Willingness to cooperate in shared natural resource management is linked to group identification through perceived efficacy and group norms

Cooperative management of shared natural resources is one of the most urgent challenges the world is facing today. While there have been advances in understanding institutional design features that enable sustainable management, there are few field studies that provide theory-based insights into social psychological predictors of willingness to cooperate around shared resources. Here, we address this issue in the context of shared land management in pastoralist community conservancies in Kenya. In a large survey of individuals from different conservancies, we test a path model that links willingness to cooperate to how strongly respondents identified with their conservancy. This relationship is mediated by how efficient conservancies were perceived to be, and to what extent other people in the community were perceived to share cooperative norms. The results also point towards several parameters that may be conducive to developing stronger conservancy identification: transparency of purpose, motive alignment, sense of ownership, and demonstrating benefit. The findings provide insights into the social psychological processes that impact whether cooperative outcomes can be achieved in real-world shared resource settings, and offer practical implications for strengthening governance within pastoralist conservancies and related shared natural resource management contexts.

In order to ensure humanity's continued existence on this planet, natural resources need to be managed in an environmentally sustainable way. However, many resources, such as forests, fisheries, or pastures are vulnerable to the commons dilemma, where prioritising individual short-term gain over collective interests is tempting, but results in over-exploitation of resources in the long-term. For example, individual pastoralists who share access to communal pastureland may face a trade-off between maximising individual benefit (e.g. by increasing herd size) versus collective benefits (e.g. by avoiding grazing their livestock in certain areas at certain times). The former may lead to pasture depletion over time, while the latter is more likely to ensure longer-term resource sustainability.
A large body of theoretical and empirical research has examined the 'design features' of social systems that enable groups to function effectively and avoid the 'tragedy of the commons' by successfully governing and managing collectively-owned resources in a sustainable manner [1][2][3]. However, it is also important to understand at the proximate, psychological level what factors affect an individual's willingness to make cooperative choices, or comply with pro-environmental rules in the face of a commons dilemma [4]. This is particularly important in situations where new organizations or new institutional rules are being established in response to pressing environmental concerns, as these create new opportunities and demands for cooperation.
Much previous research has focussed on individual-level predictors of willingness to cooperate in the collective interest, such as personal values or attitudes [e.g. [5][6][7]. However, recent work has begun to stress the importance of group processes and social contexts for understanding individual choices in shared resource dilemmas [e.g. [8][9][10][11]. The social identity approach [12][13][14] argues that a key predictor of willingness to cooperate with a group and prioritise its interests is group identification (i.e. a sense of belonging to a group and its inclusion into one's self-concept). While this approach has received a lot of empirical support in a variety of laboratory and field contexts [e.g. 15,16], there are few field studies that assess the role of group identification in driving sustainable behaviour in the face of a collective environmental challenge. In addition, little is known about factors that could increase relevant group identifications in such contexts in the first place. The present paper addresses these issues by exploring predictors of willingness to cooperate within three communityled natural resource management (CNRM) organizations in Northern Kenya where pastoralism is the main livelihood (see figure 1). By doing this, it contributes to testing and extending theoretical understanding of drivers of cooperation in a relevant field context, as well as offers practical insights for understanding how community-led sustainability initiatives may be more effectively implemented.

