Hebrew University of Jerusalem
Research on decision making in foreign policy and national security has had little interaction with the field of public policy. This review connects the two fields. We utilize a key public policy concept, the policy cycle, to provide a framework for our review of group decision making dynamics in national security and foreign policy. We describe key stages of the policy cycle followed by a review of the leading models of group decision-making dynamics. We then construct a bridge between the two, demonstrating how specific stages of the policy cycle are typically associated with specific group decision-making dynamics. To illustrate this link we provide an example of decision making dynamics within the Obama administration throughout policy stages of the 2016 campaign against the Islamic State in Raqqa, Syria.
group decision making, foreign policy, national security decision making, policy cycle, groupthink, polythink, con-div, ISIS
The fields of foreign policy and public policy have had very little interaction, yet a vast trove of scholarly work in the discipline of public policy exists alongside great contributions to foreign policy decision-making (FPDM). The purpose of this article is to provide a review of the literature within the discipline of FPDM which is focused on key group decision making models, enriched by the application of a well-known public policy concept, the policy cycle. Both public policy and foreign policy deal with group decision making analysis, yet their tool kits are quite distinct. By engaging in cross-fields efforts, scholars can fill gaps and bring new insights in the respective fields. Specifically, we demonstrate an association between stages of the policy cycle and leading models of group decision-making dynamics.
We first lay the public policy foundations for our review by providing a brief summary of the policy cycle broadly and then more precisely the stages of policy formulation, decision making, and implementation. We then highlight a new framework (the group decision-making continuum) and summarize three leading models of group decision-making dynamics within the literature of FPDM: groupthink, polythink, and con-div. We discuss the foundations and current trends in the application of each model to foreign policy and national security decision making. We will demonstrate how scholars have utilized these leading models to explain foreign policy decisions. However, these scholars have not yet linked group dynamics to stages in the policy cycle. Likewise, none of these scholars have demonstrated how group dynamics may shift through stages of a foreign policy decision. Our review constructs a bridge between stages of the policy cycle and specific group decision making models to expand on our understanding of foreign policy decision making. To illustrate the applicability of this interdisciplinary endeavor, we provide a brief example of group decision-making models within the context of policy stages of the Obama administration’s actions in the 2016 Raqqa Campaign against the Islamic State.
Since Lasswell (1956) first put forward the policy cycle model to illustrate the life cycle through which policy passes, the study of policy processes and stages has burgeoned into a diverse and thriving field. Lasswell originally identified seven stages: intelligence, promotion, prescription, invocation, application, termination, and appraisal. Hargrove (1975) further invigorated the policy process approach by identifying “implementation” as a missing stage in the policy-making process. Jann and Wegrich summarized the current differentiation of the stages as: agenda-setting, policy formulation, decision making, implementation, and evaluation (eventually leading to termination) as “the conventional way to describe the chronology of policy process" (2006, 43). As an important step in bridging between public policy and foreign policy decision making (FPDM), we utilize the policy cycle framework. This is because of the openness of the policy cycle to other theoretical and empirical studies (Schlager 1999; 239, 258). Jann and Wegrich (2007) contend that "the model itself has been highly successful as a basic framework for the field of policy studies and became the starting point of a variety of typologies of the policy process" and conclude that “the policy cycle perspective will continue to provide an important conceptual framework in policy research” due to its “receptivity for other and new approaches in the wider political science literature” (57). They further explain how scholars of policy studies seldom apply the whole policy cycle framework but rather utilize specific stages to "guide the selection of questions and variables" and that research is generally "related to particular stages of the policy process rather than on the whole cycle" (Ibid, p. 45). This review will focus on the stages of policy formulation, decision making, and implementation.
We acknowledge the major criticisms to the policy cycle, termed by one scholar as the “textbook approach” (Nakamura 1987). Noted policy scholars Sabatier and Jenkins-Smith (1993) argued in favor of more complex models that reflect complex realities. They developed the Advocacy Coalition Framework as a more interactive, inclusive, and nuanced picture of policy processes (see also Sabatier 1999 and Weible and Jenkins-Smith 2016).While the linear approach of stages in the policy process has been challenged, it is nevertheless useful, particularly in the effort to construct a bridge between foreign policy group decision making models and public policy frameworks.
Foreign policy is distinct from public policy in many aspects. Schafer and Crichlow (2010) explain that advisors around the president (1) have unique attributes providing for "more opportunities to overcome otherwise insurmountable bureaucratic conflicts or the inertia-like power of entrenched routines”; (2) are less likely to be bound by enduring norms of procedure than many other groups”; (3) have a "limited life span (they change in fundamental ways with every change in administration, and their format, design, and practices may be greatly altered during any particular government)”; (4) experience a greater frequency in change of "the nature of the rules and operational procedures that are utilized at different times." (2010, 22).
Despite the differences between public and foreign policy, we agree with Archuleta's conclusion (2016), that even though “Defense and national security subsystems are far more insular and less pluralistic than domestic policy subsystems, the construct [of the policy cycle] still applies” (2016, S51). We now address the stage-specific definition for policy formulation, decision-making, and implementation, followed by a review of small group dynamics and the link between the two. We conclude with an example of the Raqqa 2016 decisions of the Obama Administration.
