An Advocacy Coalition Framework Approach to Stakeholder Analysis: Understanding the Political Context of California Marine Protected Area Policy

There is a growing recognition that public policy controversies are driven more by value differences than by technical deficiencies. Unfortunately, we have yet to develop, test, and refine systematic approaches for understanding political systems. In this article I explain how the advocacy coalition framework (ACF) can be used as a theoretical basis for understanding political context via a stakeholder analysis. An ACF stakeholder analysis widens the attention of policy analysts toward subsystem-wide dynamics with multiple actors who are motivated by their beliefs, structure their relationships into advocacy coalitions, and try to influence policy through utilizing multiple resources and venues. I illustrate an ACF approach to stakeholder analysis in a scientifically contentious political conflict over the establishment of marine protected areas in California. I conclude with a summary of contributions to the ACF literature and the strengths and limitations of conducting an ACF stakeholder analysis. The Death of Environmentalism (Shellenberger and Nordhaus 2004) criticized the environmental movement for defining the underlying causes of environmental problems as technical deficiencies rather than value conflicts. While Shellenberger and Nordhaus’s criticism shocked the environmental community, they were not the first to argue that environmental or policy conflicts were driven more by differences in values than analytical shortcomings in technology and science (Dror 1967; Fiorino 1990; Jenkins-Smith 1990; Kingdon 1994; Mazur 1981; Meltsner 1972; Sabatier and Jenkins-Smith 1999). Mazur (1981, 41) probably said it best: ‘‘Many technical controversies are primarily disputes over political goals and only secondarily concerned with the veracity of scientific issues which are related to these goals.’’ I wish to thank Monica Gaughan, Peter May, Paul Sabatier, and three anonymous reviewers for their feedback on this manuscript. Barry Bozeman also provided comments. But since he is a disagreeable old reprobate I am not going to thank him! Indeed, any remaining errors are due entirely to his commentary. I also wish to express my gratitude for the research support from Christal Love and Maryann Hulsman and for the cooperation of many stakeholders who participated in this study. This research was supported by the California Natural Resources Grant and the Jastro-Shields Scholarship. Address correspondence to the author at [email protected]. doi:10.1093/jopart/muj015 Advance Access publication on April 26, 2006 a The Author 2006. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: [email protected]. JPART 17:95–117 at A uaria Lrary on N ovem er 3, 2010 jpart.oxjournals.org D ow nladed fom For policy analysts dealing with analytically intractable policy issues, one implication from this observation is to focus not only on conducting a high-quality, technical analysis (such as a benefit-cost analysis) but also on developing a good understanding of the political context of the problem. Most policy analysis textbooks agree that understanding political systems is an important step in recommending alternatives (Patton and Sawicki 1993; Weimer and Vining 2005). Unfortunately, there has been relatively little effort devoted to developing a theoretical framework to guide policy analysts in understanding policy disputes. The goal of this article is twofold: (1) to explain how the advocacy coalition framework (ACF) can be used as a theoretical basis for understanding policy disputes via a stakeholder analysis and (2) to demonstrate an ACF approach to stakeholder analysis with a case study of marine protected area (MPA) policy in California. I begin this article with an assessment of the typical methods of conducting a stakeholder analysis. I then explain how to conduct an ACF stakeholder analysis and why an ACF approach addresses some of the deficiencies of other approaches to stakeholder analysis. I illustrate an ACF stakeholder analysis with original data on California MPA policymaking. I conclude with a discussion of potential users and the strengths and limitations of conducting an ACF approach to stakeholder analysis. STAKEHOLDER ANALYSIS Stakeholder analysis is defined by identifying opportunities and constraints for calculating the likelihood that a strategy, venue, or alternative will be successful in initiating or preventing belief and policy change. Almost all applications of stakeholder analysis address a similar set of questions (Brugha and Varvasovsky 2000; Crosby 1991; Susskind and Thomas-Larmer 1999). These questions include: 1. Who are the stakeholders to include in the analysis? 