A subset of the consortium members received an NSF/EPA grant to study differences and similarities in how various stakeholder group members frame the conflicts. Three schools participated in this project-Texas A&M (PI for the grant), Ohio State University, and Pennsylvania State University. This study conducted both quantitative and qualitative analyses of framing data from Edwards, Voyageurs, Drake Chemical, and Ohio EPA anti-degradation advisory group. The objective of this ongoing research group was to develop profiles for understanding how stakeholders and stakeholder groups frame intractable environmental disputes.

The research found that stakeholders' repertoires do NOT correspond to natural stakeholder grouping-instead there are classifications of environmental disputants who hang together because of the frames they adopt-rather than because they have shared interests in how the dispute should be resolved. Below we describe the patterns of frame use by each of the four clusters of disputants identified through cluster analysis of their frame adoption.

Research Design.  We conducted analyses on 153 stakeholder interviews across the four sites. Stakeholders clustered into eight groups based on occupation, interest affiliation, and role in the conflict. The eight groups were business and industry, farmers and ranchers, citizen activists, elected officials, government agency officials, environmentalists, media, and neutrals. We posed two broad-based questions: 1) What patterns of framing characterize the different stakeholder groups? How are they similar or different across disputes? 2) What particular themes typify the profiles of each stakeholder group? What perceptions and views of the dispute do they raise within each frame type?

Cluster Analysis of Stakeholders Framing.  To address the first question, we conducted a quantitative analysis from the coded transcripts for the 153 stakeholders. First they transformed the frequency data in five coded frame categories into frame ratios for identity (positive and negative), characterization (positive and negative), conflict management (collaborative and non-collaborative), social control (hierarchical, egalitarian, and individualistic), and power. The researchers also calculated intercoder reliabilities for each site. From the common database, researchers conducted a cluster analysis to determine the natural groupings of the 153 stakeholders based on the ratio of their frame use. Based on this technique, the researchers found four clusters of stakeholders based on their framing of the different conflicts.

Cluster 1 consisted of 45 stakeholders who predominantly used negative characterization and non-collaborative conflict management frames; Cluster 2 consisted of 46 people who relied heavily on positive characterization, positive identity, and hierarchically control as a way to manage disputes, Cluster 3 was comprised of 18 individuals who saw the conflict through the frames of positive characterization and collaborative conflict management; and Cluster 4 consisted of 44 disputants who relied heavily on power frames to characterize the conflict. From this data, we observed that some sites contributed more to a cluster membership than did others and that some stakeholder group categories seemed more dominate in particular clusters.

Relationships Among Clusters Sites, and Stakeholder Groups.  To tease out this relationship among cluster, site, and stakeholder group, we conducted log linear comparisons and discriminant analyses. Our findings reveal a complex interrelationship between types of frames and cluster, site, and stakeholder group. Basically, two types of frames seemed dominant in contributing to stakeholders' worldviews and their alliance with other stakeholders in these framing clusters. Power substantially contributed to research site membership, and characterization (both positive and negative) contributed to the ways that stakeholder groups align in these clusters.

Research site as a variable predicted cluster membership with 68% accuracy. Some of these differences stem from involvement of different numbers of stakeholder groups across the four sites. For example, the Edwards aquifer dispute, with multiple governmental agency officials had more stakeholders in Cluster 2 and fewer in Cluster 4. Drake differed from other sites in its prevalence of stakeholders with power frames (Cluster 4) and its lack of stakeholders in Cluster 2. Ohio Antidegration differed from other research sites in its prevalence of people who held collaborative conflict management frames (Cluster 3) and its lack of stakeholders in Cluster 2. Voyageurs did not differ from other sites even though it had a large number of stakeholders in Cluster 4. Overall, Clusters 1 and 3 did not seem related to a particular research site.

Stakeholder group members predicted cluster membership with only 34% accuracy. Overall, this would suggest that membership in a particular stakeholder group is NOT the determining factor for placement in a particular cluster. However, farmers were more prevalent in Cluster 1 than Cluster 3; government officials aligned more in Cluster 2 than in other clusters; and neutrals dominated Cluster 3 more than other groups. In comparing the framing categories across stakeholder groups, farmers were more likely to hold non-collaborative views of conflict management and hierarchical views of social control than were other groups. Activists differed from other stakeholder groups in their prevalence of negative identity and egalitarian views of social control and neutrals differed from other stakeholder groups in their reliance on positive characterization and collaborative forms of conflict management.

Accounting for Cluster Findings.  A general conclusion that surfaces from the quantitative analysis is that framing is an important factor in understanding the composition and worldviews of stakeholders in intractable conflicts. Some stakeholder groups are obviously more dominant and hold similar frames for certain sites while others splinter and align with different groups in the way they make sense of a conflict. Site differences also stem in part from the type of environmental dispute. What affected these differences were the number of farmers present in the dispute and also the stage in which the researchers entered the conflict-at the end, in the middle, or during the midst of it.

In like manner, stereotypes of stakeholder groups may account for conflict framing. It is not surprising that activists frame a conflict from their negative identities and strong desires for egalitarian control. Moreover, the finding that neutrals have worldviews of conflicts rooted in collaborative conflict management and positive characterization is also very predictable. These explanations, however, are too simplistic to account for the differential framing of a wide array of stakeholders in four different types of intractable environmental conflicts. In particular, it does not account for the reasons that neutrals and activists align with farmers in Cluster 1; and environmentalists, activists, and elected officials were prevalent in Cluster 4.

