How does the context and design of participatory decision making processes affect their outcomes ? Evidence from sustainable land management in global drylands

Although the design of participatory processes to manage social-ecological systems needs to be adapted to local contexts, it is unclear which elements of process design might be universal. We use empirical evidence to analyze the extent to which context and process design can enable or impede stakeholder participation and facilitate beneficial environmental and social outcomes. To explore the role of design and minor variations in local context on the outcomes of participatory processes, we interviewed participants and facilitators from 11 case studies in which different process designs were used to select sustainable land management options in Spain and Portugal. We analyzed interview data using quantitative and qualitative approaches. Results showed that although some aspects of local context affected process outcomes, factors associated with process design were more significant. Processes leading to more beneficial environmental and social outcomes included the following: the legitimate representation of stakeholders; professional facilitation including structured methods for aggregating information and balancing power dynamics among participants; and provision of information and decision-making power to all participants. Although processes initiated or facilitated by government bodies led to significantly less trust, information gain, and learning, decisions in these processes were more likely to be accepted and implemented. To further test the role of context in determining the outcomes of participation, we interviewed facilitators from a process that was replicated across 13 dryland study sites around the world, reflecting much greater national variations in context. The similarity of outcomes across these sites suggested that the socio-cultural context in which the process was replicated had little impact on its outcomes, as long as certain design principles were fulfilled. Overall, our findings provide a solid empirical basis for good practice in the design of participatory processes in the management of social-ecological systems.


INTRODUCTION
Scientists and policy makers managing social-ecological systems are increasingly encouraged to use participatory approaches.Claims in support of participation are bold.Engaging with stakeholders offers the promise of achieving environmental goals more efficiently and effectively.It helps to reduce and resolve conflict by building trust and learning among stakeholders.This increases the likelihood that stakeholders will support project goals and implement decisions in the long term (Macnaghten and Jacobs 1997, Beierle 2002, Dietz and Stern 2008, Reed 2008).However, there is concern that many of the claimed benefits are not being realized.Some critics, for example, discuss how problems with stakeholder representation or participatory process design and implementation mean processes fail to achieve their goals.Participatory processes can also exacerbate conflict, or allow special interest groups to bias outcomes (Coglianese 1997, Cook and Kothari 2001, Gerrits and Edelenbos 2004, Scott 2011).Others emphasize the limited success of collaborative natural resource management, especially when the causes and effects of environmental problems reach beyond local boundaries.This underlines the importance of social and institutional complexity, the institutional scale, and the context in which the process is performed (Singleton 2002, Ferreyra et al. 2008, Armitage et al. 2012).
The outputs, e.g., strategies, plans, or other agreements, and ultimate outcomes, e.g., social learning, network forming, preference change, or implementation of solutions, of participatory processes in management of social-ecological systems depend highly on the selection of participants, the process design, and the socio-cultural, institutional, and environmental context in which they are conducted.Cuppen (2012) found that social learning from participatory processes (using Reed et al's.2010 definition of social learning as "a change in understanding that goes beyond the individual to become situated within wider social units or communities of practice through social interactions between actors within social networks") was dependent upon the diversity of perspectives held by those who engaged in the process.Similarly, Newig and Fritsch (2009a) found that the composition of groups engaging in participatory processes and their associated preferences strongly influenced the environmental standard of outputs and outcomes.Others point to the importance of systematically identifying and selecting stakeholders for inclusion as a prerequisite for achieving desired outputs and outcomes (Reed et al. 2009, Stanghellini 2010).Various studies have focused on the context of participation, such as cultural barriers to successful participation in former communist states (Stringer et al. 2009, Blicharska et al. 2011).Koontz (2005) found that the degree to which collaborative Ecology and Society 21(2): 24 http://www.ecologyandsociety.org/vol21/iss2/art24/planning led to policy change was determined largely by local contextual factors rather than by the design of the participatory process.
The research we present in this paper contributes to debates surrounding the challenges and opportunities linked to participation.We consider how participatory approaches to management of social-ecological systems (1) result in better informed and sustainable environmental decisions or policies, facilitate their acceptance and implementation, and thus achieve environmental goals more effectively, and (2) benefit participants in other ways linked to the participatory process, for example, through increased learning and trust, and achieve their stated goals, whether environmentally related or not.
Previous attempts to critically analyze participation have typically been based on qualitative case study approaches or on comparisons of cases in very different contexts (Newig 2012).Most studies focus on evaluating the process of participation rather than its outcomes (Renn et al. 1995, Beierle 1999).Here, we seek to make an important contribution by explicitly linking the process of participation with its outcomes, carefully considering the extent to which process design and local versus national context influence these outcomes.We use the term "outcomes" to cover outputs and outcomes as defined above.

