Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Guiding principles for evaluating the impacts of conservation interventions on human well-being

Guiding principles for evaluating the impacts of conservation interventions on human well-being Guiding principles for evaluating the impacts of conservation interventions on rstb.royalsocietypublishing.org human well-being 1,2,3 2 3 Emily Woodhouse , Katherine M. Homewood , Emilie Beauchamp , 1 4 1 3 Tom Clements , J. Terrence McCabe , David Wilkie and E. J. Milner-Gulland Research Wildlife Conservation Society, 2300 Southern Boulevard, Bronx, NY 10460, USA Department of Anthropology, University College London, 14 Taviton Street, London WC1H 0BW, UK Cite this article: Woodhouse E, Homewood Department of Life Sciences, Imperial College London, Silwood Park Campus, Buckhurst Road, KM, Beauchamp E, Clements T, McCabe JT, Ascot SL5 7PY, UK Wilkie D, Milner-Gulland EJ. 2015 Guiding 4 Department of Anthropology, University of Colorado, 233 UCB, Boulder, CO 80309-0233, USA principles for evaluating the impacts of con- servation interventions on human well-being. Measures of socio-economic impacts of conservation interventions have largely been restricted to externally defined indicators focused on income, Phil. Trans. R. Soc. B 370: 20150103. which do not reflect people’s priorities. Using a holistic, locally grounded http://dx.doi.org/10.1098/rstb.2015.0103 conceptualization of human well-being instead provides a way to under- stand the multi-faceted impacts of conservation on aspects of people’s Accepted: 13 August 2015 lives that they value. Conservationists are engaging with well-being for both pragmatic and ethical reasons, yet current guidance on how to opera- tionalize the concept is limited. We present nine guiding principles based One contribution of 16 to a theme issue around a well-being framework incorporating material, relational and sub- ‘Measuring the difference made by protected jective components, and focused on gaining knowledge needed for areas: methods, applications and implications decision-making. The principles relate to four key components of an for policy and practice’. impact evaluation: (i) defining well-being indicators, giving primacy to the perceptions of those most impacted by interventions through qualitative research, and considering subjective well-being, which can affect engage- Subject Areas: ment with conservation; (ii) attributing impacts to interventions through ecology quasi-experimental designs, or alternative methods such as theory-based, case study and participatory approaches, depending on the setting and evi- Keywords: dence required; (iii) understanding the processes of change including evidence of causal linkages, and consideration of trajectories of change development, impact evaluation, livelihoods, and institutional processes; and (iv) data collection with methods selected poverty, social impact, well-being and applied with sensitivity to research context, consideration of hetero- geneity of impacts along relevant societal divisions, and conducted by Author for correspondence: evaluators with local expertise and independence from the intervention. Emily Woodhouse e-mail: e.woodhouse@ucl.ac.uk 1. Introduction In response to both evidence of the dependence of vulnerable human popu- lations on ecosystems [1] and the costs associated with some conservation interventions for local people [2,3], the question of how to reconcile conserva- tion with human development has emerged as a key policy issue [4]. Recognition of the inadequacies of narrow economic indicators such as income and consumption in measuring social development has shifted atten- tion to a broader and multi-faceted vision of human well-being [5]. For instance, a conservation intervention may improve the local economy by pro- viding jobs and alternative livelihoods, but could negatively affect other priorities for local communities such as social relationships or autonomy. Where the specifics of social impacts are not intentionally and systematically examined, they could be misunderstood or missed entirely, with repercussions for social justice and conservation outcomes. While there is ready acknowledg- ment in the environmental literature that robust empirical evaluation is & 2015 The Authors. Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 required in order to better understand which approaches can collection of data relevant to conservation planning rather work for biodiversity [6], extension of this premise to conser- than impact evaluation. vation impacts on human lives is still rare. Across a range of We propose accepting the plurality of the concept of well- conservation strategies, there is a lack of evidence of the being, and present guiding principles based around existing impacts on human well-being that adequately capture the theoretical frameworks, an approach that allows comparable complexity on the ground [7]. but locally relevant results. Our principles are intended to Despite the variety of definitions, there is increasing support evaluators in operationalizing a multi-dimensional agreement in international policy on a conception of human conceptualization of well-being to measure and understand well-being that encompasses objective material circumstances impacts in ways that align with realities on the ground. We of people’s lives such as housing, income, livelihoods, health take a pragmatic approach to evaluation focused on gaining and the environment, social aspects such as community net- the knowledge needed for decision-making and policy devel- works, and a subjective component capturing an individual’s opment, a perspective that necessitates flexibility and an assessment of their own circumstances [5,8]. Well-being can openness to mixed methods, incorporating quantitative and therefore be defined as a positive physical, social and qualitative data and analysis [23]. Our aim is for the principles mental state [9]. Understanding the impacts of conservation to be useful for conservation practitioners and adaptable to on the multiple dimensions of people’s lives and working small-scale projects with limited budgets and technical exper- to improve them are ethical imperatives for conservation tise, as well as larger programmes that have research capacity. practitioners, as well as being important to the success of We do not advocate the use of particular tools but aim to strategies. Conservationists have responsibilities towards the guide critical thinking in applying methods in ways that sup- communities they work in, to ensure at the very least they port depth of understanding and robust results appropriate to do not harm people [10], a premise that is encapsulated in the evidence required. The principles relate to four key stages policy commitments such as the Durban Accord on protected of conducting an impact evaluation found in the literature areas [11]. Well-being is also important for policy analysis, [24–26], and take into account some of the challenges to because its pursuit is a primary driver of people’s decision- evaluation in conservation such as nonlinear response out- making [12]. Interventions that support local well-being can comes, lack of comparators, multiple outcomes and complex increase environmentally desirable behaviour, and lead to confounding factors [24,27]. The four stages are: (1) defining positive local perceptions and engagement [4]. Using a outcomes of interest and well-being indicators, including for- well-being framework provides a holistic way to incorporate mulating complex theories of change and considering goals for different values (e.g. livelihoods and the environ- confounding factors; (2) designing the evaluation to link out- ment) into decision-making, which can also help to build comes to the intervention; (3) understanding processes of political support and mobilize funding. For these reasons, it change; and (4) collecting data on selected indicators and con- is vital that well-being is taken into account both in conserva- textual factors. Step 3 reflects the increasingly recognized tion programmes that explicitly incorporate livelihoods, and need in conservation policy for richer understandings of in those with narrower biodiversity targets. mechanisms through which impacts are produced [18]. Recent studies on the impacts of protected areas on pov- erty [13–15] have used robust quasi-experimental designs but have tended to focus on externally defined asset- and 2. Conceptualizing human well-being monetary-based measures, or on human–wildlife conflict and attitudes towards parks [16]. Further, by only measuring In conceptualizing well-being, there is a tension between a average net impacts, studies do not reveal how benefits and universal approach that allows comparisons, and ensuring losses are distributed across different groups of people. local relevance [28]. Local perspectives must drive our under- With the recent exception of a more nuanced evaluation of standing of well-being, as externally derived categories may marine protected areas (MPAs) [17], these types of studies not have meaning for local people, and thus will not account do not consider multiple dimensions of well-being and its for locally significant impacts of interventions. Any universal subjective aspects. Crucially, much research does not eluci- frameworks or methods used in evaluations must be adapt- date the mechanisms through which interventions impact able to locally meaningful formulations, made relevant to well-being [18], an element of evaluation that is especially the target population by using comparable categories with important for decision-making [19]. locally specific indicators. There is also a balance to be One difficulty in embarking on a well-being impact evalu- struck between objective and subjective definitions. In devel- ation is in operationalizing such a wide and complex concept. opment economics, well-being has been conceived of as an The array of conceptual frameworks for well-being [20] can objective concept mainly focused on material assets, and in cause confusion, especially among researchers and prac- social psychology as an internal, subjective psychological titioners trained in the natural rather than social sciences. state felt by the individual. Using either approach alone is Although the Conservation Measures Partnership [21] advo- insufficient. People will have different capabilities to gain cates inclusion of human well-being targets when designing benefits from assets, whereas a person’s expressed satisfac- and adaptively managing conservation projects, it provides tion with life is a poor guide to objective valuation of no methodological guidance on how to measure and evaluate material impoverishment, as it does not account for people well-being impacts, and ignores the possibility of unexpected adapting their preferences to harsh conditions [29]. Conserva- consequences, institutional changes resulting from interven- tionists should be interested both in objective indicators, as tions and the importance of local perspectives of change. they show tangible changes and are often most sought by Stephanson & Mascia [22] build upon this work with a funders and policy-makers, and also in subjective well- broader conception of well-being, but do not consider subjec- being, because people’s feelings and experiences impact on tive experiences, and remain focused on methods for participation and social sustainability. rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 One framework that combines objective and subjective presented in table 1. These principles are summarized in valuation and gives primacy to local understandings comes table 2. from the Wellbeing in Developing Countries (WeD) project [30]. Well-being is conceptualized as an outcome and a pro- (a) Defining outcomes and indicators cess, in three interacting dimensions: the objective material circumstances of a person, subjective evaluation of people’s (i) Principle 1: put local people at the centre of the evaluation goals and the processes they engage in, and a relational One of the central benefits of using a broadly defined, locally component [31]. This last dimension acknowledges that grounded conception of well-being is that it puts at the centre individual well-being is pursued in relation to other people, those people most affected by policy changes and interven- that social connectedness is a human need and that defi- tions. Local people should be involved throughout the nitions of a good life are socially constructed [32]. Culture process of evaluation, but most crucially when initially defin- is often viewed as external in discussions on poverty and ing the scope of the evaluation. Interventions are based on a well-being, but here it forms the lens through which all theory of change (ToC), which explains the process through aspects of well-being are constituted [33]. For example, the which the intervention is thought to give rise to specific out- significance of cattle-raising goes beyond a livelihood for comes. An evaluation is effectively testing this theory. To pastoralist communities to being a culturally meaningful start it is helpful explicitly to map out the ToC causal chain way of life entailing social contracts of ownership or use from inputs to outcomes, the underlying logic and the rights over land. The WeD approach emphasizes the holistic, social, behavioural and institutional assumptions being dynamic and social nature of well-being. It brings together a made [21,25], a process that allows space for reflection on unique configuration of interdependent elements, counterba- assumptions and context, and the development of hypoth- lancing a tendency in policy to privilege material well-being eses and evaluation questions. External drivers and and underplay subjective feelings and the social dimension of pressures such as government policies, market changes, cli- people’s lives [31]. mate change and environmental shocks, and how they may Another framework—the ‘Voices of the Poor’ (VoP)—is be changing well-being and interacting with the intervention, based on empirical data and is familiar to conservationists, must be considered in order to take confounding factors into because it was used in the Millennium Ecosystem Assessment account. A ToC is best developed with the participation of as a means of conceptualizing relationships between ecosystem local stakeholders who hold highly contextual knowledge, services and aspects of well-being. The project found five com- and may well consider potential consequences and unin- ponents commonly considered to constitute well-being among tended changes that would not otherwise be addressed. For individuals across 23 countries [34]. They are material assets, single-stranded projects, tracing potential changes may be health, social relations, security and freedom of choice and relatively straightforward, but for complicated projects with action. The last component, which underpins the others, means multiple components linked together (e.g. land rights, edu- having a sense of control over one’s life and the capacity to cation and livelihoods), more thought must be given to achieve what one values doing and being. This is easily over- interactions and feedback loops [46]. looked in conventional assessments but may be especially Qualitative research, using semi-structured, informal relevant for conservation interventions, which can be rejected if interviews and participant observation, and which is flexible perceived as imposed and undermining freedom with regard and open to unexpected findings, can provide details on to environmental behaviour [35]. On the other hand, interven- local nuances in the language used to express well-being, tions that secure local land tenure and improve natural resource and the priorities and aspirations that people have. For governance could increase feelings of empowerment [17]. example, Abunge et al. [35] used focus group discussions The VoP framework provides a useful checklist of themes with different stakeholder groups (e.g. women fish vendors to consider when starting an evaluation, based on empirical and beach seine captains) connected to a Kenyan coastal research. We combine it with ideas from the WeD research, fishery to understand how well-being was expressed by thus providing a thematically based framework that allows different types of people, how and why it had changed, depth of understanding, and can be used to guide the struc- and the hopes people had for the future of the fishery. ture of evaluations (table 1). The VoP domains will be Open-ended questions such as ‘How would you describe in informative to evaluators about which aspects of well-being general a person who is doing well in this community?’ to consider, bearing in mind that WeD’s three-dimensional encouraged people to open up and discuss what constituted perspective will help to define the questions asked and the a good life in this particular context. Qualitative research is type of data collected. For example, in studying the benefits also valuable at the start of evaluation to understand the his- of a payment for ecosystem services programme, in the torical, political and cultural issues that can shape people’s material domain, a relevant (and commonly used) objective perceptions, helping in the development of locally relevant indicator could be household income. The subjective dimen- questions for any structured and standardized questionnaires sion suggests consideration of levels of satisfaction with [36]. Other useful sources of insight are past studies, ethno- income changes, and feelings of fairness about benefit distri- graphies and informed sources who understand local bution. The relational dimension suggests the relevance of politics and history. how people use income, the way it can change relationships and differential capabilities within a household to benefit. (ii) Principle 2: select multiple outcomes to measure and consider subjective components 3. Principles to guide well-being evaluation Given the multi-dimensional nature of the well-being concep- We next discuss the nine key principles to bear in mind when tualization presented here, it cannot be captured by carrying out a robust evaluation based on the framework measuring only one outcome. Rather than using a standard rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 Table 1. Theoretical framework for well-being evaluation, which links ‘Voices of the Poor’ (VoP) well-being domains with perspectives from Wellbeing in Developing Countries (WeD). ‘Voices of the Poor’ well-being domains description and examples insights provided by WeD perspective and research material secure and adequate livelihoods not only about what people have, but what they can do and be, enough food and food security and how they feel about these things assets, e.g. land, natural resources, livestock, the ways in which objective material well-being outcomes are savings and capital, goods, housing, furniture defined and satisfied are socially and culturally constructed, and tools requiring attention to local context human as well as material resources are important, including knowledge and education health feeling strong and well health is subjectively experienced access to health services mental health is as significant as physical health in well-being appearing well having a healthy physical environment e.g. fresh air social relations good relations with family, community and collective well-being is significant for individual well-being in country culturally defined ways dignity, e.g. not being a burden, feeling social structures and institutions that enable people to pursue well- listened to being in relation to one another may be impacted by ability to help others and fulfil social interventions obligations people’s ideas and strategies for pursuing well-being may not be ability to care for children (including education compatible, resulting in trade-offs that must be confronted and marriage) security confidence in the future—predictability, people’s well-being and decisions are influenced by perceptions of peace future and perceived threats safe and secure environment, e.g. safety from capabilities to achieve other aspects of well-being may increase disasters security personal physical security and safety sustained security can only be the outcome of autonomy rather than security in old age and for future generations dependency freedom of choice and sense of control and power not about independence but self-endorsement of one’s own action ability to pursue what you value doing and behaviour, i.e. feeling personal value and interest regarding being, and meet aspirations actions ability to be a good person, e.g. to help others autonomy can be evaluated with regard to different aspects of people’s lives that they value related to the ability to adapt in times of change list of indicators that may be irrelevant, evaluators can more positive perspective than the concept of poverty, it is also target pertinent components suggested during qualitative important to include negative information when assessing inter- discussions with communities and considered sensitive to ventions. External interventions may contribute as much towards the intervention actions in the theory of change. The frame- ‘ill-being’ e.g. social exclusion, conflict or malnourishment, as work in table 1 can be used to guide this inquiry. For to well-being [34]. example, health and nutrition may be prioritized in the com- The need for measurement of multiple outcomes of munity and could be improved by access to water and food interventions is highlighted by the fact that there may be sources, or social capital could be increased through the trade-offs and synergies between outcomes [40]. For instance, establishment of community-based governance of natural there could be trade-offs between different dimensions of resources. Quantitative indicators could then include number well-being such as wealth and equity, invalidating of meals eaten per day or levels of participation in community conclusions about the overall impact of interventions on activities. Some outcomes relevant to well-being, such as social well-being if one dimension was missed. Considering these relations or political change, may not be amenable to quantifi- relationships in developing a theory of change, and later in cation at all [38]. Although the concept of well-being takes a data analysis, moves the evaluation away from simplistic rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 Table 2. Summary of guiding principles for evaluating impacts of conservation interventions on well-being. references for guiding principle examples of approaches to addressing the principle further details defining outcomes and indicators (1) put local people at the centre of the start with flexible qualitative research, e.g. semi-structured interviews to [35,36] analysis explore local understanding and components of well-being qualitative research on the local context, e.g. through literature reviews and informed sources map out theory of change developed with participating communities [21,25,37] and stakeholders (2) select multiple outcomes to measure select multiple well-being indicators based on local priorities and and consider subjective components outcomes in theory of change collect qualitative data on outcomes not amenable to quantification, [38] e.g. institutional change collect data on subjective feelings about pertinent aspects of well-being [39] allow opportunities for people to voice unintended consequences, and negative outcomes consider relationships (trade-offs and synergies) between outcomes [40] evaluate impacts on security through identifying locally relevant [41] indicators evaluation design (3) match evaluation design to the setting consider quasi-experimental and before-after-control-intervention [14,15,26,42] and questions asked designs if no baseline data, consider recall interviews for simple variables [43] control-intervention designs without baselines should be supported by other data alternatives to quasi-experimental designs: theory based, case studies, [19,37,44,45] participatory methods understanding processes of change (4) provide evidence of causal linkages theory-based analysis using quantitative and qualitative methods to [36,37,46] understand the how and why of impacts quantitative data can produce estimates of the contribution of different [18] causal mechanisms (5) consider trajectories of change anticipate and acknowledge possible trajectories of change and measure [17,47,48] ex-post impacts if possible (6) investigate institutions and governance institutional analysis using secondary and primary data [49] structures participatory institutional profiling (before-and-after intervention) [50] data collection (7) select and apply methods with choose tools appropriate to the cultural context and apply with [51,52] sensitivity to context consideration to equity combine methods to take advantage of their strengths, e.g. quantitative measures with qualitative insights (8) take into account heterogeneity disaggregate data according to qualitative understandings of social [53,54] structures and livelihoods individual interviews to capture differences across age and gender [39] within households (9) ensure independence recruit locally trusted people independent of implementing institutions [36] and conservation draw upon local language skills and in-country researchers rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 narratives towards a realistic acknowledgement of potential matching techniques [26]. Controls are selected based on their gains and losses that can inform decision-making. It can similarity with the intervention targets on a suite of measurable also highlight aspects of well-being that people feel cannot covariates that are thought to affect participation in the inter- be ‘traded-off’ at all, such as cultural heritage [55]. vention and the outcomes of that intervention (according to Observable, quantitatively measured changes can provide the theory of change). For example, in the case of protected credible evidence of impact to external audiences, but areas, these covariates could be distance to a city, elevation perceived change by local communities—reflecting the subjec- and asset-based indicators [15]. Controls should be selected tive aspect of well-being—may also be significant, especially outside the zone of influence of the intervention, to avoid spil- for conservation managers on the ground wanting to take miti- lover effects of the project or contamination by other gating measures to improve elements of the project people are interventions. Clements & Milner-Gulland [42] evaluated liveli- not satisfied with, in order to gain local support. Indeed, per- hood outcomes of protected areas in Cambodia, and selected ceived well-being may be at odds with objective measures, control villages based on matching variables thought to affect highlighting where there is dissatisfaction with interventions. village-level poverty and access to natural resources prior to In villages involved in MPAs in Indonesia, for example, the intervention. They ensured that control villages were there were negative changes in perceived well-being despite more than 20 km from the protected area border. It is impor- increases in wealth (based on material assets) during the tant, however, to recognize that controls in that study did not course of the intervention, owing to inequitable sharing of reflect the groups to which people actually compared them- benefits, conflict and unmet expectations [17]. selves, which could affect subjective well-being and therefore Security—living under conditions where there is predict- interpretations of impact. A more qualitative approach to match- ability—is a key constituent of well-being (table 1), drawing ing could instead be used, by selecting controls with attention to temporal aspects of well-being. People’s current participation from local people, to identify controls that are well-being takes place in the context of past experiences, as both methodologically appropriate (i.e. deal with confounding well as expectations, fears and aspirations about the future. factors) and meaningful for local people (e.g. geographically People engage in trade-offs through time to establish security closer). It is also worth considering that there may be hetero- and reduce threats [41]. Especially when faced with rapid geneity within the treatment, such as spatial differences in changes, perhaps as a result of an external intervention, the impact of an intervention that will be disguised if only aver- uncertainty prevails, reducing security and therefore well- age effects are considered. For example, people experienced the being. Conservation may increase feelings of insecurity, same alternative livelihood project very differently across even if implemented in the hope of improving environmental different villages in Tanzania [58]. security in the longer term. For example, in Tanzania con- Using baselines in addition to controls to create a BACI cerns about future land-use restrictions were highest among (before-after-control-intervention) evaluation allows impacts households near to Tarangire National Park, and this influ- to be isolated from two biases: selection bias in which the tar- enced decisions to convert land to agriculture to secure gets of the intervention are a non-random selection of the land tenure, ultimately affecting conservation outcomes whole population (for example, when wealthier families are [56]. Alternatively, threats of large-scale land acquisitions in more equipped to participate in a PES (payment for ecosys- countries where community tenure rights are weak may tem services) scheme), and concurrent change (for example, lead to higher participation in conservation activities, such improvements in wealth affected by both a sustainable liveli- as seeking community forestry titles [57]. Evaluating impacts hoods intervention and wider economic changes) [6]. of interventions on security could involve developing indi- Selection bias is tackled by matching techniques in quasi- cators relevant to the context, such as livelihood and experimental designs, while tracking change through time income diversity, access to justice, and the functioning and in both the control and intervention (the difference in differ- membership of collective institutions [41]. ence method) takes into account differences between treatments and controls that are constant over time, including unobserved intrinsic characteristics such as levels of motiv- (b) Evaluation design: linking outcomes to the ation or optimism. A prospective evaluation, developed at the same time as the intervention is designed, is preferable intervention to a retrospective evaluation as it allows the collection of (i) Principle 3: match evaluation design to the setting and baseline data and is more likely to generate valid counterfac- questions asked tuals [26]. It is possible, however, to construct an approximate At the heart of evaluation is the process of attributing specific baseline ex-post from secondary data or carefully designed effects, in this case changes in well-being, to the intervention surveys asking participants to recall specific, easily remembe- rather than to other factors. This can be achieved by inferring red variables such as assets and link them to locally important the counterfactual—what would have happened in the absence events [59], although researchers should be aware that recall of the intervention—through the identification of controls, thus data are prone to inaccuracies [43]. Although before–after reducing bias from confounding factors [24]. Quantitative and control–intervention comparisons on their own make experimental and quasi-experimental designs that allow assess- for weaker causal inferences, where a full BACI design is not ment of the magnitude of impacts are often viewed as the ‘gold possible, these simpler designs can provide credible insights standard’ for answering the attribution question in a robust if supported by other data such as community perceptions way in development and conservation. Truly experimental that substantiate or refute the quantitative trends. designs, where the researcher randomly assigns interventions, There are real-world constraints to using controls and are not often possible in conservation [6], and instead quasi- quasi-experimental designs to attribute changes to an inter- experimental methods are used in which control groups vention; they require a large sample size, may not be (most likely households or villages) are identified through suitable for complex or broadly defined interventions such rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 as regional policy with small number of units, and require a how contextual factors may be hindering or reinforcing suitable budget and technical skills [19]. There may be ethical change in particular outcomes [62]. This is especially impor- concerns such as raising expectations or subjecting people to tant where evaluation is directed towards lesson learning. surveys that bring no benefits [51]. Controls may be difficult Theory-based approaches as described above take a deduc- to find if, for instance, the intervention focusses on small tive approach by empirically discounting alternative areas of particular