2. What are the stakeholders’ interests and beliefs? 3. Who controls critical resources? 4. With whom do stakeholders form coalitions? 5. What strategies and venues do stakeholders use to achieve their objectives? In general, stakeholder analysis helps policymakers conceptualize the dynamics of a policy subsystem. It has a broader perspective than political feasibility analysis, which tends to focus on the probability of successfully implementing a particular policy alternative for a particular problem (May 1986; Meltsner 1972; Weimer and Vining 2005). Instead, stakeholder analysis focuses on mapping the activities of multiple stakeholders employing multiple strategies in multiple venues. This broad perspective is extremely important in the United States or any other political system that is open to multiple participants, where authority is shared among actors and institutions, and where there are ample opportunities for losers of one policy battle to strike back against winners later in the policy process (Weber 1998). 1 There are many definitions of stakeholder analysis (Brugha and Varvasovsky 2000; Crosby 1991; Grimble and Wellard 1997). This definition is based mostly on May (1986) and from the premises of the ACF (Sabatier and JenkinsSmith 1999). Venues are institutional arenas within which stakeholders have the opportunity to influence policymaking. Journal of Public Administration Research and Theory 96 at A uaria Lrary on N ovem er 3, 2010 jpart.oxjournals.org D ow nladed fom To help navigate this political landscape, stakeholder analysis provides a guide to investigate stakeholders’ perceptions regarding the severity, causes, and proposals of a problem, the distribution of resources among coalitions, and the accessible political venues for influencing policy. This helps policy analysts, leaders, and other stakeholders identify political roadblocks, develop strategies for achieving objectives, and find paths to collective agreements. The beneficiaries of a stakeholder analysis tend to be policy participants in leadership positions, such as directors and program managers, who work above the street level and who need to develop a systematic map of a political community. Policy analysts can also use stakeholder analysis to make recommendations to clients if the alternatives they are considering are sufficiently broad in scope or are long-term strategies. It is somewhat surprising that policy analysis, which trains students to deal with policy problems with political ramifications, deals more with technical feasibility than with political feasibility. This might be because teaching analytical techniques is easier and has a longer pedagogical history than teaching political competency. Indeed, the leading American textbooks in policy analysis spend just a few pages or, at the most, a chapter on understanding political contexts (Bardach 2005; Patton and Sawicki 1993; Stokey and Zeckhauser 1978; Weimer and Vining 2005). This neglect is even more surprising considering that the policy analysis literature has long recognized that the effectiveness of technically astute policy analysts is limited because of their inattention to politics (Dror 1967; Jenkins-Smith 1990; May 1986; Meltsner 1972; Radin 2000; Weiss 1977). Despite its sparse coverage in policy analysis textbooks, stakeholder analysis has not been completely neglected (for a review see Brugha and Varvasovsky 2000). Good applications can be found in health policy (Glassman et al. 1999), natural resource/ environmental policy (Grimble and Wellard 1997; Ramirez 1999), strategic management of private and public organizations (Freeman 1984; Lindberg and Crosby 1981), policy evaluation (Brown et al. 2001; Gregory and Wellman 2001), and collaborative policymaking (Carpenter and Kennedy 1988; Susskind and Thomas-Larmer 1999). However, a review of this literature suggests at least three major limitations to current approaches to stakeholder analysis. First, some have noted that stakeholder analysis is easily outdated, making the implicit assumption that stakeholder coalitions, beliefs, resources, and strategies change too rapidly to make a systematic stakeholder analysis worthwhile (Brugha and Varvasovsky 2000). In this article I outline a theoretical framework to help understand the frequency and magnitude of change in a policy subsystem and the likely effects on the different parts of a stakeholder analysis. Second, many stakeholder analysts create typologies or matrices of key stakeholder variables, such as stakeholder alliances, interests, policy positions, and resources, but do not specify a theoretical basis for explaining the causal interdependence among these variables and how a combination of these