A participatory research project on health issues at the dog–human interface in seven indigenous communities in Northern Canada: how successful was participation?

Abstract

This study examines community participation in a research project on health issues at the human-dog interface conducted in seven Northern Canadian Indigenous communities. Using a realist evaluation, we analyzed the mechanisms underlying participatory successes and failures in eleven projects derived from an ecosystem approach to health. This qualitative evaluation revealed that the researchers' concerns about collaborators’ participation were largely unfounded. Experienced collaborators demonstrated resilience in the face of previous negative experiences and appreciated the participatory approach. However, over-solicitation of the same stakeholders and mistrust of researchers from the South limited wider community participation. Organizational collaborators played a crucial, but ambivalent role as intermediaries—both facilitating access to communities (boundary-spanning) and, at times, restricting researchers' contact with broader community members (gatekeeping). Although the “dog problem” was recognized as important, it was not a priority in the face of more pressing local or regional challenges. This evaluation highlights the importance of trust, realistic expectations and an understanding of cultural dynamics in optimizing participation in participatory research in Indigenous contexts.

Introduction

Participatory research is essential in addressing community-specific public health issues, yielding richer knowledge and more resilient interventions [1–4]. In Indigenous communities, participation additionally supports cultural and linguistic preservation [5].

Participation was at the heart of our research programme, which focused on the relationships between dogs, humans and their health in Northern Indigenous communities in Canada.

Anchored in ecosystem approaches to health (Ecohealth, One Health) [6–8], approaches that have been increasingly applied in Indigenous community health research in the circumpolar North [9,10] and with Indigenous communities more broadly [11,12], the programme was co-constructed and conducted with various local partners. Nevertheless, organisational partners occasionally appeared overwhelmed by research demands, while broader community participation remained limited. To understand the mechanisms behind participatory successes and failures within the projects, we conducted a Realist Evaluation of participation which we report in this paper.

Health at the human-dog interface in Northern Canada

The research programme aimed to analyse human-dog relationships, propose culturally acceptable solutions, and evaluate their effects, ultimately improving these relationships and promoting dogs' positive roles in communities [13–17]. It comprised three phases: documenting human-dog relationships, assessing and selecting possible actions, and evaluating implemented actions' effects.

Prior to applying for the main funding, the research programme was co-constructed through multiple consultations with representatives from several northern communities, including communities with which some team members had previously collaborated. Interested communities were invited to participate, and the programme's objectives were defined jointly with them. This co-construction process established the participatory foundation on which all subsequent phases were built.

The programme reaffirmed dogs' historical, cultural, practical, and emotional value in remote northern communities. However, dogs can pose physical health risks (bites, pathogen transmission) or psychological concerns (fear, reluctance to go outside) exacerbated by widespread free-roaming dog practices [18]. Disagreements about dog care practices create social tensions. Lack of local veterinary services and limited control over canine reproduction result in overpopulation. These issues interact, linking human and canine health within an epidemiologically and socially complex situation.

The programme included 11 projects across seven Northern Communities, conducted by five teams using various disciplinary methodologies (anthropology, communication, epidemiology, public health, socio-cultural geography, veterinary parasitology). All projects involved community participation. The three phases were to progress at each community's own pace. Kuujjuaq was the first community with which collaborations were established and the programme coconstructed. A Canadian Institutes for Health Research grant funded full implementation in Kuujjuaq and expansion to other communities. As COVID-19 restricted travel for nearly two years, only Kuujjuaq completed all three phases.

Participatory research in Indigenous contexts: key challenges

Participatory approaches have been widely recognised as essential in research conducted with Arctic Indigenous communities [19]. However, their implementation is conditioned by a complex socio-historical context. Colonial history has generated a widespread and documented mistrust of “southern" researchers among Indigenous populations, rooted in past exploitative practices that benefited academic careers while offering no tangible benefit to communities [20,21]. Rebuilding trust requires sustained engagement, humility, and long-term relational investment [22]. Frameworks for the co-production of knowledge in the Arctic specifically emphasise that equitable research demands Indigenous partners be involved at all stages, from agenda-setting to dissemination [23].

A persistent challenge in this context is over-solicitation: the same community members, particularly Elders, are repeatedly approached by multiple research teams, leading to research fatigue [24]. Organisational intermediaries play an ambiguous role, simultaneously enabling and constraining researcher access to community members [25]. Additionally, gender dimensions of the research relationship remain under-theorised in Arctic research, despite evidence that researcher positionality—including gender—shapes field dynamics and participant access in context-dependent ways [26–29].

Participation in the programme

Participation levels varied across communities, activities, and categories of participants. For clarity, we distinguish three categories throughout this paper. Collaborators were individuals with sustained involvement in the programme; they were either representatives of partner entities (municipal authorities, Indigenous development organisations, governmental agencies, public health professionals, educators, mushers) who participated beyond their usual professional duties—or facilitators, i.e. individuals unaffiliated with partner organisations who promoted specific projects by enabling contact with key individuals or providing logistical support. Community participants, by contrast, were community members with more limited programmatic involvement, who joined specific research activities such as surveys, interviews, focus groups, photovoice sessions, or school workshops. This paper does not involve community co-researchers or a community advisory board; rather, community engagement was embedded in the programme’s participatory design from the outset. Most activities were discussed with and accepted by community members prior to implementation.

Evaluation focus

Despite active community participation, researchers encountered challenges. Halfway through the programme, several researchers wondered if the level of participation was adequate and, based on comments on their part, worried that their partners were happy with it. Two researchers less involved in fieldwork were tasked to investigate. A project lead with long-term experience in the region was hired to supervise the process. A researcher with experience with realist evaluation, external to the project, joined the evaluation team.

Participation determinants extend beyond research activities to cultural, social, and structural contexts. Realist evaluation offered a rigorous method to identify underlying mechanisms affecting participation in the programme. This study aimed to understand community participation processes using the programme as a case study through a critical meta-research perspective. Our research questions were 1) How satisfying was participation and engagement for all parties involved? 2) What factors have limited community participation in projects?