Group identification and cooperative behaviour
Social identity theory [12][13][14] argues that strong identification with a group leads individuals to prioritise collective (as opposed to individual) outcomes, driving cooperative action within that group [17]. In addition, important group memberships motivate adherence to group norms, leading individuals to behave in line with the group 'prototype' (a set of group's central characteristics [18]). Several processes have been suggested to explain the relationship between group identification and willingness to act in the interests of that group. For example, strongly identified group members tend to believe that the group is capable of achieving its purposes, and this sense of collective efficacy makes individual contributions to collective endeavours seem meaningful and important [19]. Collective efficacy has been empirically shown to drive collective action [19][20][21][22], including in environmental sustainability contexts [e.g. 23]. On this basis, it could be expected that identification with one's CNRM group would be linked to perceptions that this group is efficient in managing shared resources, and this, in turn, would lead to stronger within-group cooperation.
Another possible path from group identification to cooperation lies through perceived support from, and trust in other group members [24,25]. Existing evidence suggests that there is a strong relationship between group identification and expected availability of social support and trust in others [e.g. [26][27][28][29]. At the same time, trust is a crucial determinant of cooperation within groups that balance individual and collective benefits [e.g. 30,31]. Trust plays a particularly important role in contexts where there is a risk that individual contributions could be taken advantage of by non-cooperating group members [e.g. 32,33]. Hence, it could be expected that CNRM group identification would be linked to stronger perceived support from and trust in its members, and that trust would translate into stronger willingness to cooperate.
In addition to collective efficacy and trust, group norms constitute another mechanism connecting group identification and cooperation. Cooperative norms are thought to enable pro-social behaviour even in contexts where traditional rational choice theories would predict more individually-oriented, or 'selfish' behaviour [34][35][36]. Such norms have been argued to be key for environmentally sustainable outcomes in resource management [37][38][39]. Previous research suggests that cooperative norms stem from strengthening group cohesion [40], and a number of studies demonstrate that developing and internalising cooperative social norms is essential for successfully managing shared resources [31,[41][42][43]. There is also evidence supporting the role of social norms in cooperation around shared environmental resources, including agricultural land [44][45][46][47]. Overall, it could be suggested that CNRM group identification would be related to stronger cooperation around resources shared by that group, and this link could be mediated by stronger sense of collective efficacy, higher trust, and cooperative norms.

Predictors of group identification in a field context
If group identification is an important factor in achieving cooperation, it would be important to understand what motivates individuals to develop this sense of identification in field contexts where they face a shared environmental dilemma. Here we focus on four possibilities: Perceived benefit. Individuals may be more likely to identify with social groups when they perceive that those groups are meeting their needs. These benefits may be psychological (e.g. distinctiveness, belonging, self-esteem, control [48][49][50]), or more practical (e.g. receiving a fair share of benefits from a shared resource, attaining better career prospects [51] or satisfying informational needs [52]). It could be expected that identification with groups that manage resources collectively is more likely to develop when such groups' actions are perceived to be beneficial by their members.
Fit between new and existing identities. Developing identification with a new group may rest on the extent to which it is compatible with existing identities [e.g. 53]. For example, new students are more likely to identify with their University when their family background is consistent with University values [51], and business mergers are more likely to lead to an overarching organizational identity development when new organizational principles are consistent with pre-merger ones [e.g. [54][55][56]. Extending this principle to CNRM, it could be expected that individuals are more likely to develop a sense of identification with a group that shares a natural resource, when motives and interests of that group are perceived to be aligned with existing identities (e.g. household or community).
Clarity of purpose. Research in organizational settings suggests that one of the key factors that determines employee identification is perceived clarity of organizational purpose. For example, Mayfield and colleagues [57] demonstrate that leader communication increases employee organizational identification by enhancing the clarity of organizational culture, and Toon and Hurth [58] provide evidence for the link between perceived clarity of purpose and employee-company identification. Clarity of purpose and working principles may provide members of an organization with a sense of meaning, which has been argued to be a key factor motivating group identification [48]. Overall, it could be suggested that when groups that manage shared resources articulate clear purpose and principles, they are more likely to inspire identification with them.
Psychological ownership. Psychological ownership (a subjective perception of possessing an entity) has been shown to be an important determinant of attitudes and choices in the area of consumer behaviour. For example, a sense of ownership enhances liking and valuing of an object [59,60]. This effect is determined by subjective (rather than factual) ownership [61], and extends to non-material objects [e.g. ideas, 62]. Importantly, psychological ownership has been shown to affect organizational citizenship behaviours when applied to employee organizational context [e.g. 63], and is related to a sense of belonging to an employer organization [64]. In the context of share resource management, it could be expected that group members who experience a sense of ownership of the group that manages the resource would be more likely to identify with it.