variables affects belief and policy change (Crosby 1991; Meltsner 1972). In this article I outline a theoretical framework that ties the essential components of stakeholder analysis together. Third, like political feasibility analysis, most applications of stakeholder analysis focus on a single alternative or venue (Crosby 1991; Ramirez 1999; Susskind and Thomas-Larmer 1999). For example, Susskind and Thomas-Larmer (1999) used conflict assessment to estimate the likely success or failure of one venue (a consensus-based process). Conducting a stakeholder analysis on one particular venue or alternative limits the substantive reach of the policy conclusions and the utility over time. To complement Weible Advocacy Coalition Framework Approach to Stakeholder Analysis 97 at A uaria Lrary on N ovem er 3, 2010 jpart.oxjournals.org D ow nladed fom political feasibility analysis, we need a theory that focuses on a wider subsystem scope, recognizing that stakeholders typically are not concerned with just one policy venue or alternative but with the outcomes of an entire policy subsystem over long periods of time. In this article I outline a theoretical framework that broadens the scope of analysis to include an entire policy subsystem, thereby increasing the utility of stakeholder analysis. AN ADVOCACY COALITION FRAMEWORK APPROACH TO STAKEHOLDER ANALYSIS One theoretical framework to ground a stakeholder analysis is the advocacy coalition framework (Sabatier 1988; Sabatier and Jenkins-Smith 1993, 1999). The advocacy coalition framework is frequently used to explain stakeholder behavior and policy outcomes in intense political conflicts over periods of a decade or more (Sabatier and Weible 2005). The ACF offers a different approach to stakeholder analysis because the ACF defines a policy subsystem—not a specific venue or alternative—as the most useful unit of analysis and has explicit assumptions and hypotheses regarding (1) the substantive and territorial boundary of the policy issue and who to include in a stakeholder analysis; (2) the structure of individual beliefs and motivations to influence policy, (3) individual motivations to form relationships (into advocacy coalitions); (4) the identification of stakeholder resources and available political venues; and (5) the factors necessary to produce major and minor policy changes. There are many theoretical descriptions of the ACF (Sabatier 1988; Sabatier and Jenkins-Smith 1993, 1999). This section contributes to the theoretical exposition of the ACF by explaining how it can be used to conduct a stakeholder analysis and by describing some of the underdeveloped components of the ACF, such as coalition resources. Policy Problems and Stakeholders Bounded by a Policy Subsystem The ACF assumes that the most useful unit of analysis for conducting a stakeholder analysis is the policy subsystem (Sabatier and Jenkins-Smith 1999). Researchers and analysts define a policy subsystem by a set of policy participants and territorial and substantive scopes. Stakeholders specialize in a policy subsystem and maintain their participation over long periods of time in order to foster, among other reasons, the institutionalization and implementation of policy objectives (Sabatier and Jenkins-Smith 1993). Stakeholders in a policy subsystem include local, state, and federal government officials, interest groups, nongovernmental organizations, community groups, researchers/ scientists, members of the media, and target groups. These stakeholders often carry out several strategies to influence the decisions in several venues. For example, stakeholders might simultaneously pressure political sovereigns, court the media, consider litigation, and try to convince opponents to support their views in public meetings. Stakeholders Motivated by Belief Systems The ACF presumes that individuals are boundedly rational with cognitive constraints (Simon 1985), filter perceptions by a hierarchical belief system (Scholz and Pinney 2 I encourage the use of other theories and frameworks of the policy process (see Sabatier 1999) in conducting a stakeholder analysis. In this article I focus on the ACF because I am familiar with the framework, the ACF has the potential to improve stakeholder analysis, and comparing multiple frameworks for conducting a stakeholder analysis is beyond the scope of this article. Journal of Public Administration Research and Theory 98 at A uaria Lrary on N ovem er 3, 2010 jpart.oxjournals.org D ow nladed fom 1995), remember losses more than gains (Quattrone and Tversky 1988), and therefore, exaggerate the influence and maliciousness of opponents (Sabatier and Jenkins-Smith 1999). The individuals in the ACF are motivated to convert their beliefs into policy but are limited in