The evaluation protocol received ethical approval from the authors’ university and was conducted accordingly. All participants provided informed consent prior to enrolment in the study. Communities deemed no further ethical approval necessary.

Methods

Realistic evaluation seeks to understand interventions through empirical observations, focusing on underlying mechanisms and contextual influences [30]. The method utilises “Context - Mechanism - Outcome” (CMO) configurations derived from a programme theory explaining intervention effects. A programme theory, central to realist evaluation, is an explicit theory of how an intervention is expected to produce its outcomes, articulated through Context–Mechanism–Outcome (CMO) configurations [31]. These configurations specify the contextual conditions under which particular mechanisms are triggered, producing specific outcomes. The realist evaluation of participation in the programme involved developing an initial programme theory and refining it through two iterations [31].

To develop the initial programme theory, research team members were invited to participate in a collective process. We asked them to reflect upon the place of participation bring in their project, their satisfaction with it and the mechanisms implemented. Additionally, they completed sections I-IV of Neef and Neubert's [32] participation evaluation framework, which helped identify contextual obstacles and favourable conditions, as well as potential mechanisms. Despite working within the same programme, contextual differences between projects resulted in varied postures and perceptions, providing deeper insights into participation processes.

Participatory research typologies, often inspired by Arnstein [33], present participation as a linear continuum from low (information gathering) to high (complete community involvement) [2,4]. Some typologies idealise maximum participation intensity as a desirable “nirvana” [32]. Taking a more pragmatic approach, Neef and Neubert propose a framework categorising participation across six dimensions: Project type, Research approach, Researchers' characteristics, Researcher–stakeholder interaction, Stakeholders' characteristics, and Stakeholders' benefits. This framework does justice to the multidimensional and dynamic nature of participation. It can serve as a basis for a more thoughtful and transparent integration of participation elements into research projects, for example in deciding for which questions and in which phases participatory elements could be used or strengthened.

The Neef and Neubert framework served two complementary purposes in our evaluation. First, research team members used it collectively to generate the initial programme theory—that is, to articulate the hypothesised CMO configurations. Second, its six dimensions structured the interview grid used to test these configurations with collaborators.

In this paper, the term “stakeholders” refers broadly to any individuals or groups with an interest in the research topic, and we have ensured consistent usage throughout the manuscript. While acknowledging legitimate critiques of the term “stakeholder” for its colonial connotations [34], we retain it throughout this paper as it is the terminology used in Neef and Neubert's ​​​​[32] framework, which structures our evaluation.

Research team members individually completed framework sections I-IV and discussed them in community-specific groups. The evaluation team analysed these data to derive an initial programme theory, which was presented to all researchers for collaborative refinement and consensus. At the end of the process team members were able to devise a common response to the different sections of the Neef and Neubert’s framework and agree upon the initial hypothesis and configurations.

We planned to test the initial programme theory with both collaborators and community participants. However, COVID-19 prevented community travel, and given digital communication challenges in Nunavik, we tested hypotheses remotely with collaborators only, who could provide perspectives on both their experiences and those of community participants with whom they are in close contact.

After compiling a comprehensive collaborator list, an external researcher conducted semi-structured phone interviews in English or French with those who consented after up to three contact attempts via email, telephone, or Facebook. All interviews were recorded.

The interview grid tested CMOs from the initial programme theory and allowed for the completion of the Neef and Neubert's framework dimensions. The interviews were designed as a conversation, with questions organised into sections to guide the interview and ensure that the dialogue covered all relevant areas. Interviews were transcribed, synthesised, and analysed thematically using both inductive and deductive approaches.

Analysis determined whether each CMO was rejected, confirmed, partially confirmed, or inconclusive. We modified existing CMOs and added new ones based on interview content, creating an intermediate programme theory discussed with the research team. A second interview grid testing modified CMOs was administered to the same collaborators following identical procedures, leading to the final realist evaluation results. The two interview rounds reflect the iterative nature of realist evaluation. Round 1 tested the initial CMO configurations derived from the research team's collective process. Round 2 tested these revised configurations, producing the results reported here.

Throughout this paper, each configuration is labelled “CMO” followed by a letter indicating the hypothesis to which it belongs (a, b, c, or d) and a number identifying the specific configuration (e.g. CMOa1 denotes the first configuration under Hypothesis A). When introduced for the first time in the results, each CMO is accompanied by a brief descriptive statement; full wording is provided in Tables 1–5.

Table 1.

Hypothesis A (slightly modified) final configurations.