importance for conservation that are plausible explanations for outcomes. For example, a theory- very different from the rest of the region (e.g. islands of natu- based analysis showed that an infant nutrition project in ral forest). Instead of, or in addition to, quasi-experimental India mis-targeted mothers who, although gaining knowl- designs there are a range of alternative approaches through edge, were not able to put it into practice as it was their which causal inferences can be made that can be divided mothers-in-law and husbands who made decisions about broadly into three types: theory-based, case studies and par- food and child-raising, an insight that was initially found ticipatory. These methods tend to be stronger than through reading anthropological studies [37]. In that quantitative counterfactual approaches in explanation and example, it was a design flaw in the project that led to poor contextualization, but weaker on estimating the magnitude results, but a theory-based approach can also detect problems of impacts [19]. Decisions about which design to use must with implementation. Ferraro & Hanauer [18] demonstrate be made on the basis of resources and skills available, and how a quasi-experimental design can incorporate quantitat- the types of evidence required. ive analysis of causal pathways to show how protected Theory-based methods consider the likely chains of areas in Costa Rica reduced poverty mainly through tourism. impact, presenting alternative hypotheses for change and The fact that the three measured causal mechanisms examining them through both qualitative and quantitative accounted for only two-thirds of impacts in this study data. These can highlight where there is a break in the serves to highlight how theories of change must incorporate causal chain to explain impacts of an intervention [46]. A in-depth and complex understanding of socio-ecological case study approach focuses on particularly successful (or systems to fully capture processes of change. unsuccessful) cases (e.g. villages or intervention types) to examine the contribution of the intervention to outcomes. (ii) Principle 5: consider trajectories of change Comparisons across the cases can elucidate the combinations Well-being is not a discrete outcome, but an ongoing of causal factors (the types of intervention, methods of dynamic process, changing through the course of an inter- implementation such as levels of capacity building, and con- vention and beyond [63]. Trajectories of change are not textual factors such as tenure regimes, wealth levels and local linear, resulting in attribution errors if well-being effects of leadership) explaining changes in well-being indicators [44]. an intervention are measured at only one point in time [47]; Participatory methods allow communities to systematically for example, there could be high initial impact owing to assess changes themselves; for example, group discussions, improved forest governance arrangements that is eroded in which causes for reported changes in well-being are through time by pre-existing power structures. Monitoring ranked and scored, can show relative perceived impacts of throughout the course of a project is ideal as it allows real- intervention and non-intervention factors, with greater time feedback for learning and adjustment [62]. Ex-post levels of agreement between groups indicating reliability assessments are rare, but may be crucial in understanding [45]. The use of ‘reflexive counterfactuals’ where participants longer-term impacts and sustainability, and for taking into compare themselves before and after the intervention, by account time lags between intervention and effect. For prioritizing perceptions, may be subject to bias, but provide example, any initial gains in aspects of well-being such as important information, for example, in protected area man- fish catch, wealth and empowerment were lost after external agement focused on improving equity and effectiveness at support for MPAs in Indonesia was withdrawn, suggest- the site level [60]. ing that interventions of this kind need to build capacity, Given the pros and cons of the quasi-experimental and gain broad-based support and sustain funding [17]. Realisti- alternative approaches to evaluation design, Roe et al. [61] cally, it may only be possible to evaluate shorter-term advocate a sensible two-track system, in which in-depth outputs or outcomes and indicate where longer-term longitudinal evaluation using controls for a selection of repre- impacts could occur [25]. The reference standards of those sentative interventions of strategic relevance is combined affected by interventions can also change, potentially as a with rapid, participatory assessment more feasible for the result of the interventions themselves [48]; for instance, majority of projects. Policy-makers and donors may empha- increasing material wealth may lead to wealth becoming a size the former approach, to gain evidence of the more important component of well-being for some, but still magnitude of impact needed for cost-effectiveness analysis. reduce well-being owing to rising aspirations not matched This may guide decisions on whether to replicate or fund by opportunities [64]. similar interventions in the future. Field managers wanting to understand people’s experiences and perspectives may focus on the latter, combined with quantitative indicators (iii) Principle 6: investigate institutions and governance structures analysed through theory-based analysis. Well-being depends on institutions—human-devised informal constraints and formal rules—which govern relationships between individuals and groups, and between humans and (c) Understanding processes of change ecosystems [65]. The choices made about the types of organiz- (i) Principle 4: provide evidence of causal linkages ations conservationists work with (state agencies, private Using a counterfactual approach that is limited to attributing corporations, customary authorities) will profoundly shape outcomes to an intervention cannot answer the fundamental the institutional landscape, affecting representation, citizen- questions of how and why a project is or is not effective, and ship, and ultimately social and environmental sustainability rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 [66]. Ill-considered interventions can subvert existing co-produced data relatively quickly and cheaply, and to institutions and cultural practices that act to regulate demonstrate local involvement in evaluation. Relevant environmental behaviour, resulting in alienation and methods include qualitative resource mapping, timelines, counter-productive conservation results [67]. Natural resource focus groups, village histories, ranking and scoring [45]. management interventions can, on the other hand, support The mainstreaming of PRA, however, has resulted in the pro- improved governance, contributing to poverty alleviation cess becoming somewhat ritualized, and the participatory [68], and act as vehicles for social change and improved partici- label is often used to mask standard extractive data collection. pation. This emphasizes the relational dimension of well- Participatory discussions can be susceptible to co-option by being, and the attention required in analyses to the relationship local elites, silencing those most affected [71]. If PRA tech- between individual and collective well-being, which is shaped niques are used, experienced and trained facilitators by dynamic institutions such as norms. During evaluations, it familiar with the local context, and independent of the inter- is important to understand the functioning of institutions and vention, should lead the work to ensure sensitive and governance structures acting within and upon communities, equitable discussion [52]. Groups should be appropriately and how conservation interventions affect it, in turn impacting constituted; for example, different gender and age groups individual well-being. This can be achieved through, for may not be comfortable in mixed groups. In some situations, example, institutional profiling with local people in which for example, where communities are suffering from research visual methods such as Venn diagrams can aid discussions fatigue, or issues are highly contested, individual or house- about key institutions, relationships and forms of power at hold interviews may be a better option. different scales that influence people [49,50]. (ii) Principle 8: take into account heterogeneity within the target (d) Data collection group (i) Principle 7: select and apply methods and toolkits with Just as different people are able to access different ecosystem sensitivity to the research context services, there will be trade-offs between the well-being There are many tools and methods available—both with a impacts of interventions on different people, between or quantitative and qualitative slant—which conservation eva- within communities [53]. Standard experimental and quasi- luators can draw upon in collecting data on well-being experimental impact evaluation methods may produce an indicators as well as on contextual and confounding factors. average effect of an intervention across households, commu- However, reflection on their appropriateness to the context nities or the whole population being investigated, but this is vital for data validity. Schreckenberg et al. [51] provide a does not address which types of people win and lose, and useful compilation of rapid social research methods that why [72], unless heterogeneity is purposefully incorporated can be used to collect data on the well-being impacts of con- into research design and data collection. It is especially servation interventions. These methods are compatible with important to ensure that vulnerable groups such as the poor- our proposed framework for well-being evaluation, but est, landless, migrants or mobile resource users (such as thought needs to be given not only to selecting particular fishers, pastoralists and forest dwelling groups) are included methods appropriate to the context, but also to the ways in evaluations as they are often invisible unless local knowl- these tools are applied to deal with culturally sensitive edge is used [73]. The impacts of interventions on issues, vested interests and equity. For example, how will households with different livelihoods may differ signifi- the use of particular local informants skew the evaluator’s cantly; for example, non-timber forest product collectors in understanding of the issues, and how will marginalized Cambodia were significantly better off in terms of basic groups be accessed? Tallying scores for different outcomes, necessities inside a protected area than outside owing to as advocated in some guidelines [69], is attractive for quick secure access to resources [54]. Impacts of interventions on and standardized assessments. But, in isolation, scoring sys- social dynamics between different groups can undermine tems run the risk of aggregating over a broad range of well-being; for example, targeting only some groups may indicators and social groups. This may lead to their falling create conflict [38]. Qualitative research can elucidate social between two stools; meaningless both locally and compara- structures, wealth and livelihood differences that can form tively. Even where carefully selected and appropriately the basis for appropriate disaggregation of data, as well as applied, a single method may be inadequate on its own. improving understanding of the ways different groups For example, the Basic Necessities Survey (BNS) is an conceive of well-being. index-based tool that assesses household poverty based on Social surveys are often carried out at the household locally defined ‘necessities’ [70]. A key benefit of the level, but there is likely to be intra-household variability, method is that it produces quantifiable results through a par- with differences in well-being according to gender and ticipatory and democratic process, but the score is aggregated age. Britton & Coulthard [39] found that the domains of at the household level and gives little detail about the pro- life important for well-being, and satisfaction with these cesses of change. A hybrid research design augmenting the domains, differed significantly between men and women in BNS with data that chime with WeD’s broader conception fishing communities in Northern Ireland. Women may lose of well-being could form a more robust way to capture the resource access under payment-based conservation interven- complexity of well-being change. For example, the incorpor- tions but receive few of the benefits, which are given to male ation of semi-structured interviews could focus on household heads [74], or they may be excluded altogether subjective experiences, and capture causal mechanisms. from participation [75]. Well-being is experienced by individ- The methods of Participatory Rural Appraisal (PRA) used uals, and so they should be the primary unit of impact by development practitioners have been enthusiastically assessment rather than the household, which is a common adopted by many conservationists, to provide locally evaluation unit in economic assessments. However, as the rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 WeD framework suggests, individual well-being is relational. evaluation, and improve their understanding of objective and Therefore, collective well-being at different scales of social subjective well-being impacts. This understanding is perhaps relationship is significant for individual well-being, although most urgently needed in materially poor areas of the Global the extent and nature of these relationships will differ across South, but well-being provides a useful way to measure social cultures and contexts [63]. impacts of conservation regardless of the wealth status of the population [77]. Mixed methods can better support causal and explanatory analysis, and conservation researchers should not (iii) Principle 9: ensure independence be reticent about using qualitative data in their own right. Although the research design and selection of methods and Far from signaling a lack of rigour, qualitative approaches are techniques are important elements to consider in an evalu- necessary to appropriately disaggregate data, identify covari- ation, the quality of research will ultimately be defined by ates, explain the processes involved in producing well-being how it is conducted and the relationships established between impacts, and allow local voices to be heard. The important the researchers and participants [76]. Evaluations conducted basis for rigour is the appropriate application of techniques, by people who are independent of actors implementing inter- either quantitative or qualitative [78]. ventions or otherwise working in the system will result in Theevidencebaseonthe impactsofconservationinter- improved validity. Although research is often labelled as ‘inde- ventions on human well-being is weak, and there is much to pendent’, it is all too often facilitated by vehicles clearly be learnt to support decision-making about the range of interven- marked with government or NGO logos, or by people who tions used in conservation under different contexts. For example, are linked to the intervention or to conservation more gener- in a recent systematic review of 136 community-based conserva- ally. Marginalized rural people are likely to find it difficult tion evaluations, 80% of the studies included were rated as poor to talk candidly with powerful individuals, which suggests quality on the basis of conflict of interest (i.e. lack of inde- the importance of considering local language, trust and ethni- pendence), data validity and other problems [79], arguably city in building a research team. Calling upon local expertise throwing the results of an otherwise meticulous statistical analy- and language skills will decrease the risk of obscuring local sis into serious doubt. There will inevitably be trade-offs between meaning and realities [36]. In-country students or young conservation outcomes and human well-being outcomes [80], researchers supported by experienced evaluators may be and between different elements of well-being itself. Only by good options for tight budgets, and this approach will also assessing well-being in a way that tackles complexity, context, contribute to capacity building within the country. politics and the wide range of impacts that conservation can bring, can stakeholders hope to openly discuss and negotiate trade-offs in a systematic and transparent way. Conservationists 4. Conclusion have a responsibility to the communities in which they carry Understanding the intricacies and dynamics of what people out their activities, and using well-implemented well-being consider to be a good life is far from straightforward. evaluations can improve accountability and lesson learning, ulti- Well-being is multi-faceted, and varies between contexts and mately improving the likelihood of successful, locally supported cultures, within communities and households, and through conservation in the long-term. time. Attributing well-being change to interventions must also take place in the context of complex and dynamic influences Authors’ contributions. All authors made contributions to the conception at multiple scales. This mayseem like an impossible task forcon- of this paper. E.W. led the drafting, and K.M.H., E.B., T.C., J.T.M., D.W. and E.J.M.