RESEARCH DESIGN
To disentangle the effects that process design and context may have on the outcomes of participatory processes, we collected and analyzed empirical data from two groups of projects tackling dryland degradation, using quantitative and qualitative techniques.In the first group, the national context in which participation took place was constant, the local context showed some variation, but the level of participation and the design of the processes differed markedly.Analysis of these cases provided insight into the effects of process design and minor variations in local context on process outcomes.In the second group of projects, the same participatory process was replicated across markedly different national decision-making contexts in different countries around the world.Analysis of these cases provided insight into the robustness of process design principles under very different contexts.
Regarding quantitative techniques, we used descriptive statistics and correlation analysis to identify the key factors influencing the environmental and social outcomes of participation.In doing this we explicitly recognize that we are analyzing people's perceptions, and that different factors combine to influence those perceived outcomes.We then take an interpretivist approach by using qualitative in-depth interview data to interpret the statistical relationships derived from the quantitative analysis, as far as possible through the eyes of the respondents.By taking a grounded theory approach to the analysis of these interviews (Corbin and Strauss 1990), whereby emergent theory comes from the data through iterative analysis, it provides us with insight into the roles of context and process in participatory practice (Cook and Wagenaar 2012).

Case study selection and characterization
First, we evaluated the role of process design and minor variations in local contextual factors in determining the outcomes of 11 cases of participatory decision making in Spain and Portugal.These cases all aimed to tackle environmental degradation, soil and water conservation, and sustainable land management (SLM) of agriculture and forestry in semiarid to dry subhumid environments.These cases reflect different levels, methods, and design of participation, illustrated by three classification schemes (Table 1 and Box 1).In these cases, we expected the outcomes of participation to be dominated by differences in process design and local context, rather than by differences in national context (Newig and Fritsch 2009b).By analyzing interviews with participants and facilitators of these cases, we highlight the effects of process design and minor variations in local context on the outcomes of participation.

Box 1:
Classification schemes of stakeholder participation.
We used three classification schemes to characterize the type and level of participation in the Spanish and Portuguese cases (Table 1).First, Pretty's (1995) scheme describes a continuum of increasing stakeholder involvement, from passive dissemination of information, to active engagement.Although recently, Hurlbert and Gupta (2015) have conceptualized this continuum as a "split ladder" of participation, recognizing that low levels of participation may be acceptable or even desirable in contexts where little disagreement exists and decisions can be made easily without reference to stakeholders.Collins and Ison (2009) suggest that we need to "jump off the ladder" to focus on social learning among multiple actors.Second, Rowe and Frewer (2000) classify stakeholder engagement according to the direction of communication between parties.They labelled information dissemination to passive recipients as "communication," gathering information from participants as "consultation," while "participation" was conceptualized as a two-way communication and learning process between all participants and process organizers.Third, Fung (2006) distinguished three dimensions of participation: (1) who participates, (2) how participants communicate with one another and make decisions, and (3) how process outcomes are linked with policy or public action.Who participates is ranked from more restrictive to more inclusive approaches, forms of communication are ranked from more to less intense, and the links with policy or public action are grouped from least to most authority.
Second, we examined the role of national context in determining outcomes of participation by evaluating the outcomes of a participatory process replicated in 13 dryland study sites around the world as part of the DESIRE research project (Table 2).The DESIRE project aimed to (1) combine local and scientific knowledge to select effective and socially accepted SLM options for land managers, (2) facilitate mutual learning through dialogue between stakeholder groups to achieve awareness, understanding, and ownership over land degradation problems and SLM solutions, and (3) implement, monitor, and demonstrate the effectiveness and feasibility of SLM to strengthen social acceptance.In each of the DESIRE sites, the same participatory process design, as detailed in other publications (Schwilch et al. 2009), was used in each country (with minor variations), but the http://www.ecologyandsociety.org/vol21/iss2/art24/contexts in which the process was carried out differed much more markedly than in the Spanish and Portuguese cases.Such variation between sites included socio-cultural, political, economic, and environmental differences, as well as variation in types of land degradation experienced.Projects in each site tackled issues ranging from salinity in Russia to overgrazing in Botswana.Evidence from the DESIRE cases was based on analysis of interviews with the participatory process facilitators in each site.
Table 2. DESIRE study sites and the main land degradation problem considered (see also Figure 1).To obtain a more comprehensive view and represent differences in perception, we interviewed people playing a variety of roles in participatory projects in Spain and Portugal, including facilitators, researchers, and a representative range of process participants, e.g., land owners, representatives of farmers' organizations, local and regional governments.In Spain, we interviewed 18 people across 6 projects.Several respondents participated in more than 1 project and for 2 projects (DESERTLINKS and SoCo) we found only 4 respondents willing to be interviewed.For the Portuguese cases we conducted 22 interviews.We found only 2 participants of MEDACTION to be interviewed and 1 of them answered only the open questions (Valente 2013).For the 13 DESIRE cases, we interviewed only the process facilitator from each study site, giving a sample size of 13 full interviews.