Hypothesis A (slightly modified) - Stakeholders’ past experience with research, the development of trusting relationships with researchers and the intensity of the request influence the involvement and level of participation of local players in the research programme.
	Context	Mechanism	Outcome
CMOa1 Rejected for collaborators Confirmed for community participants (with modifications)	Negative past experiences and over-solicitation of indigenous stakeholders by other research teams (exhaustion). Feeling that research leads to results for researchers (only), there are many research projects involving southern researchers that have no spin off for communities. Over-solicitation of the same stakeholders by other research teams.	A priori disillusionment on the part of indigenous stakeholders with the project's researchers, leading to little or no willingness to participate.	Indigenous stakeholders lack interest in participating (difficulty for project researchers in recruiting them to take part in research activities).
CMOa2 Rejected for indigenous collaborators Led to CMOa7	Tensions in relations between Indigenous and non-indigenous people. Mistrust of research, which is perceived as benefiting only researchers. The small size and geographical isolation of these communities mean that everything is known, and rumours are rife.	Indigenous stakeholders don't feel comfortable participating/collaborating and being seen with non-indigenous because they might feel as if they betrayed their community members (anticipated social stigmatisation). Exacerbated feelings of embarrassment about what others might say.	Indigenous stakeholders refuse to participate/collaborate in the project (difficulty for project researchers to recruit them to participate/collaborate in the project).
CMOa3 Partly confirmed Led to new hypothesis D	Negative past experiences and over-solicitation of other stakeholders by other research teams. No recognition of the power imbalances between the professional staff of organisations and other local players, and the former's absolute control over research projects.	Perception of the organisations' collaborators that other stakeholders are indifferent or disillusioned with the research and will therefore not participate in the activities of the present research programme; that the participation of other local stakeholders is pointless.	The people who work for the organisations are not in favour of involving other stakeholders in research activities (project researchers have difficulty recruiting them due to lack of support).
CMOa4 Non conclusive	Previous positive experiences of collaboration between researchers in this programme and stakeholders, foster the development of bonds of trust that encourage exchanges.	Sense of trust among stakeholders, who are ready to embark on the new project and share their perspectives. They feel at ease with members of the research team, and confident enough to share their personal views on the question at hand.	Stakeholders participate easily and fully in research activities (easy recruitment and access to intimate and varied perspectives).
CMOa4—2nd round Non conclusive	Previous positive experiences and the development of bonds of trust over time between the research team in the current programme and collaborators, encouraging exchanges.	Collaborators feel at ease with the researcher and confident enough to share more of their perspectives and contact networks.	Collaborators share their perspectives fully and offer support in involving other stakeholders in research activities, notably by giving researchers access to new networks to recruit them.
CMOa5 Rejected for collaborators Led to CMOa8	Previous experience of collaborators with research projects using a traditional scientific approach.	Collaborators don't understand the new participatory approach; they have their own ways of interacting, participating and collaborating on research projects, and feel disconcerted by the new ways of doing.	Collaborators are reluctant to get involved in the project (researchers find it difficult to get them to collaborate in a non-traditional approach, which requires a relatively high level of commitment).
CMOa6 Confirmed for collaborators	Long-term working relationships (prior to the current programme) between researchers and collaborators enable mutual trust to be established. Collaborators are familiar with the research process and the application of research methods and have acquired an appreciation of the value of research.	Collaborators are interested in the project, feel a sense of belonging and autonomy, and feel a certain freedom granted by the researchers in decision-making.	Collaborators feel inspired and take the initiative in the project, drawing on their expertise and research know-how (towards co-construction).
CMOa7 Output from round 1 Rejected in round 2	Researchers arrive with a lack of sensitivity and knowledge of the socio-cultural and historical context (colonialism and its impacts), as well as insufficient social skills, which can make stakeholders feel disrespected.	People working with organisations are concerned about the potential negative impacts of research projects or the behaviour of researchers. They don't want to be associated with researchers whose practices are deemed unsuited to the local context, practices that could discredit their organisation in the eyes of the residents of their community	The people working in the organisations are suspicious of the researchers and are not fully committed to working together on the project.
CMOa8 Output from round 1 Rejected in round 2	Previous experience of collaborators with research projects using only a traditional scientific approach (ways of interacting and collaborating on research projects and relationship to expert knowledge). Participatory nature of the research, with a higher expectation of commitment from collaborators than with traditional research.	Collaborators are met with unusual demands on their time from the researchers, which was disconcerting when compared with the usual practice in research. For example, they consider researchers to be the only ones with expertise on the subject and expect them to be the ones to decide.	Collaborators are reluctant to become more involved in the participatory project (difficulty for researchers to get them to collaborate in a non-traditional approach, which requires a relatively high level of commitment).

Table 2.

Hypothesis B (slightly modified) final configurations.

Hypothesis B - In the relationship between researchers and collaborators, understanding and considering each other's roles and expectations promote collaborators' commitment to the research programme.
	Context	Mechanism	Outcome
CM0b1 Rejected for collaborators	The expectations of the researchers and collaborators were fully or partially named. On the other hand, explicit expectations were not formalised at the outset and were not monitored throughout the project (in particular, the nature of what the research programme could offer and how it would go about it). There was no agreement on what the project would deliver.	The collaborators did not feel that their expectations were considered by the researchers in the project and felt frustrated as a result. Specifically, they are disappointed because they expected the researchers to take responsibility for and provide solutions to the problems associated with dogs.	Disenchantment among collaborators, with a loss of confidence in the project, possibly leading to their disengagement.
CMOb2 Rejected for collaborators	Difficulty in ensuring continuity and follow-up in the programme, given the funding, the remoteness of the northern communities (logistical constraints and costs) and the respective calendars (academic and community).	A feeling of abandonment and disappointment among collaborators, who can't see the research programme progressing and the continuity of actions (discontinuous project).	Disengagement of project collaborators, with reduced motivation to participate. Difficulty for researchers to maintain confidence in the project.
CMOb3 Output from round 1 Rejected in round 2	Difficulty of collaboration in a non-traditional (participatory) research approach, which requires a relatively high level of commitment. Expectations and roles in relation to the project were not clearly defined at the outset, and there was no agreement in place, nor monitoring of expectations and roles throughout the project.	Collaborators perceive that researchers have too many expectations of them, and they feel that what is asked of them goes beyond their collaborative role (e.g. perception that researchers rely too much on them for solutions).	Limited involvement of organisational collaborators (researchers find it difficult to get them to participate, as they slow down and complicate the project's progress).
CMOb4 Output from round 1 Rejected in round 2	Gap between the operational expectations of collaborators for researchers and what researchers are able or competent to do.	Collaborators do not see the usefulness of certain aspects of the research and are disappointed or even frustrated by the lack of concrete, rapid action to address the issues they perceive.	Disengagement of collaborators, with a drop in confidence and a potential abandonment of collaboration in the project.

Table 3.

Hypothesis C (slightly modified) final configurations.

