Learning and the ‘change enterprise’: inclusion and ambivalences in an educational action research and development project

ABSTRACT

Educational action research in development projects applies participative methods to include people with diverse backgrounds, experiencing learning collaboratively. In this article, I explore the learning activities, and their outcomes, unfolding in those settings. Based on project documents, 34 interviews with project staff and smallholder farmers, and participant observations for three months in Nepal, I analyse the action research and development project ‘Strengthening Adaptive Farming in Bangladesh, India and Nepal’ (SAF-BIN), and show how inclusive activities allow for critical reflection, individual experience, and dialogue. I find dialectics in this process of adjustment, such as resistant social and cultural arrangements that are activated. The inclusion of specific social groups, for example, goes along with exclusionary dynamics. This case illustrates that inclusion and action research can be conceptualised at least threefold: from a dialectical, institutional, and learning perspective. I conclude that action research in development provides a fruitful basis for learning, allows for legitimacy, and temporarily leads to social change at a micro-level, for better or worse.

KEYWORDS:

• Learning
• inclusion
• development
• farming
• global south
• institutions

Introduction

This study concerns educational action research (EAR) in the field of development cooperation. Organisations in this field have a long history of dealing with societal challenges under complex socio-cultural conditions. Complexity refers to the pressure of working effectively and, at the same time, the need to integrate different norms, values, expectations, and conventions, which typically meet and, at times, clash. Development endeavours have been criticised for not acting sustainably and imposing western values in a postcolonial manner (cf. Castro Varela and Dhawan 2015). As organisations have increasingly come under pressure to achieve legitimacy granted by both western and local audiences, development projects increasingly apply participative, open and inclusive forms of organising (e.g. Gegenhuber and Dobusch 2017) and learning (e.g. Rauch 2020). Resulting from an increased quest to adhere to the principle of sovereignty of ‘partner countries’ and ‘beneficiaries’, it is nowadays ‘good practice’ to organise learning in a participative and inclusive manner. Hence, action research projects experience tensions in navigating the local institutional context (Drori, Höllerer, and Walgenbach 2014) and integrating ‘globally shared cultural beliefs within an increasingly isomorphic world society’ (Goldenstein and Walgenbach 2019, 289).

Against this backdrop, I explore learning in action research in development, examining activities and elements of learning, the role of inclusion as a core quality of action research, related outcomes and possible ambivalences. The theoretical purpose of exploring this question in this complex setting is to gain a deeper understanding of the manifold ways in which inclusion and learning, particularly in EAR, can be conceptualised. Institutional literature (e.g. Goldenstein and Walgenbach 2019), and learning literature (Mezirow and Taylor 2009) can help us to understand what is going on in those projects. More practically, this study helps to identify conditions for learning and inclusion in action research.

Empirically, I explore the action research and development project ‘Strengthening Adaptive Farming in Bangladesh, India and Nepal’ (SAF-BIN). The data comprise project documents, 34 semi-structured interviews with project staff and participating smallholder farmers, and a participant observation of project activities for three months in Nepal. Methodologically, I did not apply action research, but rather the EAR SAF-BIN was the object of analysis. Within SAF-BIN, the NPO Caritas collaborates with local partners from civil society, policymakers, research institutes, and thousands of smallholder farmers, who are organised in groups called smallholder farmers collectives (SHFC), aiming at building resilience towards climate change through agricultural innovations. I show that inclusive and regular project activities facilitate elements of learning, such as individual experience, dialogue and critical reflection, leading to instrumental and communicative learning outcomes in the sense of Mezirow and Taylor (2009). I find dialectics in this process of adjustment, such as resistant social and cultural arrangements that are activated, as the inclusion of specific social groups (e.g. disadvantaged ethnic groups, people from lower castes, widows) goes along with exclusion.

The core theoretical insights of exploring learning in this setting is that inclusion and EAR can be conceptualised and understood in at least three ways. First, from a dialectical perspective – that is: with inclusion comes exclusion. Second, from an institutional perspective – inclusion, as an institutional rule and modern way of organising, inevitably involves clashes. Inclusion here serves to justify the organisations’ practices, and at the same time, must be justified to diverse audiences to achieve their acceptance. Third, from a learning perspective, inclusion and EAR provides a fruitful basis for learning and social change. I conclude that EAR allows for organisational legitimacy and survival, and leads to learning, adoption, and social change at a micro-level.

The paper is structured as follows: in the literature review, I first depict the basic premises of EAR, and its specifics in the development context. I then start, by looking at relevant literature, conceptualising inclusion as a set of practices, a powerful institution, and a core quality of action research. Based on that, I outline the methodological approach in the methodology section, in which I depict the case, research site, and the data work in detail. After presenting the findings, focusing on learning activities, outcomes and tensions related to inclusion, I discuss how the findings relate to the literature, defining three main perspectives on EAR and inclusion.

