “sanctions,” but Powell stood alone in urging for “inspections.” Rumsfeld (supported by Powell) again pressed for a “regime change.” Cheney and Wolfowitz stressed the “risk of inaction,” whereas the President reaffirmed that “all options are on the table.” Figure 6. View largeDownload slide Affiliation network of the “policy-discourse” in the second quarter 2002 (agreement “yes”). Figure 6. View largeDownload slide Affiliation network of the “policy-discourse” in the second quarter 2002 (agreement “yes”). As the congruence networks and the values for modularity (0.189 in the first and 0.375 in the second quarter of 2002) illustrate, the initial consensus and the homogeneity of the group broke completely apart. In those six months, discussion within the administration shifted incrementally from a nonmilitary discourse (i.e., “sanction” and “inspections”) toward a military discourse (i.e., “risk of inaction,” “regime change,” and “all options are on the table”). This finding supports Woodward's (2004, 108 and 119) argument that the military pillar of the “coercive diplomacy” policy gained the upper hand in the first half of 2002. According to Powell, the “[m]omentum within the administration was building toward military action” (Powell and Koltz 2012, 209) in August 2002. Rice (2011, 158159) agrees with Powell and sees Cheney and Rumsfeld as proponents of this realignment. She is convinced that the president became increasingly sympathetic to this shift. From the third quarter of 2002 until the invasion in March 2003 (see Figure 7) the “risk of inaction” and “war” became the dominant categories, shared by all. Sanctions were not an issue any more. The decision group suddenly became homogeneous again (modularity of zero), and the policy discourse was solely characterized by military options. Diplomacy as the first pillar of the coercive-diplomacy strategy was abandoned in favor of military pressure. This shift unleashed a dynamic that gathered momentum and inevitably led to the war in March, as Tenet (2008, 333) argues. By turning to the military pillar of the strategy, according to Bush (2010, 250252), the United States could not back down when Iraq refused to fulfill US claims. The United States was forced to go to war in order not to lose its credibility. The crucial period of time, thus, was the first six months of 2002, as a new consensus centering around military options replaced the initial consensus on nonmilitary policy options. The dominant actors in this phase of the decision-making process were Rumsfeld and Powell, who had the highest values for betweenness centrality, 4.33 and 2.0, respectively. These two actors served as brokers between all the others. This, again, indicates that the decision process in this crucial phase was shaped by an intense controversy between Powell, who favored nonmilitary options and thus the diplomatic pillar of the coercive-diplomacy strategy, and Rumsfeld, who demanded a tougher stance towards Iraq and the pursuit of the military pillar. Figure 7. View largeDownload slide Affiliation network of the “policy-discourse” from 1 October 2002 until 31 March 2003 (agreement “yes”). Figure 7. View largeDownload slide Affiliation network of the “policy-discourse” from 1 October 2002 until 31 March 2003 (agreement “yes”). This controversy was furthered by an increasingly dysfunctional group with personal rivalries, missing norms, and the lack of an impartial leader who could structure the decision process. Bush (2010, 65–66) acknowledges in his biography that he knew of the importance of an effective and heterogeneous decision group as a precondition for good advice. He, therefore, appointed Rumsfeld as secretary of defense to balance Powell, the highly admired secretary of state and potential dominant foreign and security policy figure in the administration. This decision was the starting point for a bureaucratic fight between members of the State Department and the Department of Defense, as well