Public policy literature defines formulation as a stage that "involves identifying and/or crafting a set of policy alternatives to address a problem, and narrowing that set of solutions in preparation for the final policy decision" (Sidney 2006, 79). According to policy cycle literature, policy formulation precedes and is distinct from decision-making (ibid). The demand placed on the group is to identify and craft a set of policy alternatives. Based on these factors from the public policy literature above, the decision unit is faced with at least two major dilemmas as they work to fulfill the demand. Ideally, the group must first think expansively, providing for a full spectrum of options. Scholars Bobrow and Dryzek (1987) recommend that policy formulation should offer options that include no intervention, the status quo, and solutions not in line with current practice. In other words, a team of policy makers are expected to demonstrate to the decision maker/s that they have thought broadly and deeply about all viable options and can adeptly present the costs and benefits of each. Secondly, there is a dilemma resulting from the demand to reduce. Policy formulators must select "from among [viable options] a smaller set of possible solutions from which decision makers actually will choose [by applying a] set of criteria to the alternatives, for example judging their feasibility, political acceptability, costs, benefits, and such (Sidney 2006, 79). Schattschneider (1960) reminds us that, “the definition of the alternatives is the choice of conflicts, and the choice of conflicts allocates power” (1960, 68). Clearly reducing options involves conflicts as well as heuristics and biases. Yet, in comparison to the stage of decision-making, the stage of policy formulation is generally more conducive to expansive thinking and is therefore likely to foster a different group dynamic.
Both policy cycle literature and policy network theories provide a framework in which formulation and decision making are distinct processes warranting individual scrutiny. In the study of group decision making, the unit responsible for this stage has been called the “final decision unit” (Schafer and Crichlow 2010, 18). At the decision-making stage the demand is to analyze a reduced set of options in preparation for selecting the most optimal path for achieving the strategic goal. Unlike policy formulation, there is not a demand for a comprehensive set of alternatives, but rather a demand for reducing and finally excluding all options, save one. The dilemmas for the decision unit is how to reduce the set of options to a single course of action. Individuals within the group must "order [their] world, making its complexities somewhat simpler… unconsciously strip[ping] the nuances, context, and subtleties out of the problems they face…" (Stein 2016, 133). There are numerous theories on how decisions are made. Stein (2016, 2017) provides a comprehensive review.
Implementation in public policy literature is defined as "what happens between the establishment of an apparent intention on the part of the government to do something, or to stop doing something, and the ultimate impact in the world of action" (O’Toole 2000, 266). Jann and Wegrich (2007, 52) outline three core demands of implementation: (1) Specification of program details (i.e., how and by which agencies/organizations should the program be executed? How should the law/program be interpreted?); (2) Allocation of resources (I.e., how are budgets distributed? Which personnel will execute the program? Which units of an organization will be in charge for the execution?); and (3) Decisions (i.e., how will decisions of single cases be carried out?).
These demands require process-oriented decision making involving numerous, real-time, micro-level decisions. During previous stages, it has been assumed that there are appropriate means (capabilities and tools) for the selected option, thus at the stage of implementation the demand "presupposes not only the capacity to pursue goals with effective means, but more generally the ability of governments to extract and mobilize resources from their audiences, both material and immaterial, and channel them into the pursuit of given objectives (Mastanduno et al. 1989). These are complex demands and are directly put to the test during implementation, a major differentiation from the two preceding stages and is likely to facilitate a distinct group dynamic. Each of the stages of the policy cycle discussed above, and their unique demands and dilemmas, is associated with specific models of group decision making dynamics (Mintz and Wayne 2016). In the next section we review the major models.
Within the discipline of Foreign Policy Decision Making (FPDM), there are several research agendas (Hudson 2016). This review is focused on the scholarship surrounding small-group dynamics.1 Within this genre, scholars are concerned with the consequences of the psychological dynamics within small groups on the outcomes of foreign policy decisions. Groupthink, summed up as premature consensus seeking, is possibly “the best known group-level phenomenon affecting foreign-policy decision making” (Schafer and Crichlow 2010, 23). The term has infiltrated the nomenclature of multiple academic, business, and even social spheres. However, since the term was first introduced by Janis (1972), as well as the formal inquiry into the effects of group dynamics on foreign policy decision making, the field has undergone significant development in terms of contributions and advancements. The next section will address what we identify as three waves of research. These will be discussed within the context of the group decision-making continuum and the specific models of groupthink, polythink, and con-div. We utilize this construct in this review because it conceptualizes and categorizes the literature of small group dynamics more broadly and effectively than existing frameworks of group dynamics. The next section briefly describes the group decision-making continuum, followed by a more in-depth discussion of each model, and how it is identified by particular symptoms, as well as each model’s impact on specific waves of research.
Mintz and Wayne (2016) introduced the group decision making continuum, arguing that a balanced decision making process lies between two opposite extremes that are equally dysfunctional and destructive. They conceptualized a framework in which "completely cohesive" (groupthink) anchors one end of the spectrum and "completely fragmented" (polythink) anchors the other. They explain that con-div exists on the group decision-making continuum within a central range where neither groupthink nor polythink dominates. They explain that it is within this range where there is a greater possibility for optimal decision-making to be crafted. Each group dynamic on the continuum exhibits particular symptoms, allowing for scholars and policy analysts to identify and diagnose which group dynamic is at work within a given decision-making unit. We now review the concept of groupthink within the context of the first wave of research into group dynamics and their impact on foreign policy decision-making.