EAR in development

Action research as an umbrella term for a range of approaches abandoning the supposed opposition between research and practice, follows a normative agenda, based on a constructivist ontology, aiming at social change by promoting reflective practice, action inquiry and participation (cf. Rauch 2020).¹ Accordingly, action research has a long history in development cooperation, as organisations in those fields aim to achieve change in socio-cultural contexts that typically vary from those of the organisations’ origin. It can be challenging to balance the adherence to western quests, values, and the standards of the ‘donor’ countries on the one hand, and to respond to local expectations on the other. Learning activities, their qualities and outcomes are well observable as those organisations are pressured to work effectively, while tensions tend to unfold due to complex conditions and conflicting logics.

Learning in development shifted in the last century from approaches focusing on the transfer of knowledge and technology (Percy 1999) in the second half of the 20th century, targeting individuals rather than collectives as the catalysts for social change, to more participative learning approaches (Duveskog, Friis-Hansen, and Taylor 2011; Najaar, Spaling, and Sinclair 2013). Approaches such as training and visits (T&V) extension systems (Percy 1999), conducted by extension agents acting as providers of knowledge, passing on technologies to rather passively framed recipients, did not bring the expected benefits, such as technologies spreading to others, as assumed by representatives of the diffusion of innovation theory (Rogers 1995). Rather, those approaches were found to increase inequalities (Chambers 1997). Transfer of technology (ToT) approaches were found unsuitable for agricultural systems due to local specificities and the complexity of farming (Berg and Jiggins 2007). Linked with critique of the framing of development as a linear process of progressive transition assumed by representatives of modernisation theory, the focus shifted to more participatory practices (Duveskog, Friis-Hansen, and Taylor 2011; Najaar, Spaling, and Sinclair 2013).

Action research as a set of participative practices aims at fostering collaborative planning, implementing, evaluating, decision-making, and shifting bargaining power and control over resources to the participants. Action research in development is typically conceptualised as a collective and democratic process, aiming at the improvement of peoples’ life (cf. Gustavsen 2014) rather than on theory development. The participants often collaborate with ‘genuine’ researchers and take on the role of researchers themselves, incorporating local knowledge and indigenous practices (Gustavsen 2014; Kamali 2007). By formal and informal learning, the participants’ empowerment (Duveskog, Friis-Hansen, and Taylor 2011), as well as the acceptance and adoption of practices should be achieved. In this context, the ‘extension agents’ typically act as moderators and facilitators of learning rather than as experts, with farmers seen as active participants rather than simply as recipients or adopters.

In the field of farming, critical thinking should foster farmers’ self-reliance in responding to exogenous challenges (Berg and Jiggins 2007). Thus, action research in development typically also seeks to drive social transformation, through changing skills and values, expressed in changed action and thinking patterns at an individual and group level. Change can also refer to the attitudes of rural extension workers and bureaucrats towards a more participatory decision-making process (Kamali 2007). As a form of action research in agriculture, Farmer Field Schools (FFS) are non-formal, community-based institutional platforms and approaches to adult learning, developed by farmers, where farmers study locally issues (Duveskog, Friis-Hansen, and Taylor 2012). First implemented in Indonesia as a way to deal with widespread pest outbreaks in rice using integrated pest management, they have been shown to support agriculture and rural development and gender empowerment (Westendorp 2012). However, in the past, FFS in Nepal were also found to exclude people with a low social status, such as ‘the majority of the poor, the untouchables or Dalits and Janajatis, thereby increasing the social divide between the groups (Westendorp 2012, 5).

Interacting with the environment and questioning established structures is perceived as key to learning (John-Steiner and Mahn 1996; Wu and Wu 2015). Within the development sector, interactions are often facilitated through the provision of resources and inputs e.g. in the form of exposure visits. Najaar, Spaling, and Sinclair (2013) explore learning in a participatory development project in Kenya, showing that learning was achieved by practical application and training of knowledge, observations and dialogue. In a different context, McGee (2008) identifies ambivalences in action research used for professional development in a Middle Eastern Gulf State, showing the loss of democratic values underpinning action research, highlighting power relations, control, and ownership of research.

Inclusion as a set of practices and powerful institution

From the perspective of new institutional theory, inclusion and educational action can be conceptualised as powerful institutions and rationalised myths that serve organisations to adopt structures that mirror the cultural beliefs of the broader society to ensure organisational legitimacy and survival (cf. Gegenhuber and Dobusch 2017; Goldenstein and Walgenbach 2019; Meyer 2010; Meyer and Bromley 2013). Institutions hereby are understood as ‘a system of established and embedded social rules that structure social interaction’ (Hodgson 2006, 18), including for example, language, money, law, manners, organisations. (Development) organisations can be conceptualised as single actors, agents, players that (re)produce institutions and depend on them, or as a special type of institutions (Hodgson 2006). Social change is driven by global norms towards which modern organisations are oriented, as they act as key players of social transformation (Meyer, Drori, and Hwang 2006). The embeddedness of (development) organisations and projects in a globalised world fosters rationalised models of organising (Djelic and Quack 2003), incorporating evidence-driven modes of decision-making. At the same time, organisations are constantly confronted by the specificities of local contexts (Drori, Höllerer, and Walgenbach 2014; Goldenstein and Walgenbach 2019), leading to clashes arising from the co-constitution of the global and local. In this sense global isomorphism is going along with local variations (ibid.). From a relational perspective, inclusion and exclusion are a mutually dependent pair (cf. Dobusch 2014), held in place by a set of practices (Reckwitz 2002). This allows to consider practices of exclusion to ‘achieve more convergence of openness ideals and openness in practice’ (Dobusch and Dobusch 2019, 14), to ask for the legitimate combination of openness and closure, and to address how openness might import, reproduce or even reinforce social inequalities (Ahmed 2012; Dobusch, Dobusch, and Müller-Seitz 2019). Within this arena, literature on EAR can benefit by applying heterogeneous theoretical perspectives on action research, as this study illustrates.