Community conservancies in Northern Kenya
In the present paper, we explore the theoretical principles outlined above in the context of community conservancies in Northern Kenya, where pastoralism is the predominant livelihood. In the recent decades, pastoralists have been facing a number of challenges, including climate change, population growth, state interference and land grabbing, and mobility restrictions [65][66][67]. In some instances, these challenges made retaining sustainable levels of grazing impossible and led to land degradation and threats to ecosystem health [68,69]. In Kenya, one initiative that is being promoted to address these problems is the establishment of conservancies, which are legally recognised land management systems that support wildlife conservation and other compatible land uses to help improve livelihoods. This is an attractive option, since extensive pastoralism and wildlife conservation both depend on the same underlying resource (i.e. healthy rangelands), and wildlife tourism may enable pastoralists to diversify livelihoods and generate income. The predominant model in northern Kenya is that of community conservancies, and most are members of an umbrella organization known as the Northern Rangelands Trust (NRT, see figure 1). These conservancies have been established voluntarily by communities on communally-held land, and communities are intended to be the dominant decision-makers through the establishment of representative boards (see SM for history, governance structure, and benefits of conservancies).
Conservancies are new systems of organization that are attempting to integrate with previous systems of collective land access and management, and traditional ethnic and kin-based forms of collective organization and identity (see SM). Traditional norms and customary institutions that have governed land management evolved over long periods of time. In contrast, the conservancy model is being relatively rapidly adopted from outside a community, and is based on skills, knowledge, or values that may be very different from those of traditional pastoralist societies. In this respect, it cannot be taken for granted that conservancies will function effectively, or that communities will trust this new system of organization. Indeed, despite the potential benefits [70][71][72], the extent of willingness to cooperate within conservancies, and, as a result, conservancy effectiveness seems to be variable [73]. Understanding what factors underlie this variation could be extremely valuable for improving the functioning of conservancies and establishing more sustainable land use systems in this region.

Testing pathways linking conservancy identification and cooperative intentions
There is a lack of studies in field contexts (and, in particular, non-western ones) assessing the links between social psychological mechanisms and willingness to cooperate around shared environmental resources. However, one recent study explored predictors of pastoralists' cooperation around shared land in Tanzania [74]. Focusing on cooperation around protecting communal pasture land from soil erosion, this work suggested that community identification was significantly related to willingness to cooperate with other community members, and this link was primarily mediated by cooperative community norms. The present study builds on this work by assessing whether group identification is also an important parameter in the context of conservancies, which (a) are larger-scale systems of organization that involve multiple settlements, (b) have a complex governance structure that has been deliberately adopted from external models, and (c) aim to produce multiple collective benefits.
The evidence reviewed above describes a number of social psychological processes that may be linked to willingness to cooperate in the context of conservancies. Based on this evidence, we expect that social identification with a conservancy is likely to be related to stronger cooperation and willingness to prioritise conservancy-wide benefits. This relationship is expected to be mediated by perceived support from the conservancy and trust in its actions, perception that the conservancy manages collective resources effectively (i.e. collective efficacy), and by the development of cooperative group norms [74]. We also assess several factors that could strengthen conservancy identification in the first place. Specifically, it could be expected that conservancy members would feel a stronger sense of belonging to their conservancy when they (a) perceive that its actions bring benefit to their households and communities, (b) perceive an alignment between the conservancy's and their own interests, (c) have a clear understanding of the conservancy's aims and processes, and (d) experience a sense of psychological ownership of the conservancy. Table 1 lists the variables measured, and an overall path diagram that summarises the predicted relationships is presented in figure 2.

Participants and design
The study used a survey design. Participants were recruited from three pastoralist conservancies that are members of the NRT in Kenya. The conservancies Rabinovich et al [74] 0.96 Note: Where no source is specified, the items were developed for the purpose of this study. For the measure of motive alignment which is comprised of two items, a correlation coefficient is reported instead of Cronbach's alpha.
were selected to ensure that the three main ethnic groups in the region are represented (Borana, Maasai, Samburu), that the selection covers both older and more recently established conservancies, and includes conservancies with higher and lower level of success. Given the criteria listed above, the selected conservancies are representative of the wide range of conservancies in the region. The sample consisted of 607 participants (58% female, mean age = 38.38, SD = 15.08). We aimed to recruit 200 participants per conservancy, which was determined by budgetary constraints, and satisfied the minimal overall sample size requirements (the rule-of-thumb of ten participants per estimated parameter of the proposed model). Within each selected conservancy, all settlements were sampled (seven, five, and eight respectively). A stratified random sampling was used, where participants within each settlement were approached randomly, and the number of participants recruited from each settlement was proportional to its size.