Profile Analysis of Stakeholders.  To tease out the nature of this framing and to address research question #2, we conducted qualitative profile analysis of the dominant themes and patterns that unite stakeholders in each cluster. Thematic analysis was conducted first through a site-specific profile analysis on each cluster in which researcher synthesized the overall composition of the site, explicated particular themes linked to the dominant frames for a given cluster, and then drew general conclusions and observations about the stakeholders in this cluster for a particular site. Then we came together to prepare a summary profile for each cluster across the four sites.

This analysis led to the following conclusions about the four clusters.

• Cluster 1 Profile. Stakeholders in Cluster 1 held frames dominated by negative characterization and noncollaborative approaches to conflict management. Dominated by farmers, business people, and activists, they felt personally impacted by the conflict through loss of identity or economic gain. Some disputants emerged as victims while others were clearly cynical, bruised, and discouraged about any prospects for resolution. They seemed to have emotional investment in the conflict and to describe institutions and entities with adjectives linked to individuals, e.g., "hard-headed, greedy, arrogant." Facilitators and mediators who fell into this cluster had prior experience with this type of dispute or were badly burned in the process of wrestling with it.
  
  
• Cluster 2 Profile. Cluster 2 stakeholders comprised primarily by governmental officials (both elected and appointed) and some environmentalists framed the conflicts through the lens of positive characterization of other disputants, positive identity, and hierarchical views of social control. Holding a strong institutional bias to the conflict, they seemed detached and trusted authorities to make appropriate decisions, based on technical solutions. They were accommodating, believed that proper rules and regulations would resolve the problems, and wore their institutional hats to speak dispassionately about sets of events. Governmental officials who became vested personally in the conflicts or "were burned" through past events were more likely to appear in Clusters 1 or 3.
  
  
• Cluster 3 Profile. Stakeholders in Cluster 3 were idealistic and optimistic. Adopting worldviews rooted in positive characterizations of other stakeholders and holding collaborative views of conflict management, these individuals wanted the dispute resolution processes to work and placed almost a blind faith in fact finding and joint problem solving. Disputants in this cluster crossed all stakeholder groups (even though neutrals were prevalent in this cluster). Individuals in this cluster rarely made negative comments about other stakeholders and when they did, they qualified or countered their negative remarks. Although folks in this cluster were important players, they never emerged as strong leaders and were not able to exert influence in turning the conflict in a constructive direction.
  
  
• Cluster 4 Profile. Cluster 4 disputants saw the conflict through a power frame, particular concerns for voice, resources, and coalition building. The primary groups in Cluster 4 were environmentalists, activists, and elected officials. Agency government officials were not prominent in Cluster 4. Stakeholders in this cluster shared the goal of gaining power or voice. They engaged in a constant comparison with other stakeholders and often saw their own constituents as disadvantaged or the underdog, particularly as the power shifted. They saw power in the actions of other stakeholders and resented the influence that other parties had.

General Conclusions and Implications.  Overall, stakeholders in intractable environmental conflicts represent a complex picture of similarities and differences in conflict framing. They often depart from the stereotypes of their constituent groups. Farmers, industrialists, and environmentalists may share in their roles as victims who experience personal loss and are battered or discouraged from past environmental struggles. Activists and environmentalist may oppose each other on basic positions but share in their concern about loss of power and feelings of being underdogs. This complex picture suggests new conflict roles that disputants hold based on their framing of the situation, ones not rooted in the typical oppositional stances found in interest-based approaches, such as attack-defend and antagonist-proponent. Basically, four types of roles surfaced in this study and grew out of conflict framing alignment-1) individuals personally bruised, discouraged, or impacted by the dispute; 2) stakeholders who are detached from the conflict and speak from their strong institutional identities; 3) optimistic and idealistic stakeholders who believe in fact finding and collaborative conflict management; and 4) disputants who see conflicts as a way to gain voice, resources, and/ or coalitions that will enhance their power in these situations.

What are the implications of these findings? One implication is that government agencies and neutrals are stakeholders in these conflicts and fall into these patterns of framing. Thus, their relationship with other constituents often comes from an agency worldview about the conflict and the way other stakeholders differ from their own framing. A second implication is that third parties and interventionists need to move away from stereotypes of stakeholder groups and avoid treating all members of one group as if they have similar views of the conflict. In fact, these different types of framing become ways to intervene in a conflict and bring disputants who hold different substantive positions together around similar worldviews. It provides a basis to rise above the specific issues and positions to pursue broader concerns for suffering, personal loss, power, and links to authority. A third implication is that many disputes have Cluster 3 individuals who are necessary to make any progress in conflict management. Finding out how to work with them so that their leadership can be felt and their voices heard could aid practitioners. Finally, practitioners and agency officials need to develop ways to uncover disputants' framing of the conflict rather than to continue to dwell on issues and positions. Framing holds the key to similarities in how parties make sense of conflicts as well as how they align with other stakeholders in subtle, yet positive ways.