Inter-community participatory social network analysis: re-envisioning humanitarian accountability with climate-related displaced communities in Bangladesh Open Access

In Bangladesh, the internal displacement associated with impacts of hazardous weather events such as floods and cyclones account for millions of people each year (IDMC, 2024). Flash floods in 2022 and Cyclone Remal in 2024 are the recent instances displacing approximately 480,000 and 800,000, respectively, in the country. The lives of those displaced are characterised by a convergence of sufferings, including severe economic disruptions, limited access to public services, heightened physical and psychological insecurity, gender-based violence and unmet basic needs such as food, shelter and sanitation, among others (United Nations Bangladesh, 2024). In this acute situation, humanitarian actions are critical to reshaping lives of displaced people by alleviating sufferings through emergency aid and contribution to disaster preparedness and reconstruction of affected livelihoods.

However, the predominant humanitarian approach has been criticised for the marginalisation of displaced people in its governing process, particularly at operational level. This phenomenon is rooted in the national humanitarian system established in 1980s, which centralises control in the hands of a few elites (Faaland et al., 1981; Hartmann and Boyce, 1983). A participatory approach has been widely applied to fix this exclusive governing process; yet, it has resulted in “participatory exclusion,” keeping displaced people at the margin while empowering only local power holders into decision-making process (Nadiruzzaman and Wrathall, 2015). According to recent reports (see Aase, 2020; Practical Action, 2021), the exclusive nature of local humanitarian governance is still pervasive, and as a result, humanitarian action risks serving the interests of local power holders rather than those of disaster-affected people.

The rise of humanitarian accountability offers a promising paradigm to reform the local humanitarian governance arrangement by shifting displaced people from the margins to a position where they can hold humanitarian agencies accountable for their actions or lack thereof. The multiplicity of international humanitarian accountability initiatives and mechanisms compels a transformation of humanitarian system, aiming to invite people in need into the decision-making process and secure their space to raise their voice (e.g. Accountability to Affected People, Core Humanitarian Standard). Yet, their implementation frequently remains technocratic and tokenistic, failing to drive substantive change in the very exclusive system (Aijazi, 2022; Anstorp and Horst, 2021; Martin et al., 2021). Taking this disjunction into consideration, this study interrogates the practice of humanitarian accountability for climate-related displacement in Bangladesh.

In this inquiry, I pursued methodological reflexivity to construct contextual knowledge centring the perspectives of displaced communities. This approach is established on disaster scholarship criticising the implications of Western-centric knowledge for misreading disaster and mis-implementing collective actions on it (Chmutina et al., 2024; Gaillard, 2021). From this, this study contributes to reflexive disaster studies that can realise the standpoints of affected people in cultivating nuanced understanding and actions (e.g. Alburo-Cañete, 2021; Gurung and McGowran, 2022). This article presents an inter-community participatory social network analysis (PSNA), specially designed and conducted for this study, resulting in a re-envisioning of humanitarian accountability with climate-related displaced communities in Bangladesh.

Humanitarian action, primarily aimed at assisting those in need, is inherently a social process involving the interaction of various social actors in planning and implementation. Actors from government, non-governmental organisations (NGOs) and other sectors of civil society engage with those affected by disasters (Coppola, 2015). These actors collaborate and contest planning and implementing humanitarian actions, bringing different ideas and pursuing distinct interests in governing the collective efforts to help people in crisis. This dynamic interaction is encapsulated in the concept “humanitarian arena,” as proposed by Hilhorst and Jansen (2010).

Social accountability offers normative guidance on the interactions among those actors, denoting a relationship “in which an actor can be held to account by another and face consequences” (Papadopoulos, 2016, p. 205). This typically involves the accountability holder assessing the performance of the duty bearer, potentially leading to rewards or sanctions. This assumption captures vertical accountability which represents the relationship between two entities in the hierarchical position of principal and agent. It is widely debated with two sub-directional concepts (Lindberg, 2013): when the more powerful party (the patron) holds the less powerful party (the client) accountable, this is referred to as upward accountability; conversely, when the less powerful party holds the more powerful one accountable, it is termed downward accountability. In the context of Bangladesh, the downward dimension has attracted researchers’ attention to examine the relational dynamics between the state and civil society in their pursuits of transparent and responsible public service (Panday and Chowdhury, 2023; Shahan et al., 2021).