Hypothesis C - The research topic and approach (methodology, epistemology, research design and process) influence the degree of participation of local stakeholders.
	Context	Mechanism	Outcome
CMOc1 Rejected for collaborators Confirmed for community participants	Dog-related issues are often given a lower priority than other issues faced by indigenous stakeholders.	Indigenous stakeholders do not feel that the project is a priority in their community, leading to indifference towards the research programme.	Indigenous stakeholders lack interest in participating in the research programme (difficulty for project researchers in recruiting them to take part in research activities).
CMOc2 Confirmed for collaborators with additions	The lack of a mandate for organisations to deal with dog-related issues leads to a lack of human, logistical or financial resources, and a lack of cohesion in the resources devoted to them. The multiplicity of research projects in indigenous communities means that employees are overstretched and have to prioritise their activities according to their organisational mandate.	Collaborators feel a lack of legitimacy in participating in the research programme, or a sense of limitation in getting involved because it's not part of their mandate, which makes them feel limited in their commitment. They feel overwhelmed by the demands of all researchers and struggle to respond.	Limited involvement of people working in the organisations (researchers find it difficult to get them to participate, as they slow down and complicate the project's progress) by their lack of availability (to respond to researchers' requests within a reasonable timeframe).
CMOc3 Partly confirmed for community participants	The research process is short, with attractive activities to involve stakeholders and immediate expected spin-offs (research product delivered in a tempting format).	Stakeholders perceive a short-term benefit from participating in research activities.	Limited involvement collaborators (researchers find it difficult to get them to participate, as they slow down and complicate the project's progress).
CMOc4 Rejected for collaborators. Partly confirmed for community participants	Sometimes, several researchers work in the same community with the same stakeholders and present their respective projects, with little integration in their various communications.	Stakeholders perceive redundancy between projects, feeling frustrated and weary to participate/collaborate or offer their support for activities (demotivation).	Disengagement of stakeholders, with a drop in confidence and potentially in participation/collaboration in the overall project.
CMOc5 Output from 1st round Inconclusive in round 2	Researchers communicate to stakeholders the activities and spin-offs of the research programme in other communities, particularly those that have led to actions/solutions.	Stakeholders, seeing the activities and spin-offs of the research programme in other communities, feel interested and motivated to participate and collaborate with the researchers.	Stakeholders are keen to participate in the project and are open to collaboration with the researchers.

Table 4.

Hypothesis D final configurations.

Hypothesis D - The perceptions and power relations of collaborators vis-à-vis other local stakeholders influence researchers' access to them.
	Context	Mechanism	Outcome
CMOd1 Output from round 1 Rejected by respondents in round 2	No recognition of the power imbalances between professional organisations and other local stakeholders, and the former's absolute control over research projects. The participatory approach of the research programme to gain access to a diversity of perspectives is not necessarily well understood, and the expected role of collaborators as community liaisons has not been explained by researchers.	The perception of the people working for the organisations that they represent the population and that the participation of other local stakeholders is therefore irrelevant.	Organisational collaborators are not in favour of involving other local stakeholders in research activities, and offer little or no support in this regard. Since the collaborators do not play their role as intermediaries, the project researchers have difficulty recruiting other local stakeholders due to a lack of support.
CMOd2 Output from round 1 Rejected in round 2	Socio-cultural and communication gap between indigenous stakeholders and researchers, making it difficult to establish contact and exchange, and thus to choose activities that encourage participation. Over-solicitation of indigenous stakeholders by other researchers to participate in various activities (exhaustion).	Perception of collaborators that local indigenous stakeholders will not be interested in participating in the activities of the present research programme, due to activities deemed of little relevance, or ill-adapted in a context of exhaustion in the face of research in general.	Collaborators offer little or no support in involving other local players in research activities. As collaborators do not play their role as intermediaries, project researchers have difficulty recruiting them due to a lack of support.

Table 5.

Configuration X.

	Context	Mechanism	Outcome
CMO X Partly confirmed	Male-female relations in native communities. The small size and geographical isolation of these communities mean that everything is known and rumours fly.	Indigenous stakeholders feel uncomfortable participating/collaborating and being seen with a woman/man (may give rise to jealousy). Exacerbated feelings of embarrassment about what others might say.	Indigenous stakeholders refuse to participate/collaborate in the project (difficulty for project researchers to recruit them to participate/collaborate in the project).

Main findings

After the presentation of the research participants, a first results section is organised around the Neef and Neubert dimensions and a second report on the testing of hypothesis and CMOs.

Research participants

All twelve members of the research team (researchers and students) participated in the study. They also established a list of 30 collaborators (20 organisation’s representatives and 10 facilitators) who were invited to participate. Nine of them accepted and participated in the first round of interviews while only two participated in the second round. These nine people came from six different communities (see map in Appendix 1), six were collaborators and three facilitators, two are Inuit. The mean duration of the interviews was 50 minutes for the first round, and 20 minutes for the second.

The reasons for non-participation were: unreachable (no valid contact details), no response, no response after an initial contact phase, refusal due to lack of time. The interviewer reported on the difficulties to reach out to people from a distance given that many do not consult their emails often. Some also expressed they would have preferred to be contacted in person and did not wish to do an interview over the phone.

Dimensions of participation

This section follows Neef and Neubert's participation dimensions, presenting results from the research team (dimensions I-IV) and collaborators (dimensions IV-VI). Team members' contributions analysis generated an intermediate programme theory with three hypotheses and 12 CMO configurations, denoted as “CMO-hypothesis(a, b, c)-number” (e.g. CMOb1). See Tables 1–5.

Dimension 1 - type of research

This dimension examines expected participation levels based on research type, objectives, beneficiaries, risks, and context ([32], p. 183).

Participation expectations were high, as the programme was designed participatory from inception. The lead investigator consulted local organisations about needs and interests before establishing the research programme. Throughout, other concerns from local stakeholders emerged. The solution-focused implementation approach created space for mutual expectations, but these remained implicit rather than formalised, potentially affecting participation (CMOb1).