-G. critically reviewed drafts for intellectual content. servation practitioners, especially with limited resources and All authors gave final approval for publication. expertise. In the face of this complexity, however, formulaic Competing interests. We declare we have no competing interests. methods will not work. Conservationists should not use pre- Funding. This research was funded by the ESRC-DFID joint fund for scriptive designs and methods without thinking about their poverty alleviation (ES/J018155/1). applicability to the case, how best to apply to them and what Acknowledgements. We would like to thank Allister McGregor for sort of data they will produce. By engaging with the principles comments on an earlier draft of this paper. This paper is a contri- and concepts set out here, and summarized in table 2, conserva- bution to Imperial College’s Grand Challenges in Ecosystems and tionists can hope to untangle the complexity of social impact the Environment initiative. References 1. Millennium Ecosystem Assessment. 2005 Ecosystems 4. Coulthard S, Johnson D, McGregor JA. 2011 Poverty, conservation investments. PLoS Biol. 4, e105. and human well-being: synthesis. Washington, DC: sustainability and human wellbeing: A social (doi:10.1371/journal.pbio.0040105) Island Press. wellbeing approach to the global fisheries crisis. 7. Leisher C, Sanjayan M, Blockhus J, Larsen SN, 2. West P, Igoe J, Brockington D. 2006 Parks and Glob. Environ. Chang. 21, 453 – 463. (doi:10.1016/j. Kontoleon A. 2012 Does conserving biodiversity peoples: the social impact of protected areas. Annu. gloenvcha.2011.01.003) work to reduce poverty? A state of knowledge Rev. Anthropol. 35, 251 – 277. (doi:10.1146/ 5. Stiglitz J, Sen A, Fitoussi J. 2009 Final report of the review. In Biodiversity conservation and poverty annurev.anthro.35.081705.123308) commission on the measurement of economic alleviation: exploring the evidence for a link 3. Coad L, Campbell A, Miles L, Humphries K. 2008 performance and social progress. Paris, France. (eds D Roe, J Elliot, C Sandbrook, MJ Walpole), The costs and benefits of protected areas for local (www.stiglitz-sen-fitoussi.fr/documents/ pp. 143 – 159. Chichester, UK: John Wiley & Sons. livelihoods: a review of the current literature. rapport_anglais.pdf ) 8. OECD (Organisation for Economic Cooperation and Cambridge, UK: UNEP World Conservation 6. Ferraro PJ, Pattanayak SK. 2006 Money for nothing? Development). 2011 How’s life? Measuring well- Monitoring Centre. A call for empirical evaluation of biodiversity being. Paris, France: OECD Publishing. rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 9. Summers JK, Smith LM, Case JL, Linthurst RA. 2012 data in conservation planning. Conserv. Biol. 28, 39. Britton E, Coulthard S. 2014 Assessing the social A review of the elements of human well-being with 1236 – 1248. (doi:10.1111/cobi.12357) wellbeing of Northern Ireland’s fishing society an emphasis on the contribution of ecosystem 23. Greene JC, Benjamin L, Goodyear L. 2001 The using a three-dimensional approach. Mar. services. Ambio 41, 327 – 340. (doi:10.1007/s13280- merits of mixing methods in evaluation. Evaluation Policy 37, 28 – 36. (doi:10.1016/j.marpol.2012. 012-0256-7) 7, 25 – 44. (doi:10.1177/13563890122209504) 04.011) 10. Roe D, Oviedo G, Pabon L, Painter M, Redford K, 24. Ferraro PJ. 2009 Counterfactual thinking and impact 40. Agrawal A, Benson CS. 2011 Common property Siegele L, Springer J, Thomas D, Walker Painemilla evaluation in environmental policy. New Dir. Eval. theory and resource governance institutions: K. 2010 Conservation and human rights: the need for 122, 75 – 84. (doi:10.1002/ev.297) strengthening explanations of multiple outcomes. international standards. London, UK: IIED. 25. Leeuw F, Vaessen J. 2009 Impact evaluations and Environ. Conserv. 38, 199 – 210. (doi:10.1017/ 11. World Parks Congress. 2003 The Durban Accord. See development: NONIE guidance on impact evaluation. S0376892910000925) http://cmsdata.iucn.org/downloads/durbanaccorden. Washington, DC: NONIE (The Network of Networks 41. Wood G. 2007 Using security to indicate wellbeing. pdf. on Impact Evaluation). In Wellbeing in developing countries: from theory to 12. Deci EL, Ryan RM. 2010 The ‘what‘ and ‘why‘ 26. Gertler PJ, Martinez S, Premand P, Rawlings LB, research (eds I Gough, JA McGregor), pp. 109 – 132. of goal pursuits: human needs and the self- Vermeersch CMJ. 2011 Impact evaluation in practice. Cambridge, UK: Cambridge University Press. determination of behavior. Psychol. Inq. 11, Washington, DC: The World Bank. 42. Clements T, Milner-Gulland EJ. 2015 The impact of 227 – 268. (doi:10.1207/S15327965PLI1104_01) 27. Baylis K et al. 2015 Mainstreaming impact payments for environmental services and protected 13. Andam KS, Ferraro PJ, Sims KRE, Healy A, Holland evaluation in nature conservation. Conserv. Lett. areas on local livelihoods and forest conservation in MB. 2010 Protected areas reduced poverty in Costa (doi:10.1111/conl.12180) Northern Cambodia. Conserv. Biol. 29, 78 – 87. Rica and Thailand. Proc. Natl Acad. Sci. USA 107, 28. Boarini R, Kolev A, McGregor JA. 2014 Measuring (doi:10.1111/cobi.12423) 9996 – 10 001. (doi:10.1073/pnas.0914177107) wellbeing and progress in countries at different 43. Papworth SK, Rist J, Coad L, Milner-Gulland EJ. 14. Naughton-Treves L, Alix-Garcia J, Chapman CA. 2011 stages of development: towards a more universal 2009 Evidence for shifting baseline syndrome in Lessons about parks and poverty from a decade of conceptual framework. OECD Working Paper No. conservation. Conserv. Lett. 2, 93 – 100. (doi:10. forest loss and economic growth around Kibale 325: OECD. 1111/j.1755-263X.2009.00049.x) National Park, Uganda. Proc. Natl Acad. Sci. USA 29. Sen A. 1999 Development as freedom. New York: 44. Byrne D, Ragin C. 2009 The SAGE handbook of case- 108, 13 919 – 13 924. (doi:10.1073/pnas. Knopf Press. based methods. London, UK: Sage. 1013332108) 30. University of Bath. 2002 Wellbeing in Developing 45. Catley A, Burns J, Abebe D, Suji O. 2008 15. Canavire-Bacarreza G, Hanauer MM. 2013 Estimating Countries (WeD) ESRC Research Group [Online]. Participatory impact assessment: a guide for the impacts of Bolivia’s protected areas on poverty. See http://www.welldev.org.uk/research/ practitioners. Medford, MA: Feinstein International World Dev. 41, 265 – 285. (doi:10.1016/j.worlddev. research.htm. Center. 2012.06.011) 31. McGregor A, Sumner A. 2010 Beyond business as 46. Rogers PJ. 2009 Matching impact evaluation design 16. Pullin AS et al. 2013 Human well-being impacts of usual: what might 3-D wellbeing contribute to MDG to the nature of the intervention and the purpose terrestrial protected areas. Environ. Evid. 2, 19. momentum? IDS Bull. 41, 104 – 112. (doi:10.1111/j. of the evaluation. J. Dev. Effect. 1, 217 – 226. (doi:10.1186/2047-2382-2-19) 1759-5436.2010.00111.x) (doi:10.1080/19439340903114636) 17. Gurney GG, Cinner J, Ban NC, Pressey RL, Pollnac R, 32. Deneulin S, McGregor JA. 2010 The capability 47. Woolcock M. 2009 Towards a plurality of methods Campbell SJ, Tasidjawa S, Setiawan F. 2014 Poverty approach and the politics of a social conception of in project evaluation: a contextualized approach to and protected areas: an evaluation of a marine wellbeing. Eur. J. Soc. Theory 13, 501 – 519. (doi:10. understanding impact trajectories and efficacy. integrated conservation and development project in 1177/1368431010382762) J. Dev. Effect. 1, 1 – 14. (doi:10.1080/ Indonesia. Glob. Environ. Chang. 26, 98 – 107. 33. White S, Ellison M. 2007 Wellbeing, livelihoods and 19439340902727719) (doi:10.1016/j.gloenvcha.2014.04.003) resources in social practice. In Wellbeing in 48. Milner-Gulland EJ et al. 2014 Accounting for the 18. Ferraro PJ, Hanauer MM. 2014 Quantifying causal developing countries: from theory to research (eds impact of conservation on human well-being. mechanisms to determine how protected areas I Gough, JA McGregor), pp. 157 – 175. Cambridge, Conserv. Biol. 28, 1160 – 1166. (doi:10.1111/cobi. affect poverty through changes in ecosystem UK: Cambridge University Press. 12277) services and infrastructure. Proc. Natl Acad. Sci. USA 34. Narayan D, Chambers R, Shah MK, Petesch P. 2000 49. Petitt J. 2013 Power analysis: a practical guide. 11, 4332 – 4337. (doi:10.1073/pnas.1307712111) Voices of the poor: crying out for change. New York, Stockholm, Sweden: SIDA (Swedish International 19. Stern E, Stame N, Mayne J, Forss K, Davies R, Befani NY: Oxford University Press for the World Bank. Development Cooperation). B. 2012 Broadening the range of designs and 35. Abunge C, Coulthard S, Daw TM. 2013 Connecting 50. Messer N, Tomsley P. 2003 Local institutions methods in impact evaluation. Working Paper 38. marine ecosystem services to human well-being: and livelihoods: guidelines for analysis. Rome, Italy: London, UK: DFID (Department for International insights from participatory well-being assessment in FAO. Development). Kenya. Ambio 42, 1010 – 1021. (doi:10.1007/ 51. Schreckenberg K, Camargo I, Withnall K, Corrigan C, 20. Agarwala M, Atkinson G, Palmer Fry B, Homewood s13280-013-0456-9) Franks P, Roe D, Scherl LM, Richardson V. 2010 KM, Mourato S, Rowcliffe JM, Wallace G, Milner- 36. Camfield L, Crivello G, Woodhead M. 2009 Social assessment of conservation initiatives: a Gulland EJ. 2014 Assessing the relationship between Wellbeing research in developing countries: review of rapid methodologies. London, UK: IIED. human wellbeing and ecosystem services: a review reviewing the role of qualitative methods. Soc. 52. Mayoux L, Chambers R. 2005 Reversing the of frameworks. Conserv. Soc. 4, 437 – 449. (doi:10. Indic. Res. 90, 5 – 31. (doi:10.1007/s11205-008- paradigm: quantification, participatory methods 4103/0972-4923.155592) 9310-z) and pro-poor impact assessment. J. Int. Dev. 17, 21. Conservation Measures Partnership. 2012 Addressing 37. White H. 2009 Theory-based impact evaluation: 271 – 298. (doi:10.1002/jid.1214) social results and human wellbeing targets in principles and practice. J. Dev. Effect. 1, 271 – 284. 53. Daw T, Brown K, Rosendo S, Pomeroy R. 2011 conservation projects. Draft Guidance. See http:// (doi:10.1080/19439340903114628) Applying the ecosystem services concept to poverty cmp-openstandards.org/wp-content/uploads/2014/ 38. Devereux S, Roelen K, Be´ne´ C, Chopra D, Leavy J, alleviation: the need to disaggregate human well- 03/Guidance-on-HWT-and-Social-Results-in- McGregor JA. 2013 Evaluating outside the box: an being. Environ. Conserv. 38, 370 – 379. (doi:10. Conservation-Projects-v2012-06-27.pdf. alternative framework for analysing social protection 1017/S0376892911000506) 22. Stephanson SL, Mascia MB. 2014 Putting people on programmes. Brighton, UK: Institute of Development 54. Clements T, Suon S, Wilkie DS, Milner-Gulland EJ. the map through an approach that integrates social Studies. 2014 Impacts of protected areas on local livelihoods rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 in Cambodia. World Dev. 64, S125 – S134. (doi:10. 63. McGregor JA. 2007 Researching wellbeing: from 72. Ravallion M. 2009 Should the randomistas 1016/j.worlddev.2014.03.008) concepts to methodology. In Wellbeing in rule? Econ. Voice. 6, 1 – 5 (doi:10.2202/1553-3832. 55. Hirsch PD, Adams WM, Brosius JP, Zia A, Bariola N, developing countries: from theory to research (eds 1368). Dammert JL. 2011 Acknowledging conservation trade- I Gough, JA McGregor), pp. 316 – 350. Cambridge, 73. Homewood K, Coast E, Thompson M. 2004 In- offs and embracing complexity. Conserv. Biol. 25, UK: Cambridge University Press. migrants and exclusion in East African rangelands: 259–264. (doi:10.1111/j.1523-1739.2010.01608.x) 64. Copestake J, Guillen-Royo M, Chou W-J, Hinks T, access, tenure and conflict. Africa 74, 567 – 610. 56. Baird TD, Leslie PW, McCabe JT. 2009 The effect of Velazco J. 2009 The relationship between economic (doi:10.3366/afr.2004.74.4.567) wildlife conservation on local perceptions of risk and subjective wellbeing indicators in Peru. Appl. 74. Bedelian C. 2014 Conservation, tourism and and behavioral response. Hum. Ecol. 37, 463 – 474. Res. Qual. Life 4, 155 – 177. (doi:10.1007/s11482- pastoral livelihoods: wildlife conservancies in the (doi:10.1007/s10745-009-9264-z) 009-9070-1) Maasai Mara, Kenya. PhD thesis, University College 57. Neef A, Touch S, Chiengthong J. 2013 The politics 65. Butler CD, Oluoch-Kosura W. 2006 Linking future London. and ethics of land concessions in rural Cambodia. ecosystem services and future human well-being. 75. Corbera E, Brown K, Adger NW. 2007 The equity and J. Agric. Environ. Ethics. 26, 1085 – 1103. (10.1007/ Ecol. Soc. 11, 30. legitimacy of markets for ecosystem services. Dev. s10806-013-9446-y) 66. Ribot JC. 2006 Choose democracy: environmentalists’ Change 38, 587 – 613. (doi:10.1111/j.1467-7660. 58. Sainsbury K, Burgess ND, Sabuni F, Howe C, Puis E, socio-political responsibility. Glob. Environ. Chang. 16, 2007.00425.x) Killenga R, Milner-Gulland EJ. 2015 Exploring 115–119. (doi:10.1016/j.gloenvcha.2006.01.004) 76. White S, Petitt J. 2004 Participatory methods and stakeholder perceptions of conservation outcomes 67. Goldman MJ, de Pinho JR, Perry J. 2013 Beyond the measurement of wellbeing. Particip. Learn. from alternative income generating activities in ritual and economics: Maasai lion hunting and Action 50, 88 – 96. Tanzanian villages adjacent to Eastern Arc Mountain conservation politics. Oryx 47, 490 – 500. (doi:10. 77. Adams VM, Pressey RL, Stoeckl N. 2014 Navigating forests. Biol. Conserv. 191, 20 – 29. (doi:10.1016/j. 1017/S0030605312000907) trade-offs in land-use planning: integrating human biocon.2015.06.001) 68. Brown D, Schreckenberg K, Shepherd G, Wells A. well-being into objective setting. Ecol. Soc. 19, 53. 59. Bamberger M. 2010 Reconstructing baseline data for 2002 Forestry as an entry point for governance (doi:10.5751/ES-07168-190453) impact evaluation and results measurement. reform. London, UK: Overseas Development Institute 78. White H. 2002 Combining quantitative and Washington, DC: The World Bank. Policy Briefing No. 1. qualitative approaches in poverty analysis. World 60. Franks P, Roe D, Small R, Schneider H. 2014 Social 69. Cahyat A, Go¨nner C, Haug M. 2007 Assessing household Dev. 30, 511 – 522. (doi:10.1016/S0305-750X(01) assessment of protected areas: early experiences and poverty and wellbeing – a manual with examples from 00114-0) results of a participatory, rapid approach. London, Kutai Barat, Indonesia. Bogor, Indonesia: Center for 79. Brooks J, Waylen K, Mulder M. 2013 Assessing UK: IIED. International Forestry Research (CIFOR). community-based conservation projects: a systematic 61. Roe D, Grieg-Gran M, Mohammed E. 2013 Assessing 70. Davies R, Smith W. 1998 The basic necessities survey: review and multilevel analysis of attitudinal, the social impacts of conservation policies: rigour versus the experience of ActionAid Vietnam.Hanoi, Vietnam: behavioral, ecological, and economic outcomes. practicality. Briefing October 2013.London, UK:IIED. ActionAid. Environ. Evid. 2, 2. (doi:10.1186/2047-2382-2-2) 62. Bamberger M, Rao V, Woolcock M. 2010 Using 71. Blaikie P. 2006 Is small really beautiful? 80. McShane TO et al. 2010 Hard choices: making trade- mixed methods for monitoring and evaluation: Community-based natural resource management in offs between biodiversity conservation and human experiences from international development. Malawi and Botswana. World Dev. 34, 1942 – 1957. well-being. Biol. Conserv. 144, 966 – 972. (doi:10. Washington, DC: The World Bank. (doi:10.1016/j.worlddev.2005.11.023) 1016/j.biocon.2010.04.038) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Philosophical Transactions of the Royal Society B: Biological Sciences Pubmed Central

Guiding principles for evaluating the impacts of conservation interventions on human well-being

Philosophical Transactions of the Royal Society B: Biological Sciences , Volume 370 (1681) – Nov 5, 2015

Loading next page...