Materials and procedure
Materials were developed in English, translated into the local languages, and back-translated into English to ensure a precise retention of meaning. The data were collected by trained enumerators who spoke participants' native languages. Enumerators described the purpose of the study, explained the response scales, and read out each item, recording participants' verbal responses. Participants responded to all items on a five-point Likert scale (1 = 'strongly disagree' , 5 = 'strongly agree'). Items within each scale were averaged to compute a single score. Demographic data were collected in the beginning of the questionnaire, followed by the measures of conservancy identification, psychological ownership, perceived support, perceived clarity of purpose and decision-making, trust in the conservancy board and grazing committee, perceived alignment of motives, perceived benefit from conservancy actions, conservancy's efficacy in managing the land, perceived community norms of cooperation with the conservancy, and intentions to cooperate with the conservancy and follow its rules. Sample items and scale reliabilities are shown in table 1 (see SM for complete list of items).

Results
Means, standard deviations, and bivariate correlations for all variables are shown in table 2 (see table  S2 for mean differences between conservancies). The predicted model was tested using path analysis with Mplus 8.4 [78]. To account for interdependence of observations within each conservancy, we created two dummy variables coding membership of a specific conservancy and included these as covariates in all analyses. The initial analysis suggested that the model did not meet established criteria for good fit: Analysis of the indirect effects showed that there were significant indirect paths between conservancy identification and intentions via community norms (β = .11, SE = .02, p < .001), and via conservancy efficacy (β = .09, SE = .02, p < .001). All indirect two-step links from the predictors of conservancy identification to cooperative intentions were statistically significant (see table S1). Multiple group analysis suggested that the model was robust across the conservancies (see table S3). Note: All correlations are statistically significant at p < .001. Note: * * * p < .001, * * p < .01, * p < .05. [79] recommends to interpret path estimates values of .10 as 'small' , .30 as 'medium' , and .50 and above as 'large' effect sizes.