Materials and methods

Case selection and research site

The action research and development project in focus is an exceptional case (Yin 2013) because it has a wide reach, affecting, according to the SAF-BIN Website, about 40,000 people in 165 villages across 21 districts in Bangladesh, India, Nepal, and Pakistan as of 7 July 2020 (See SAF-BIN website at: http://www.safbin.org/project-overview#click). It involves multiple qualities of action research in development, comprising inclusion, transparency, participation, and addresses multidimensional challenges (see also George et al. 2016) by operating across sectors and nation-states. Within SAF-BIN, the NGO Caritas cooperates with the University of Natural Resources and Life Sciences, Vienna, and other partners,² including research institutes and policymakers, and has done so since 2011.

The inclusion of marginalised groups has become a ‘cross-cutting’ issue (Ottacher and Vogel 2015) in development, playing a role in projects beyond those explicitly focusing on the reduction of inequalities. SAF-BIN primarily aims at promoting food security and resilience through agricultural innovation in smallholder farms, but also fosters ‘inclusive and responsive policy’ and pursues these goals ‘with a focus on gender, cultural identity, and equity to achieve a maximum benefit for marginalised smallholder farmers and their communities.’ (SAF-BIN, n.d). SAF-BIN is funded by the European Union Global Program on Agricultural Research for Development (ARD). The headquarters of development projects are located in donor countries, while operationalising units are in the global south. Projects are expected to incorporate western development policies, and, speaking in managerial terms, addressing these policies by including ‘the marginalised’, for example, is considered a ‘critical success factor’ (see also Ottacher and Vogel 2015).

The geographical focus of this paper is Nepal, a diverse country in agro-ecological and socio-economic terms. It is divided into three agro-ecological zones in which the climate varies from subtropical to alpine within small distances. People from different casts, ethnic groups, religions, and genders are part of the complex, hierarchically structured society. Smallholder farms (with less than 2 hectares of land) are the basis of the food supply for the growing population, the majority of whom live in rural areas. Agriculture is subsistence-oriented and farms have limited access to external resources and increasingly suffer from the impacts of climate change. In the project regions, farmers face problems such as drought, irregular rainfall, an increase in temperature, and extreme climatic events. Agriculture in Nepal is further characterised by feminization due to men migrating because of their fragile economic situation and political conflicts (Westendorp 2012).

SAF-BIN’s core activity is to identify, select, and further develop local agricultural innovation, such as crop diversification and integrated resource management, through action research. In the case of the SHFC, this was achieved by conducting field trials collaboratively. Trial plots were established to perform on-farm trials, with farmers using part of their farmland for the establishment of these plots. The SHFC established and monitored the plots, collected data, and analysed the results within the groups (on-farm adaptive research) in collaboration with the project staff (http://www.safbin.org/project-overview#click). Thereby, ‘candidate models’ (selected innovations) are further tested by the farmers in cooperation with project staff and researchers for their efficiency, sustainability, and acceptability. The models were then collaboratively refined, adapted, and evaluated annually.

Data collection

I collected the data between 2014 and 2017. First, I drew on project documents comprising a baseline study, including presentation slides, project concepts, reports, the SAF-BIN website, and secondary data, such as news coverage. To get familiar with the project, I studied these documents and discussed them with a project coordinator in Vienna. These talks were the basis for getting an understanding of the project structure and networks.

Second, I observed the project activities from March to May 2014 in Nepal. I participated in the group meetings of the SHFC, in the field trials conducted by the SHFC and the project staff, and in the everyday life of the farmers. I studied and compared six SHFC, each consisting of 15 farmers and located in two districts in Nepal (Bardiya District and Kaski District). I observed the on-farm adaptive research trials, mostly in the form of a tour guided by the Village Research Assistants (VRAs) employed in the project, and farmers holding a leadership role. I also observed the agricultural activities, the everyday life of the farmers and the group meetings of the project staff, including the District Project Officer (DPOs) and VRAs in an informal way. Additionally, I participated in a SAF-BIN conference in 2015 in New Delhi, where farmers and other stakeholders, including myself, reflected on project activities. I documented the observations by writing a research diary, taking pictures, and recording conversations.