 
/lp/pubmed-central/guiding-principles-for-evaluating-the-impacts-of-conservation-h5Q21wynMj

References (87)

Publisher
Pubmed Central
Copyright
© 2015 The Authors.
ISSN
0962-8436
eISSN
1471-2970
DOI
10.1098/rstb.2015.0103
Publisher site
See Article on Publisher Site

Abstract

Guiding principles for evaluating the impacts of conservation interventions on rstb.royalsocietypublishing.org human well-being 1,2,3 2 3 Emily Woodhouse , Katherine M. Homewood , Emilie Beauchamp , 1 4 1 3 Tom Clements , J. Terrence McCabe , David Wilkie and E. J. Milner-Gulland Research Wildlife Conservation Society, 2300 Southern Boulevard, Bronx, NY 10460, USA Department of Anthropology, University College London, 14 Taviton Street, London WC1H 0BW, UK Cite this article: Woodhouse E, Homewood Department of Life Sciences, Imperial College London, Silwood Park Campus, Buckhurst Road, KM, Beauchamp E, Clements T, McCabe JT, Ascot SL5 7PY, UK Wilkie D, Milner-Gulland EJ. 2015 Guiding 4 Department of Anthropology, University of Colorado, 233 UCB, Boulder, CO 80309-0233, USA principles for evaluating the impacts of con- servation interventions on human well-being. Measures of socio-economic impacts of conservation interventions have largely been restricted to externally defined indicators focused on income, Phil. Trans. R. Soc. B 370: 20150103. which do not reflect people’s priorities. Using a holistic, locally grounded http://dx.doi.org/10.1098/rstb.2015.0103 conceptualization of human well-being instead provides a way to under- stand the multi-faceted impacts of conservation on aspects of people’s Accepted: 13 August 2015 lives that they value. Conservationists are engaging with well-being for both pragmatic and ethical reasons, yet current guidance on how to opera- tionalize the concept is limited. We present nine guiding principles based One contribution of 16 to a theme issue around a well-being framework incorporating material, relational and sub- ‘Measuring the difference made by protected jective components, and focused on gaining knowledge needed for areas: methods, applications and implications decision-making. The principles relate to four key components of an for policy and practice’. impact evaluation: (i) defining well-being indicators, giving primacy to the perceptions of those most impacted by interventions through qualitative research, and considering subjective well-being, which can affect engage- Subject Areas: ment with conservation; (ii) attributing impacts to interventions through ecology quasi-experimental designs, or alternative methods such as theory-based, case study and participatory approaches, depending on the setting and evi- Keywords: dence required; (iii) understanding the processes of change including evidence of causal linkages, and consideration of trajectories of change development, impact evaluation, livelihoods, and institutional processes; and (iv) data collection with methods selected poverty, social impact, well-being and applied with sensitivity to research context, consideration of hetero- geneity of impacts along relevant societal divisions, and conducted by Author for correspondence: evaluators with local expertise and independence from the intervention. Emily Woodhouse e-mail: e.woodhouse@ucl.ac.uk 1. Introduction In response to both evidence of the dependence of vulnerable human popu- lations on ecosystems [1] and the costs associated with some conservation interventions for local people [2,3], the question of how to reconcile conserva- tion with human development has emerged as a key policy issue [4]. Recognition of the inadequacies of narrow economic indicators such as income and consumption in measuring social development has shifted atten- tion to a broader and multi-faceted vision of human well-being [5]. For instance, a conservation intervention may improve the local economy by pro- viding jobs and alternative livelihoods, but could negatively affect other priorities for local communities such as social relationships or autonomy. Where the specifics of social impacts are not intentionally and systematically examined, they could be misunderstood or missed entirely, with repercussions for social justice and conservation outcomes. While there is ready acknowledg- ment in the environmental literature that robust empirical evaluation is & 2015 The Authors. Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 required in order to better understand which approaches can collection of data relevant to conservation planning rather work for biodiversity [6], extension of this premise to conser- than impact evaluation. vation impacts on human lives is still rare. Across a range of We propose accepting the plurality of the concept of well- conservation strategies, there is a lack of evidence of the being, and present guiding principles based around existing impacts on human well-being that adequately capture the theoretical frameworks, an approach that allows comparable complexity on the ground [7]. but locally relevant results. Our principles are intended to Despite the variety of definitions, there is increasing support evaluators in operationalizing a multi-dimensional agreement in international policy on a conception of human conceptualization of well-being to measure and understand well-being that encompasses objective material circumstances impacts in ways that align with realities on the ground. We of people’s lives such as housing, income, livelihoods, health take a pragmatic approach to evaluation focused on gaining and the environment, social aspects such as community net- the knowledge needed for decision-making and policy devel- works, and a subjective component capturing an individual’s opment, a perspective that necessitates flexibility and an assessment of their own circumstances [5,8]. Well-being can openness to mixed methods, incorporating quantitative and therefore be defined as a positive physical, social and qualitative data and analysis [23]. Our aim is for the principles mental state [9]. Understanding the impacts of conservation to be useful for conservation practitioners and adaptable to on the multiple dimensions of people’s lives and working small-scale projects with limited budgets and technical exper- to improve them are ethical imperatives for conservation tise, as well as larger programmes that have research capacity. practitioners, as well as being important to the success of We do not advocate the use of particular tools but aim to strategies. Conservationists have responsibilities towards the guide critical thinking in applying methods in ways that sup- communities they work in, to ensure at the very least they port depth of understanding and robust results appropriate to do not harm people [10], a premise that is encapsulated in the evidence required. The principles relate to four key stages policy commitments such as the Durban Accord on protected of conducting an impact evaluation found in the literature areas [11]. Well-being is also important for policy analysis, [24–26], and take into account some of the challenges to because its pursuit is a primary driver of people’s decision- evaluation in conservation such as nonlinear response out- making [12]. Interventions that support local well-being can comes, lack of comparators, multiple outcomes and complex increase environmentally desirable behaviour, and lead to confounding factors [24,27]. The four stages are: (1) defining positive local perceptions and engagement [4]. Using a outcomes of interest and well-being indicators, including for- well-being framework provides a holistic way to incorporate mulating complex theories of change and considering goals for different values (e.g. livelihoods and the environ- confounding factors; (2) designing the evaluation to link out- ment) into decision-making, which can also help to build comes to the intervention; (3) understanding processes of political support and mobilize funding. For these reasons, it change; and (4) collecting data on selected indicators and con- is vital that well-being is taken into account both in conserva- textual factors. Step 3 reflects the increasingly recognized tion programmes that explicitly incorporate livelihoods, and need in conservation policy for richer understandings of in those with narrower biodiversity targets. mechanisms through which impacts are produced [18]. Recent studies on the impacts of protected areas on pov- erty [13–15] have used robust quasi-experimental designs but have tended to focus on externally defined asset- and 2. Conceptualizing human well-being monetary-based measures, or on human–wildlife conflict and attitudes towards parks [16]. Further, by only measuring In conceptualizing well-being, there is a tension between a average net impacts, studies do not reveal how benefits and universal approach that allows comparisons, and ensuring losses are distributed across different groups of people. local relevance [28]. Local perspectives must drive our under- With the recent exception of a more nuanced evaluation of standing of well-being, as externally derived categories may marine protected areas (MPAs) [17], these types of studies not have meaning for local people, and thus will not account do not consider multiple dimensions of well-being and its for locally significant impacts of interventions. Any universal subjective aspects. Crucially, much research does not eluci- frameworks or methods used in evaluations must be adapt- date the mechanisms through which interventions impact able to locally meaningful formulations, made relevant to well-being [18], an element of evaluation that is especially the target population by using comparable categories with important for decision-making [19]. locally specific indicators. There is also a balance to be One difficulty in embarking on a well-being impact evalu- struck between objective and subjective definitions. In devel- ation is in operationalizing such a wide and complex concept. opment economics, well-being has been conceived of as an The array of conceptual frameworks for well-being [20] can objective concept mainly focused on material assets, and in cause confusion, especially among researchers and prac- social psychology as an internal, subjective psychological titioners trained in the natural rather than social sciences. state felt by the individual. Using either approach alone is Although the Conservation Measures Partnership [21] advo- insufficient. People will have different capabilities to gain cates inclusion of human well-being targets when designing benefits from assets, whereas a person’s expressed satisfac- and adaptively managing conservation projects, it provides tion with life is a poor guide to objective valuation of no methodological guidance on how to measure and evaluate material impoverishment, as it does not account for people well-being impacts, and ignores the possibility of unexpected adapting their preferences to harsh conditions [29]. Conserva- consequences, institutional changes resulting from interven- tionists should be interested both in objective indicators, as tions and the importance of local perspectives of change. they show tangible changes and are often most sought by Stephanson & Mascia [22] build upon this work with a funders and policy-makers, and also in subjective well- broader conception of well-being, but do not consider subjec- being, because people’s feelings and experiences impact on tive experiences, and remain focused on methods for participation and social sustainability. rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 One framework that combines objective and subjective presented in table 1. These principles are summarized in valuation and gives primacy to local understandings comes table 2. from the Wellbeing in Developing Countries (WeD) project [30]. Well-being is conceptualized as an outcome and a pro- (a) Defining outcomes and indicators cess, in three interacting dimensions: the objective material circumstances of a person, subjective evaluation of people’s (i) Principle 1: put local people at the centre of the evaluation goals and the processes they engage in, and a relational One of the central benefits of using a broadly defined, locally component [31]. This last dimension acknowledges that grounded conception of well-being is that it puts at the centre individual well-being is pursued in relation to other people, those people most affected by policy changes and interven- that social connectedness is a human need and that defi- tions. Local people should be involved throughout the nitions of a good life are socially constructed [32]. Culture process of evaluation, but most crucially when initially defin- is often viewed as external in discussions on poverty and ing the scope of the evaluation. Interventions are based on a well-being, but here it forms the lens through which all theory of change (ToC), which explains the process through aspects of well-being are constituted [33]. For example, the which the intervention is thought to give rise to specific out- significance of cattle-raising goes beyond a livelihood for comes. An evaluation is effectively testing this theory. To pastoralist communities to being a culturally meaningful start it is helpful explicitly to map out the ToC causal chain way of life entailing social contracts of ownership or use from inputs to outcomes, the underlying logic and the rights over land. The WeD approach emphasizes the holistic, social, behavioural and institutional assumptions being dynamic and social nature of well-being. It brings together a made [21,25], a process that allows space for reflection on unique configuration of interdependent elements, counterba- assumptions and context, and the development of hypoth- lancing a tendency in policy to privilege material well-being eses and evaluation questions. External drivers and and underplay subjective feelings and the social dimension of pressures such as government policies, market changes, cli- people’s lives [31]. mate change and environmental shocks, and how they may Another framework—the ‘Voices of the Poor’ (VoP)—is be changing well-being and interacting with the intervention, based on empirical data and is familiar to conservationists, must be considered in order to take confounding factors into because it was used in the Millennium Ecosystem Assessment account. A ToC is best developed with the participation of as a means of conceptualizing relationships between ecosystem local stakeholders who hold highly contextual knowledge, services and aspects of well-being. The project found five com- and may well consider potential consequences and unin- ponents commonly considered to constitute well-being among tended changes that would not otherwise be addressed. For individuals across 23 countries [34]. They are material assets, single-stranded projects, tracing potential changes may be health, social relations, security and freedom of choice and relatively straightforward, but for complicated projects with action. The last component, which underpins the others, means multiple components linked together (e.g. land rights, edu- having a sense of control over one’s life and the capacity to cation and livelihoods), more thought must be given to achieve what one values doing and being. This is easily over- interactions and feedback loops [46]. looked in conventional assessments but may be especially Qualitative research, using semi-structured, informal relevant for conservation interventions, which can be rejected if interviews and participant observation, and which is flexible perceived as imposed and undermining freedom with regard and open to unexpected findings, can provide details on to environmental behaviour [35]. On the other hand, interven- local nuances in the language used to express well-being, tions that secure local land tenure and improve natural resource and the priorities and aspirations that people have. For governance could increase feelings of empowerment [17]. example, Abunge et al. [35] used focus group discussions The VoP framework provides a useful checklist of themes with different stakeholder groups (e.g. women fish vendors to consider when starting an evaluation, based on empirical and beach seine captains) connected to a Kenyan coastal research. We combine it with ideas from the WeD research, fishery to understand how well-being was expressed by thus providing a thematically based framework that allows different types of people, how and why it had changed, depth of understanding, and can be used to guide the struc- and the hopes people had for the future of the fishery. ture of evaluations (table 1). The VoP domains will be Open-ended questions such as ‘How would you describe in informative to evaluators about which aspects of well-being general a person who is doing well in this community?’ to consider, bearing in mind that WeD’s three-dimensional encouraged people to open up and discuss what constituted perspective will help to define the questions asked and the a good life in this particular context. Qualitative research is type of data collected. For example, in studying the benefits also valuable at the start of evaluation to understand the his- of a payment for ecosystem services programme, in the torical, political and cultural issues that can shape people’s material domain, a relevant (and commonly used) objective perceptions, helping in the development of locally relevant indicator could be household income. The subjective dimen- questions for any structured and standardized questionnaires sion suggests consideration of levels of satisfaction with [36]. Other useful sources of insight are past studies, ethno- income changes, and feelings of fairness about benefit distri- graphies and informed sources who understand local bution. The relational dimension suggests the relevance of politics and history. how people use income, the way it can change relationships and differential capabilities within a household to benefit. (ii) Principle 2: select multiple outcomes to measure and consider subjective components 3. Principles to guide well-being evaluation Given the multi-dimensional nature of the well-being concep- We next discuss the nine key principles to bear in mind when tualization presented here, it cannot be captured by carrying out a robust evaluation based on the framework measuring only one outcome. Rather than using a standard rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 Table 1. Theoretical framework for well-being evaluation, which links ‘Voices of the Poor’ (VoP) well-being domains with perspectives from Wellbeing in Developing Countries (WeD). ‘Voices of the Poor’ well-being domains description and examples insights provided by WeD perspective and research material secure and adequate livelihoods not only about what people have, but what they can do and be, enough food and food security and how they feel about these things assets, e.g. land, natural resources, livestock, the ways in which objective material well-being outcomes are savings and capital, goods, housing, furniture defined and satisfied are socially and culturally constructed, and tools requiring attention to local context human as well as material resources are important, including knowledge and education health feeling strong and well health is subjectively experienced access to health services mental health is as significant as physical health in well-being appearing well having a healthy physical environment e.g. fresh air social relations good relations with family, community and collective well-being is significant for individual well-being in country culturally defined ways dignity, e.g. not being a burden, feeling social structures and institutions that enable people to pursue well- listened to being in relation to one another may be impacted by ability to help others and fulfil social interventions obligations people’s ideas and strategies for pursuing well-being may not be ability to care for children (including education compatible, resulting in trade-offs that must be confronted and marriage) security confidence in the future—predictability, people’s well-being and decisions are influenced by perceptions of peace future and perceived threats safe and secure environment, e.g. safety from capabilities to achieve other aspects of well-being may increase disasters security personal physical security and safety sustained security can only be the outcome of autonomy rather than security in old age and for future generations dependency freedom of choice and sense of control and power not about independence but self-endorsement of one’s own action ability to pursue what you value doing and behaviour, i.e. feeling personal value and interest regarding being, and meet aspirations actions ability to be a good person, e.g. to help others autonomy can be evaluated with regard to different aspects of people’s lives that they value related to the ability to adapt in times of change list of indicators that may be irrelevant, evaluators can more positive perspective than the concept of poverty, it is also target pertinent components suggested during qualitative important to include negative information when assessing inter- discussions with communities and considered sensitive to ventions. External interventions may contribute as much towards the intervention actions in the theory of change. The frame- ‘ill-being’ e.g. social exclusion, conflict or malnourishment, as work in table 1 can be used to guide this inquiry. For to well-being [34]. example, health and nutrition may be prioritized in the com- The need for measurement of multiple outcomes of munity and could be improved by access to water and food interventions is highlighted by the fact that there may be sources, or social capital could be increased through the trade-offs and synergies between outcomes [40]. For instance, establishment of community-based governance of natural there could be trade-offs between different dimensions of resources. Quantitative indicators could then include number well-being such as wealth and equity, invalidating of meals eaten per day or levels of participation in community conclusions about the overall impact of interventions on activities. Some outcomes relevant to well-being, such as social well-being if one dimension was missed. Considering these relations or political change, may not be amenable to quantifi- relationships in developing a theory of change, and later in cation at all [38]. Although the concept of well-being takes a data analysis, moves the evaluation away from simplistic rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 Table 2. Summary of guiding principles for evaluating impacts of conservation interventions on well-being. references for guiding principle examples of approaches to addressing the principle further details defining outcomes and indicators (1) put local people at the centre of the start with flexible qualitative research, e.g. semi-structured interviews to [35,36] analysis explore local understanding and components of well-being qualitative research on the local context, e.g. through literature reviews and informed sources map out theory of change developed with participating communities [21,25,37] and stakeholders (2) select multiple outcomes to measure select multiple well-being indicators based on local priorities and and consider subjective components outcomes in theory of change collect qualitative data on outcomes not amenable to quantification, [38] e.g. institutional change collect data on subjective feelings about pertinent aspects of well-being [39] allow opportunities for people to voice unintended consequences, and negative outcomes consider relationships (trade-offs and synergies) between outcomes [40] evaluate impacts on security through identifying locally relevant [41] indicators evaluation design (3) match evaluation design to the setting consider quasi-experimental and before-after-control-intervention [14,15,26,42] and questions asked designs if no baseline data, consider recall interviews for simple variables [43] control-intervention designs without baselines should be supported by other data alternatives to quasi-experimental designs: theory based, case studies, [19,37,44,45] participatory methods understanding processes of change (4) provide evidence of causal linkages theory-based analysis using quantitative and qualitative methods to [36,37,46] understand the how and why of impacts quantitative data can produce estimates of the contribution of different [18] causal mechanisms (5) consider trajectories of change anticipate and acknowledge possible trajectories of change and measure [17,47,48] ex-post impacts if possible (6) investigate institutions and governance institutional analysis using secondary and primary data [49] structures participatory institutional profiling (before-and-after intervention) [50] data collection (7) select and apply methods with choose tools appropriate to the cultural context and apply with [51,52] sensitivity to context consideration to equity combine methods to take advantage of their strengths, e.g. quantitative measures with qualitative insights (8) take into account heterogeneity disaggregate data according to qualitative understandings of social [53,54] structures and livelihoods individual interviews to capture differences across age and gender [39] within households (9) ensure independence recruit locally trusted people independent of implementing institutions [36] and conservation draw upon local language skills and in-country researchers rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 narratives towards a realistic acknowledgement of potential matching techniques [26]. Controls are selected based on their gains and losses that can inform decision-making. It can similarity with the intervention targets on a suite of measurable also highlight aspects of well-being that people feel cannot covariates that are thought to affect participation in the inter- be ‘traded-off’ at all, such as cultural heritage [55]. vention and the outcomes of that intervention (according to Observable, quantitatively measured changes can provide the theory of change). For example, in the case of protected credible evidence of impact to external audiences, but areas, these covariates could be distance to a city, elevation perceived change by local communities—reflecting the subjec- and asset-based indicators [15]. Controls should be selected tive aspect of well-being—may also be significant, especially outside the zone of influence of the intervention, to avoid spil- for conservation managers on the ground wanting to take miti- lover effects of the project or contamination by other gating measures to improve elements of the project people are interventions. Clements & Milner-Gulland [42] evaluated liveli- not satisfied with, in order to gain local support. Indeed, per- hood outcomes of protected areas in Cambodia, and selected ceived well-being may be at odds with objective measures, control villages based on matching variables thought to affect highlighting where there is dissatisfaction with interventions. village-level poverty and access to natural resources prior to In villages involved in MPAs in Indonesia, for example, the intervention. They ensured that control villages were there were negative changes in perceived well-being despite more than 20 km from the protected area border. It is impor- increases in wealth (based on material assets) during the tant, however, to recognize that controls in that study did not course of the intervention, owing to inequitable sharing of reflect the groups to which people actually compared them- benefits, conflict and unmet expectations [17]. selves, which could affect subjective well-being and therefore Security—living under conditions where there is predict- interpretations of impact. A more qualitative approach to match- ability—is a key constituent of well-being (table 1), drawing ing could instead be used, by selecting controls with attention to temporal aspects of well-being. People’s current participation from local people, to identify controls that are well-being takes place in the context of past experiences, as both methodologically appropriate (i.e. deal with confounding well as expectations, fears and aspirations about the future. factors) and meaningful for local people (e.g. geographically People engage in trade-offs through time to establish security closer). It is also worth considering that there may be hetero- and reduce threats [41]. Especially when faced with rapid geneity within the treatment, such as spatial differences in changes, perhaps as a result of an external intervention, the impact of an intervention that will be disguised if only aver- uncertainty prevails, reducing security and therefore well- age effects are considered. For example, people experienced the being. Conservation may increase feelings of insecurity, same alternative livelihood project very differently across even if implemented in the hope of improving environmental different villages in Tanzania [58]. security in the longer term. For example, in Tanzania con- Using baselines in addition to controls to create a BACI cerns about future land-use restrictions were highest among (before-after-control-intervention) evaluation allows impacts households near to Tarangire National Park, and this influ- to be isolated from two biases: selection bias in which the tar- enced decisions to convert land to agriculture to secure gets of the intervention are a non-random selection of the land tenure, ultimately affecting conservation outcomes whole population (for example, when wealthier families are [56]. Alternatively, threats of large-scale land acquisitions in more equipped to participate in a PES (payment for ecosys- countries where community tenure rights are weak may tem services) scheme), and concurrent change (for example, lead to higher participation in conservation activities, such improvements in wealth affected by both a sustainable liveli- as seeking community forestry titles [57]. Evaluating impacts hoods intervention and wider economic changes) [6]. of interventions on security could involve developing indi- Selection bias is tackled by matching techniques in quasi- cators relevant to the context, such as livelihood and experimental designs, while tracking change through time income diversity, access to justice, and the functioning and in both the control and intervention (the difference in differ- membership of collective institutions [41]. ence method) takes into account differences between treatments and controls that are constant over time, including unobserved intrinsic characteristics such as levels of motiv- (b) Evaluation design: linking outcomes to the ation or optimism. A prospective evaluation, developed at the same time as the intervention is designed, is preferable intervention to a retrospective evaluation as it allows the collection of (i) Principle 3: match evaluation design to the setting and baseline data and is more likely to generate valid counterfac- questions asked tuals [26]. It is possible, however, to construct an approximate At the heart of evaluation is the process of attributing specific baseline ex-post from secondary data or carefully designed effects, in this case changes in well-being, to the intervention surveys asking participants to recall specific, easily remembe- rather than to other factors. This can be achieved by inferring red variables such as assets and link them to locally important the counterfactual—what would have happened in the absence events [59], although researchers should be aware that recall of the intervention—through the identification of controls, thus data are prone to inaccuracies [43]. Although before–after reducing bias from confounding factors [24]. Quantitative and control–intervention comparisons on their own make experimental and quasi-experimental designs that allow assess- for weaker causal inferences, where a full BACI design is not ment of the magnitude of impacts are often viewed as the ‘gold possible, these simpler designs can provide credible insights standard’ for answering the attribution question in a robust if supported by other data such as community perceptions way in development and conservation. Truly experimental that substantiate or refute the quantitative trends. designs, where the researcher randomly assigns interventions, There are real-world constraints to using controls and are not often possible in conservation [6], and instead quasi- quasi-experimental designs to attribute changes to an inter- experimental methods are used in which control groups vention; they require a large sample size, may not be (most likely households or villages) are identified through suitable for complex or broadly defined interventions such rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 as regional policy with small number of units, and require a how contextual factors may be hindering or reinforcing suitable budget and technical skills [19]. There may be ethical change in particular outcomes [62]. This is especially impor- concerns such as raising expectations or subjecting people to tant where evaluation is directed towards lesson learning. surveys that bring no benefits [51]. Controls may be difficult Theory-based approaches as described above take a deduc- to find if, for instance, the