Quantitative and qualitative analysis
We used a combined quantitative and qualitative approach to evaluate all cases using a questionnaire consisting of two parts.
The first part comprised five open questions around which discussion took place as part of a semistructured interview.These questions asked respondents to identify the most important outcomes of participation, factors leading to the successful achievement of these outcomes, a description of their project, and the main challenges for participation that they faced.The second (structured) part contained 51 closed questions, where we asked respondents to provide an integer score between 0 and 4 or between -4 and 4, during the same interview (see Appendix 1).Questions addressed issues related to context, process design, the actual process, and process outcomes.Questions about context referred to aspects like the existing policy, legal and institutional framework, social networks, trust between stakeholder groups, and the type of environmental problem under consideration.
Questions about process design referred to characteristics such as who initiated and facilitated the process, who participated, and how knowledge exchange and decision making were facilitated.
The second part of the questionnaire was based on the "Scheme for the Comparative Analysis of Public Environmental decisionmaking" (SCAPE) prepared as part of a larger research project that aims to assess if participatory governance improves environmental policy delivery, drawing on a meta-analysis of 250 published case studies (Newig et al. 2013).We transcribed the answers to the open questions and used grounded theory analysis to identify the main outcomes and challenges of participation and to highlight the key principles of best practice in a participatory process.Grounded theory analysis is a qualitative method used to systematically analyze large bodies of text, to construct theoretical models that are grounded in the text (Corbin and Strauss 1990), or in our case, the interviewees' responses to our questions.Using the answers from the second part of the questionnaire, we explored significant statistical correlations (p < 0.01) between process design, context parameters, and process outcomes, based on nonparametric Spearman correlation coefficients calculated in SAS statistical software (SAS 2011).We used the combined results from the quantitative and qualitative assessment to support the discussion, conclusions, and recommendations for the design of participatory processes.

RESULTS
The role of process design: findings from the analysis of participatory processes in Spain and Portugal Statistical analysis of answers to structured interviews with people engaged in processes with varying levels of participation in Spain and Portugal showed that 73% of the evaluated process design variables significantly affected at least one of the perceived outcomes (p < 0.01).Seventy-eight percent of the outcomes significantly correlated to at least one process design variable.Despite potential differences in perception between people involved in the processes, we found no significant differences between outcomes perceived by facilitators and other participants.We did find significant correlations between various process design variables and process outcomes, meaning that respondents' perceptions were not diverse.Table 3 shows all significant correlations (p < 0.01) between process design variables and process outcomes, which are summarized below: . Legitimate representation of stakeholders, including opinion leaders and those who would need to implement decisions, significantly increased the likelihood of learning among participants; development of mutual gains and sustainable solutions that addressed socioeconomic and environmental concerns; and attaining the goals specified at the start of the process; . Conversely, where those who initiated the process controlled participant selection, mutual gains and win-win solutions were less likely; the process was less likely to reach its originally stated goals; conflict resolution was less likely; participants were less likely to learn from each other; and the process was less likely to foster trust between nonstate actors and researchers; . Processes preceded by a stakeholder analysis were significantly more likely to lead to information gain and learning among participants and enhanced trust among different stakeholders who did not represent the government ("nonstate actors"), between nonstate actors and the http://www.ecologyandsociety.org/vol21/iss2/art24/governmental institution responsible for the decisionmaking process, and between nonstate actors and researchers.
. Professionally facilitated processes that included structured methods for eliciting and aggregating information from participants, e.g., sorting and prioritization methods, were more likely to lead to information gain and learning among participants; win-win solutions; conflict resolution; and enhanced trust among nonstate actors, between nonstate actors and the governmental institution responsible for the decision-making process, and between nonstate actors and researchers; . In processes where nonstate actors received relevant information as part of the process and were able to have a significant influence on decision making within the process there was a significant likelihood that participants would gain information and learn from each other; win-win solutions would be reached; and the original goals of the process would be reached; . If information exchange occurred through face-to-face contact between participants, there was also a significantly greater likelihood that the process would lead to sustainable solutions; conflict resolution; increased trust among nonstate actors and between nonstate actors and researchers; . In processes where all participants could freely make statements and participate in discussion and decision making, the process would lead to win-win and sustainable solutions that are deemed to be socially equitable; achieve its original goals; contribute toward conflict resolution; facilitate learning between participants; and increase trust among all participant groups; . Participatory processes that were initiated with the goal of empowering stakeholders were highly likely to achieve their stated goals, while processes initiated to achieve specific environmental outcomes or that had an underlying rationale of providing democratic legitimacy to decisions were associated with beneficial social outcomes, in particular, increased trust among participants; . Participatory processes initiated and/or facilitated by government bodies were less likely to lead to information gain, learning by participants, flexible solutions that could adapt to new knowledge, or to contribute toward trust between nonstate actors and researchers.However, in these processes or those where government bodies were present as participants, the decisions that emerged were more likely to be accepted by the government and by those who had to implement them on the ground, and decisions were more likely to be implemented.
We further explored these relationships between process design variables and process outcomes identified through quantitative analysis by means of qualitative analysis of answers to open questions.The qualitative analysis echoed many of the findings from the quantitative analysis while also providing additional insights.For example, many respondents mentioned the importance of selecting relevant participants in relation to the process objectives (Box 2).Where the aim of a participatory process was to create scenarios or develop innovative ideas, respondents emphasized the importance of involving civil servants, researchers, land owners, and technicians.If these scenarios or ideas were intended to have an impact on policy, then more strategic stakeholders like farmers' organizations and highlevel policy makers needed to take part and commit themselves to the process.