Third, I conducted 34 semi-structured interviews with project staff and smallholder farmers between 2014 and 2017 in Nepal and in Austria. All but one of the interviews were conducted face-to-face interviews: 32 interviews were held in Nepal with project staff (#10) and farmers (#22) from six SHFC either in the capital city, Kathmandu, or in the districts of Bardiya and Kaski. The project staff provided access to interview partners, and I selected the study sites in agreement with Caritas Nepal. I interviewed three to five farmers from every SHFC. The interviewed project staff comprised VRAs from the study villages, DPOs, project coordinators in Nepal and Vienna, and one evaluator, hired by Caritas Nepal to conduct a mid-term evaluation (cf. Table 1).

Table 1. Case study database: Overview of the interview partners by role and year.

Year	Interview partner (role at the time of the interview)	Region and Country	Number Interviews #
2014	Farmers of SHFC 1	Bardiya/Nepal	5
2014	Farmers from SHFC 2 in Bardiya	Bardiya/Nepal	4
2014	Farmers from SHFC 3 in Bardiya	Bardiya/Nepal	4
2014	Farmers from SHFC 4 in Kaski	Kaski/Nepal	3
2014	Farmers from SHFC 5 in Kaski	Kaski/Nepal	3
2014	Farmers from SHFC 6 in Kaski	Kaski/Nepal	3
2014	Village Research Assistants (VRA)	Nepal	3
2014	District project officers (DPO)	Bardiya and Kaski/Nepal	2
2014	Project coordinators Nepal		3
2014	Project coordinator Vienna (Caritas)	Vienna/Austria	1
2014	Evaluator	Kathmandu/Nepal	1
2017	Project coordinator Vienna (Caritas)	Vienna/Austria	1
2017	Project coordinator Vienna (BOKU)	Skype	1
Total			34

Summary of used data

Interviews with project staff

Interviews with a project evaluator

Interviews with smallholder farmers

Ethnography from March-May 2014, Bardia, Kaski and Kathmandu, Nepal

Participant observation at a SAF-BIN conference 2015, New Delhi, India

Participant observation at SAF-BIN events 2016, Vienna, Austria

Project Documents

The interview partners had diverse socio-economic backgrounds in terms of gender, ethnic group, cast, and official roles. The farmers had a low levels of formal education, were mostly married, and aged between 20 and 65. An overview of some of the socio-economic characteristics of the interview partners is provided in Table 2.

Table 2. Some socio-economic factors of interviewed smallholder farmers.

Socio-economic factors	Kalika Interviewees		Dhikurpokhari Interviewees
	Female n = 8	Male n = 5	Female n = 6	Male n = 3
Educational Level Uneducated 1–4 years 5–9 years School Leaving Certificate	3 1 4 -	- 1 1 3	- 2 4 -	1 1 1 -
Marital status Married Unmarried Widowed	6 - 2	4 1 -	6 - -	3 - -
Age Average Range	31 24–59	31 20–51	37 27–48	57 51–65

I choose the interview partners in those cases in which the observation had been conducted before the interviews. This was possible for all groups in Bardiya. However, in Kaski, I observed only two of the three group meetings, so the VRAs selected the interview partners. In 2017, the data was complemented with two interviews with project staff from Austria. Narrations rotated with dialogue sections with ad-hoc-questions. I pretested the interviews with the Nepalese translator, who translated from Nepali to English as required. The participants gave their consent to the use of their data (sometimes using fingerprints as their signature). The interviews with the farmers were conducted in Nepali, while the interviews with the Nepalese project staff were in English, and the interviews with the Austrian project staff in German. The interviews with the farmers took place in the fields or at the respondents’ home, whereas the interviews with the project staff were conducted at their workplace or, at their request, in a bar, always taking care to ensure the necessary privacy and tranquillity. The interview duration varied from 20 to 120 minutes, depending mainly on the availability of the interview partners. The first interviews lasted longer as they were conducted in quite an open manner, leaving room for all kinds of topics; however, as the research progressed, the focus in the interview settings was narrowed down.

At the time I was a research fellow at the BOKU University and had an interest in the organisation of development projects. My role as a researcher exploring social phenomena in farming projects granted me a kind of exclusive status in the communities, with extra tours of the fields and households provided for me. The people met me with extraordinary hospitality, interest in the research, and openness towards sharing their experiences with me. The interviews were conducted with no monetary incentives for the participants, which is relevant to ensure the validity of the data. Hence, the interview partners provide their limited time resources voluntarily. However, methodological limitations arose concerning the translation of the data between the three languages, which can be accompanied by loss of information, misunderstandings, and selective bias. While the first non-professional translator was provided by the project and had a background in agriculture, the second non-professional translator had just studied English. It is possible, therefore, that the first translator put a focus on the farming activities when translating due to his professional background.