As the theory of social accountability expanded, more concepts denoting diverse relationship emerged. Horizontal dimension of it reflects the relationship between equal institutions holding each other accountable (Sakib, 2020). Diagonal accountability connotates the role of non-state actors, such as media and CSOs, in holding the government accountable (Lührmann et al., 2020). In contemporary discourse, accountability has become interconnected with other concepts, encompassing deliberative democracy (Mulgan, 2000), good governance (Bovens, 2010) or others. It has also developed into sub-concepts, with “NGO accountability” being one of the most widely applied in Bangladesh to reveal the diverse relationships NGOs have with donors, government and affected people (e.g. Mir and Bala, 2015; Parnini, 2006). Overall, these provide unrigid frameworks that capture the relational dimension of governance, subject to variations in concerned actors, their associations and pursued values.

The concept of accountability has garnered increasing attention within the humanitarian sector to seek “the responsible use of power in an unbalanced setting” between humanitarian agencies and affected populations become particularly evident (Turunen, 2020, p. 1). Their relationship is broadly abstracted as that of an aid provider and recipient (Damousi, 2022; Saez and Bryant, 2023). In essence, humanitarian accountability promotes that “crisis-affected people should be able to hold humanitarian organisations to account” (Owl RE, 2022, p. 8). Complaint boxes or 24/7 hotlines are the typical instances of humanitarian accountability practice led by international organisations. The underlying assumption within these practices is that inviting affected people to the arena can empower marginalised people to express their needs and feedbacks on received services (Humanitarian Practice Network, 2011; Owl RE, 2022). These efforts are directed to achieve downward accountability, challenging the tenets of marginalisation in humanitarian governance.

While it is ideally preferred, its implementation is not straightforward. Rubenstein (2007) aptly notes the challenges in an unequal world, stating “accountability of more powerful actors to less powerful actors is difficult to achieve” (p. 621). Theoretical limitation of humanitarian accountability lies in its narrowly defined focus on the responsibility of humanitarian agencies towards donors (upward accountability) rather than towards those affected (downward accountability) (Harris, 2023; van Zyl and Claeyé, 2019). Moreover, the current emphasis by humanitarian agencies on transparency and feedback mechanisms, though well-intentioned, often falls short of fostering the deeper accountability required for participatory programming and genuine responsibility-taking (Hilhorst, 2015; Hilhorst et al., 2021).

Considering this discussion, the present study treats humanitarian accountability as a fluid concept. Its meanings evolve and transform through the engagement of multiple social actors in knowledge processes. Varied theoretical backgrounds and conceptual foci broaden its meaning rather than constraining it to a fixed definition. If so, then in seeking and transforming its meaning, the selection of standpoint becomes a matter of practicing it where certain perspectives are primed by de-priming others. In this deliberate knowledge-power dynamic, plural epistemologies rooted in the lives of marginalised groups are often neglected in the real unequal world (Allen, 2017; Collyer et al., 2019; Haraway, 1988). As further elaborated, the application of PSNA in inter-community study embodied the emergence of contextual knowledge from the viewpoints of displaced community largely neglected in public domain.

Along with the forementioned theoretical stance, I appreciate all knowledge is contextually constructed within and through particular social, cultural, historical, spatial, environmental and other contexts. The social positionality is primarily concerned in this study as the context shaping the particularity in experience and interpretation on the research subject by climate-related displaced community. Haraway’s (1988) notion on “situated knowledge” inspired me: “The only way to find a larger vision is to be somewhere in particular” (p. 590). In this sense, climate-related displaced community, rather than focus on isolated individuals, is where the socio-historically embodied knowledge on humanitarian accountability is shaped and can be assessed.

For this study, I selected riparian and coastal communities where fieldwork was conducted in 2023, with one local research assistant. “Kalapara” Upazila in Patuakhali District is the one located in the southern coastal region, especially susceptible to floods and cyclones; another site is “Ulipur” Upazila in Kurigram District, often affected by floods and riverbank erosion. I used constructivist grounded theory as a methodology which guided an iterative process of data collection and analysis to construct knowledge grounded in empirical data. Its inductive nature aligned with my constructivist epistemological position, seeking “a process of describing voices hidden from public view” (Apramian et al., 2017, p. 359).