Box 2:
Illustrative quotes concerning the impact of process design on process outcomes "If you want to have an impact, for example, organize a 1-day workshop, because you want the higher level policy makers, who will not come for more than 1 day.Respondents further argued that policy makers with decisionmaking power needed to be included in the process for short-term implementation and impact because they have the means to implement decisions in policy and provide incentives for their wider scale implementation.However, the power imbalance could limit active participation and the emergence of new ideas from other participants.To foster trust, the language, form, and location of communication must be adapted to each stakeholder group.Respondents reported that field visits or village meetings were often more effective than seminar presentations at Universities or government buildings.By using the local knowledge of those directly dependent on the threatened resources, participation led to solutions that fitted local conditions and were therefore more effective than top-down decisions.Most respondents also stressed that social and political change processes take time.Therefore, successful participation requires long-term commitment from all participants.

The impact of minor variations in local context on process outcomes: findings from participatory processes in Spain and Portugal
Table 4 shows all significant correlations (p < 0.01) between local context variables and perceived process outcomes in the Spanish and Portuguese cases.Twenty-six percent of the context parameters were significantly associated with at least one of the perceived outcomes of participation.Thirty-seven percent of the outcomes significantly correlated to at least one context variable.These findings are summarized as follows: . Participatory processes taking place in an autocratic institutional culture, with little decision making autonomy for individuals, are less likely to result in socially equitable outcomes; .Participatory processes initiated in response to an agenda set by nonstate actors are more likely to lead to socially equitable outcomes and enhance trust between nonstate actors and the competent authority, i.e., the governmental institution responsible for decisions arising from the process; . Processes with strong cooperation from private enterprises are more likely to lead to solutions that are accepted by government bodies and deemed sustainable, addressing socioeconomic as well as environmental concerns.However, solutions to a problem were significantly less likely to be deemed sustainable when policy regulating the problem under consideration already existed; . Strong cooperation from civil society organizations and landowners and managers was likely to lead to mutual gains and win-win solutions regarding economic and conservation objectives.In addition, cooperation from civil society organizations correlated positively with the delivery of solutions that are flexible and able to adapt to new knowledge as it becomes available; . Implementation of outputs identified during participatory processes was not strongly associated with any contextual parameters.We found weaker associations (p < 0.05, not indicated in Table 4) between implementation of solutions and a good understanding of the environmental problem by individuals participating in the process, a government that cooperates in the process, trust among nonstate actors, and trust between nonstate actors and the competent authority.
Qualitative interview analysis provided more detailed insights into the role of local context (Box 3).Most respondents stressed the importance of personal motivation of stakeholders to participate and the sense that their contributions will be acted upon.This in turn often related to available funding and the extent to which a process could feed directly into policy.Motivation to participate was strongly reduced if there was no funding for implementation of solutions, or if participation was merely a minor part of a research project.These challenges may be exacerbated because environmental problems like land degradation are often gradual or intangible processes, so many stakeholders do not give them priority.

The role of national context parameters: findings from an international process
Table 5 lists the mean scores of process outcomes as ranked by interviewees from the 13 DESIRE cases.These showed great similarity with the top eight outcomes identified by interviewees from the Spanish and Portuguese cases (see footnote Table 5).
There was considerable overlap between the top ranked outcomes for the majority of DESIRE cases, despite the participatory process being conducted in very different national contexts.The Table 5. Ranking outputs from 13 DESIRE cases indicating mean and standard deviations of 13 scores.High scores refer to a higher importance of that output.Negative scores refer to a negative or decreased output.