as the vice president and his office, over the dominance of the foreign and security policy process (Mann 2004, 271–75). Bush (2010, 89) concedes that the group became more and more ineffective as time passed: “What started as creative tension, had turned destructive.” The other group members shared this view of an increasingly dysfunctional National Security Council (NSC). Rumsfeld (2011, 319 and 328) laments that the whole decision process lacked structuring norms. Furthermore, Bush never made clear where he stood at the end of an intensive discussion and what the results had been. Rumsfeld (2011, 325–26) blames Rice and the president for this failure. Rice not only failed in structuring the process, according to Rumsfeld, but also failed to act as an impartial leader. She tried to forge a consensus that offered President Bush watered down compromises instead of distinctive policy choices. In contrast to Rumsfeld, Powell complains that his advice was no longer in demand and that crucial decisions were often made in his absence. Consequently, Bush did not get the full spectrum of opinions (Powell and Koltz 2012, 36 and 55). Rice confirms both Rumsfeld's as well as Powell's view. She tried to be the honest broker that Rumsfeld and Powell objected to for very different reasons (Rice 2011, 18–19). Tenet (2008, 138) also blames Rice for the dysfunction of the decision body. She was neither willing nor able to put up a fight against such dominant actors as Cheney and Rumsfeld, who began to marginalize her and take over the decision-making process. This finding is in line with other analyses of the decision-making process in the run-up to the 2003 Iraq War, which also found: the increase of influence of actors like Cheney and Rumsfeld and their allies within the inner circle of advisers; the creation of alternative centers for intelligence analysis to bolster their preferred policy choices and strategies; the pressure put on dissenters; and the neglect of a comprehensive and realistic postwar and contingency planning (Pfiffner 2004; Woodward 2004; Hersh 2005; Isikoff and Corn 2007; Badie 2010). These consequences of the groupthink syndrome also underline how the actors struggled for discourse dominance and why Cheney, Rumsfeld, and others succeeded in doing so and convinced President Bush, the crucial actor in this process, of their approaches. What these networks further demonstrate is that Secretary of State Powell, who is often regarded as an opponent of the war, also shifted to the dominant military discourse beginning in the third period of 2002. This finding confirms Rumsfeld's (2011, 457) claim that the decision to go to war was unanimous, with no one verbally dissenting. Although Bush (2010, 251) is convinced that Powell had serious doubts about the war, the secretary of state did not question the decision in public (an indicator of self-censorship). Conclusion The literature on the decision-making process in the run-up to the 2003 Iraq War faces a serious dilemma. It builds its argument largely on investigative journalism, sources for which are neither scientifically verifiable nor replicable. Primary documents, that is, speeches, interviews, and autobiographies, are often neglected in these analyses because they are either seen as useless for deciphering private decision-making processes or too biased to be reliable. Consequently, scholars do not take advantage of significant sources. In contrast, this article relies upon them, as well as on other primary sources to explain foreign-policy decision-making. They are corrective to accounts that rely upon investigative journalism and, beyond that, are verifiable and replicable. The article also shows the value of integrating Discourse Network Analysis into analyses of foreign-policy decision-making (i.e., groupthink in the case of this study). First, DNA permits political scientists to broaden their sources of information by bringing speeches and interviews into their