their ability to do so. Since individuals’ identities are closely tied to their beliefs, they tend to filter or ignore dissenting information or events that challenge their beliefs and readily accept information that bolsters their beliefs. People are also very suspicious of people with dissimilar beliefs and remember lost policy battles more than previous gains. This makes individuals highly susceptible to exaggerating the influence and maliciousness of their opponents, which in turn strengthens their ties with others who have similar beliefs (Sabatier and McLaughlin 1987). Following the public attitude and opinion research (Hurwitz and Peffley 1987; Putnam 1976), the ACF assumes that the defining characteristic of individuals is their three-tiered hierarchical belief system (Sabatier and Jenkins-Smith 1993, 1999). On the top tier are deep core beliefs, which are normative/fundamental beliefs that span multiple policy subsystems and are very resistant to change (for example, political conservatism). In the middle tier are policy core beliefs, which are normative/empirical beliefs that span an entire policy subsystem. The ACF identifies eleven categories of policy core beliefs, including perceptions of the severity and causes of subsystem-wide problems, orientation on basic value priorities directly related to the policy subsystem, the effectiveness of policy instruments, and the proper distribution of authority between the market and government (Sabatier and Jenkins-Smith 1999, 133). Policy core beliefs are still resistant to change but are more pliable than deep core beliefs. On the bottom tier are secondary beliefs, which are empirical beliefs that relate to a subcomponent (either substantively or territorially) of a policy subsystem. Of the three layers of beliefs, secondary beliefs are most susceptible to change in response to new information and events. For stakeholder analysis, measures of the different components of deep core and policy core beliefs are likely to remain stable for long periods of time—making this part of a stakeholder analysis applicable over time. Stakeholders Use Policy Core Beliefs to Structure Advocacy Coalitions The ACF assumes that stakeholders are primarily motivated to convert their beliefs into actual policy and thereby seek allies to form advocacy coalitions to accomplish this objective. Advocacy coalitions include actors of similar policy core beliefs who engage in a nontrivial degree of coordination (Sabatier and Jenkins-Smith 1999, 120). Since policy core beliefs are resistant to change and structure participation in advocacy coalitions, coalition membership is predicted to remain stable for a decade or more, extending its utility for a stakeholder analysis (Zafonte and Sabatier 2004). Stakeholders Utilize Resources Most stakeholder analyses identify a set of resources available to coalitions for achieving their objectives (Meltsner 1972). Based partly on Kelman (1987) and Sewell (2005), the ACF assumes that individuals employ available resources that enable them to carry out strategies in a variety of venues to influence policy. The categories of resources include (Sabatier and Weible 2005): Access to legal authority to make policy decisions. The ACF views many agency officials, legislators, and even some judges as members of advocacy coalitions. When this happens, direct political access and influence become a major resource for an advocacy coalition Weible Advocacy Coalition Framework Approach to Stakeholder Analysis 99 at A uaria Lrary on N ovem er 3, 2010 jpart.oxjournals.org D ow nladed fom (Sabatier and Pelkey 1987). In fact, an important feature of a dominant coalition, in comparison to minority coalitions, is that it has more of its members in positions of formal authority. Public opinion. Opinion polls documenting support for the policy views of advocacy coalitions are a major resource. A coalition with strong public support can argue that it represents the public interests, which can be used to lobby new legislation, to press for changes in rules and regulations, and to shift resources to the coalition’s cause. A supportive public is also likely to elect coalition supporters to positions of legal authority. Consequently, stakeholders within the subsystem typically spend a lot of time trying to convince the public to support their advocacy coalition’s policies and candidates. Information. For the most part, the ACF views information as political salvo to win policy disputes. The ACF assumes that information is utilized by stakeholders to buttress their coalition’s membership, to argue against the policy views of an opposing coalition, to convince decision-making sovereigns to support their positions, and to influence public opinion. In this