Process outcome
Average score (possible range)
four most important perceived outcomes from the DESIRE cases (average score above 3) were: (i) information gain, (ii) identification of sustainable solutions, (iii) acceptance of solutions, and (iv) learning by participants.Most interviewees also reported increased trust among participants and the development of economically rational and socially equitable solutions as important outcomes.Similar themes emerged from the qualitative analysis of interview transcripts.Table 6 identifies five broad types of outcomes that emerged from this qualitative analysis for all 13 DESIRE cases.
As outcomes of the participatory process were similar across all sites, despite the large contextual variations, it appears that national context had little influence on outcomes.This is not to say that context had no effect on outcomes whatsoever.The limited amount of variation in outcomes could be explained by a small proportion of the contextual factors (13%) that were significantly (p < 0.01) correlated with 21% of the outcome variables (Table 7), specifically the following: . In three specific types of context, participants were more likely to gain information from their participation in the participatory process (i) if the issues being discussed had been proposed by the competent authority, i.e., the governmental institution responsible for the issue; (ii) if the issue had been put on the agenda as a result of a previous government policy; or (iii) if the issues being dealt with in the process related to nature conservation, as opposed to other issues such as health or the exploitation of scarce natural resources.
. If the issues being tackled had previously received media attention, then the DESIRE participatory process was more likely to enhance trust between (i) government representatives, i.e., state actors, and other stakeholders, i.e., nonstate actors; (ii) nonstate actors and researchers involved in the process; and (iii) between different nonstate actors involved in the process.
Further national contextual factors emerged during the qualitative analysis of interview transcripts.The most important factors that negatively affected the outcomes of the DESIRE process were apathy of stakeholders toward participation in some countries, previous negative experiences with participation, or because power imbalances reduced engagement.

Environmental and social outcomes of well-designed participatory processes
Those who argue the case for participation do so primarily for normative, pragmatic, or instrumental reasons (Chambers 1983, Fiorino 1990, Reed 2008).In our study, all interviewees emphasized pragmatic benefits of participation, for example, arriving at decisions that are better informed and feasible.They also noted decisions were more durable and flexible and observed a higher level of acceptance and ownership of decisions by participants and decision makers.Our findings support literature claiming that participation can enable groups to go beyond the acquisition of factual knowledge, to collectively and creatively develop new solutions through genuine deliberation and reflection, an inspiring group atmosphere, and the multiplicity of perspectives involved (Pahl-Wostl andHare 2004, Siebenhüner andSuplie 2005).Overall, our research confirms that well designed participatory processes are likely to lead to more beneficial environmental outcomes through better informed, sustainable decisions, and win-win solutions regarding economic and conservation objectives.
We did not find a direct link between most process design characteristics and acceptance or implementation of solutions.However, the use of professional facilitation and structured methods of knowledge aggregation during participatory processes correlated positively with social outcomes, including information gain, conflict resolution, learning, and trust among stakeholder groups.At the same time, transparency and trust, Table 6.Outcomes from the participatory process conducted as part of the DESIRE project across 13 study sites internationally, including a description of the outcome and a representative quote (from qualitative interview analysis).

Outcome Description Example Quote
Enhanced social networks, collaboration, and trust among participants Most interviewees mentioned that participation builds and strengthens relationships between individuals and organizations leading to new collaborations, both within and outside the participatory process.Successful processes were felt to increase trust between land users and institutional participants.
"For the stakeholders this kind of experiences is very useful, because it develops contacts and further collaboration among the stakeholders."[Facilitator, Portugal] Learning and knowledge exchange among participants Participatory processes provide a forum for discussion among stakeholders that otherwise do not often meet or may find it difficult to communicate with one another.By learning from each other during the process, it may be possible to better appreciate the motives and rationale behind the attitudes and behaviour of different participants.By sharing knowledge with each other, it may be possible to identify and critique new solutions to land management challenges."Before we could not understand why farmers don't apply certain technologies.Participation allows us to learn from the stakeholders about the limitations and problems."[Facilitator, Chile] Better problem identification and awareness Stakeholders start thinking more holistically by identifying direct and indirect effects of their actions.Interaction between researchers and other stakeholders can force researchers to question the assumptions that underlie their work and lead to more rigorous research.
"Participation makes people think about the problem as a whole."[Facilitator, China] Consensus, acceptance, and implementation of tailor-made solutions is more likely Participants are more likely to reach consensus over, or accept a group decision by deliberating together over its likely impacts.As people feel increased ownership over the problem and the solutions that have been identified, they are more motivated to actually implement agreed solutions.This may be further enhanced by observing proposed solutions during field demonstrations.By combining the expertise of researchers and local communities in the decision-making process, solutions are tailor-made to the local context."Even if people do not agree at the beginning, they can accept change if they can see the benefits."[Facilitator, Morocco] More confident and motivated stakeholders Participants gain confidence as they feel that their views are taken seriously in the participatory process.This makes them feel that their participation in the decision-making process really does make a difference, and they are more likely to actively engage in it.Local land users and groups that are otherwise often not heard are given a voice through participation."It was the first time the experts and farmers came together to discuss their own problems.It helped give the farmers more self-esteem."[Facilitator, Turkey] especially between land users and government bodies, are needed to make these structured forms of participation successful.This is consistent with other studies that have identified mutual trust among the participants as both a precondition for social learning (Wenger 2000, Leach andSabatier 2005) and an outcome of social learning in participatory processes (Pahl-Wostl et al. 2007).Our results further show that participation provides opportunities for participants to get to know each other and each other's concerns.This increases trust and development of alliances and other networks.Outcomes from decision-making processes that are perceived as fair and legitimate are more likely to be accepted by stakeholders over the long term (Lind andTyler 1988, Young et al. 2013).
Increased trust and learning were perceived as major outcomes of participation.The reason for this may be that participatory processes tend to seek and value a range of perspectives, leading to the appreciation of different stakeholder positions.By enabling participants to listen to a wider range of perspectives with less prejudice, participation may enable learning at a number of levels: from better understanding the decision on a cognitive level, to deeper learning that can enable participants to re-evaluate the assumptions and values underlying their positions.This can foster changes in attitudes that shift their positions so that they are more in line with their values in relation to the environment (Argyris and Schön 1978, Fazey et al. 2005, Keen and Mahanty 2006, Kenter et al. 2014).New understandings, attitudes, and positions that arise from this learning process may then diffuse from those directly participating in the process to those in their social network by social learning (Reed et al. 2010).