Janis (1972, 1982) conceived of groupthink as a linear model of broadly defective group processes that flow from seven antecedents which included group cohesiveness, four types of structural faults of the organization, and two types of provocative situational contexts. This, he argued, leads to problematic concurrence seeking tendencies and groupthink. The symptoms included the illusion of invulnerability, belief in inherent group morality, collective rationalization, stereotypes of outsiders, self-censorship, the illusion of unanimity, pressure on dissenters, and self-appointed mind guards. These symptoms resulted in specific defects including an incomplete survey of alternatives and objectives, failure to re-examine preferred choices, failure to re-examine rejected alternatives, poor information search, selective bias in processing information, and the failure to develop contingency plans). The final result being a "low probability of a successful outcome" (Janis 1972, 1982).
The first wave of scholarship following Janis’s model is characterized by scholars applying, reinforcing, and amending or expanding the groupthink model. Those applying the theory focused on single case studies confirming, and in some cases excluding, groupthink as a significant cause of policy failure. The term groupthink became expansive and often disconnected from the entirety of Janis’s model, prompting scholars Schafer and Crichlow to note that "If one reads 25 different articles on groupthink, one is likely to see 25 different definitions of the concept" (2017, 1).
Other scholars raised some questions about the soundness of the groupthink model. In their review of groupthink literature, Schafer and Crichlow note one category of scholars as “those who modified Janis’s framework somewhat, or placed more emphasis on one of Janis’s concepts as key” (2017, 4). We identify these scholars as being part of the first wave. They were the first to interact with the model and while they did challenge many aspects of the concept, they did not offer something wholly “other” than groupthink but rather noted weaknesses and offered corrective insights by adding or changing emphasis on specific variables or antecedent conditions. The general focus become centered on groupthink avoidance and how to engage in more effective group decision making. For example, Janis 1982; 't Hart 1990; and Aldag and Fuller 1993, placed heavy emphasis on one of Janis’s key antecedent conditions, group isolation. Gladstein (1984) and Aldag and Fuller (1993) reinforced Janis’s concern with group homogeneity, demonstrating that heterogeneity among group members yields a better quality of group interactions, perspectives, and information in group problem solving. Callaway et al. (1985) stressed another antecedent condition, the importance of the role of the leader. Herek, Janis, and Huth (1987) quantitatively tested and confirmed parts of Janis’s causal chain.
Key scholars developed strategies for avoiding groupthink dynamics. George (1972) designed the multiple advocacy model in which leaders encourage debate and "harness diversity of views and interests in the interest of rational policy making" (751). Johnson (1974) advanced the Competitive Advisory System, in which the leader conceals his views through the process and engineers strenuous debate among group members.
Other scholars expanded on variables Janis neglected to emphasize. White (1998) introduced the term collective efficacy, wherein a group overestimates their likelihood of being effective leading to an increase in high level risk taking. McCauley (1998) raised the problem of uncertainty, arguing that uncertainty introduces the threat of failure in a given situation which induces group members to seek support, increasing premature group compliance and concurrence. Vertzberger (1990) focused on the impact of cultures and subcultures on the decision process. Stern and Sundelius (1997) argued that New Group Syndrome explains conformity because new or ad hoc groups often lack established procedural norms.
The literature on groupthink has become extensive and dominant in the study of small-group foreign policy decision-making (FPDM). While the contributions of the first wave advanced our understanding of many aspects of FPDM, they also formed an unfortunate bias in analysis at the group level such that poor group processes became almost synonymous with Janis’ concept of groupthink.
First wave researchers were instrumental in identifying tensions and inconsistencies in the groupthink model. This was an important step in revealing heterogeneity across situations and group characteristics. However, the baseline continued to be groupthink. While individual scholars challenged and/or refined several aspects of the theory, some presented the dynamics of group decision making as Janis (1972) did in his introduction of the theory such that there are good dynamics (vigilant decision making) and the opposite of good dynamics (groupthink). To illustrate this, McCauley’s 1998 review of groupthink literature stated that "ideal decision making emerges as the inverse of the definition of groupthink" (143). In their 2017 review of groupthink literature, scholars Schafer and Crichlow explained that "What Janis described fundamentally was a limited, circumscribed decision making process that was not thorough or complete or unbiased” (4). They concluded, “We can call all of those parts and pieces something like the syndrome of groupthink, or we could simply call it the opposite of careful, thorough, rational decision making" (ibid). The success of the concept of groupthink had a rather astonishing blinding effect to the possibility of its polar opposite. This is likely because its opposite was conceived of, by many leading scholars, as effective group processing. The dualistic framework of groupthink vs. good began to shift with second wave scholars which we discuss under the term polythink.
Mintz, Mishal, & Morgan (2005) presented the group decision-making model of polythink, which was further developed by Mintz and Wayne (2014, 2016). These authors defined polythink as “a group dynamic whereby different members in a decision-making unit espouse a plurality of opinions and offer divergent policy prescriptions, and even dissent, which can result in intra-group conflict and a fragmented, disjointed decision-making process” (2016, 3). They presented the symptoms of polythink as: greater likelihood of intragroup conflict, greater likelihood of leaks, confusion and lack of communication; greater likelihood of framing effects; adoption of positions with the lowest common denominator; decision paralysis; limited review of policy options; and no room for reappraisal of previously rejected policy options. They outline five key explanations of the dynamic as: (1) the institutional 'turf wars' explanation, (2) the political explanation, (3) the normative explanation, (4) the expert/novice explanation, and (5) the leader/followers explanation. Mintz and Wayne noted that polythink is a contingent phenomenon "dependent on a variety of factors such as government structure (e.g., parliamentary vs. presidential" and advisory group structure (competitive vs. collegial)" as well as leadership style. (2016b, 11). Similar to Janis, they also list and stress the importance of specific contextual variables.