Analytical framework: transformative learning

Transformative learning relates to adult learning and means learners transforming or expanding their presumptions (Mezirow and Taylor 2009). It can be purposefully attained, or result from a transformative experience, which is sometimes described as a ‘disorienting dilemma’ e.g. a personal life crisis (Mezirow and Taylor 2009). The authors differ between ‘instrumental’ and ‘communicative learning’, which serves analytical purposes here as these dimensions feed into another (cf. Najaar, Spaling, and Sinclair 2013). While ‘instrumental learning’ is about learning to control, organise, or change one’s environment and about how to achieve desired ends, ‘communicative learning’ involves abstract conceptualisation, and negotiating meaning, values and ideals (Mezirow and Taylor 2009; Najaar, Spaling, and Sinclair 2013). As conditions for transformative learning, Mezirow (2000) mentions the availability of accurate and complete information, safety and freedom from coercion, openness to other points of view, and equal opportunities to participate. Those conditions help farmers not only to find and adopt new practices and technologies, but also to question and break oppressive relations (Duveskog, Friis-Hansen, and Taylor 2011). The facilitators thereby act as mediators, enabling learning conditions and stimulating individual experiences (ibid.).

Thus, an effective learning process should facilitate three interrelated elements of transformative learning (Mezirow and Taylor 2009): dialogue, reflection and individual experience. While individual experiences consist of the learners’ prior experiences and those within the learning setting, dialogue relates to the medium through which transformation is enabled (face-to-face) and rational discourse and argumentation, and critical reflection relates to reflection on content, process and premise and comprises questioning of oppressive structures and ways of overcoming those structures (ibid.). These normative elements, as well as the participatory and process-oriented character, are core in action research in development (Duveskog, Friis-Hansen, and Taylor 2011).

Data analysis

I analysed the data based on an iterative interplay in a content-analytical manner, going back and forth between emerging theoretical concepts and data (Corbin and Strauss 1990; Lueger 2010), using the software Atlas.ti. I identified the main activities, their qualities and main characteristics, and linked them with the elements of transformative learning (critical reflection, dialogue and individual experience), instrumental and communicative learning outcomes. In the next analytical step, I reconstructed from the narrations and observations ambivalences, tensions, discrepancies (See Table 3).

Table 3. Analytical framework.

Project activities and elements of transformative learning	Learning outcomes	Ambivalences/tensions/discrepancies
Code 1: Activities: Which activities are conducted in the project? Code 2: Elements of transformative learning: In which ways do the activities allow for critical reflection, dialogue and individual experience?	Code 3: Instrumental learning outcomes: What kind of skills, information and techniques do the concerned learn related to climate change and farming? Code 4: Communicative learning outcomes: What kind of management, analytical and communication skills and concepts do they learn? In which way are abilities fostered related to abstract concepts and negotiate meaning, values and ideals?	Code 5: What kind of impacts result from the learning outcomes? In which way do the concerned experiences ambivalences?

The content analysis allowed structuring issues that appeared during the data collection process. Each interview was categorised using codes of the pre-defined analytical framework (See Table 4).

Table 4. Codes and sub-codes.

CODES	Sub-code	Explanation
Project activities	Activities conducted in the two regions	Which activities were conducted in Kaski and Bardiya by SAF-BIN?
Transformative learning	Critical Reflection, dialogue and individual experience	In which ways do the activities allow critical reflection, dialogue, and individual experience?
Instrumental learning	Climate Change	Which skills, information and techniques regarding climate change are learned by smallholder farmers?
	Cultivation practices & inputs	Which skills, information and techniques regarding cultivation practices and inputs are learned by smallholder farmers?
	Human Health	Which skills, information and techniques regarding human health are learned by smallholder farmers?
	Farming as business	In which ways is farming perceived as business?
Communicative learning	Management skills	In which ways have management skills been achieved? Management skills refer here to f.e. planning, (re)organising, forming, administering.
	Communication	In which ways have communication skills been achieved?
	Analytical capacities	In which ways have analytical skills been achieved?
	Concepts	In which ways new concepts have been learned?
Ambivalences/tensions/discrepancies	Related to social structure and social change	What kind of roles do SHF with different backgrounds hold and how does it change? How does the roles differ in regard to gender, disadvantaged ethical groups and casts?

Source: Author’s Own data.

Results

As the data showed, elements of transformative learning unfolded in a range of project activities in both study sites, Kalika and Dhikurpokhari, occurring in a cyclical manner (cf. Table 5).

Table 5. Project activities in the study sites.

PROJECT ACTIVITIES	Dhikurpokhari Kalika
Baseline study	X	X
Multi-stakeholder forums	X	X
Using resource centres	-	X
Group meetings	X	X
On-farm adaptive research	X	X

At the early stages, the project staff conducted a ‘baseline survey’ to assess smallholder households and farmers’ situations in cooperation with external researchers from private and public research institutes. Collaboratively, they collected qualitative and quantitative data to draw a picture of the socio-economic structure and the agricultural (land holding and land use patterns, cropping patterns, food handling and livestock) and climatic challenges, and made the results publicly available. The survey included qualitative methods such as ‘participatory rural appraisal’, an approach building on Paulo Freire’s argument that rural and local knowledge, including local farmers, needs to be integrated into rural development processes. The learning outcomes of those studies were of relevance for the project staff and the projects’ further conceptualisation. Farmers, who participated in this process, reflected on their current lives and recent challenges related to farming and climate change that had emerged in the last years.