How important is process design in determining the outcomes of participatory processes?
Our research explicitly addresses many of the concerns in the literature about stakeholder participation in management of social-ecological systems.Interviews with those who facilitated or were engaged in participatory processes highlighted the dangers of poorly designed or implemented participatory processes.Respondents pointed specifically to inadequate representation of relevant stakeholders, and inattention to power dynamics.Without effective design and strong facilitation, more powerful individuals have the potential to limit the engagement of others and bias the outcomes.
Although acknowledging potential for failure, this research identifies factors that are most likely to lead to failure, and those aspects of process design that will reduce the risk of failure.Table 8 summarizes seven recommendations for the design of participatory processes in management of social-ecological systems, based on the evaluation of all interviews from all cases presented here, and relating them to their expected outcomes.These recommendations are not to be seen as individual tools, but as an integral approach required to achieve the indicated outcomes.They may be considered design considerations for participatory processes that always need to be enacted by practitioners in their given context.In line with the literature (Newig and Fritsch 2009a, Reed et al. 2009, Stanghellini 2010), our results suggest that participant selection strongly determines the outcomes of a participatory process.Legitimate representation of stakeholders, including opinion leaders and implementers, facilitates learning and trust among participants, leading to selection of sustainable solutions and fulfilment of project objectives.Respondents stressed the importance of performing a stakeholder analysis to identify key players and assist participant selection.However, too restrictive, controlled participant selection may deliver fewer positive social outcomes, i.e., less learning, conflict resolution, and trust building, and may increase conflict.Indeed, Reed et al. (2009) warn against the use of stakeholder analysis to marginalize less influential stakeholders.Careful selection of participants is needed to guarantee a heterogeneous group of people who represent all stakeholders' interests, including those with high levels of interest and low levels of influence in the decision, as well as strategic stakeholders with power, motivation, and means to implement decisions.Our findings suggest that to increase acceptance and implementation of solutions, governmental institutions responsible for the decision-making process and those who have to implement the solutions, such as land owners and managers, should be involved early in the decision-making process because this will give them ownership and greater acceptance of selected solutions (cf.Stringer and Dougill 2013).
Our respondents stressed that a diverse group of well-informed stakeholders will provide most comprehensive and innovative ideas.However, this is not simply about providing more comprehensive information inputs to the decision-making process.Information gain, learning, selection of sustainable solutions, and actual implementation of solutions depend on the group composition, the process design, and equal opportunities to start or participate in discourse.A diverse group of civil servants, land owners, researchers, and technicians is likely to provide sustainable and flexible solutions that are responsive to new knowledge, but these are not necessarily accepted and implemented by government bodies or those responsible for implementation.Processes where government bodies act as participants or facilitators provided fewer social outcomes, that is, less information gain, learning, and trust among stakeholder groups, and less flexible solutions, most likely because of a perceived power imbalance.However, these processes resulted in better acceptance and actual implementation by those who have to implement solutions.This stresses the importance of process design and skilled facilitation in dealing with power relations and differences in background and education among stakeholders.This may also point toward the need for selection or grouping of participants according to the process objectives.These findings and especially the need for government involvement to achieve wider stakeholder acceptance are in line with results reported by Young et al. (2012) and Sandström et al. (2014).
Our respondents emphasized that to be effective, participation must be easy and attractive.This means it should be clear what is in it for participants and how their contributions will be acted upon.Participants must gain returns for their time investment, for example, knowledge, contacts, or influence on decision making.Therefore, respondents suggested that a clear description of the problem and of the process objectives is crucial, so that stakeholders easily identify with the problem, irrespective of their background or education level.For example, stakeholders will find it easier to identify with specific tangible aspects of land degradation, e.g., water shortage or reduced crop productivity, than with the relatively slow process of land degradation itself.
Respondents further indicated that continuous participation in problem definition, identification, selection, implementation, and evaluation of solutions is preferred over ad hoc participation or lower levels of participation like consultation and communication.Decision making through dialogue, deliberation, and face-to-face contact is preferred because this increases trust and ownership over solutions.However, the level of participation may vary throughout the process, and to prevent unsatisfactory compromises, there may be need for cut-off points where someone has to make a decision, respecting but not necessarily fulfilling all stakeholder opinions.Communication and transparency of decisions are crucial in this phase to maintain trust and involvement.Irrespective of the level of participation, participants need to receive regular feedback on implementation of the outputs.Without this, stakeholder fatigue and distrust can ensue.
Where problematic relationships between individuals or organizations emerge during stakeholder analysis, it may be possible to design parallel processes through which trust can be built more slowly.Handing responsibility for running the process to an independent facilitator that is perceived to be neutral and trusted by participants can also be useful.Moreover, providing participants with sufficient independent background information, using structured forms of information aggregation, and adapting language and location to participants, will also increase trust and thereby acceptance and possible implementation of solutions.In line with Diduck and Sinclair (2002), a high likelihood of an actual impact on policy, e.g., evidence of buy-in to the process by high level policy makers, also increases stakeholders' motivation to participate.http://www.ecologyandsociety.org/vol21/iss2/art24/Increased trust and implementation of solutions 7. Adapt language, location, and process design to the participants Use accessible language and forms of information adapted to the education level of participants.Field visits or village meetings are often more effective than seminar presentations at universities or government buildings.