Through the conceptualizing of the term polythink and the broad application and testing of it in multiple foreign policy case studies, scholars have demonstrated the soundness of the concept and the need for this paradigm shift in the study of group dynamics. The model has been effective in shedding light on defective group processes in foreign policy decision making that have been concealed by the great shadow cast by the dominance of the groupthink model.
In what we identify as the second wave of group dynamics scholarship, several scholars began to challenge Janis' central concept of group consensus seeking as the dominant cause of dysfunctional group dynamic. Some of the earliest work in this vein we find in Callaway et al. (1985) who demonstrated that groups composed of high dominance members with dominant leaders make "better decisions in faster time". Diehl and Stroebe (1987) demonstrated results that group productivity loss is associated with increased group participation due to group members’ having to wait to express an idea thus becoming distracted and unable to remember their own ideas. They were also less capable of generating new ideas. Aldag and Fuller (1993) touched on the possibility that groups may become burdened by a deluge of poor suggestions from individual members. Scholars ‘t Hart (1998) and Stern and Sundelius (1997) focused on uncovering different group interaction patterns, acknowledging that groupthink should be considered as one of many explanations of failed group decision making.
‘t Hart (1998b) emphasized the need to move beyond the groupthink model and consider that small groups function in many different ways. He encouraged researchers to consider that, in some cases, competing values within a decision making unit may result in “the inability to resolve conflict and to reach any collective decision at all” (ibid, 312). In response to Janis's method for preventing groupthink, as well as George's multiple advocacy (1972) model and Johnson's competitive advisory system (1974), 't Hart argued that the tactics would “at best…slow down the policymaking process; at worst, they may inadvertently result in a breakdown of collegiality in government"(ibid, 316). He was concerned that "By institutionalizing dissent and opening up group deliberation to a wide array of outside forces, the decision making process may break down under political factionalism and bureaucratic in-fighting” (ibid, 316). His perspective helped to broaden the scope of academics regarding the multiple possibilities for group decision making failures. These groupthink avoidance tactics are distinct from the polythink phenomenon. Mintz and Wayne note that "polythink is more a reflection of the group dynamic. Leaders do not seek or engineer it. It is a description of reality" (2016, 34). They explain how groupthink avoidance strategies can "descend into destructive polythink if not effectively managed by the leader" (ibid).
Thus, in the second wave, several scholars ebbed away at the notion of groupthink as the governing concept of dysfunctional group dynamics. Mintz, Mishal, & Morgan (2005) took an important step in removing undue emphasis on consensus seeking as the vanguard of group dysfunctions by identifying and conceptualizing polythink, the diametric opposite of groupthink. The dominant question in the second wave became “not whether groupthink fits or can be made to fit ‘the facts’ of the case in question, but rather which of a number of competing or complementary group interaction patterns best captures what went on in the decision group or groups under study” (Stern and Sundelius 1997, 125). Stern and Sundelius reviewed patterns in the literature and concluded that “because these interaction patterns are drawn from differing discourses and idiosyncratic conceptualizations by individual authors, together they do not yet meet the standard typological criteria of mutual exclusiveness and collective exhaustiveness. (1997, 125). This begins to shift in the third wave which we associate with and discuss under the concept of con-div and the decision making continuum.
Any theory of how decision making can go wrong “must contain at least the seeds of a theory of how decision making can go right” (McCauley 1998, 143). Mintz and Wayne (2016) developed the concept of a successful group dynamic called “con-div” in which “the convergence and divergence of group members' viewpoints are more or less balanced and in equilibrium” (2016, 9). They explained that in a con-div dynamic, the group does not suffer from a dominating dynamic of either groupthink or polythink but is able to balance and utilize converging and diverging view-points in the process of making a decision. The dynamic, like groupthink and polythink, exhibits particular symptoms which these authors identify as: a clearer policy direction than in polythink with little or no confusion over the policy direction, fewer group information processing biases than in Groupthink, less likelihood of ignoring critical information than in groupthink, operating in one voice, too much harmony that may hinder real debate, less likelihood of decision paralysis, and finally a greater likelihood of a “good” decision compared with groupthink or polythink (ibid).
Mintz and Wayne (2016) demonstrated symptoms of the con-div dynamic in the decision making of the Bush administration regarding the 2006 Surge in Iraq. Other articles have referenced the con-div dynamic and the group decision-making continuum (Mintz and Wayne 2016b, Mintz and Schneiderman 2017, Sofrin 2017, Barr and Mintz 2018). The model has fewer case study examples than the groupthink or polythink models due to (1) the newness of the model. It was introduced in 2016. And (2) the model deals with effective decision making and many case studies in FPDM deal with situations of defective decision-making. However, what sets the con-div model apart from the other models of how decision making can go right is the more holistic decision making process it advocates. Its presence at the center of the group decision-making continuum, and within academic literature on group decision-making dynamics, is essential for its normative as well as practical utility. An excellent avenue for future research would be to define and clarify where the boundaries lie between groupthink and con-div and con-div and polythink. We expand upon this below in terms of the direction for the third wave of group dynamics research.
For the better part of the last half century of academic research into group decision making processes, the field has lumbered under the dichotomy of good versus groupthink. This resulted in what Schafer and Crichlow (2017) identified as "major tensions in the groupthink literature" including contradictory findings related to core elements of Janis's theory such as the effects of too little consensus versus too much, too little information versus too much, and leaders who are too dominant versus leaders who are too passive.