analyses. Second, DNA contributes to the understanding of decision-making processes in groups. It reveals struggles between different coalitions of actors over discursive hegemony. It does so by relating the individual discourses of actors to the group discourse and thereby showing how coalitions are constructed and deconstructed. Network metrics help us to comprehend these dynamics by displaying the properties of groups and coalitions, such as density and modularity. DNA, therefore, provides a methodological means to make the consensus building efforts (in form of concurrence-seeking) visible and to display the phases of this process in more detail. In particular, DNA provided substantial empirical evidence for examining the intermediate steps in the decision-making process and for answering why the United States invaded Iraq, who the driving forces behind this decision were, and when the decision was made. First, the decision-making process in the run-up to the war occurred in two phases. In the first phase, decision-makers agreed on a common threat perception. Although there was no consensus on the dangers Iraq posed at the beginning of the administration, the terrorist attacks of 9/11 changed that. This consensus building process featured the three antecedent conditions of groupthink: a cohesive group of actors; a dysfunctional group lacking norms and impartial leadership that could structure the decision process; and a provocative situational context. Given the homogeneous threat-perception that was an outcome of the first phase, the second phase required the decision group to find a consensus concerning the right policy towards Iraq. This process was characterized by the incremental collapse of the originally dominant discourse, the focus on diplomatic means and nonmilitary policies, and the rise of a new, more hawkish military approach. After redirecting US policy towards Iraq on to the military path, the administration could not back down. It had to walk this path until the end in order not to lose face. Second, the driving forces behind this policy shift were a group of actors around Vice President Cheney and Secretary Rumsfeld. It was especially the latter who, as a broker, had the most influence on the deliberation process according to the discourse networks. Rumsfeld's discourse was more compatible with that of the other actors. Consequently, the discourse coalition surrounding him was more successful in convincing President Bush than his principal opponent, Secretary of State Powell. This finding supports those accounts that see the Iraq War decision as the result of a fierce fight between “hawks” (Cheney, Rumsfeld, and Wolfowitz) and “doves” (especially Powell). Finally, the replacement of the former diplomatic discourse with a more military one took place in the first half of 2002. In the third quarter of 2002, Bush and his inner circle (including Powell) once again shared a common approach towards Iraq. This finding suggests that the decision to go to war or at least the trend-setting decision to focus solely on the military path of the policy of coercive diplomacy, which led to the war, was taken in the last weeks of the second or early weeks of the third quarter of 2002. The use of DNA reveals why actors and groups follow certain policy options instead of others. Future research should not just apply this concept to case studies with small decision groups but to larger communities, including actors like bureaucracies and organizations, parliaments, pressure groups and media. Although DNA provides valuable insights into small group dynamics, it is much more powerful when the number of actors involved increases. 1 Following Holzscheiter (2014, 144) I define “discourse“ and “discourse analysis” as follows: “[D]iscourse analysis is engagement with meaning and the linguistic and communicative processes through which social reality is constructed. Discourse can therefore be defined as, basically, the space where intersubjective meaning is created, sustained, transformed and, accordingly, becomes constitutive of social reality.” Furthermore, I use a micro-interaction approach, focusing on the communicative and deliberative exchange of meanings between agents in the struggle for the better argument (Holzscheiter 2014, 146–150). 