Team members hypothesised that high participation expectations paradoxically created pressure on collaborators, who may have been unfamiliar with the participatory approach or misunderstood researcher intentions (CMOa5).

Beneficiaries of the programme include heterogeneous local populations and organisations, with communities differing in geography, size, and ethnicity. Within communities, roles and attitudes toward dogs vary. The research community also benefits through publications, scientific advancements, and training.

Research inherently carries risks, including failure to achieve objectives or implement solutions, potentially disappointing community members. Community remoteness created research discontinuities, exacerbated by COVID-19. Differences between community and academic calendars may have negatively impacted participation (CMOb2).

Dimension II - approach

This dimension examines the programme's epistemology, methodology, and process ([32], p. 184).

The programme acknowledged diverse knowledge systems despite incorporating disciplines with positivist foundations. This epistemological tension manifested methodologically through collaborator inclusion at various research stages, even when tools required minimal participation (e.g. inviting feedback on surveys, analysis, and dissemination strategies).

Team interdisciplinarity generated varied research designs with different participation levels. Data collection methods included interviews, surveys, focus groups, participatory mapping, and ethnographic investigation. While all projects shared the programme's overarching goal, specific objectives addressed local stakeholder concerns, allowing development of more engaging activities that positively impacted participation (CMOc3). Examples include photovoice implementation with a contest resulting in a calendar showcasing animal care, and Kuujjuaq's sled dog demonstrations for students, highlighting traditional practices.

Projects lacked methodological integration, although methods were shared through a common repository, regular team meetings, and joint academic conference participation. This fragmentation generated confusion among collaborators, who requested projects and personnel mapping (CMOc4).

Dimension III - researchers’ characteristics

The third dimension of Neef and Neubert's framework [32], p. 186) examines researchers' participatory experience, attitudes toward local populations, and problem-solving commitment.

The interdisciplinary team ranged from experienced researchers to master's students, with varied participatory research backgrounds. All expressed enthusiasm for this approach, aligning with their values and belief in incorporating multiple voices and knowledge systems. Despite this, many acknowledged insufficient academic training in participatory methods and expressed concern about creating “false” community expectations.

The team agreed that researchers should contribute to solution implementation while recognising that implementation decisions belong to communities. This philosophy manifested in Kuujjuaq through a multi-criteria decision-making analysis (MCDA), a participatory process in which community members—including diverse people beyond organisational representatives—identified, discussed, and prioritised interventions, ultimately shaping the community's “Dog programme”. Following the MCDA, community participants, supported by the research team, secured funding to implement the selected measures, including a pilot veterinary services project, school workshops reviving Inuit practices related to dog care and preventing dog bites and rabies transmission, and awareness-raising social media content. Collaborators subsequently took ownership of the programme's continuation. Similar analyses planned for other communities were cancelled due to the pandemic.

Dimension IV - researcher– stakeholder interaction

This dimension examines stakeholder involvement and control over research.

Regarding the research team, the sheer diversity of interested groups and organisations has favoured participation, but this was tempered by that for all groups, the “dog problem” was far from their main concerns (CMOc1 and c2). Despite general satisfaction with participation levels, team members identified several challenges affecting researcher-stakeholder interactions. Prior research relationships created both opportunities and barriers. Positively, established relationships with some collaborators fostered mutual trust and participation motivation (CMOa4 and a6). Negatively, these long-term partners became intermediaries who “filtered” researchers' access to broader community members, possibly stemming from negative past experiences between Indigenous and non-Indigenous people (CMOa3). The mistrust of Indigenous stakeholders toward researchers from the South is a possible explanation why some parents refused to give their consent for an assessment of their children’s knowledge and attitudes following a workshop that had been organised at the school. However, while this pattern was initially interpreted as reflecting mistrust toward southern researchers (CMOa1), it may equally reflect lack of interest in the evaluation component, competing demands, or parental preferences unrelated to researcher mistrust (CMOa1). Project continuity challenges arose from geographical remoteness, incompatible calendars (academic/community), and COVID-19. To prevent discontinuities, it was initially planned, that each project/community had a (paid) local coordinator, but their recruitment was difficult in some communities which was attributed to mistrust towards southern researchers (CMOa2).

Regarding the collaborators, they identified several community participation barriers: inadequate recruitment strategies; limited outreach beyond immediate contacts; stakeholder unavailability or disinterest; and unperceived short-term benefits. Activity attractiveness influenced participation—parasitic surveys involving faeces collection deterred respondents while photovoice activities were well-received. Some suggested home visits rather than centralised meetings, with one noting that northern timelines differ from southern expectations, potentially disappointing researchers. Communication challenges included confusion about research stages (perceived as repetitive), scientific jargon, and academic constraints affecting project implementation. Despite challenges, mutually beneficial exchanges were acknowledged and much appreciated.

Dimension V - stakeholders’ characteristics

This dimension examines stakeholders' experience with the research, their perception of the programme and team, and their “scope for action” ([32], p. 188).

Respondents judged the project highly relevant, acknowledging real dog-related community problems. The project effectively highlighted these issues while considering cultural dimensions, meeting a genuine need.

The research approach received strong appreciation, with respondents highlighting researcher respect, confidentiality, dialogue openness, strong presence despite remoteness, transparency, seriousness, action-orientation, results sharing, resident inclusiveness (primarily Indigenous), and the project evaluation component. Among the positive impacts and spin-offs for the communities, the benefits of coupling veterinary services and research activities were highlighted, notably the support of the demand for a veterinarian in the North. Despite project heterogeneity, respondents highly valued the research team, emphasising their long-term commitment through frequent northern visits over several years, which built essential trust bonds.

Dimension VI - stakeholder benefits

This final dimension addresses “the crucial question of expected benefits or outcomes of participatory research” ([32], p. 189). The programme catalysed law reinforcement, human resource recruitment, and funding access. One respondent described evolving from initial consultant to deeply involved participant, ultimately developing project ownership. Another collaborator noted that project networking unexpectedly provided access to additional resources.