Scholars catching the third wave of group dynamics research conceive of group dynamics as existing on a continuum where good processes and interactions are found as the balance between extremes. This is reflective of Aristotle's golden mean wherein a virtue lies between the excess of a particular vice on one end of the spectrum and the deficiency of the particular vice on the other end. For example, con-div is the balance between too much group harmony on one end (cohesive groupthink) and too little harmony on the other end (conflictive polythink).
Third wave scholars are demonstrating symptoms and cases along the entire continuum of group decision-making, indicating the variations of group dynamics possible yet within a comprehensible framework for analysis. The group decision making continuum provides an effective framework allowing scholars to identify important anomalies or puzzles such as: what level of group conflict is productive and at what level does it become counter-productive; at what point does information overload the process and at what point is it too constricted; what type of information is productive and what type is counterproductive? These and other threshold questions are more easily asked and understood in light of the continuum framework (ibid). There is a need within the field for scholars to construct and apply a scale to the continuum framework. For example, the absence of X or Y or Z, or the presence of A or B or C, as well as how the “balanced” area of con-div is distinct from the extremes of groupthink and polythink. Answering these questions will offer more effective and applicable advice to practitioners by helping them to identify when and how dysfunctional dynamics (across a broad spectrum) are likely to be reversed and moved into a more functional direction. And as Stein (2017) noted, "the better the baseline, the better the choice of puzzles and the more productive the research agenda" (S259).
The present wave of scholarship continues to include characteristics of the first and second wave. For example, Dubé and Thiers (2017) utilized the groupthink model to explain contradictory patterns related to integration processes in Latin America. Forsberg and Pursiainen (2017) also utilize the groupthink dynamic as part of their explanation of Russia's annexation of the Crimean Peninsula in 2014. Other scholars are engaged in the "groupthink or its opposite" debate. For example, Monroy and Sánchez (2017) challenged findings of the groupthink model in explaining the Latin American foreign policy decision processes while negotiating with the United States about the major financial aid package, Plan Columbia. They found that decision makers benefited from group cohesiveness and concurrence-seeking tendencies. Kelman, Sanders, and Pandit (2017) argue that "what distinguishes outstanding executives from others is not vigilance but decisiveness. They argue that, more than groupthink, the danger within government decision making may be “paralysis by analysis”.
Several recent publications are demonstrating the effectiveness and broad applicability of the polythink model. Scholars have yielded key insights about the negative impact of polythink in key U.S. foreign policy failures such as Pearl Harbor and 9/11 as well as the Israeli foreign policy fiasco of the Yom Kippur War (Mintz and Schneiderman, 2017). Sofrin (2017) analyzed the group dynamic within the sub-groups of foreign policy decision-making in the Israeli government. Barr and Mintz (2018) analyzed group dynamics in the U.S. administration and in the U.S. led coalition against ISIS. Expanding beyond U.S. foreign policy cases, Maor, Tonsun, and Jordan (2017) point to how dysfunctional group dynamics of groupthink and polythink can be linked to disproportionate policy responses to climate change. Kelman, Sanders, and Pandit (2017) pointed to one of the symptoms of polythink "paralysis by analysis" in their challenge of groupthink model assumptions. These findings have been important in confirming the polythink theory as well as the group dynamic continuum framework.Assessment: The first wave of research focusing on groupthink avoidance and how to engage in more effective group decision making was key in demonstrating the importance of small group processes in the outcome of foreign policy decisions. The second wave of group dynamic research expanded the concept of dysfunctional group dynamics beyond groupthink by presenting effects in the opposite direction. This was an important step in identifying tensions and contradictions in theories and models of group dynamics. Third wave scholars have moved beyond the old tensions of "competing perspectives" and are poised to address more complex questions relating to thresholds as well as situational contexts. We now move from the connecting stages of the policy cycle to the analysis of the FPDM group decision making models.
All policy groups and all policy contexts are obviously not identical. Stern and Sundelius (1997) noted that despite this, “the bulk of small group literature, is directed at the case where a single, stable, easily identifiable, and (in the case of foreign policy literature) top-level decision group, is the relevant decision unit" (146). They posited that “a succession of radically different interaction patterns might well emerge if one tracked group decision making over time through the course of a particular crisis or policy problem.” The policy cycle framework offers an excellent framework to trace decision making through different contexts. Stein recently expressed the need for this in FPA stating that while scholarship has given great weight to the decisions of the presidents who reserve to themselves the central decisions in foreign policy that “it is in the framing of decisions and in the implementation phase that more complex models are necessary (2016, 144). Consequently, it is incumbent upon scholars to engage extensively, not only with "the decision" but also with the pre and post decision stages. The policy cycle is formulated precisely for this type of analysis. As discussed above, each stage of the cycle has specific demands and dilemmas. We now demonstrate how these stage-specific factors are associated with symptoms of particular group decision-making models. Specifically, we focus on policy formulation, decision making, and implementation.