2 The only exception was the second quarter of 2002. However, in this case modularity reached 0.065, which is almost zero and thus not significant. Supplemental Information Supplemental information is available at the Foreign Policy Analysis data archive. References Allison Graham T., Zelikow Philip. 1999. Essence of Decision: Explaining the Cuban Missile Crisis . 2nd ed. New York, NY: Longman. Badie Dina. 2010. “Groupthink, Iraq, and the War on Terror: Explaining US Policy Shifts toward Iraq.” Foreign Policy Analysis  6: 277– 96. Google Scholar CrossRef Search ADS   Bamford James. 2005. A Pretext for War: 9/11, Iraq, and the Abuse of American Intelligence Agencies . New York, NY: Anchor Books. Baron Robert S. 2005. “So Right It's Wrong: Groupthink and the Ubiquitous Nature of Polarized Group Decision Making.” Advances in Experimental Social Psychology  37: 219– 53. Google Scholar CrossRef Search ADS   Beasley Ryan K., Kaarbo Juliet. 2014. “Explaining Extremity in the Foreign Policies of Parliamentary Democracies.” International Studies Quarterly  58: 729– 40. Google Scholar CrossRef Search ADS   Bennett Andrew, Checkel Jeffery T.. 2015. “Process Tracing: From Philosophical Roots to Best Practices.” In Process Tracing: From Metaphor to Analytic Tool , edited by Bennett Andrew, Checkel Jeffery T.. Cambridge: Cambridge University Press. Borgatti Stephen P., Mehra Ajay, Brass Daniel J., Labianca Giuseppe. 2009. “Network Analysis in the Social Sciences.” Science  323: 892– 95. Google Scholar CrossRef Search ADS PubMed  Burgos Russell A. 2008. “Origins of Regime Change: ‘Ideapolitik’ on the Long Road to Baghdad, 1993–2000.” Security Studies  17: 221– 56. Google Scholar CrossRef Search ADS   Burke John P. 2005. “Condoleezza Rice as NSC Advisor: A Case of the Honest Broker Role.” Presidential Studies Quarterly  35: 554– 75. Google Scholar CrossRef Search ADS   Bush George W. 2010. Decision Points . New York, NY: Crown. Cheney Dick, Cheney Liz. 2011. In My Time: A Personal and Political Memoir . New York, NY: Simon & Schuster. Dyson Stephen Benedict. 2006. “Personality and Foreign Policy: Tony Blair's Iraq Decision.” Foreign Policy Analysis  2: 289– 306. Google Scholar CrossRef Search ADS   Dyson Stephen Benedict. 2009. “‘Stuff Happens’: Donald Rumsfeld and the Iraq War.” Foreign Policy Analysis  5: 327– 47. Google Scholar CrossRef Search ADS   Elman Miriam Fendius. 2000. “Unpacking Democracy: Presidentialism, Parliamentarism, and Theories of Democratic Peace.” Security Studies  9: 91– 126. Google Scholar CrossRef Search ADS   Erisen Elif, Erisen Cengiz. 2012. “The Effect of Social Networks on the Quality of Political Thinking.” Political Psychology  xx: 1– 27. Feith Douglas J. 2009. War and Decision: Inside the Pentagon at the Dawn of the War on Terrorism . New York: Harper. Fitzgerald Michael, Lebow Richard Ned. 2006. “Iraq: The Mother of all Intelligence Failure.” Intelligence and National Security  21: 884– 909. Google Scholar CrossRef Search ADS   Flibbert Andrew. 2006. “The Road to Baghdad: Ideas and Intellectuals in Explanations of the Iraq War.” Security Studies  15: 310– 52. Google Scholar CrossRef Search ADS   Franz Timothy M. 2012. Group Dynamics and Team Interventions: Understanding and Improving Team Performance . Malden, MA: Wiley-Blackwell. Fuller Sally Riggs, Alday Ramon J.. 1997. “Challenging the Mindguards: Moving Small Group Analysis Beyond Groupthink.” In Beyond Groupthink: Political Group Dynamics and Foreign Policy-making , edited by Hart Paul ’t, Stern Eric K., Sundelius Bengt. Ann Arbor, MI: The University of Michigan Press. George Alexander L. 1997. “From Groupthink to Contextual Analysis of Policy-Making Groups.” In Beyond Groupthink: Political Group Dynamics and Foreign Policy-making , edited by ’t Hart Paul, Stern Eric K., Sundelius Bengt. Ann Arbor, MI: The University of Michigan Press. George Alexander L., Bennett Andrew. 