Following the tradition of methodology, I conducted in-person interviews with 35 participants including local leaders, government officers and NGO practitioners as well as villagers. The data collected were analysed through coding, a technique to identify patterned insights to develop codes, themes that construct emerging knowledge. In this case, social network emerged as a preliminary central theme. The network is represented by displaced communities as a fundamental social fabric through which they find meanings of everyday life (see also Masud-All-Kamal et al., 2021; Qayum, 2022). Furthermore, social network, as “the structural conditionality of social support” (Vonneilich, 2022, p. 31), offered a lens to develop nuanced understanding of humanitarian accountability in the given relational context. The fieldwork continued interview and coding to develop the identified key theme by articulating its multi-faceted aspects on humanitarian accountability.

One day in Kalapara, I paused and considered the methodological limitations in two dimensions. Firstly, I observed that the typical facilitation of one-and-one interviews was less effective in a local communication culture that is more rooted in group discussions or collective dialogue. Second, the collection of data formed divergent rather than convergent insights on the pattern of social network and accountability, making my role disproportionately prominent in distilling insights through coding. This was the moment methodological reflexive emerges from the field, drawing the notion of “rich contexts” by Olmos-Vega et al. (2023) and “methodological maze” by Raven (2006). This means that researchers are deliberately and constantly responsive to particularities of the research sites.

The central motivation of methodological reflexivity in this inquiry was to realise essential values in constructivist approach to constructing knowledge by encompassing plural epistemological positions (Charmaz, 2017) and positioning myself as a “reflexive participants” rather than “discoverer” of knowledge (Hernandez and Andrews, 2012). The process brought me to focus on incorporating PSNA as further illustrated in the follow.

Social network analysis is widely applied in social research to examine network dynamics which indicate structure of relationship and interaction between social actors (Li et al., 2021; Yang et al., 2017). PSNA advances this by developing a collective approach in which participants and researchers jointly identify key actors, map relational structure and interpret relationships by unpacking interactional aspects of networks. This participatory method can be selectively applied in quantitative, qualitative or mixed-methods research (Campbell et al., 2019; Längler et al., 2019; Reyhani and Grundmann, 2021). For this study, I adopted qualitatively driven PSNA to foster mutual learning process between four climate-related displaced communities from different locales, referred to here as inter-community PSNA.

It began with network mapping with a total 60 villagers across 4 selected villages. In each village, we invited villagers with experience of climate-related displacement, ensuring diversity in terms of gender and age. Additionally, neighbours who, although not directly displaced, had witnessed the displacement and subsequent humanitarian responses participated in it. For the mapping exercise, participants were asked to select a facilitator from their group. The facilitator, pen in hand, led the dialogue among participants to collaboratively draw a map. The type of knowledge I pursued in this practice is thus collective rather than individual.

Unlike in interviews, which tend to result in a reporting-style conversation between participants and me, the PSNA created a relaxed environment where participants freely exchanged opinions with the facilitator and each other. I observed that these interactions, often filled with laughter and gestures, resembled everyday conversations—an atmosphere conducive to eliciting authentic contextual knowledge grounded in local epistemology. At certain points, participants engaged in more serious discussions. For example, when one participant described a religious leader as an important figure in operating assistance for displaced villagers, others pushed back, arguing that religious leaders only offer support intermittently and primarily to those close to them. The photos in Figure 1 illustrate the implementation of the key steps in the inter-community PSNA.

Through these exchanges, qualitative insights from participants’ interpretations of their social networks enriched the nuanced understanding of humanitarian accountability. To support this process step by step, I provided guide questions, pre-designed on the basis of preliminary insights from interviews:

• (1)

  Who are the key actors involved in assisting displaced villagers?

• (2)

  How do these key actors interact to assist displaced villagers with informational, financial or other resources?

• (3)

  What are the key forms of assistance displaced villagers receiving from these actors? (e.g. life-saving measures, immediate alleviation of suffering or long-term livelihood support)

• (4)

  Who are the most important actors in assisting displaced villagers? Could you rate the importance from 0 to 3 stars?