During formulation, the demand placed on the group is to identify and craft a set of policy alternatives and to narrow that set of solutions in preparation for the final policy decision (Sidney 2006, 79). The requirement to provide multiple and diverse options at the stage of policy formulation facilitates the more positive and constructive con-div dynamic within a decision unit. There are circumstances where specific individuals, known as policy entrepreneurs (Kingdon 1995; Smith and Larimer 2009; Zahariadis 2014, David 2015), manipulate the process even in this early stage. However, the demand for inclusive options makes this far more difficult at this stage than at the stage of decision-making. The demand for broad, yet feasible, solutions means that there will be fewer group information processing biases and less likelihood of ignoring critical information at this stage than at the stages of decision-making and implementation. This is reinforced by the general practice of policy makers to include specialists for the solicitation of advice. Because information in today's highly complex world is no longer centralized, architects of policy increasingly invite specialists and institutions to play an increasingly important role in communicating knowledge (Albaek, Christiansen, and Togeby 2003). There may be a tendency for the group to operate in one voice, however, this is less likely at this stage than other stages. Contrary to another con-div symptom, we do not expect that there is likely to be too much harmony due to the demand for a choice set across a spectrum of options (Bobrow and Dryzek 1987). We identify five of six con-div symptoms as more likely at this stage than during decision making or implementation. Scholars associated the formulation of the Bush administration's Surge decision in Iraq with the con-div dynamic, explaining that "Unlike the re-invasion period…the Administration's decision makers strongly benefited from the diverse and conflicting points of view regarding the best strategy for moving forward in the Iraq War" (Mintz and Wayne 2016, 58).
Several studies have demonstrated that at the stage of decision making, the dominant group dynamic is reflective of the groupthink model. In an effort to meet the stage-specific demand to analyze a reduced set of options in preparation for selecting the most optimal course of action to accomplish the strategic goal, the group becomes more susceptible to groupthink dynamics. As expressed by Janis' model, in order to reach a collective decision, group members are more likely to suffer from groupthink symptoms than during formulation and implementation. These include premature consensus, close-mindedness, overestimation of the group's power and morality, rationalization to discount warnings, pressure toward uniformity such as self-censorship, the illusion of unanimity, and pressure on dissenters through self-appointed mind guards. Six of seven symptoms of groupthink are more associated with the decision-making stage. Stereotyped views may be just as likely in the stage of formulation. However, no symptoms of groupthink are "more" likely during formulation or implementation. Most of the groupthink literature is centered on this stage of the policy cycle. Janis's case studies relating to groupthink decision-making dynamics related to Japan’s attack on Pearl Harbor (1941), the escalation of the war in Korea (1950), Bay of Pigs operation (1961) and the involvement of the U.S. in the Vietnam conflict under President Johnson focus on the decision stage. More recent examples of groupthink during this stage include the 2003 decision of the Bush administration to invade Iraq (Mintz and Wayne 2016), the 2014 decision of the Obama administration to fight ISIS in Iraq and Syria (ibid), the 2014 decision by Russia to annex the Crimean Peninsula (Forsberg and Pursiainen 2017), and the policy decisions related to integration processes in Latin America (Dubé and Thiers 2017).
During the stage of implementation, the demand is for the group to translate macro-level strategic objectives into micro-level tactical demands within an environment in flux. This stage-specific demand, and resulting dilemmas make the decision unit during implementation more prone to a polythink dynamic. We associate five of seven polythink symptoms as more dominant during implementation, including: intragroup conflict, leaks, confusion; lack of communication; adopting the lowest common denominator positions; decision paralysis; and limited review of policy options. Two of the symptoms of polythink may be also be likely in both the preceding stages of policy formulation and decision-making. These include: no room for reappraisal of previously rejected policy options and framing effects. There is existing literature demonstrating the polythink dynamic during the process of implementation. Sofrin (2017) documents the polythink dynamic within the Israeli cabinet's decision making during the implementation of operation Protective Edge. Mintz and Wayne (2016) uncover the polythink syndrome during the implementation of the Obama administration's initial decision to combat ISIS in Syria wherein decisions related to the tactics of whether to use force, sanctions, or support the rebels resulted in a dominating polythink dynamic. The decision of Israel on October 6, 1973 not to pre-empt with an attack on Syria and Egypt was also a product of the polythink syndrome (Mintz and Schniderman 2017).
The previous section highlighted how scholars have utilized three leading models of group decision making dynamics - groupthink, con-div, and polythink — to explain foreign policy decisions. However, these scholars have not linked group dynamics to stages of the policy cycle or demonstrated how group dynamics may shift through stages of the same foreign policy decision. In the next section we illustrate how this could be done. While our illustration only presents an example of connecting the policy cycle framework with group decision making models, it reveals how it can yield fruitful insights in this direction. Mintz and Barr (2018) demonstrated how polythink in the Obama administration led to paralysis in the 2016 campaign against ISIS in Raqqa, Syria. We illustrate here how this case study actually focused on the stage of implementation. We further illustrate that in this same decision regarding the 2016 campaign against Raqqa there is evidence of a con-div dynamic during the stage of policy formulation, whereas during the actual decision-making stage, the group dynamic is more reflective of groupthink. This example demonstrates how dominant group dynamics can be related to the policy cycle and how these group dynamics can shift from stage to stage.
Members of President Obama’s National Security Decision-Making Unit relating to operations against ISIS included, but were not limited to, Vice President Joe Biden, Secretary of State John Kerry, National Security Council (NSC) Advisor Susan Rice, Secretary of Defense Ashton Carter, Chairman of the Joint Chiefs of Staff General Joseph Dunford, Director of the Central Intelligence Agency (CIA) John Brennan, Director of National Intelligence James Clapper, White House Chief of Staff Denis McDonough and Middle East director on the NSC Robert Malley, also known in the press as the ISIS Czar (Chuck 2015) . During the stage of implementation there is also evidence that Ambassador to the United Nations Samantha Power and Ambassador to Turkey John Bass were active and influential advisors within the decision unit (Entous, et al. 2017).