2005. Case Studies and Theory Development in the Social Sciences . Cambridge, MA: MIT Press. George Alexander L., Stern Eric K.. 2002. “Harnessing Conflict in Foreign Policy Making: From Devil's to Multiple Advocacy.” Presidential Studies Quarterly  32: 484– 508. Google Scholar CrossRef Search ADS   Hafner-Burton Emilie M., Kahler Miles, Montgomery Alexander H.. 2009. “Network Analysis for International Relations.” International Organization  63: 559– 92. Google Scholar CrossRef Search ADS   Hajer Maarten A. 1993. “Discourse Coalitions and the Institutionalization of Practice: The Case of Acid Rain in Britain.” In The Argumentative Turn in Policy Analysis and Planning , edited by Fisher Frank, Forester John. Durham, NC and London: Duke University Press. Google Scholar CrossRef Search ADS   Hajer Maarten A. 1997. The Politics of Environmental Discourse: Ecological Modernization and the Policy Process . Oxford: Oxford University Press. Google Scholar CrossRef Search ADS   Halperin Morton H., Clapp Priscilla A., Kanter Arnold. 2006. Bureaucratic Politics and Foreign Policy . 2nd ed. Washington, DC: Brookings Institution Press. Hart Paul ’t. 1991. “Irving L. Janis’ Victims of Groupthink.” Political Psychology  12: 247– 78. Google Scholar CrossRef Search ADS   Hart Paul ’t. 1994. Groupthink in Government: A Study of Small Groups and Policy Failure . Baltimore, MD: The Johns Hopkins University Press. Hart Paul ’t, Stern Eric K., Sundelius Bengt. 1997a. “Foreign Policymaking at the Top: Political Group Dynamics.” In Beyond Groupthink: Political Group Dynamics and Foreign Policy-Making , edited by Hart Paul ’t, Stern Eric K., Sundelius Bengt. Ann Arbor, MI: The University of Michigan Press. Hart Paul ’t, Stern Eric K., Sundelius Bengt, eds. 1997b. Beyond Groupthink: Political Group Dynamics and Foreign Policy-making . Ann Arbor, MI: The University of Michigan Press. Hersh Seymour M. 2005. Chain of Command: The Road from 9/11 to Abu Ghraib . New York, NY: Harper. Holzscheiter Anna. 2014. “Between Communicative Interaction and Structures of Signification: Discourse Theory and Analysis in International Relations.” International Studies Perspectives  15: 142– 62. Google Scholar CrossRef Search ADS   Houghton David Patrick. 2008. “Invading and Occupying Iraq: Some Insights From Political Psychology.” Peace and Conflict  14: 169– 92. Google Scholar CrossRef Search ADS   Hudson Valerie M. 2007. Foreign Policy Analysis: Classic and Contemporary Theory . Lanham: Rowman & Littlefield. Isikoff Michael, Corn David. 2007. Hubris: The Inside Story of Spin, Scandal, and the Selling of the Iraq War . New York, NY: Three Rivers Press. Janis Irving L. 1982. Groupthink: Psychological Studies of Policy Decisions and Fiascoes . 2nd ed. Boston, MA: Wadsworth. Jervis Robert. 2006. “Reports, Politics, and Intelligence Failures: The Case of Iraq.” The Journal of Strategic Studies  29: 3– 52. Google Scholar CrossRef Search ADS   Kaarbo Juliet. 2008. “Coalition Cabinet Decision Making: Institutional and Psychological Factors.” International Studies Review  10: 57– 86. Google Scholar CrossRef Search ADS   Kaarbo Juliet, Beasley Ryan K.. 2008. “Taking It to the Extreme: The Effect of Coalition Cabinets on Foreign Policy.” Foreign Policy Analysis  4: 67– 81. Google Scholar CrossRef Search ADS   Kadushin Charles. 2012. Understanding Social Networks: Theories, Concepts, and Findings . Oxford - New York, NY: Oxford University Press. Kaplan Fred. 2008. Daydream Believers: How a Few Grand Ideas Wrecked American Power . Hoboken, NJ: John Wiley & Sons. Kassop Nancy. 2003. “The War Power and Its Limits.” Presidential Studies Quarterly  33: 509– 29. Google Scholar CrossRef Search ADS   Kaufmann Chaim. 