SAF-BIN also organised ‘multi-stakeholder forums’ which served to identify and involve key players in the project in order to include different decision-makers in the process. The District Development Committee,³ project staff, researchers, and the local farmers discussed which geographical districts to select for project implementation. These meetings and the baseline survey provided the basis for decisions related to the inclusion of SHFC with different socio-economic backgrounds from four districts with diverse agro-ecological conditions, ranging from alpine to tropical to overall pursuing diversity of cropping systems. Additionally, the project staff ensured that women, as a marginalised group, were able to participate:

We have addressed the gender aspects […] because women are the main aware, women are the persons who have been exploited, held back (…) so we have given priority to women joining the group so that we have 72% (Interview Project Staff)

As the project staff explained, these approaches enabled participants to learn collectively and to make ‘evidence-based’ and rational decisions related to which problems should be tackled. Including the District Development Committee in particular was meant to ensure political acceptance and strengthen the prioritisation of environmental and farming issues.

Resource centres were established at the district level, providing demonstrative information concerning climate change and its consequences for farming. The learning material comprised brochures and picture books for younger people and adults who were illiterate. The project staff typically had access to this material and forwarded it to the groups or opened up the centres, enabling the distribution of seeds as crucial inputs. As the interview partners perceived it, these settings serve not only as administrative units, as spaces to collect information and store inputs, but also as areas for social exchange, and hence, foster learning, and a sense of togetherness and belonging. One of the resource centres further served as accommodation for me during the stay, indicating the centres’ openness and its role as potential meeting points for various audiences.

Dialogue facilitated through formalised SHFC meetings

Within the farmer groups, dialogue is particularly facilitated through regular (weekly in Kalika and monthly in Dhikurpokhari), structured and formalised meetings. These are held in easily accessible public or private locations in the village centres, such as in front of a shop or the house of a participant, varying in duration depending on the growing season, project phase and agenda (See Figure 1). The meetings’ agenda, suggested by the group leader, the so-called ‘president’, mostly relates to the field trials, cultivation practices and inputs, climate change and the organisation of the farming system, but also to issues such as human health or cooking.

Figure 1. Farmers presenting and interpreting data in Kalika.

Source: Taken by the author

The participants assign specific ‘functions’ such as ‘president’, ‘secretary’ and ‘treasurer’, and post holders are elected by the group members. Mostly, the ‘president’ is in charge of the meetings’ opening and the moderation. The moderator tries, as he describes, to let everyone have their say and motivates rather shy participants to engage in the sessions to facilitate learning. People with a lower status tend to fulfil the role of ‘secretary’ and ‘treasurer’ while the responsible job of the ‘president’ is typically filled by an older, ‘experienced’ or ‘wise’ person, who is in charge of suggesting solutions for upcoming problems. The opinions of those older, more experienced participants is considered especially enriching by the farmers, due to their long-time experiences in farming and, possibly, also in prior action research and development projects.

The VRAs also regularly participate in the meetings, acting as gatekeepers, enabling dialogue between the SHFC and other stakeholders, including scientists. They work in close contact with the farmers, speak both English and Nepali, and span the boundaries between the farmers and other stakeholders. VRAs therefore act as gatekeepers, facilitators and mediators and have a crucial role in managing the interface between audiences, such as between scientists and farmers, for example, in facilitating the field trials. Learning therefore depends very much on the VRAs qualities and practices. According to the interview partners, SAF-BIN intentionally employed mediators from diverse social and educational backgrounds to capture a wide variety of expertise regarding cropping systems, and VRAs, working with farmers, other project staff at the district and national level, and other stakeholders, are crucial facilitators of learning.

Individual experience and critical reflection unfolding in on-farm adaptive research

Individual experience was particularly facilitated through on-farm adaptive research, formally divided into different stages: planning, installing, monitoring, nurturing, collecting and analysing data. The setup of the field trials was designed based on previous activities, such as the baseline survey, multi-stakeholder forums and group meetings, by Caritas Nepal and the farmers in consultation with Nepal Agricultural Research Council (NARC), providing technical inputs and assistance. The SHFC focused on different trials depending on the requirements and climatic conditions in the sites. For instance, in Kalika, a region that is prone to drought, drought-tolerant seed varieties were tested and compared with local seeds within ‘varietal wheat trials’ by the farmers, the project staff and scientists (See Figure 2). A new seed variety called Gautam, released by NARC and distributed by Caritas Nepal, formed part of the trials.

Figure 2. Farmers presenting the wheat trials (‘local varieties’) in Kalika.