intervention focusses on small tive approach by empirically discounting alternative areas of particular importance for conservation that are plausible explanations for outcomes. For example, a theory- very different from the rest of the region (e.g. islands of natu- based analysis showed that an infant nutrition project in ral forest). Instead of, or in addition to, quasi-experimental India mis-targeted mothers who, although gaining knowl- designs there are a range of alternative approaches through edge, were not able to put it into practice as it was their which causal inferences can be made that can be divided mothers-in-law and husbands who made decisions about broadly into three types: theory-based, case studies and par- food and child-raising, an insight that was initially found ticipatory. These methods tend to be stronger than through reading anthropological studies [37]. In that quantitative counterfactual approaches in explanation and example, it was a design flaw in the project that led to poor contextualization, but weaker on estimating the magnitude results, but a theory-based approach can also detect problems of impacts [19]. Decisions about which design to use must with implementation. Ferraro & Hanauer [18] demonstrate be made on the basis of resources and skills available, and how a quasi-experimental design can incorporate quantitat- the types of evidence required. ive analysis of causal pathways to show how protected Theory-based methods consider the likely chains of areas in Costa Rica reduced poverty mainly through tourism. impact, presenting alternative hypotheses for change and The fact that the three measured causal mechanisms examining them through both qualitative and quantitative accounted for only two-thirds of impacts in this study data. These can highlight where there is a break in the serves to highlight how theories of change must incorporate causal chain to explain impacts of an intervention [46]. A in-depth and complex understanding of socio-ecological case study approach focuses on particularly successful (or systems to fully capture processes of change. unsuccessful) cases (e.g. villages or intervention types) to examine the contribution of the intervention to outcomes. (ii) Principle 5: consider trajectories of change Comparisons across the cases can elucidate the combinations Well-being is not a discrete outcome, but an ongoing of causal factors (the types of intervention, methods of dynamic process, changing through the course of an inter- implementation such as levels of capacity building, and con- vention and beyond [63]. Trajectories of change are not textual factors such as tenure regimes, wealth levels and local linear, resulting in attribution errors if well-being effects of leadership) explaining changes in well-being indicators [44]. an intervention are measured at only one point in time [47]; Participatory methods allow communities to systematically for example, there could be high initial impact owing to assess changes themselves; for example, group discussions, improved forest governance arrangements that is eroded in which causes for reported changes in well-being are through time by pre-existing power structures. Monitoring ranked and scored, can show relative perceived impacts of throughout the course of a project is ideal as it allows real- intervention and non-intervention factors, with greater time feedback for learning and adjustment [62]. Ex-post levels of agreement between groups indicating reliability assessments are rare, but may be crucial in understanding [45]. The use of ‘reflexive counterfactuals’ where participants longer-term impacts and sustainability, and for taking into compare themselves before and after the intervention, by account time lags between intervention and effect. For prioritizing perceptions, may be subject to bias, but provide example, any initial gains in aspects of well-being such as important information, for example, in protected area man- fish catch, wealth and empowerment were lost after external agement focused on improving equity and effectiveness at support for MPAs in Indonesia was withdrawn, suggest- the site level [60]. ing that interventions of this kind need to build capacity, Given the pros and cons of the quasi-experimental and gain broad-based support and sustain funding [17]. Realisti- alternative approaches to evaluation design, Roe et al. [61] cally, it may only be possible to evaluate shorter-term advocate a sensible two-track system, in which in-depth outputs or outcomes and indicate where longer-term longitudinal evaluation using controls for a selection of repre- impacts could occur [25]. The reference standards of those sentative interventions of strategic relevance is combined affected by interventions can also change, potentially as a with rapid, participatory assessment more feasible for the result of the interventions themselves [48]; for instance, majority of projects. Policy-makers and donors may empha- increasing material wealth may lead to wealth becoming a size the former approach, to gain evidence of the more important component of well-being for some, but still magnitude of impact needed for cost-effectiveness analysis. reduce well-being owing to rising aspirations not matched This may guide decisions on whether to replicate or fund by opportunities [64]. similar interventions in the future. Field managers wanting to understand people’s experiences and perspectives may focus on the latter, combined with quantitative indicators (iii) Principle 6: investigate institutions and governance structures analysed through theory-based analysis. Well-being depends on institutions—human-devised informal constraints and formal rules—which govern relationships between individuals and groups, and between humans and (c) Understanding processes of change ecosystems [65]. The choices made about the types of organiz- (i) Principle 4: provide evidence of causal linkages ations conservationists work with (state agencies, private Using a counterfactual approach that is limited to attributing corporations, customary authorities) will profoundly shape outcomes to an intervention cannot answer the fundamental the institutional landscape, affecting representation, citizen- questions of how and why a project is or is not effective, and ship, and ultimately social and environmental sustainability rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 [66]. Ill-considered interventions can subvert existing co-produced data relatively quickly and cheaply, and to institutions and cultural practices that act to regulate demonstrate local involvement in evaluation. Relevant environmental behaviour, resulting in alienation and methods include qualitative resource mapping, timelines, counter-productive conservation results [67]. Natural resource focus groups, village histories, ranking and scoring [45]. management interventions can, on the other hand, support The mainstreaming of PRA, however, has resulted in the pro- improved governance, contributing to poverty alleviation cess becoming somewhat ritualized, and the participatory [68], and act as vehicles for social change and improved partici- label is often used to mask standard extractive data collection. pation. This emphasizes the relational dimension of well- Participatory discussions can be susceptible to co-option by being, and the attention required in analyses to the relationship local elites, silencing those most affected [71]. If PRA tech- between individual and collective well-being, which is shaped niques are used, experienced and trained facilitators by dynamic institutions such as norms. During evaluations, it familiar with the local context, and independent of the inter- is important to understand the functioning of institutions and vention, should lead the work to ensure sensitive and governance structures acting within and upon communities, equitable discussion [52]. Groups should be appropriately and how conservation interventions affect it, in turn impacting constituted; for example, different gender and age groups individual well-being. This can be achieved through, for may not be comfortable in mixed groups. In some situations, example, institutional profiling with local people in which for example, where communities are suffering from research visual methods such as Venn diagrams can aid discussions fatigue, or issues are highly contested, individual or house- about key institutions, relationships and forms of power at hold interviews may be a better option. different scales that influence people [49,50]. (ii) Principle 8: take into account heterogeneity within the target (d) Data collection group (i) Principle 7: select and apply methods and toolkits with Just as different people are able to access different ecosystem sensitivity to the research context services, there will be trade-offs between the well-being There are many tools and methods available—both with a impacts of interventions on different people, between or quantitative and qualitative slant—which conservation eva- within communities [53]. Standard experimental and quasi- luators can draw upon in collecting data on well-being experimental impact evaluation methods may produce an indicators as well as on contextual and confounding factors. average effect of an intervention across households, commu- However, reflection on their appropriateness to the context nities or the whole population being investigated, but this is vital for data validity. Schreckenberg et al. [51] provide a does not address which types of people win and lose, and useful compilation of rapid social research methods that why [72], unless heterogeneity is purposefully incorporated can be used to collect data on the well-being impacts of con- into research design and data collection. It is especially servation interventions. These methods are compatible with important to ensure that vulnerable groups such as the poor- our proposed framework for well-being evaluation, but est, landless, migrants or mobile resource users (such as thought needs to be given not only to selecting particular fishers, pastoralists and forest dwelling groups) are included methods appropriate to the context, but also to the ways in evaluations as they are often invisible unless local knowl- these tools are applied to deal with culturally sensitive edge is used [73]. The impacts of interventions on issues, vested interests and equity. For example, how will households with different livelihoods may differ signifi- the use of particular local informants skew the evaluator’s cantly; for example, non-timber forest product collectors in understanding of the issues, and how will marginalized Cambodia were significantly better off in terms of basic groups be accessed? Tallying scores for different outcomes, necessities inside a protected area than outside owing to as advocated in some guidelines [69], is attractive for quick secure access to resources [54]. Impacts of interventions on and standardized assessments. But, in isolation, scoring sys- social dynamics between different groups can undermine tems run the risk of aggregating over a broad range of well-being; for example, targeting only some groups may indicators and social groups. This may lead to their falling create conflict [38]. Qualitative research can elucidate social between two stools; meaningless both locally and compara- structures, wealth and livelihood differences that can form tively. Even where carefully selected and appropriately the basis for appropriate disaggregation of data, as well as applied, a single method may be inadequate on its own. improving understanding of the ways different groups For example, the Basic Necessities Survey (BNS) is an conceive of well-being. index-based tool that assesses household poverty based on Social surveys are often carried out at the household locally defined ‘necessities’ [70]. A key benefit of the level, but there is likely to be intra-household variability, method is that it produces quantifiable results through a par- with differences in well-being according to gender and ticipatory and democratic process, but the score is aggregated age. Britton & Coulthard [39] found that the domains of at the household level and gives little detail about the pro- life important for well-being, and satisfaction with these cesses of change. A hybrid research design augmenting the domains, differed significantly between men and women in BNS with data that chime with WeD’s broader conception fishing communities in Northern Ireland. Women may lose of well-being could form a more robust way to capture the resource access under payment-based conservation interven- complexity of well-being change. For example, the incorpor- tions but receive few of the benefits, which are given to male ation of semi-structured interviews could focus on household heads [74], or they may be excluded altogether subjective experiences, and capture causal mechanisms. from participation [75]. Well-being is experienced by individ- The methods of Participatory Rural Appraisal (PRA) used uals, and so they should be the primary unit of impact by development practitioners have been enthusiastically assessment rather than the household, which is a common adopted by many conservationists, to provide locally evaluation unit in economic assessments. However, as the rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 WeD framework suggests, individual well-being is relational. evaluation, and improve their understanding of objective and Therefore, collective well-being at different scales of social subjective well-being impacts. This understanding is perhaps relationship is significant for individual well-being, although most urgently needed in materially poor areas of the Global the extent and nature of these relationships will differ across South, but well-being provides a useful way to measure social cultures and contexts [63]. impacts of conservation regardless of the wealth status of the population [77]. Mixed methods can better support causal and explanatory analysis, and conservation researchers should not (iii) Principle 9: ensure independence be reticent about using qualitative data in their own right. Although the research design and selection of methods and Far from signaling a lack of rigour, qualitative approaches are techniques are important elements to consider in an evalu- necessary to appropriately disaggregate data, identify covari- ation, the quality of research will ultimately be defined by ates, explain the processes involved in producing well-being how it is conducted and the relationships established between impacts, and allow local voices to be heard. The important the researchers and participants [76]. Evaluations conducted basis for rigour is the appropriate application of techniques, by people who are independent of actors implementing inter- either quantitative or qualitative [78]. ventions or otherwise working in the system will result in Theevidencebaseonthe impactsofconservationinter- improved validity. Although research is often labelled as ‘inde- ventions on human well-being is weak, and there is much to pendent’, it is all too often facilitated by vehicles clearly be learnt to support decision-making about the range of interven- marked with government or NGO logos, or by people who tions used in conservation under different contexts. For example, are linked to the intervention or to conservation more gener- in a recent systematic review of 136 community-based conserva- ally. Marginalized rural people are likely to find it difficult tion evaluations, 80% of the studies included were rated as poor to talk candidly with powerful individuals, which suggests quality on the basis of conflict of interest (i.e. lack of inde- the importance of considering local language, trust and ethni- pendence), data validity and other problems [79], arguably city in building a research team. Calling upon local expertise throwing the results of an otherwise meticulous statistical analy- and language skills will decrease the risk of obscuring local sis into serious doubt. There will inevitably be trade-offs between meaning and realities [36]. In-country students or young conservation outcomes and human well-being outcomes [80], researchers supported by experienced evaluators may be and between different elements of well-being itself. Only by good options for tight budgets, and this approach will also assessing well-being in a way that tackles complexity, context, contribute to capacity building within the country. politics and the wide range of impacts that conservation can bring, can stakeholders hope to openly discuss and negotiate trade-offs in a systematic and transparent way. Conservationists 4. Conclusion have a responsibility to the communities in which they carry Understanding the intricacies and dynamics of what people out their activities, and using well-implemented well-being consider to be a good life is far from straightforward. evaluations can improve accountability and lesson learning, ulti- Well-being is multi-faceted, and varies between contexts and mately improving the likelihood of successful, locally supported cultures, within communities and households, and through conservation in the long-term. time. Attributing well-being change to interventions must also take place in the context of complex and dynamic influences Authors’ contributions. All authors made contributions to the conception at multiple scales. This mayseem like an impossible task forcon- of this paper. E.W. led the drafting, and K.M.H., E.B., T.C., J.T.M., D.W. and E.J.M.