effort, stakeholders might spin or even distort information to their advantage. Having better information than opponents does not guarantee a policy victory, but it does force opponents to expend additional resources to neutralize the advantage (Sabatier and Weible 2005). This is one of the reasons why the ACF spotlights the role of analysts, researchers, and consultants within coalitions. It also explains why analysis at the policy subsystem level is helpful in explaining the use and effect of policy analyses, reports, and scientific and technical information (Sabatier and Zafonte 2001). Mobilizable troops. The political outcomes often hinge on the mobilization of public supporters and the expansion of conflict (Baumgartner and Jones 1993). Coalitions periodically ask supporters from the general public to participate in events to help achieve objectives. Supporters might be asked to engage in letter-writing campaigns, to provide labor in electoral and fund-raising campaigns, and to participate in public demonstrations and other activities. In fact, coalitions with minimal financial resources typically rely very heavily on mobilizable troops, especially when a coalition’s interests are threatened by reductions in government-sponsored benefits or additional government oversight and regulations. Financial resources. Money is a valuable resource because it can be used to purchase other resources. A coalition with deep pockets can bankroll sympathetic candidates, thereby gaining inside access to legislators and political appointees. A coalition might also finance research and think tanks to generate information to alter the policy process, to sway public opinion, and to mobilize their supporters. Skillful leadership. The literature on policy entrepreneurs demonstrates how skillful leaders can help navigate a coalition toward policy victories (Kingdon 1994). In fact, public policy research describes how most antecedents to policy change—for example, external shocks—dispose a political system to change, but skillful entrepreneurs are needed to bring about actual changes in policy (Kingdon 1994; Mintrom and Vergari 1996). Coalition leaders help articulate a coherent belief system for other coalition members, thereby strengthening their resolve and focus. Leaders also attract additional resources to their coalition, which offer more strategic choices and open venues to influence policy (Mintrom and Vergari 1996; Muller 1995). Stakeholders Attempt to Influence Policy in Venues The ACF predicts that stakeholders will strategically use their resources to influence policy in multiple venues. Venues are institutional arenas within which stakeholders have the Journal of Public Administration Research and Theory 100 at A uaria Lrary on N ovem er 3, 2010 jpart.oxjournals.org D ow nladed fom opportunity to influence policymaking. Stakeholders spend considerable amounts of time venue shopping, looking for institutional access where they might have a competitive advantage. They often launch offenses in several venues and defend their interests in several venues simultaneously. Based partly on Sabatier and Pelkey (1987), potential venues include elections, public referenda and decisions in legislatures, chief executives, courts, and agencies. The ACF predicts that coalitions will choose to employ resources to influence venues based on attaining the most policy benefits per incurred costs (Sabatier and Jenkins-Smith 1999, 142). Initiating or Preventing Major and Minor Belief and Policy Change One of the underlying goals of stakeholder analysis is to help analysts evaluate the political response to policy recommendations and the likelihood that recommendations will lead to the desired outcome, which typically include major changes in policy. The ACF predicts two precursors to major policy change: changes in beliefs of a dominant coalition or changes in available resources and venues. Changes in beliefs or in available resources and venues are brought about by external shocks, policy-oriented learning, or hurting stalemate. External shocks are events that occur outside of a policy subsystem. Examples include changes in socioeconomic conditions, changes in governing coalitions, and impacts from other subsystems (Sabatier and Jenkins-Smith 1999). External shocks dispose a policy subsystem to policy change in two ways. First, external shocks may open or close venues or shift resources because of renewed attention of the public or key sovereigns. This adjusts the power among coalitions, thereby tipping the advantage to a different coalition with different policy core beliefs. Second, external shocks may change components of the policy core beliefs of a dominant advocacy coalition in the policy subsystem. During an