2004. “Threat Inflation and the Failure of the Marketplace of Ideas: The Selling of the Iraq War.” International Security  29: 5– 48. Google Scholar CrossRef Search ADS   Lazer David. 2011. “Networks in Political Science: Back to the Future.” PS: Political Science & Politics  44: 61– 68. Google Scholar CrossRef Search ADS   Lee Su-Mi. 2014. “Understanding the Yalta Axioms and Riga Axioms through the Belief Systems of the Advocacy Coalition Framework.” Foreign Policy Analysis  DOI: 10.1111/fpa.12040:1-21. Leifeld Philip. 2011. Discourse Network Analyzer Manual . Technical report EAWAG Konstanz: URL: http://www.philipleifeld.de/cms/upload/Downloads/dna-manual1.30.pdf. Lind Michael. 2012. “Neoconservativism and American Hegemony.” In Why Did the United States Invade Iraq?  edited by Cramer Jane K., Thrall A. Trevor. Routledge Global Security Studies London - New York, NY: Routledge. Mann James. 2004. Rise of the Vulcans: The History of the Bush's War Cabinet . New York, NY: Penguin Books. Mazaar Michael J. 2007. “The Iraq War and Agenda Setting.” Foreign Policy Analysis  3: 1– 23. Google Scholar CrossRef Search ADS   McCauley Clark. 1979. “The Nature of Social Influence in Groupthink: Compliance and Internalization.” Journal of Personality and Social Psychology  57: 250– 60. Google Scholar CrossRef Search ADS   McClurg Scott D., Young Joseph K.. 2011. “Political Networks: Editor's Indroduction - A Relational Political Science.” PS: Political Science & Politics  44: 39– 43. Google Scholar CrossRef Search ADS   Mearsheimer John J., Walt Stephen M.. 2007. The Isreal Lobby and U.S. Foreign Policy . New York, NY: Farrar, Straus and Giroux. Metselaar Max V., Verbeek Bertjan. 1997. “Beyond Decision Making in Formal and Informal Groups: The Dutch Cabinet and the West New Guinea Conflict.” In Beyond Groupthink: Political Group Dynamics and Foreign Policy-making , edited by Hart Paul ’t, Stern Eric K., Sundelius Bengt. Ann Arbor, MI: The University of Michigan Press. Mitchell David, Massoud Tansa George. 2009. “Anatomy of Failure: Bush's Decision-Making Process and the Iraq War.” Foreign Policy Analysis  5: 265– 86. Google Scholar CrossRef Search ADS   Mulcahy Kevin V. 1995. “Rethinking Groupthink: Walt Rostow and the National Security Advisory Process in the Johnson Administration.” Presidential Studies Quarterly  25: 237– 50. Newman M. E. J. 2012. Networks: An Introduction . Oxford - New York, NY: Oxford University Press. Packer George. 2007. The Assassin's Gate: America in Iraq . London: Faber & Faber. Pfiffner James P. 2004. “Did President Bush Mislead the Country in His Arguments for War with Iraq?” Presidential Studies Quarterly  34: 25– 46. Google Scholar CrossRef Search ADS   Pfiffner James P. 2009. “Decision Making in the Bush White House.” Presidential Studies Quarterly  39: 363– 84. Google Scholar CrossRef Search ADS   Pillar Paul R. 2011. Intelligence and U.S. Foreign Policy: Iraq, 9/11, and Misguided Reform . New York, NY: Columbia University Press. Powell Colin, Koltz Tony. 2012. It Worked for Me: In Life and Leadership . New York, NY:   Harper. Renshon Jonathan. 2009. “When Public Statements Reveal Private Beliefs: Assessing Operational Code at a Distance.” Political Psychology  30: 649– 61. Google Scholar CrossRef Search ADS   Rice Condoleezza. 2011. No Higher Honor: A Memoir of My Years in Washington . New York, NY: Crown. Rumsfeld Donald. 2011. Known and Unknown: A Memoir . London: Sentinel. Sabatier Paul A. 1988. “An Advocacy Coalition Framework of Policy Change and the Role of Policy-Oriented Learning Therein.” Policy Sciences  21: 129– 68. Google Scholar CrossRef Search ADS   Sabatier Paul A., Weible Christopher M.. 2007. “The Advocacy Coalition Framework.” In Theories of the Policy Process , edited by Sabatier Paul A.. Cambridge, MA: Westview Press. Schafer Mark, Crichlow Scott. 1996. “Antecedents of Groupthink: A Quantitative Study.” Journal of Conflict Resolution  40: 415– 35. Google Scholar CrossRef Search ADS   Schafer Mark, Crichlow Scott. 