In the second phase of my fieldwork, I revisited each village to facilitate participatory comparisons using previously drawn social network maps. I unfolded the four maps, one per village and research assistant invited participants to share reflections on their village’s networks in comparison to the other three. As Boeije (2002) noted, comparison is the most crucial intellectual tool in grounded theory, crystallising the development of knowledge. However, this inquiry resists researcher-dominated comparisons; instead, it recognised displaced people as those most capable of comparing their own social networks for development of knowledge. This approach enabled to reorient the research aim from generating contextual knowledge grounded in perspectives of participants towards generating with participants.

In adopting participatory comparison, as opposed to the typical PSNA focus on a single community, this study applied an “inter-community” approach. This allowed participants to engage in comparative analysis across villages, identifying and validating shared relational structures and distinct interactional patterns. Participants also drew valuable lessons from analysing other network types. A key insight was the recognition of unique social capital within their networks, a resource previous underappreciated. Additionally, observing distinct network patterns in other communities prompted participants to consider ways to reshape their own networks similarly. Synthesised and developed into contextual knowledge on humanitarian accountability are these key lessons from inter-community PSNA, presented in the following.

Before presenting research insights from inter-community PSNA, I will here briefly elaborate why I considered social network as the primary social situation in drawing nuanced understanding on humanitarian accountability. When participants illustrated their lives in climate-related displacement, they often highlighted interpersonal networking, social ties and connectedness as the most important value. According to one:

When he lived on the mainland, his neighbours collectively voiced their concerns to pressure village leaders into addressing flood-related issues. In the Char, however, residents tend to appeal individually to the leaders. He attributes this individualistic aid-seeking behaviour to the fact that Char inhabitants, displaced from different villages, are aware of the threat of further displacement due to the region’s vulnerability to erosion during the monsoon. This precarious situation discourages collaboration among the displaced, perpetuating their vulnerability to cyclical displacement with diminished networks with both neighbours and those in mainland. For him and other many, climate-related displacement means transformation of relationships and interactions with neighbours, leaders and others, left them “neglected.”

In contrast, displaced people observe the centralisation of village leaders in cases of climate-related displacement. Specifically, once displacement situations occur, village leaders—such as the elected Chairman and Members representing the community—participate in formal meetings with government and NGOs, in accordance with national policy guidelines. As alluded in introduction of this paper, village leaders are authorised to list beneficiaries, playing brokerage role between displaced people and humanitarian agencies. Their unchallenged power often leads to the misuse of public aid resources, marked by favouritism, bias and corruption. In a strongly sarcastic tone, a local researcher referred to these local power holders as “Dewany” (দেওয়ানি in Bengali)—a term rooted in the colonial era, denoting those who collect money (Interview-19).

I analysed similar narratives from participants and represented the studied area as a networked society. Climate-related displacement marginalises displaced individuals by weakening their ties and connectedness with other actors, whereas it has the opposite effect for village leaders. As humanitarian accountability is not practiced in the vacuum, understanding studied areas through social network angle served as a critical departure to deepen contextual knowledge on humanitarian accountability. How is humanitarian accountability practiced in a networked society? Inter-community PSNA drew two key insights which I conceptualise as co-opted downward humanitarian network and accountable humanitarian network, summarised in Figure 2 and discussed further in what follows.

Among the various forms of assistance, displaced people consistently identify emergency food supplies—often represented by 5 kg of rice—and occasional financial subsidies for rehabilitation as the most prevalent. In coastal region, the government’s Ashrayan Project was highlighted, which provides housing to selected landless and homeless families, as preferred significant support by displaced people. These activities typically operate through a targeting mechanism, wherein village leaders, acting as brokers, identify recipients for food, cash, housing, or other resources provided by local government and NGOs. A senior government official in Kalapara elaborated the importance of this intermediary role:

In the interviews, government officials and NGO workers asserted that this operational mechanism becomes accountable through routine practices to increase communication channel. Commonly cited practices include (1) holding meetings with villagers (2) posting beneficiary lists in public spaces (e.g. local markets), to allow villagers to request corrections for those improperly listed, and (3) providing complaint boxes and hotlines in the targeted villages. This approach assumes that enhancing communication channels empowers displaced people to voice concerns, enabling enforceability and sanctions to operationalise downward accountability.