economic recession, for example, a proregulatory advocacy coalition may reconsider the adverse economic effects on target populations from stringent controls (Zafonte and Sabatier 2004). The second mechanism of belief change is learning over long periods of time from the gradual accumulation of information, such as a scientific study, policy analysis, and experiences of various local stakeholders (Bennett and Howlett 1992; May 1992; Sabatier 1987). The ACF defines policy-oriented learning as ‘‘relatively enduring alterations of thought or behavioral intentions that result from experience and/or new information that are concerned with the attainment or revision of policy objectives’’ (Sabatier and JenkinsSmith 1999, 123). Policy-oriented learning affects the beliefs of actors within the policy subsystem, which can lead to major policy change. Learning is inhibited, however, because individuals face cognitive constraints and filter or avoid belief-conflicting information. Whereas external shocks can lead to rapid changes in individual policy core beliefs and, consequently, the policy core aspects of a policy subsystem, policy-oriented learning may take ten years or more to generate similar changes. Policy-oriented learning has a larger effect on secondary beliefs, which are more pliable to information, than on policy core beliefs. A third mechanism of policy change is a hurting stalemate (Zartman 1991). A hurting stalemate is a situation in which all parties involved in the dispute view a continuation of 3 The ACF defines major policy change as an alteration of a policy that is subsystem-wide (Sabatier and JenkinsSmith 1999). In contrast, the ACF defines minor policy change as modification of a specific policy, which is a subcomponent of a policy subsystem (changes in secondary aspects of the policy subsystem). Minor policy change occurs more frequently but with a smaller magnitude. Weible Advocacy Coalition Framework Approach to Stakeholder Analysis 101 at A uaria Lrary on N ovem er 3, 2010 jpart.oxjournals.org D ow nladed fom the status quo as unacceptable and run out of alternate venues to achieve their objectives (Sabatier et al. 2005). The assumption is that individuals satisfied with the status quo have little incentive to give up anything in negotiations. Individuals are also unlikely to compromise in negotiations when they have access to other venues where they can win outright. In both situations, negotiating toward a collective agreement is probably a waste of time. Only when both coalitions are out of options and dissatisfied with the current situation are they willing to compromise and negotiate for major shifts from the status quo. Successful negotiations among coalitions require the existence or the establishment of a professional forum (or collaborative institution) based on fair procedural rules and norms (Sabatier et al. 2005). The ACF offers an alternate approach to conducting a stakeholder analysis. It broadens the substantive scope away from a single alternative or venue by focusing attention at the policy subsystem level. It explains why conflict is not about technical deficiencies but about value differences because stakeholders are motivated to convert their beliefs into policy, filter out dissonant information, and structure their interactions within homogeneous advocacy coalitions. It specifies that policy core beliefs are the most important stakeholder attribute for understanding individual behavior. It also provides a theoretical justification for the longevity of stakeholder analysis. For instance, in the absence of external shocks, the ACF predicts that measures of coalition structure and stakeholder policy core beliefs should be stable for a decade or more. After an external shock, resources may be redistributed and venues may change. Even after an external shock, the ACF predicts that stakeholder deep core beliefs and most components of policy core beliefs will remain stable, and basic coalition structure will remain the same (Zafonte and Sabatier 2004). In the next section I conduct a partial ACF stakeholder analysis in a case study of the California Marine Life Protection Act policy subsystem. AN ACF STAKEHOLDER ANALYSIS OF THE CALIFORNIA MARINE PROTECTED AREA POLICY SUBSYSTEM The 1999 California Marine Life Protection Act (MLPA) charged the Department of Fish and Game (DFG) with developing a plan to establish marine protected areas (MPAs) along the coast. MPAs are a space-based management strategy that restrict access to areas of the ocean and limit, to various degrees, extractive activities. The DFG has attempted to implement the MLPA twice and is currently attempting it a third time. The first attempt to implement the MLPA involved a master plan team of