Hypotheses and CMO

The three hypotheses were tested with collaborators through two interview rounds. We present here the final results, indicating which CMO configurations were retained, rejected, modified, added, or remained inconclusive. Collaborators often confirmed hypotheses while adding nuances, these appear in italics in the final versions (Tables 1–5). For reasons explained previously, hypotheses were tested only with collaborators. Therefore, confirmation or rejection of CMOs regarding community participants reflects collaborators' opinions.

In total, 20 configurations concerned the collaborators themselves (11 in the first interview cycle, nine new or modified in the second). Ultimately, two configurations were confirmed, two partially confirmed, 12 rejected, three remained inconclusive, and one could not be tested (Figure 1). Respondents' perceptions of community participants' opinions were collected for six configurations in the first cycle and five in the second. In the end, two were confirmed, three partially confirmed, two rejected, three remained inconclusive, and one was not tested (Figure 1).

Figure 1.

Testing of CMOs with end-results.

Hypothesis A: stakeholders' past research experience

Six configurations were grouped under this hypothesis (Table 1). Respondents emphasised that trusting relationships with researchers were crucial to their involvement, even leading to initiative-taking. They suggested that participation barriers stemmed more from over-solicitation than negative research perceptions. Hypothesis A was modified accordingly.

Respondents rejected CMOa1 for themselves. Most of them have had unsuccessful experiences of collaboration with researchers, which in their opinion, prevents disillusionment. None felt reluctance to participate as they were motivated by their personal interest in dogs or their professional obligations. However, they believed other Indigenous stakeholders experienced disillusionment and fatigue from over-solicitation, affecting their motivation to participate. They noted that researchers repeatedly engage the same individuals, creating availability issues. Colonial history and negative research perceptions were identified as contextual factors hindering trust-building.

Indigenous respondents disagreed with CMOa2. Their extensive collaboration with non-Indigenous researchers caused no discomfort regarding others' perceptions. One respondent expressed concern that researchers lacking cultural understanding might “do damage” through disrespectful behaviour, potentially affecting their organisation's reputation. This concern generated a new configuration (CMOa7), not confirmed in the second round.

Respondents agreed that community participants might be disillusioned by previous research experiences (CMOa3), complicating recruitment. While most respondents favoured community involvement, one admitted providing little support for activity organisation and blocking stakeholder access, believing these individuals wouldn't participate. Respondents cited socio-cultural factors (shyness, unreliability, lack of initiative, addiction problems) and culturally inappropriate activities or recruitment methods as Indigenous participation barriers.

Though respondents generally supported community participant involvement, scepticism about participation willingness, especially for unattractive activities, limited their support. This important configuration explaining participation deficits was divided into two configurations under a new hypothesis D, tested in the second round.

CMOa4 results were inconclusive in the first round, as most respondents were first-time research team collaborators. Also, the wording of the question made it unclear whether it referred to previous experiences within this programme or with researchers in general. This CMO was reformulated to focus on trust-building over time but remained inconclusive (see COMa4 - 2nd round). And, contrary to CMOa5, respondents unanimously viewed the participatory approach positively, though some felt highly solicited, generating a new configuration (CMOa8) that was tested and rejected in the second round, as collaborators described solicitation as appropriate. One respondent felt limited research influence given the team's expertise.

In communities with pre-existing working relationships with team members, respondents confirmed CMOa6, highlighting connections between trust-building and collaborator-led initiatives. The research programme became a springboard for Kuujjuaq collaborators who, with research team support, submitted a dog programme funding application. Respondents noted increased independence and shared expertise.

Hypothesis B: expectation and roles

Both CMOs under this hypothesis were rejected by respondents regarding their collaborator roles, though their responses led to new configurations. The hypothesis wording was modified to incorporate the concept of roles (Table 2).

Concerning CMOb1, most respondents felt the project met or exceeded their expectations, particularly regarding dog importance awareness and care promotion. One respondent reported sensing researchers' desire for them to play a larger role, with solution implementation falling on their shoulders, causing some detachment from the project. CMOb2 reflected researchers' perceived difficulties in ensuring project follow-up, potentially disappointing collaborators. Respondents contradicted this, emphasising that researchers' sustained presence ensured continuity. Several noted their expectations were long-term given the northern context, with most acknowledging pandemic-related delays. Respondents appreciated regular project updates and results feedback, which strengthened researcher-collaborator trust.

Overall, the programme met collaborators' expectations. However, one collaborator's concern about researchers' high implementation expectations generated two new CMOs (CMOB3-4), both tested and rejected in the second interview round. Second-round respondents didn't perceive unusual demands (CMOb3), instead finding the approach appropriate, pressure-free, and aligned with expectations. Disappointments (CMOb4) stemmed not from researchers' solution implementation limitations but from difficulties engaging northern community members, preventing resolution of dog-related problems.

Hypothesis C: research topic and approach

Indigenous and non-Indigenous respondents disagreed with CMOc1. While finding the project relevant and dog-related problems important, they acknowledged competing community priorities affecting stakeholder participation (Table 3). Long-term programme collaborators confirmed the impact of lacking clear organisational mandates for dog-related issues (CMOc2). They noted concurrent project demands and the tendency to prioritise local initiatives over “Southern” projects. Interviews clarified the mechanism, leading to configuration modifications.

Configuration CMOc3 primarily concerned community participants. Respondents indicated certain activities were less appreciated by community participants. Team discussions generated CMOc5, better reflecting their experience of how benefits obtained in one community influenced another. This CMO remained inconclusive after round 2.

Respondents rejected CMOc4 in round 1. They attributed redundancy perceptions to scientific methods rather than project duplication. Respondents believed people didn't fully comprehend scientific approach requirements, potentially causing frustration when progress seemed stalled.