-G. critically reviewed drafts for intellectual content. servation practitioners, especially with limited resources and All authors gave final approval for publication. expertise. In the face of this complexity, however, formulaic Competing interests. We declare we have no competing interests. methods will not work. Conservationists should not use pre- Funding. This research was funded by the ESRC-DFID joint fund for scriptive designs and methods without thinking about their poverty alleviation (ES/J018155/1). applicability to the case, how best to apply to them and what Acknowledgements. We would like to thank Allister McGregor for sort of data they will produce. By engaging with the principles comments on an earlier draft of this paper. This paper is a contri- and concepts set out here, and summarized in table 2, conserva- bution to Imperial College’s Grand Challenges in Ecosystems and tionists can hope to untangle the complexity of social impact the Environment initiative. References 1. Millennium Ecosystem Assessment. 2005 Ecosystems 4. Coulthard S, Johnson D, McGregor JA. 2011 Poverty, conservation investments. PLoS Biol. 4, e105. and human well-being: synthesis. Washington, DC: sustainability and human wellbeing: A social (doi:10.1371/journal.pbio.0040105) Island Press. wellbeing approach to the global fisheries crisis. 7. Leisher C, Sanjayan M, Blockhus J, Larsen SN, 2. West P, Igoe J, Brockington D. 2006 Parks and Glob. Environ. Chang. 21, 453 – 463. (doi:10.1016/j. Kontoleon A. 2012 Does conserving biodiversity peoples: the social impact of protected areas. Annu. gloenvcha.2011.01.003) work to reduce poverty? A state of knowledge Rev. Anthropol. 35, 251 – 277. (doi:10.1146/ 5. Stiglitz J, Sen A, Fitoussi J. 2009 Final report of the review. In Biodiversity conservation and poverty annurev.anthro.35.081705.123308) commission on the measurement of economic alleviation: exploring the evidence for a link 3. Coad L, Campbell A, Miles L, Humphries K. 2008 performance and social progress. Paris, France. (eds D Roe, J Elliot, C Sandbrook, MJ Walpole), The costs and benefits of protected areas for local (www.stiglitz-sen-fitoussi.fr/documents/ pp. 143 – 159. Chichester, UK: John Wiley & Sons. livelihoods: a review of the current literature. rapport_anglais.pdf ) 8. OECD (Organisation for Economic Cooperation and Cambridge, UK: UNEP World Conservation 6. Ferraro PJ, Pattanayak SK. 2006 Money for nothing? Development). 2011 How’s life? Measuring well- Monitoring Centre. A call for empirical evaluation of biodiversity being. Paris, France: OECD Publishing. rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 9. Summers JK, Smith LM, Case JL, Linthurst RA. 2012 data in conservation planning. Conserv. Biol. 28, 39. Britton E, Coulthard S. 2014 Assessing the social A review of the elements of human well-being with 1236 – 1248. (doi:10.1111/cobi.12357) wellbeing of Northern Ireland’s fishing society an emphasis on the contribution of ecosystem 23. Greene JC, Benjamin L, Goodyear L. 2001 The using a three-dimensional approach. Mar. services. Ambio 41, 327 – 340. (doi:10.1007/s13280- merits of mixing methods in evaluation. Evaluation Policy 37, 28 – 36. (doi:10.1016/j.marpol.2012. 012-0256-7) 7, 25 – 44. (doi:10.1177/13563890122209504) 04.011) 10. Roe D, Oviedo G, Pabon L, Painter M, Redford K, 24. Ferraro PJ. 2009 Counterfactual thinking and impact 40. Agrawal A, Benson CS. 2011 Common property Siegele L, Springer J, Thomas D, Walker Painemilla evaluation in environmental policy. New Dir. Eval. theory and resource governance institutions: K. 2010 Conservation and human rights: the need for 122, 75 – 84. (doi:10.1002/ev.297) strengthening explanations of multiple outcomes. international standards. London, UK: IIED. 25. Leeuw F, Vaessen J. 2009 Impact evaluations and Environ. Conserv. 38, 199 – 210. (doi:10.1017/ 11. World Parks Congress. 2003 The Durban Accord. See development: NONIE guidance on impact evaluation. S0376892910000925) http://cmsdata.iucn.org/downloads/durbanaccorden. Washington, DC: NONIE (The Network of Networks 41. Wood G. 2007 Using security to indicate wellbeing. pdf. on Impact Evaluation). In Wellbeing in developing countries: from theory to 12. Deci EL, Ryan RM. 2010 The ‘what‘ and ‘why‘ 26. Gertler PJ, Martinez S, Premand P, Rawlings LB, research (eds I Gough, JA McGregor), pp. 109 – 132. of goal pursuits: human needs and the self- Vermeersch CMJ. 2011 Impact evaluation in practice. Cambridge, UK: Cambridge University Press. determination of behavior. Psychol. Inq. 11, Washington, DC: The World Bank. 42. Clements T, Milner-Gulland EJ. 2015 The impact of 227 – 268. (doi:10.1207/S15327965PLI1104_01) 27. Baylis K et al. 2015 Mainstreaming impact payments for environmental services and protected 13. Andam KS, Ferraro PJ, Sims KRE, Healy A, Holland evaluation in nature conservation. Conserv. Lett. areas on local livelihoods and forest conservation in MB. 2010 Protected areas reduced poverty in Costa (doi:10.1111/conl.12180) Northern Cambodia. Conserv. Biol. 29, 78 – 87. Rica and Thailand. Proc. Natl Acad. Sci. USA 107, 28. Boarini R, Kolev A, McGregor JA. 2014 Measuring (doi:10.1111/cobi.12423) 9996 – 10 001. (doi:10.1073/pnas.0914177107) wellbeing and progress in countries at different 43. Papworth SK, Rist J, Coad L, Milner-Gulland EJ. 14. Naughton-Treves L, Alix-Garcia J, Chapman CA. 2011 stages of development: towards a more universal 2009 Evidence for shifting baseline syndrome in Lessons about parks and poverty from a decade of conceptual framework. OECD Working Paper No. conservation. Conserv. Lett. 2, 93 – 100. (doi:10. forest loss and economic growth around Kibale 325: OECD. 1111/j.1755-263X.2009.00049.x) National Park, Uganda. Proc. Natl Acad. Sci. USA 29. Sen A. 1999 Development as freedom. New York: 44. Byrne D, Ragin C. 2009 The SAGE handbook of case- 108, 13 919 – 13 924. (doi:10.1073/pnas. Knopf Press. based methods. London, UK: Sage. 1013332108) 30. University of Bath. 2002 Wellbeing in Developing 45. Catley A, Burns J, Abebe D, Suji O. 2008 15. Canavire-Bacarreza G, Hanauer MM. 2013 Estimating Countries (WeD) ESRC Research Group [Online]. Participatory impact assessment: a guide for the impacts of Bolivia’s protected areas on poverty. See http://www.welldev.org.uk/research/ practitioners. Medford, MA: Feinstein International World Dev. 41, 265 – 285. (doi:10.1016/j.worlddev. research.htm. Center. 2012.06.011) 31. McGregor A, Sumner A. 2010 Beyond business as 46. Rogers PJ. 2009 Matching impact evaluation design 16. Pullin AS et al. 2013 Human well-being impacts of usual: what might 3-D wellbeing contribute to MDG to the nature of the intervention and the purpose terrestrial protected areas. Environ. Evid. 2, 19. momentum? IDS Bull. 41, 104 – 112. (doi:10.1111/j. of the evaluation. J. Dev. Effect. 1, 217 – 226. (doi:10.1186/2047-2382-2-19) 1759-5436.2010.00111.x) (doi:10.1080/19439340903114636) 17. Gurney GG, Cinner J, Ban NC, Pressey RL, Pollnac R, 32. Deneulin S, McGregor JA. 2010 The capability 47. Woolcock M. 2009 Towards a plurality of methods Campbell SJ, Tasidjawa S, Setiawan F. 2014 Poverty approach and the politics of a social conception of in project evaluation: a contextualized approach to and protected areas: an evaluation of a marine wellbeing. Eur. J. Soc. Theory 13, 501 – 519. (doi:10. understanding impact trajectories and efficacy. integrated conservation and development project in 1177/1368431010382762) J. Dev. Effect. 1, 1 – 14. (doi:10.1080/ Indonesia. Glob. Environ. Chang. 26, 98 – 107. 33. White S, Ellison M. 2007 Wellbeing, livelihoods and 19439340902727719) (doi:10.1016/j.gloenvcha.2014.04.003) resources in social practice. In Wellbeing in 48. Milner-Gulland EJ et al. 2014 Accounting for the 18. Ferraro PJ, Hanauer MM. 2014 Quantifying causal developing countries: from theory to research (eds impact of conservation on human well-being. mechanisms to determine how protected areas I Gough, JA McGregor), pp. 157 – 175. Cambridge, Conserv. Biol. 28, 1160 – 1166. (doi:10.1111/cobi. affect poverty through changes in ecosystem UK: Cambridge University Press. 12277) services and infrastructure. Proc. Natl Acad. Sci. USA 34. Narayan D, Chambers R, Shah MK, Petesch P. 2000 49. Petitt J. 2013 Power analysis: a practical guide. 11, 4332 – 4337. (doi:10.1073/pnas.1307712111) Voices of the poor: crying out for change. New York, Stockholm, Sweden: SIDA (Swedish International 19. Stern E, Stame N, Mayne J, Forss K, Davies R, Befani NY: Oxford University Press for the World Bank. Development Cooperation). B. 2012 Broadening the range of designs and 35. Abunge C, Coulthard S, Daw TM. 2013 Connecting 50. Messer N, Tomsley P. 2003 Local institutions methods in impact evaluation. Working Paper 38. marine ecosystem services to human well-being: and livelihoods: guidelines for analysis. Rome, Italy: London, UK: DFID (Department for International insights from participatory well-being assessment in FAO. Development). Kenya. Ambio 42, 1010 – 1021. (doi:10.1007/ 51. Schreckenberg K, Camargo I, Withnall K, Corrigan C, 20. Agarwala M, Atkinson G, Palmer Fry B, Homewood s13280-013-0456-9) Franks P, Roe D, Scherl LM, Richardson V. 2010 KM, Mourato S, Rowcliffe JM, Wallace G, Milner- 36. Camfield L, Crivello G, Woodhead M. 2009 Social assessment of conservation initiatives: a Gulland EJ. 2014 Assessing the relationship between Wellbeing research in developing countries: review of rapid methodologies. London, UK: IIED. human wellbeing and ecosystem services: a review reviewing the role of qualitative methods. Soc. 52. Mayoux L, Chambers R. 2005 Reversing the of frameworks. Conserv. Soc. 4, 437 – 449. (doi:10. Indic. Res. 90, 5 – 31. (doi:10.1007/s11205-008- paradigm: quantification, participatory methods 4103/0972-4923.155592) 9310-z) and pro-poor impact assessment. J. Int. Dev. 17, 21. Conservation Measures Partnership. 2012 Addressing 37. White H. 2009 Theory-based impact evaluation: 271 – 298. (doi:10.1002/jid.1214) social results and human wellbeing targets in principles and practice. J. Dev. Effect. 1, 271 – 284. 53. Daw T, Brown K, Rosendo S, Pomeroy R. 2011 conservation projects. Draft Guidance. See http:// (doi:10.1080/19439340903114628) Applying the ecosystem services concept to poverty cmp-openstandards.org/wp-content/uploads/2014/ 38. Devereux S, Roelen K, Be´ne´ C, Chopra D, Leavy J, alleviation: the need to disaggregate human well- 03/Guidance-on-HWT-and-Social-Results-in- McGregor JA. 2013 Evaluating outside the box: an being. Environ. Conserv. 38, 370 – 379. (doi:10. Conservation-Projects-v2012-06-27.pdf. alternative framework for analysing social protection 1017/S0376892911000506) 22. Stephanson SL, Mascia MB. 2014 Putting people on programmes. Brighton, UK: Institute of Development 54. Clements T, Suon S, Wilkie DS, Milner-Gulland EJ. the map through an approach that integrates social Studies. 2014 Impacts of protected areas on local livelihoods rstb.royalsocietypublishing.org Phil. Trans. R. Soc. B 370: 20150103 in Cambodia. World Dev. 64, S125 – S134. (doi:10. 63. McGregor JA. 2007 Researching wellbeing: from 72. Ravallion M. 2009 Should the randomistas 1016/j.worlddev.2014.03.008) concepts to methodology. In Wellbeing in rule? Econ. Voice. 6, 1 – 5 (doi:10.2202/1553-3832. 55. Hirsch PD, Adams WM, Brosius JP, Zia A, Bariola N, developing countries: from theory to research (eds 1368). Dammert JL. 2011 Acknowledging conservation trade- I Gough, JA McGregor), pp. 316 – 350. Cambridge, 73. Homewood K, Coast E, Thompson M. 2004 In- offs and embracing complexity. Conserv. Biol. 25, UK: Cambridge University Press. migrants and exclusion in East African rangelands: 259–264. (doi:10.1111/j.1523-1739.2010.01608.x) 64. Copestake J, Guillen-Royo M, Chou W-J, Hinks T, access, tenure and conflict. Africa 74, 567 – 610. 56. Baird TD, Leslie PW, McCabe JT. 2009 The effect of Velazco J. 2009 The relationship between economic (doi:10.3366/afr.2004.74.4.567) wildlife conservation on local perceptions of risk and subjective wellbeing indicators in Peru. Appl. 74. Bedelian C. 2014 Conservation, tourism and and behavioral response. Hum. Ecol. 37, 463 – 474. Res. Qual. Life 4, 155 – 177. (doi:10.1007/s11482- pastoral livelihoods: wildlife conservancies in the (doi:10.1007/s10745-009-9264-z) 009-9070-1) Maasai Mara, Kenya. PhD thesis, University College 57. Neef A, Touch S, Chiengthong J. 2013 The politics 65. Butler CD, Oluoch-Kosura W. 2006 Linking future London. and ethics of land concessions in rural Cambodia. ecosystem services and future human well-being. 75. Corbera E, Brown K, Adger NW. 2007 The equity and J. Agric. Environ. Ethics. 26, 1085 – 1103. (10.1007/ Ecol. Soc. 11, 30. legitimacy of markets for ecosystem services. Dev. s10806-013-9446-y) 66. Ribot JC. 2006 Choose democracy: environmentalists’ Change 38, 587 – 613. (doi:10.1111/j.1467-7660. 58. Sainsbury K, Burgess ND, Sabuni F, Howe C, Puis E, socio-political responsibility. Glob. Environ. Chang. 16, 2007.00425.x) Killenga R, Milner-Gulland EJ. 2015 Exploring 115–119. (doi:10.1016/j.gloenvcha.2006.01.004) 76. White S, Petitt J. 2004 Participatory methods and stakeholder perceptions of conservation outcomes 67. Goldman MJ, de Pinho JR, Perry J. 2013 Beyond the measurement of wellbeing. Particip. Learn. from alternative income generating activities in ritual and economics: Maasai lion hunting and Action 50, 88 – 96. Tanzanian villages adjacent to Eastern Arc Mountain conservation politics. Oryx 47, 490 – 500. (doi:10. 77. Adams VM, Pressey RL, Stoeckl N. 2014 Navigating forests. Biol. Conserv. 191, 20 – 29. (doi:10.1016/j. 1017/S0030605312000907) trade-offs in land-use planning: integrating human biocon.2015.06.001) 68. Brown D, Schreckenberg K, Shepherd G, Wells A. well-being into objective setting. Ecol. Soc. 19, 53. 59. Bamberger M. 2010 Reconstructing baseline data for 2002 Forestry as an entry point for governance (doi:10.5751/ES-07168-190453) impact evaluation and results measurement. reform. London, UK: Overseas Development Institute 78. White H. 2002 Combining quantitative and Washington, DC: The World Bank. Policy Briefing No. 1. qualitative approaches in poverty analysis. World 60. Franks P, Roe D, Small R, Schneider H. 2014 Social 69. Cahyat A, Go¨nner C, Haug M. 2007 Assessing household Dev. 30, 511 – 522. (doi:10.1016/S0305-750X(01) assessment of protected areas: early experiences and poverty and wellbeing – a manual with examples from 00114-0) results of a participatory, rapid approach. London, Kutai Barat, Indonesia. Bogor, Indonesia: Center for 79. Brooks J, Waylen K, Mulder M. 2013 Assessing UK: IIED. International Forestry Research (CIFOR). community-based conservation projects: a systematic 61. Roe D, Grieg-Gran M, Mohammed E. 2013 Assessing 70. Davies R, Smith W. 1998 The basic necessities survey: review and multilevel analysis of attitudinal, the social impacts of conservation policies: rigour versus the experience of ActionAid Vietnam.Hanoi, Vietnam: behavioral, ecological, and economic outcomes. practicality. Briefing October 2013.London, UK:IIED. ActionAid. Environ. Evid. 2, 2. (doi:10.1186/2047-2382-2-2) 62. Bamberger M, Rao V, Woolcock M. 2010 Using 71. Blaikie P. 2006 Is small really beautiful? 80. McShane TO et al. 2010 Hard choices: making trade- mixed methods for monitoring and evaluation: Community-based natural resource management in offs between biodiversity conservation and human experiences from international development. Malawi and Botswana. World Dev. 34, 1942 – 1957. well-being. Biol. Conserv. 144, 966 – 972. (doi:10. Washington, DC: The World Bank. (doi:10.1016/j.worlddev.2005.11.023) 1016/j.biocon.2010.04.038)

Journal

Philosophical Transactions of the Royal Society B: Biological SciencesPubmed Central

Published: Nov 5, 2015

There are no references for this article.