In contrast, displaced people commonly observe the futility of such accountability practices, largely due to the nature of networked societies where sensitive information is difficult to conceal, creating a dynamic where a reversed direction of punishment is imposed. Namely, the right of displaced people to monitor and sanction humanitarian actors is not upheld but rather objectified and punished when they challenge local power order. The existing local power imbalance barriers downward accountability through a learned passivity rooted in fear of repercussions. As one conveyed: ‬‬‬‬‬‬‬‬

The abuse of power by local leaders is a well-known reality, yet it is treated as if it does not exist within the practice of downward accountability by formal humanitarian agencies. Based on the reasoning of local humanitarian actors, I understood this phenomenon as not merely the result of individual-level deviation but rather as linked to two institutional norms. First, local government and NGO workers are mandated to deliver “emergency” aid within a constrained timeframe, where reliance on village leaders is explained as a pragmatic approach to cover wider affected areas and large affected population. Second, these practitioners are pressured to commit to “localisation” in response to directives from headquarters or donors in Dhaka, prompting local practitioners to argue that their close connections with village leaders constitutes a form of localised aid.

These two institutional norms foster an environment for rhetorical and pseudo-downward accountability, expanding channels without enhancing genuine communication. Eventually, local humanitarian agencies co-opt this approach to downward accountability for upward accountability, focussing on executing aid resources within set timelines under the guise of localisation. This co-opted humanitarian accountability further marginalises displaced people in governing process.

Social networks in rural Bangladeshi society are multifaceted; while they can be instrumentalised in a way of undermining humanitarian accountability, they also play a crucial role in facilitating it. Focussing on its supportive role, this section presents the concept of an accountable humanitarian network, denoting a locally driven relational commitment to uphold their rights to receive indiscriminate care and nurturing their agency in governing process against marginalisation. Research participants articulated this concept through two distinct strategic patterns.‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬

The construction of dyadic relationships between displaced people and humanitarian agencies was identified as a straightforward strategy to remove the intermediary role of village leaders from the operational process. Research participants often highlighted the military as a symbolic actor in implementing transparent and responsive assistance across both riparian and coastal regions. Although the military provided the same basic food supplies and emergency kits as other humanitarian agencies, it was their direct, unmediated connection with the displaced villagers—bypassing village leaders—that established trust in the eyes of the people. This observation challenged my preconceived notion of the military as an oppressive force. One participant expressed this view succinctly:

In comparison to other NGOs, the Bangladesh Red Crescent Society (BDRCS, hereafter referred to as Red Crescent) was distinguishable as presenting the broker-free network with displaced people particularly in the coastal region. Most of all, the physical presence of volunteers of Red Crescent in the villages before and during cyclone-triggered emergency situations was highly valued; in contrast, government officials and other NGO practitioners often arrived weeks after a cyclone had already displaced people. Furthermore, the volunteers conduct door-to-door surveys to compile beneficiary lists without relying on village leaders, as aspired by the displaced people. While participants expressed dissatisfaction with the items (e.g. food, cash) and services (e.g. early warning, protection in cyclone shelter) provided by the Red Crescent, as they are intermittent and insufficient to address the economic poverty caused by displacement, these communities appreciated the direct and strong connections fostered by the volunteers, which address the relational aspects of climate-related displacement.

Similarly, research participants in four villages consistently identified neighbours as the most trustable actors in assisting displaced villagers, recognising their direct and constant support. One participant from a Char in Ulipur recounted his experience of requesting assistance from the local government to plant vegetation around the island as an adaptive measure to reduce the river’s flow velocity, which contributes to displacement during the monsoon. However, the government explicitly declined his request. Neighbours in another Char, who understood the challenges faced by displaced people, freely shared the plants so that he can plant them around his houses in the island. Other examples demonstrate that the support provided by neighbours—whether in the form of food, cash, shelter or psychological support through simply being present and sharing everyday conversations—is timely more than any assistances offered by external actors. One said that:

From the perspective of displaced people, humanitarian accountability is understood as a relational and emotional mechanism rather than narrowly a material or objective one. These broker-free networks reflect the desire of displaced people on accountability by upholding the rights to non-discrimination within the oppressive social and local humanitarian structure. Ensuring uncompromised assistance and nurturing existence and emotional well-being of displaced people are therefore critical to upholding humanitarian accountability in a networked society.