In the fall of 2015, Secretary of Defense Ashton Carter presented the President with options for how to deal the final blow to ISIS in Syria and deprive them of territorial holdings in their last major stronghold in Raqqa (Department of Defense [DOD], 2016). The con-div dynamic characterizes the policy formulation process within the Obama administration. Secretary Carter, a recent replacement of defense secretary Chuck Hagel, presented balanced, careful and comprehensive options to the President. Press briefings reveal several options across a broad spectrum including: (1) going it alone with American troops on the ground, (2) refraining from any intervention, and (3) recruiting, training, and facilitating local troops on the ground to defeat ISIS (ibid).These options were reviewed by multiple agencies. Typically, presidents hold National Security Council meetings at the White House. However, in an effort to create and “illustrate the multifaceted U.S. approach to defeating IS” (Lederman 2016), President Obama held NSC meetings at the Pentagon, CIA, and State Department. While official plans were drawn up by the Pentagon by Secretary Carter and the Joint-Chiefs, officials from the Department of State and the CIA were engaged in the process. Defense officials directly referenced intelligence estimates in connection to the urgency of the campaign against ISIS in Raqqa (DOD 2016b, 2016c). CIA director Brennan was very vocal in his discussion of the external threat emanating from Raqqa as the center for plotting and planning of global terror attacks (Starr 2016).
The con-div dynamic facilitated the inclusion of multiple opinions. Beyond the inter-agency cooperation, Secretary Carter invested months soliciting advice from multiple and diverse actors including top generals down to combat soldiers in the field as well as international leaders within the U.S. led coalition (DOD 2016). The inclusion of multiple perspectives made it far less likely for the president to ignore critical information than under a groupthink dynamic. And though it took time to consult with individuals and groups outside of the decision unit, this did not significantly delay action or result in paralysis. Also, the stage-specific demand for the inclusion of multiple options regarding tactical operations made decision paralysis less likely. Based on evidence available of the administration’s efforts to involve advisors of multiple government agencies, as well as input from generals and troops on the ground, perspectives were clearly balanced and productive, yet the administration was able to speak for the most part in one voice with regard to the broad mission against ISIS. Secretary Carter has written about his extensive effort to shift the mission focus away from the “degrade” aspect with a focus on a “lasting defeat” of ISIS. (Carter, 2017). This exact wording is evident in the 2016 messaging of administration officials: in a speech by the President (White House 2016b), comments by NSC Advisor Susan Rice (White House 2016), remarks by Vice President Biden (White House 2016c), messaging of the White House Press Secretary Josh Earnest (2016d), as well as many other top officials and military leaders (DOD 2016c, 2016d). Yet not all con-div symptoms fit the formulation stage of the Raqqa campaign. Anonymous sources revealed that there were top advisors warning against toppling ISIS in Raqqa within the established time frame for fear that it would lead to increased regional instability (Rogin 2016). However, the policy formulation process reveals a purposeful decision unit engaged in a balanced, careful, and comprehensive effort to formulate broad policies based on multiple, diverse, and diverging opinions. Overall, six of the seven symptoms of con-div fit the formulation phase of the Raqqa Campaign. We now turn to the stage of decision-making.
Group dynamics during the decision-making stage in the Raqqa campaign were very different from the stage of policy formulation. Unlike the balanced and careful con-div dynamic that characterized the formulation of the policy, the group during the decision-making stage reflected a pattern of groupthink symptoms. To begin with, the antecedent conditions of groupthink were textbook. There was strong group cohesiveness. Once the U.S. had engaged in Syria, there was unusual bipartisan support in the highly polarized Congress. In fact each side was pressuring Obama to “increase the pace and intensity of the campaign against the Islamic State following the recent terrorist attacks in San Bernardino, California, and Paris” (Jaffe 2015). Secondly, Obama acted as an impartial leader, “direct[ing] aides to examine all proposals that could accelerate the fight against the Islamic State,” explaining that he wanted an offensive "well underway before he left office” (Reuters Staff 2016). Thirdly, there was an intense provocative situational context with each international terror attack by ISIS. Secretary Carter asserted that the attacks “underscored, for anybody who doubted it, the necessity of this campaign” (DOD 2016). In October 2016, Carter announced the timeframe for the Raqqa Campaign, saying that it would take place "within weeks", and affirmed the administration's decision to “enable local, motivated forces" to fight against ISIS in Raqqa (ibid). He justified the urgency, citing intelligence estimates of the CIA (Starr 2016).
In accordance with the symptoms of groupthink, it is clear that administration officials demonstrated an overestimation of their capabilities to maneuver proxies to accomplish the strategic objectives of the U.S. This is exemplified in Carter's comment that “the United States is not just the finest fighting force the world has ever seen…, but as Ramadi showed, our unmatched capabilities can enable and multiply the power and force of our local partners (ibid). Reflective of the groupthink dynamic, individuals within the decision unit exhibited overconfidence about U.S. capabilities. Secretary Carter's statement reveals that the administration firmly believed the U.S. had the power to maneuver the situation and bring the victory in their established timeframe (DOD 2016). Administration officials locked into a preferred course of action, to utilize an essentially Arab force to liberate Raqqa (Wintour 2016). The administration discounted warnings about the battle readiness of the local Arab forces. An earlier attempt at using local Arab fighters to fight ISIS in Syria provided graphic evidence of the dangers of this option. In July 2015 the first U.S.-backed Arab fighters in Syria were dead or missing just days after being commissioned (Miklaszewski 2015). The administration discounted this event, selectively interpreting information in favor of their strategic objective, as is predicted by the groupthink model.