Hypothesis D: emerging themes

Respondents in round 2 rejected CMOd1. They strongly favoured involving local stakeholders in research activities, particularly to facilitate knowledge acquisition (Table 4). Regarding organisational intermediaries, one respondent noted that projects supported by trusted local organisations generate greater participation compared to initiatives introduced by outsiders lacking established community trust.

CMOd2 was also rejected in round 2. Respondents considered activities appropriately designed for participation (contextually adapted with standard commitment requirements). They identified over-solicitation as the primary participation barrier rather than activity design. They noted motivation variations between groups, with younger generations showing greater interest and increasing advocacy for dog welfare.

Configuration X

Results feedback to the research team after the first series of interviews led to a new CMO configuration, not linked to the hypotheses, which was tested in the second interview cycle.

One respondent believed it to be more difficult for a female researcher to work in collaboration with a local indigenous stakeholder for it may create reactions of jealousy. The other respondent declared not having met such a dynamic but mentioned that a more junior woman could be less intimidating, more accessible (soft approach) than a very experienced and technical woman.

Discussion

This evaluation addressed two questions: (1) how satisfying was participation and engagement for all parties involved, and (2) what factors limited community participation in projects? Realist evaluation, chosen for its capacity to uncover the mechanisms underlying intervention effects [30], allowed us to explore these questions in depth. Despite the limitations detailed below, the evaluation achieved its objectives by deepening our understanding of participation's role in the programme.

Limits

The pandemic, occurring midway through the programme, imposed travel restrictions for close two years in Nunavik and significantly impacted research. Only the first phase was completed in all communities, but Kuujjuaq. While some activities continued remotely (communication programme development for school children, realist evaluation), most were postponed or cancelled.

Despite digital communication availability in Nunavik, it remains neither preferred nor optimal. Consequently, we limited the realist evaluation to collaborators, excluding community participants. Even reaching these collaborators proved challenging. Additionally, mediated interviews potentially compromised quality and hindered sensitive information sharing [35,36]. Of the nine respondents, two are Inuit. This reflects the composition of the collaborator pool, many of whom are non-Indigenous professionals working in northern communities. As a result, the perspectives of Inuit community members may be underrepresented in this evaluation, particularly regarding how trust, intermediation, and research fatigue are experienced from within Indigenous communities.

Hence, travel restrictions prevented access to certain perspectives, notably the voices of people who did not participate. However, most CMOs concerned collaborators, and we could indirectly test those related to community participants.

Despite these limitations, this realist evaluation revealed previously unrecognised processes, thanks to our respondents' generous contributions. We hope these findings benefit other researchers. Most importantly, respondents were very appreciative of the realist evaluation, as expressed in the words of one of them: “They want to make sure they are doing things properly, so they are taking the time to invest time, energy and money in an evaluation of how they do things. … it is the first time I have experienced this type of thing. So, I think it is important to underscore the fact that it is not everybody who takes the time and energy and cares enough for the work and their northern partners to do that”.

A collective process of hypotheses building

The research team process of constructing hypotheses gave rise to an open interdisciplinary conversation. The guidance of the Neef and Neubert framework was innovative and offered a conversational pathway towards a finer understanding of what the research team meant by participation. This collective process allowed junior and senior scientists to voice their concerns, hopes and fears and to put into words feelings and impressions. This gave substance to the apprehensions of team members and enabled us to share them with our partners.

The process was conducted at a time at which most of the projects were well advanced. The richness of the conversations it stimulates convinced us the team would have benefited from such an exercise earlier, even independently of the realist evaluation. Among others, given that team members’ experience in participatory projects varied greatly, an earlier conversation would have shown the relevance of training or mentorship for those with little experience in participatory research.

Collectively, team members devised three hypotheses and 12 configurations. It turned out that participation in the team was very much conceptualised as partnership with various local organisations, most of which were involved in the planning of the research programme. Team members believe this led to a dependence towards those organisations as they became intermediaries between them and non-organisational stakeholders. An early diagnostic of this situation would have allowed us to address the issue before it crystallises (notwithstanding the pandemic).

The multiple-criteria decision analysis (MCDA) conducted in Kuujjuaq has been an opportunity to listen to multiple voices including that of people not tied to our partners and led to very positive and direct benefits. The pandemic prevented us from conducting similar activities in other communities. Would we have been telling a different story without COVID-19? Certainly, but as we discuss further, the issue of “organisation representatives as intermediaries” might still be present.

Realist evaluation

Many team concerns regarding collaborator participation proved unfounded. This gap may partly stem from team members limited participatory research experience combined with high enthusiasm, creating elevated expectations and self-efficacy doubts. Additionally, our respondents likely included primarily those with positive programme perceptions.

Trust, expectations and motivation to participate

As anticipated, long-term collaborators emphasised trust in the research team built on established relationships, aligning with studies identifying trust as foundational to successful participation [2,37]. However, contrary to team beliefs, previous negative research experiences didn't diminish collaborator participation intention; rather collaborators reported that such experiences created resilience against disillusionment. Collaborators appreciated the participatory approach and valued opportunities to conceive their own activities.

When formulating hypotheses, researchers overlooked collaborators' research experience and maturity as contextual variables. Experienced collaborators didn't require explicit roles to participate effectively. Researchers may have overinterpreted certain partner comments regarding confusion about multiple projects—the team's response of creating a project map with descriptions, locations, and team member images likely sufficiently addressed this concern.

The story is different for community participants, especially Indigenous ones. Consistent with team hypotheses and collaborator feedback, repeated solicitation of the same stakeholders (particularly elders) by numerous research teams created exhaustion. This pattern reflects a well-documented dynamic in Indigenous health research: colonial history has generated a deep and lasting mistrust toward outside researchers [38–41], rooted in past practices that extracted data while offering no tangible benefit to communities—what is often termed “helicopter research”. Building trust under these conditions requires sustained relational investment, humility, and long-term commitment rather than procedural compliance alone [21,22]. The pandemic-induced researcher absence, by interrupting this relational continuity, potentially heightened this pre-existing mistrust.