The active role of displaced people in operating humanitarian actions was examined through their creation of grassroots humanitarian networks in which three actors are involved; displaced people disseminate their situations and needs through their surrogates those who amplify information to civic humanitarians in remote areas and beyond the immediate social circles of displaced people. A case of creating grassroots humanitarian network in Ulipur illustrates how displaced people in Char exercise their relational power to actively construct the accountable humanitarian network:

In this narrative, the identity of displaced people emerges as agents rather than victims passively waiting aid or accountability driven by external actors; displaced people exercise their power to mobilise surrogates and civic humanitarians to fulfil their roles. In another example in another Char, local journalists played the role of surrogate while university students in other district and Dhaka appeared as civic humanitarians. Reflecting the precise understanding of journalist’s informational and relational power, research participants explained how they actively span the boundary of their social networks:

One notable lesson from the cited analysis above is that village leaders are potential contributor to humanitarian accountability by assisting displaced people in spanning their network boundary. Is this against the previous analysis that village leaders abuse their powers? To address this question, I had conversation with several village leaders and realised that these leaders’ brokerage role is not inherently given but has been institutionalised as humanitarian agencies have conditioned them to list limited number of beneficiaries with limitedly allocated aid resources for the village. Therefore, it is clearer to argue the problematic brokerage role of leaders in co-opted accountability is the issue of humanitarian norm but not simply personal immorality.

The cases of constructing grassroots network inspired research participants in coastal region whose social networks are more simplistic relying on the top-down assistance. Participants drew the lesson that diversifying accountable humanitarian networks even can be critically used to realise downward accountability:

This perspective of displaced people resonates their agency which can be realised through humanitarian network. This stands in contrast to the self-perception of powerlessness often observed in the narratives of co-opted accountability:

To summarise, boundary-spanning grassroots networks is the process of expanding accountable humanitarian relationship. It is initiated by affected communities, reshaping their identity from passive aid recipients to critical agents in governing humanitarian actions. By leveraging both existing and potential social networks, displaced people develop interactional strategies to build accountable humanitarian networks, engaging with their fellow surrogates, civic humanitarians and, potentially, formal local humanitarian agencies.

Clearly, humanitarian accountability is socially situated, and affected communities are uniquely positioned to understand and analyse its practice. This study embraced this notion by revising research design during fieldwork to centralise the perspectives of displaced people. The application of inter-community PSNA was not simply a data collection and analysis technique; it exemplified methodological reflexivity as a continuous inquiry into the research design, reflecting the researcher’s political and epistemological stance. Consequently, research sites in disaster studies become spaces where the relationship between researcher and participants is constituted through a reflexive process.

As a result, this article presents two contextual insights on humanitarian accountability. First, it critiques the conventional vision of humanitarian accountability by revealing that universally practiced downward accountability is subject to co-optation, paradoxically, exacerbating the marginalisation of displaced people. Second, it introduces the nuanced concept of accountable humanitarian network which illuminates alternative practices driven by local actors that prioritise the rights and agency of displaced people in operating humanitarian assistance. When asocial approach to humanitarian accountability remains prevalent in formal institution, these alternative practices create socially embedded forms of accountability grounded in grassroots social networks.

These findings prompt a critical question: from whose points of view do we understand and practice humanitarian accountability? This inquiry underscores that humanitarian accountability is not a philosophy but a political process through which knowledge and actions are shaped from particular standpoints among others. Reflexive disaster studies solidify this political dimension by centring local epistemology and resisting Western hegemony, thus offering the analytical lens to challenge structural issues in the status quo. Furthermore, this approach re-envisions humanitarian accountability by respecting and incorporating the role of displaced people, their surrogates and citizen humanitarians that pluralise practices and are essential to manoeuvre power imbalance that remains unchallenged.

This study was supported by Rawnak Jahan Khan Ranon from the International Centre for Climate Change and Development (ICCCAD), whose contributions were invaluable in navigating the research context. This study was conducted in association with the project “Humanitarian Governance: Accountability, Advocacy, Alternatives” supported by the European Research Council (ERC) Horizon 2020 programme (Grant number: 884139).