Other key groupthink symptoms are evident at the decision-making stage. With the resignation of Defense Secretary Hagel, rumors swirled in the press that there was pressure toward conformity within the administration (Goldberg 2014). And as Janis predicted, this leads to self-censorship and the illusion of unanimity. While there was near unanimity related to the strategic goal to defeat ISIS in Raqqa, the specifics regarding tactical questions were anything but settled. This is evident in the statement of General Votel, commander of U.S. Special Operations who testified to the Senate Armed Services Committee, that he had "concerns" and when questioned about the capabilities of the U.S. to recapture both Mosul and Raqqa, and he testified that “there have yet to be plans developed to take away ISIS's Syrian capital of Raqqa” (Guardian staff 2016). Obama's team of advisors demonstrated a pattern of groupthink wherein the strategic objective was quickly agreed upon without due diligence to the tactical complexities. This becomes more apparent in the stage of implementation, to which we now turn.
In the stage of implementation, the demand of a decision-unit is to translate strategic objectives into tactical operations. This often involves addressing the complex questions that are often brushed aside in earlier stages of the policy process (O’Tolle, 2000). Despite public announcements by U.S. and international coalition leaders that the liberating force would be a local Arab force, the internal decision making during implementation was still focused on what composition of forces could actually accomplish the strategic objective within the timeframe and without repeating the embarrassing and costly debacle of 2015 (Schmitt 2016). This led to serious disagreements (Entous, et. al. 2017). The tactical options included liberating Raqqa with the battle ready YPG Kurdish fighters, the broadly accepted but untested Arab only force, or an Arab force led by Turkish troops. These tactical options were discussed during formulation, but no immediate decision was required. During the decision-making phase the administration chose to use the term "local forces" avoiding the tactical choice of the exact composition of forces (DOD 2016, 2016c). However, in the stage of implementation, soldiers needed to be deployed and the decision of which forces would best achieve the strategic objective needed to be made.
Within Obama’s administration there were long and contentious debates over the force structure (Entous et al. 2017). Two sub-groups emerged. Secretary Carter as well as Chairman of the Joint Chiefs of Staff, General Dunford, were in favor of arming the Syrian Kurds directly. The competing perspective came from National Security Advisor Susan Rice, Ambassador to the UN, Samantha Power, and Ambassador to Turkey, John Bass. These advisors to the President adamantly opposed the use of Kurdish YPG fighters on the grounds that it would severely compromise U.S. relations with Turkey. They argued for local Arab fighters reinforced by soldiers promised by Turkey (ibid).
The policy was fought out between two distinct institutional lines, the State Department and the Pentagon, each perceiving their needs to be in direct competition with the needs of the other. The result would be “dozens of meetings of President Obama’s top national security team, scores of draft battle plans and hundreds of hours of anguished, late-night debates” (ibid). Former Secretary of Defense Hagel’s account of meetings directed by Susan Rice sheds some light on the process inside the National Security Council which typified polythink. He complained that "everybody had a chance to talk. We rarely got to a conclusion or a decision, [with] too many people talking, and I think that always leads to an ineffective process.” (Kirk et al. 2016). Pointing to another symptom of polythink, he discussed a fear of leaks, noting that there were unknown staffers at White House meetings making it difficult for him to express his views noting that “The more people you have in a room, the more possibilities there are for self-serving leaks to shape and influence decisions in the press” (de Luce 2015). As predicted by the Polythink model, there was a significant delay and decision paralysis. On January 17, 2017, the team placed the decision in the hands of the Trump administration, handing over a document with an official recommendation to arm the Kurds and a memo on how to explain this to Turkey (Entous et. al. 2017). As predicted by the polythink model, the president’s foreign policy team had been fragmented and conflictive over the implementation of the initial decision and unable to advance a key objective of the President that had initially been widely accepted within his administration, by the American public and a broad coalition of global leaders (DOD 2016b, Ensor 2016).
In the Raqqa offensive example, we illustrated group decision-making dynamics in the context of policy cycle stages. We demonstrated how dominant group dynamics in this particular foreign policy decision-making process shifted throughout stages of the policy cycle. The more successful group dynamic of con-div was dominant during policy formulation. However this quickly shifted to the sub-optimal group dynamic of groupthink at the decision-making stage. Finally, the polythink dynamic dominated the stage of implementation when tactical questions required action. A single illustration has little empirical, external validity, but the example should encourage scholars to connect the fields of public policy and foreign policy decision making and continue to test the assertions presented in this article.
Kasey Barr is a doctoral student at the Hebrew University of Jerusalem and contributing scholar in the Behavioral Political Science Lab of the Interdisciplinary Center (IDC) Herzliya.
Alex Mintz is provost of the Interdisciplinary Center (IDC) Herzliya, in Israel, and President of the Israeli Political Science Association.
We would like to thank the reviewers for their constructive comments which aided in the improvement of our review. We also would like to thank the editorial team of Hank Jenkins-Smith, Nina Carlson, and Julie Krutz from the University of Oklahoma for the opportunity to contribute to the Policy Studies Journal.
1 We acknowledge the relevance and importance of other models of group decision-making, such as organizational processes and bureaucratic politics. However, we have not included such models because they are not included within the conformity-plurality continuum (Mintz and Wayne 2016) and are categorically different. See Hudson (2016, 19-21) for a discussion of the unique research questions and agendas in areas focused on group decision-making dynamics and those areas of inquiry into organizational processes and bureaucratic politics.