2002. “The Process-Outcome Connection in Foreign Policy Decision Making: A Quantitative Study Building on Groupthink.” International Studies Quarterly  46: 45– 68. Google Scholar CrossRef Search ADS   Schafer Mark, Crichlow Scott. 2010. Groupthink Versus High-Quality Decision Making in International Relations . New York, NY: Columbia University Press. Schimmelfennig Frank. 2001. “The Community Trap: Liberal Norms, Rhetorical Action, and the Eastern Enlargement of the European Union.” International Organization  55: 47– 80. Google Scholar CrossRef Search ADS   Shannon Vaughn P., Keller Jonathan W.. 2007. “Leadership Style and International Norm Violation: The Case of the Iraq War.” Foreign Policy Analysis  3: 79– 104. Google Scholar CrossRef Search ADS   Simon Herbert A. 1985. “Human Nature in Politics: The Dialogue of Psychology with Political Science.” The American Political Science Review  79: 293– 304. Google Scholar CrossRef Search ADS   Smith Steve. 1984. “Groupthink and the Hostage Rescue Mission.” British Journal of Political Science  15: 117– 26. Google Scholar CrossRef Search ADS   Snyder Richard C., Bruck H. W., Sapin Burton M., Hudson Valerie M., Chollet Derek H., Goldgeier James M.. 2002. Foreign Policy Decision-Making (Revisited) . New York, NY: Palgrave Macmillan. Google Scholar CrossRef Search ADS   Stern Eric K. 1997. “Probing the Plausibility of Newgroup Syndrome: Kennedy and the Bay of Pigs.” In Beyond Groupthink: Political Group Dynamics and Foreign Policy-making , edited by Hart Paul ’t, Stern Eric K., Sundelius Bengt. Ann Arbor, MI: The University of Michigan Press. Stern Eric K., Sundelius Bengt. 1997. “Understanding Small Group Decisions in Foreign Policy: Process Diagnosis and Research Procedure.” In Beyond Groupthink: Political Group Dynamics and Foreign Policymaking , edited by Hart Paul ’t, Stern Eric K., Sundelius Bengt. Ann Arbor, MI: The University of Michigan Press. Sunstein Cass R., Hastie Reid. 2015. Wiser: Getting Beyond Groupthink to Make Groups Smarter . Boston, MA: Harvard Business Review Press. Suskind Ron. 2004. The Price of Loyalty: George W. Bush, the White House, and the Education of Paul O'Neill.  New York, NY: Simon & Schuster Paperbacks. Suskind Ron. 2007. The One Percent Doctrine: Deep Inside America's Pursuit of Its Enemies Since 9/11 . New York, NY: Simon & Schuster Paperbacks. Tenet George. 2008. At the Center of the Storm: The CIA during America's Time of Crises . New York, NY: Harper. Tetlock Philip E. 1979. “Identifying Victims of Groupthink From Public Statements to Decision Makers.” Journal of Personality and Social Psychology  37: 1314– 24. Google Scholar CrossRef Search ADS   Tetlock Philip E., Peterson Randall S., McGuire Charles, Chang Shi-jie, Feld Peter. 1992. “Assessing Political Group Dynamics: A Test of the Groupthink Model.” Journal of Personality and Social Psychology  63: 403– 52. Google Scholar CrossRef Search ADS   Vanderbush Walt. 2009. “Exiles and the Marketing of U.S. Policy toward Cuba and Iraq.” Foreign Policy Analysis  5: 287– 306. Google Scholar CrossRef Search ADS   Vertzberger Yaacov Y. I. 1997. “Collective Risk Taking: The Decision-making Group.” In Beyond Groupthink: Political Group Dynamics and Foreign Policy-making , edited by Hart Paul ’t, Stern Eric K., Sundelius Bengt. Ann Arbor, MI: The University of Michigan Press. Woodward Bob. 2004. Plan of Attack . New York, NY: Simon & Schuster. Yetiv Steven A. 2003. “Groupthink and the Gulf Crisis.” British Journal of Political Science  33: 419– 42. Google Scholar CrossRef Search ADS   Yetiv Steven A. 2011. Explaining Foreign Policy: U.S. Decision-Making and the Persian Gulf Wars . 2nd ed. Baltimore, MD: The Johns Hopkins University Press. © The Author (2017). Published by Oxford University Press on behalf of the International Studies Association. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Journal

Foreign Policy AnalysisOxford University Press

Published: Oct 11, 2017

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off