Collaborators as intermediaries

Research team and respondents recognised collaborators' crucial intermediary role between researchers and community participants. Researchers hypothesised that collaborators sometimes restricted community participant access. Respondent answers revealed ambiguity—they expressed willingness to serve as intermediaries while acknowledging controlling which activities should be proposed based partly on beliefs about sociocultural factors, some potentially reflecting colonial legacy stereotypes.

This ambiguity mirrors literature depicting intermediaries as both boundary spanners and gatekeepers. For Indigenous collaborators, concerns about researchers' cultural insensitivity may reduce intermediary intentions due to potential social capital damage [25].

The shift to remote engagement during the pandemic may have intensified collaborators' intermediary role, as researchers could no longer interact with community members directly. However, the boundary-spanning/gatekeeping dynamic was already identified in the pre-pandemic period through the research team's collective reflection, suggesting a structural feature of the collaboration rather than solely a pandemic artefact.

Also, this dynamic is compounded by the structural conditions of Indigenous organisations: lacking explicit mandates around dog-related issues, collaborators navigated competing priorities and an overload of research demands, which shaped how—and how much—they facilitated access to other community members [21]. Addressing the boundary-spanner versus gatekeeper tension requires early, explicit negotiation of the intermediary role, including transparent discussion of what access to the broader community entails and why it matters to the research.

Epistemology of knowledge and the neutrality of science

Despite positive programme views, some respondents perceived researchers as “knowledge holders”. This likely reflects society's “absolute epistemic privilege” accorded to science [42,43] rather than programme-specific issues. Despite the team's ecosystem health approach favoring transdisciplinarity, prevailing social representations privileging Western scientific knowledge complicate achieving truly participatory research with equitable knowledge valuation.

Interviews revealed unanticipated gender dynamics potentially affecting community participation and researcher access. Despite the Ecohealth approach explicitly incorporating gender equity, team members insufficiently considered how their own gendered positionality might shape interactions in the field. As noted in the introduction, this gap reflects a broader pattern in Arctic research. A major limitation of this study is that it did not examine the discrimination experienced by women in greater depth. The results suggest that such dynamics are at play, sometimes overlapping with other factors (notably age). These dynamics are themselves situated within a specific context and influenced by the colonial past, resulting in a heterogeneity across communities whose full understanding goes beyond the scope of our study [27,28].

Priority of the dog issue

Although collaborators acknowledged the “dog issue” importance, it wasn't prioritised in their work or advocacy given the region's numerous challenges such as access to potable water [44], housing [45] or mental health [46], and limited resources to address them. While researchers recognised this for collaborators, they may have underestimated its impact on community participants. Despite incorporating mushers and connecting research with cultural activities like the Ivakkak sled-dog competition, these efforts may have insufficiently demonstrated the team's holistic view of dogs' role. According to respondents, certain activities deterred community participation and potentially reinforced perceptions of “southern researchers doing their research”.

Nevertheless, as demonstrated elsewhere with participatory research [47], the programme's existence created a discussion space for the issue. As one respondent noted, the programme served as a catalyst.

Strengths and future directions

Despite limitations, this evaluation provided valuable insights. Towards a collective team members representation of the initial programme theory, we successfully transformed the Neef and Neubert framework into a participative tool. Participatory research challenges all stakeholders by confronting diverse thought-styles [48]. In Indigenous contexts with predominantly settler-descendant teams (only one student had Indigenous ancestry), colonial history compounds these challenges [49]. Furthermore, insufficient familiarity with local culture, customs, and timelines on the part of some team members—particularly more junior researchers—likely contributed to the participation challenges encountered. This confirms the point we made earlier: early training, pre-departure workshops, and in-depth mentoring are absolutely essential in participatory and intercultural research projects. Looking back, this project has allowed us to re-examine our own practices. Admittedly, administrative and academic constraints (particularly the university academic calendar) are difficult to circumvent, but we must constantly strive to find innovative solutions that adhere to the best practices required when working in Indigenous communities.

The team recognised these issues, establishing from the outset that the programme would proceed only with community representatives' interest and participation. This commendable stance couldn't eliminate broader cultural contexts or geographic barriers, further exacerbated by the pandemic. This approach fostered trust relationships with collaborators, both Indigenous and non-Indigenous. However, the cultural gap between researchers and collaborators may have been less significant than with other community members who did not necessarily have as much experience working with a research team. The boundary-spanners vs gatekeeper roles of collaborators is clearly an issue that needs to be addressed early in the research process.

Ultimately, respondents perceived researchers' participation concerns as disproportionate, likely amplified by high expectations.

Beyond confirming findings already documented in the participatory research literature, this study makes three specific contributions. First, it demonstrates the value of realist evaluation as a tool for assessing participatory processes themselves—a methodological application that remains rare. Second, it documents the ambivalent intermediary role of organisational collaborators as simultaneously boundary-spanners and gatekeepers, a dynamic that is acknowledged in the literature but seldom examined empirically. Third, it shows that researchers' concerns about partners' satisfaction with participation were largely unfounded, a finding with practical implications for managing researcher expectations and self-doubt in participatory projects.

This realist evaluation explored underlying participation determinants between stakeholders in a human-canine health research project following an ecosystemic approach. It enhanced understanding of researcher and local stakeholder perspectives, refining perceptions of conditions for successful mutual participation. Though partial and context-dependent, these clarifications should benefit other researchers and local stakeholders in enhancing their collaborative project participation.

Appendix 1: Map with the location of participating communities

Funding Statement

This work is supported by the Canadian Institutes of Health Research (153250).

Disclosure statement

The Authors declare no conflict of interest.

AI assistance

During the preparation of this work, the authors used Claude AI to synthetise the document. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.