Increased trust Increased participation Learning
How important are local and national context in determining the outcomes of participatory processes?Several studies emphasize the role of context in determining the outcomes of participatory processes (Koontz 2005, Stringer et al. 2007, Blicharska et al. 2011).Most of this research has focused on the socioeconomic, cultural, and institutional contexts within which participation is enacted (Delli Carpini et al. 2004).For example, the distribution of power may impact the nature of the decision that is made, as well as its acceptance, because those who feel disadvantaged by the process may choose to delay or prevent implementation of the decision, for example by taking legal action (Turner and Weninger 2005).
Contrary to the literature, our findings suggest that although a number of mainly localized contextual factors play an important role in determining the outcomes of a participatory process, these are outnumbered by factors linked to process design.This is not to say that context is unimportant in determining the outcomes of participatory processes.The entire discipline of environmental anthropology is devoted to understanding the role of culture in the relationship between people and their environment (Milton 1996, Sanga andOrtalli 2007), while social and institutional complexity and scale need to be taken into account during process design (Singleton 2002, Ferreyra et al. 2008, Armitage et al. 2012).However, our findings suggest that well designed participatory processes, which by definition are flexible and sensitive to cultural factors, are largely transferable across local and national contexts.
The adaptability and quality of facilitation in well-designed processes enables them to adjust to context.This was demonstrated by the success and the similar outcomes of a http://www.ecologyandsociety.org/vol21/iss2/art24/process replicated across a wide range of socioeconomic and environmental national contexts in the DESIRE project (Schwilch et al. 2012), and the relatively small number of local contextual factors that significantly correlated with process outcomes in Spain and Portugal, compared with process design factors.
Nevertheless, some local and national contextual factors that affected process outcomes may be difficult to change, e.g., stakeholder fatigue, the lack of a "participation culture" in the society where the process is being implemented, a lack of funding to implement decisions, or distrust between stakeholders and those running the process or those with the power to implement decisions.Some of the most commonly cited reasons for distrust and stakeholder fatigue in our study were lack of immediate personal benefit from participating, frequently changing policies, and bad past experiences of participatory processes that did not effect change.Appreciating these context related factors may facilitate the design of more effective participatory processes to fit specific contexts.For example, where stakeholder fatigue is identified at the outset, it may be possible to alter the timing of events to reduce demands on people's time, or collaborate with other initiatives to reduce the number of different approaches that are made to stakeholders (cf.Sandström et al. 2014), and identify participants who have not been invited to previous workshops.