Discussion
The results provide support for the hypothesis that conservancy identification is linked to willingness to follow conservancy rules and cooperate around shared land management. Overall, the results suggest that members of pastoralist communities who perceive their conservancy as a beneficial, well-meaning, and transparent organization, tend to feel a stronger sense of belonging to it, and this sense of belonging, in turn, increases perceptions of conservancy efficacy and strengthens cooperative community norms, leading to a greater willingness to cooperate with conservancy land management decisions. On a theoretical level, the findings are consistent with the principle that links group identification with willingness to take group-benefitting action [e.g. [17,18]. The present study responds to the calls for extending research on cooperation in shared resource dilemmas beyond laboratory contexts [4], and demonstrates that the link between collective identification and cooperation persists in a field setting that involves larger groups with nested structures, where consequences of (non-)cooperation loom large for the communities involved. The results also provide some support for the processes that were expected to mediate the relationship between conservancy identification and willingness to cooperate around its rules. In particular, in line with the predictions, we found that community norms supportive of cooperation at a conservancy level, and perceived conservancy efficacy in managing shared land mediated the link between conservancy identification and cooperation. These findings are consistent with the existing evidence for the role of group norms in motivating sustainable behaviour [e.g. 44,47], and with the literature that demonstrates the impact of efficacy on willingness to take collective action [19][20][21][22]. The evidence for the mediating role of community norms is consistent with the earlier work in pastoralist contexts in Tanzania [74], and extends it to demonstrate that community norms are strongly related to cooperation over large areas incorporating several local settlements. On the other hand, while previous work in pastoralist communities [74] failed to find evidence for the role of collective efficacy as a mediating process, the present study provides such evidence in the context of conservancy cooperation. Unlike local communities in Tanzania, conservancies in Kenya are newly created organizations that participating communities have chosen to join voluntarily. Given that conservancies have been established with an explicit goal of managing shared land in an effective way, their efficacy may be more salient or keenly monitored, and may be particularly important for motivating ongoing cooperation.
In the present study, we do not find consistent evidence for the role of trust in mediating the relationship between conservancy identification and cooperation with its rules. This is not consistent with the wider research on the impact of trust on cooperation [e.g. 30,31]. A potential explanation is that 'trust' in a conservancy board relates to aspects of conservancy functioning that are not directly linked to land management and grazing plans. For example, mistrust may stem from mishandling of financial issues by certain individuals, or suspicion about their goals, yet participants may still feel there is value in following grazing rules, especially if many others will do the same. While the present study measures a generalised state of trust, future research would benefit from exploring impacts of different types of trust [80,81]. For example, it is possible that identification with conservancy is linked to affiliative trust, but it is rational or procedural types of trust that translate into rule compliance.
The present findings may have a number of practical implications for practitioners who aim to establish and manage effective pastoralist conservancies. On the most general level, they suggest that it may not be sufficient to establish conservancy rules and expect conservancy members to cooperate with these. It is important to account for group processes taking place at both community and conservancy levels, as these may produce significant variability in individual willingness to cooperate. More specifically, conservancy management boards may be able to strengthen willingness to cooperate by developing a sense of identification with conservancy among its members. Our work suggests that several factors may be conducive to this. First, practitioners may want to ensure that information about conservancy purpose, aims, and decision-making processes is widely accessible and that opportunities for feedback and discussion are offered to ensure its understanding. Second, it is important to align conservancy motives and principles with those of the communities involved, ensuring visible consistency between conservancy and communities' interests [82]. Third, management boards could take steps to widen participation in conservancy decision-making, thereby strengthening the sense of psychological ownership of the conservancy. Finally, it would be essential to ensure that conservancy actions bring demonstrable benefits to communities involved, and that these benefits are clearly communicated to and discussed with conservancy members. A combination of these steps could encourage development of community-level norms supporting conservancy cooperation and strengthen perceptions of conservancy efficacy, which are essential for motivating cooperative action.
While the relationships explored in the present study are based on theory and previous research, the cross-sectional design does not allow for confirming causal pathways. To address this, future research should explore longitudinal or experimental approaches. For example, it would be beneficial to capture group processes that take place during the early stages of establishing a conservancy, and to explore how these develop over time, as well as the extent to which they determine longer-term cooperation outcomes. Another productive approach could be based on comparing cooperation rates between conservancies with different management approaches (e.g. with more versus less transparency and member involvement, or different levels of monitoring rule compliance, see 83,84). Future research could also develop and test intervention methods with the aim of enhancing cooperation rates. These could be inspired, for example, by existing discussionbased approaches to group identification strengthening in collective action contexts [e.g. 20], or by techniques from organizational psychology developed for aligning subgroup goals and strengthening identification with overarching structures [85]. Future research would also benefit from more directly measuring cooperative behaviour. The measure of intentions used in the present research may overestimate willingness to cooperate due to the fact that expressing willingness to cooperate in a survey is not costly. Combining measures of intentions to cooperate with experimental measures of cooperation [86,87] and direct measures of conservancy performance could address these issues. Finally, the focus of the present study has been on factors at an individual level that affect intentions to act cooperatively. However, there may also be many relevant factors at a conservancy level (such as cultural parameters, specific systems of organization, or rule compliance monitoring and enforcement). Future research needs to engage with the multilevel nature of CNRMs in addressing these factors.
In conclusion, the present work is one of the few field studies that assess how group identification is related to willingness to cooperate around managing shared resources in a sustainable way (see [4]). Being the first social psychological analysis of group processes associated with cooperation in the context of pastoralist conservancies, it focuses on a globally important but under-researched context, where the consequences of (non-)cooperation are strongly significant for the communities involved. Our results strengthen the idea that collective identification is essential for enabling cooperation in real-world settings, and have important practical implications. They suggest that cultivating a sense of belonging to a conservancy as a group may play a crucial role in ensuring members' participation and rules adherence, and that this sense of belonging could be encouraged by highlighting householdand community-level benefits of participation, aligning conservancy motives with community interests, and ensuring transparency of purpose and decisionmaking.

Data availability statement
The data that support the findings of this study are available upon reasonable request from the authors.