Source: Taken by the author

After the setup of trial plots, the farmers in collaboration with the VRAs, were in charge of monitoring, nurturing and evaluating the plots. In the ‘varietal wheat trial’, yield height, plant height and insect-based problems were analysed. In the group meetings, the data was presented via charts or presentations and collectively discussed. Gautam was found to be more drought-tolerant, facilitating cultivation despite increasing water scarcity. In comparison to local varieties, the ‘improved practice’ resulted in higher yields, higher plants, more spikes, less insect-based problems and easier manual weeding. Decisions on adaptations of the plots were made after analysis of the data and partly after consultation with project staff. Other groups in Kalika tested different mulching techniques, such as plastic mulches (thin vs. thick) and mustard and straw mulches (See Figure 3). Again, the farmers discussed the results of the trials and together reflected on the next steps. The activities were conducted in a cyclical and iterative manner due to the recursive nature of farming practices, but also due to the conception of the trials. Between the cycles of the trials, time and space for reflection (e.g. in the group meetings) allowed for the refinement and further development of the trials.

Figure 3. Farmers testing mulching techniques.

Source: SAF-BIN

Other SHFC in Kalika systematically compared the farming practices of broadcasting and line sowing of wheat seeds, analysing the impact of soaking the seeds or conducting irrigation trials (drip irrigation and plastic mulches) in the drought areas. In Dhikurpokhari, which is located in a more humid region, the focus was on vegetable trials. Tomato trials were conducted by two SHFC in the Dhikurpokhari, with tomato tunnels built by the farmers on easily accessible land belonging to a group member. Due to the regulation of humidity and temperature, an increase in tomato production, a reduction in diseases, and an improvement in fruit quality were expected by the farmers (See Figure 4).

Figure 4. A farmer in her tomato tunnel in Dhikurpokhari.

Source: SAF-BIN

In on-farm adaptive research processes, the farmers interpreted the data and reflected on the trials, including their suitability and sustainability. Some questioned the uncertain input supply (especially seeds) in the ‘varietal trials’ after the project. Inputs like drought-tolerant seeds and irrigation systems could not be afforded in the past, and it remains uncertain if the new practices can be maintained after the programme. In this sense, the farmers questioned project activities critically, reflected on their potential futures, as well as on their gender roles, as inequalities in action research related decision-making become apparent e.g. men make the decisions, even if they live and work abroad.

Tensions and ambivalences emerging through social heterogeneity

A project centrepiece was the formation of the heterogeneous SHFC. In the composition of all 28 SHFCs in Nepal, the stakeholders made particular efforts to include representatives of marginalised groups. People from lower castes, different ethnic groups, and female-headed households were intentionally included in the groups, after having problematised social inequalities in the conception phase. However, some farmers from other socio-economic groups expressed serious doubts about working collaboratively with farmers of a different socio-economic status. In some cases, farmers holding a higher social status had not interacted with members of the ethnic group of the Dalits, meaning the ‘broken/scattered’, before, as they have been kept repressed within the hierarchical social system. These doubts were mostly finally smoothed out by the project staff. SAF-BIN not only provided fertilisers and seeds, but also childcare, ‘exposure tours’, and celebrations for participants, supporting the interaction of farmers with different backgrounds. Over time, the heterogeneous participants collaborated intensively on agricultural innovations, and the otherwise deeply grounded idea of class separation became increasingly blurred within the groups – at least in the context of the project. In this sense, the project supported the recombination of norms and social transformation within the groups and the villages within the context of the project. Learning in this sense refers not only to a change in skills, but penetrates deeper, affecting and disrupting deeply rooted norms and worldviews of the participants. However, it remains to be seen how long lasting these changes and effects have been for those involved.

While some farmers initially expressed doubts or explicitly did not want to participate, others were not able to. When asked why her husband did not participate in the project, one farmer answered, ’we people are dependent on one person, the head person, for feeding us. I also depend upon the income of the husband. If the husband participates, who will work?’ and further explained that her husband was doing construction work in India. Most of the women interviewed explained that their husbands work in low paid jobs, such as trekking guides, or in the army, or depicted them as not being interested in farming and preferring to ‘play cards and have drinks in their free time’, which resulted in disputes between participating and non-participating family members.

Another centrepiece of the early project implementation was the collaborative selection of agricultural innovations that were further tested by the SHFC on trial plots. Farmers initially expressed doubts about participating, as they feared potential disadvantages due to unfamiliar practices, as the following examples illustrate. Farmers regularly measured the increase of yields in their trial plots, kept notes to identify the production increase, and commonly discussed and interpreted the findings. However, these rationalised endeavours did not find acceptance on the side of non-participating farmers.