scientists. Working behind closed doors, the master plan team created a proposal for the placement of MPAs based only on the natural sciences and with priority toward protecting habitat. During the summer of 2001, the DFG organized ten public meetings to present the science-based proposal in a traditional public hearing format. Most of the fishing interests and local government officials reacted with outrage. They were upset about the proposed size and locations of the proposed MPAs and indignant for not being consulted by the DFG. In response to political pressure from the fishing community, the California DFG abandoned the master plan team process six months after the public meetings. In the summer of 2002, the DFG attempted to implement the MLPA a second time. The second effort included seven regional working groups. Each working group consisted 4 After developing a master plan, the California Department of Fish and Game would submit the plan to the California Fish and Game Commission for authorization. Journal of Public Administration Research and Theory 102 at A uaria Lrary on N ovem er 3, 2010 jpart.oxjournals.org D ow nladed fom of fourteen to eighteen stakeholders representing a range of interests. The stakeholder working groups were asked to provide recommendations for the placement of MPAs to the DFG. Negotiations began in the summer of 2002, but the working group process ended in the spring of 2003 because of insufficient funding during the California budget crisis. The third attempt to implement the MLPA began in 2004. Financed by private sources, the third attempt shifted authority up to the California Resources Agency (the agency in charge of the DFG). To help guide the third attempt, the secretary for resources appointed a new statewide ‘‘Blue Ribbon Task Force,’’ which consisted of a mix of stakeholder leaders. The Blue Ribbon Task Force organized another collaborative process with just one stakeholder group on the central California coast, postponing negotiations for the northern and southern coasts until a later time. The third attempt is ongoing. I illustrate a partial ACF stakeholder analysis using data collected after the master plan team process but before the start of the stakeholder working group process (that is, between the first and the second attempts to implement the MLPA). The goal of this ACF stakeholder analysis is to use existing data to examine and explain the current political state of the MLPA policy subsystem. Using data that are approximately three years old has limitations. There has been at least one shock (a budget crisis) to this policy subsystem. The ACF predicts that most policy core beliefs and coalition membership have probably remained the same and can still be useful for identifying impediments to conflict resolution in the ongoing attempt to implement the MLPA. Therefore, I spend most of the analysis studying commonalities and differences in policy core beliefs, with some emphasis on identifying areas of conflict mitigation and consensus. Less time is spent discussing and drawing generalizations from the data regarding resources and venue. This is because a limited amount of data on resources and venues were collected back in 2002 and because this information will likely have changed over the past few years. Policy Subsystem Scope and Stakeholders I define the spatial boundary of the MLPA policy subsystem by California ocean waters and the topical boundary by a specific governing tool: MPAs. These boundaries are largely based on the legal specification of the legislation. Major industries that depend on coastal marine resources include commercial fisheries, recreational fisheries, and various nonconsumptive recreational activities, such as touring and scuba diving. To identify stakeholders either directly or indirectly involved in influencing the MLPA process, I conducted over fifty informal interviews in the late spring and early summer of 2002. These informal interviews were used to help understand the history of this policy subsystem, to design and pretest a survey, to establish a stakeholder advisory committee to help guide the research design, and to start a snowball sample of additional stakeholders involved in the MLPA process (Singleton and Straits 1999, 163). I modified the snowball technique with additional individuals or organizations quoted in newspapers, by identifying individuals who have published on the California MLPA process, and by 5 See http://www.dfg.ca.gov/mrd/mlpa (accessed 27 March 2006). 6 This is a prediction that is worth testing rather than assuming, but another survey of this subsystem is beyond the scope of this article. See Zafonte and Sabatier (2004) for empirical evidence of the longevity of policy core beliefs