CONCLUSIONS
We explicitly linked the process of participation with its outcomes, carefully considering the extent to which context versus process design influences these outcomes.Based on empirical evidence from quantitative and qualitative evaluation of interviews with facilitators and stakeholders engaged in participatory processes in 11 cases from Spain and Portugal and 13 international dryland sites, we identified a range of social and environmental outcomes of stakeholder participation.A small number of local contextual factors influenced the outcomes of participatory processes in Spain and Portugal.However, in the international comparison of an almost identical participatory process replicated across 13 dryland study sites around the world, the outcomes of the participatory process were similar across all sites, despite large variations in context between those sites.Although this finding suggests that variation in national context had little influence on outcomes, this is not to say that context had no effect on outcomes whatsoever.The limited amount of variation in outcomes that was observed across national contexts could be explained by a small number of contextual factors.We therefore conclude that well-designed participatory processes that consider the recommendations from this research, can lead to well-informed, durable, and flexible outcomes across a wide range of contexts.Moreover, through increased trust and ownership over problems and solutions, decisions taken in these processes are more likely to be accepted and implemented, helping to achieve environmental goals more effectively.
Our analysis of cases in widely contrasting local and national contexts suggests that the most important factors determining process outcomes are who participates, how the process and communication among participants is organized, and how the outputs of the process are linked to policy and implementation of solutions.These three aspects reflect Fung's (2006) approach to characterizing stakeholder participation.Local factors such as lack of experience or interest in participation, stakeholder fatigue, and power imbalances, are the principal context-related challenges that may need to be overcome.Professional facilitation and higher levels of participation by a heterogeneous group of stakeholders lead to learning, acceptance, and trust.However, implementation of solutions requires participation of government institutes, while this negatively correlates with learning and trust due to perceived power imbalances.This stresses the need for skilled facilitation and supports the need for separate processes for the participation of government representatives and other stakeholders.In addition to group composition and flexible process design, implementation of solutions depends on a good understanding of the environmental problem by individuals, a cooperative government, and trust between stakeholders.
These findings have important implications when it comes to upscaling of participatory processes because there could be a number of generic "good practices" in terms of design, implementation, and facilitation that should be shared between local and national contexts.Based on the empirical evidence, we identified seven recommendations for participatory processes in management of social-ecological systems: (1) select participants carefully, (2) make participation attractive and easy, (3) foster trust, (4) provide participants with information and decisionmaking power, (5) use professional independent facilitation and structured methods of information aggregation, (6) promote long-term commitment, and (7) adapt language, location, and design to the participants.Although participatory processes will always be challenging, following these recommendations will provide better-informed and sustainable environmental decisions and beneficial social outcomes in a range of decision-making contexts where stakeholders are engaged in management of social-ecological systems.
Although our study provides empirical evidence for a range of perceived outcomes of participation, the measurable impacts of participation on environmental quality and social benefits still require further evaluation.Consideration of a wider range of case studies and variables, perhaps building on the meta-analysis approach used by Newig et al. (2013), may also provide further insight into key factors influencing the outcomes of participatory processes.Alternatively, involving a wider range of participants and beneficiaries in research may provide more in-depth insights into the reasons why and under which conditions these factors are important.The empirical work reported in this paper goes some way toward explaining why participation works (or not), and as such, may provide the building blocks for a more complete and nuanced understanding of participation.
Responses to this article can be read online at: http://www.ecologyandsociety.org/issues/responses.php/8053 EDGE project as grant to JN (grant 263859-EDGE), and in the framework of the EC-DG .3)-Research on Desertificationproject DESIRE (037046): Desertification Mitigation and Remediation of land -a global approach for local solutions.JDV acknowledges funding from the Spanish Ministry of Science andInnovation (RYC-2012-10375 &CGL2013-42009-R) and the Séneca foundation (18933/JLI/13).We warmly thank all the interviewees who were so kind to share their experiences and views with us.Thanks also to Steven Vella for useful feedback on an earlier draft of this paper.Argyris, C., and D. A. Schön. 1978.Organizational learning: a theory of action perspective series on organization development.Addison-Wesley, Reading, Massachusetts.

Fig. 1 .
Fig. 1.Location and global coverage of the 13 evaluated DESIRE study sites (world image obtained from MODIS Earth-composite).

Table 1 .
Overview and general characterization of the evaluated cases from Spain and Portugal, including the level of participation, the participatory methods used, and the number of respondents.Four of the Portuguese cases focused on preventing and restoring land from forest fires, while the fifth case was developed at Mindelo village, located in northern Portugal, where a local Agenda 21 initiative developed an action plan for the sustainable development and environmental conservation of a dry subhumid natural area. ‡

Table 3 .
Spearman correlations between process design variables and process outcomes for the 11 Spanish and Portuguese cases (only correlations with p < 0.01 are shown).See methods section and questionnaire in annex for detailed description of process design and outcome variables.
1: Information gain; 2: Mutual gains; 3: Attainment of initial goals compared to final decision; 4: Flexible solutions adaptive to new knowledge; 5: Sustainable solutions; 6: Conflict resolution; 7: Acceptance by those who have to implement and apply solutions; 8: Acceptance by the competent authority; 9: Learning by participants; 10: Trust among nonstate actors; 11: Trust between nonstate actors and competent authority; 12: Trust between competent authority and scientists; 13: Trust between nonstate actors and scientists; 14: Socially equitable output; 15: Implementation of solutions.

Table 4 .
Spearman correlations between context variables and process outcomes for the 11 Spanish and Portuguese cases (only correlations with p < 0.01 are shown).See methods section and questionnaire in annex for detailed description of context and process outcome variables.
. Power imbalances among participants significantly reduce the likelihood of mutual gains and win-win solutions;

Table 7 .
Spearman correlations between context variables and main process outcomes for the 13 DESIRE cases (only correlations with p < 0.01 are shown).See methods section and questionnaire in Appendix 1 for detailed description of context and process outcome variables.

Table 8 .
Summary of recommendations for the design of participatory processes in environmental management to achieve more beneficial environmental and social outcomes.