The [non-participating] farmers they were teasing other farmers, like, for example, ‘Ah, is this the way you are doing research kind of things’. Like for example, we are cultivating rice in a different kind of way. In a different way, you know, not broadcasting, we are doing it in a line. Line sewing. (…) With chemical fertilisers and more farm manuals like that. The [non-participating] farmers were teasing other [participating] farmers who were there in our project. (Farmer Interview)

In this example, new practices induced tensions between the participating and non-participating farmers, and, when confronted with rationalised farming techniques, participating farmers again expressed doubts about the measures. However, the same interviewee later explained that, over the course of the project, these tensions were resolved:

But after seeing the results of the first year, more and more farmers wanted to be part of the project. But we do not have this, you know, privilege to include all the farmers. We can share the information with other farmers who are not there in the project. (Farmer Interview)

This quote illustrates the project finally gained acceptance because the sceptical farmers saw increased economic output. The legitimacy of exclusionary effects was achieved by ensuring transparency and sharing of information as a management practice with non-participating farmers by the project staff and the participating farmers.

Discussion and conclusion

Based on 34 interviews with smallholder farmers and project staff and a three-month ethnography of an action research and development project in Nepal, I explored learning in development cooperation, its outcomes and frictions, tensions, and ambivalences. In particular, inclusion, a core quality of participative learning in action research, is found to be both a source of and a way to resolve tensions in adult learning.

Empirically, the action research and development project in focus in this qualitative study is SAF-BIN, funded by the European Union Global Program on ARD, with the NGO Caritas in charge of the project. Its aim is to improve smallholder farmers’ resilience towards changing climatic conditions by collaboratively identifying, testing, and spreading local agricultural innovation in Nepal, Bangladesh and India. To do so, thousands of smallholder farmers with different socio-economic backgrounds formed as groups called smallholder farmers collectives (SHFC) collaborate with the project staff, scientists from public and private research institutes, and policymakers.

The findings show that learning activities in the SAF-BIN project range from conducting a baseline study, to opening resource centres, organising regular and formalised farmer group meetings, and participative on-farm adaptive research. Those activities, and particularly their variety, combination, cyclicity, inclusivity, and participation (cf. Rauch 2020) were found to facilitate learning in the sense of Mezirow and Taylor (2009) by fostering dialogue, individual experience, and the critical reflection of the participants. The action research activities (e.g. the baseline survey and the on-farm adaptive research trials) were conducted in a cyclical and iterative manner, going back and forth between data and concepts, and compared across project regions. The learners’ experiences lead to instrumental outcomes (e.g. new knowledge about climate change was gained, new inputs and cultivation practices were used, farming was increasingly diversified and professionalised) and communicative learning outcomes (e.g. meaning, values and ideals were negotiated), which were identified as being important to achieve more resilience towards climate change in farming. In this sense, action research and inclusion supported the integration and rearrangement of institutions and is a source of learning, adaptation, and social change.

The results show that inclusion does not prevent the organisation and project participants from experiencing ambivalences and struggles. In fact, the inclusion of heterogeneous audiences can be both the source and a key to the (dis)entanglement of those ambivalences. Action research brings forth tensions, as collaboration in multilateral development work is a melting point for diverging beliefs, values and ideas. While including the marginalised is perceived as key for development by donors, it is not always seen as such by the local population, as it is uncommon that farmers from different backgrounds work intensely together across social groups such as ethnic groups, castes, and genders. The inclusion of marginalised groups and topics is hence controversially negotiated, fostering reflection on social structures that were typically taken for granted. This case illustrates that inclusion of one social group, for example, inevitably goes along with the exclusion (e.g. if a distinct ethnic group is selected for participation, another is excluded) and hence requires legitimacy in the first place (cf. Dobusch and Dobusch 2019). Management practices such as action research and inclusion thus foster legitimacy through signalling sovereignty, the locus of control, and citizenship (cf. Meyer and Bromley 2013). In this sense, inclusion as a model of organising and a powerful institution does allow for achieving legitimacy and survival, as argued by institutional scholars (Goldenstein and Walgenbach 2019), as well as for social change.

The project appears as a ‘responsible’ actor through the rationalisation (cf. Djelic and Quack 2003) of activities, when farmers, for example, measure the status quo and communicate the progress of their field trials. Mechanisms of rationalisation include the quantification of project outcomes, but also the selection of project participants from the very start, and the closeness to science in the identification and selection of agricultural innovations for further development. Similar to the individuals participating in the project, the action research project as a whole is equipped with civil duties and constructed as a moral actor (Drori and Meyer 2006), signalling citizenship (Meyer and Bromley 2013) by highlighting their social responsibility.

While action research in development typically aims at the improvement of peoples’ life (cf. Gustavsen 2014), this study proves that studying EAR allows for theoretical contributions. In summary and as argued above, inclusion and EAR can be conceptualised and theorised (at least) in the following three ways:

1. from a dialectical perspective – Whereas inclusion is often treated as an ethical and economic norm, goal, or purpose, it can also be conceptualised as binary and a matter of degree in the inclusion/exclusion continuum. Thereby, inclusion and exclusion are seen as relational (context-dependent), constituting, and dialectical, that is, conceived as two sides of the same coin. From this perspective, there is no inclusion without exclusion: their coexistence is a necessary condition to be able to talk and think about them. This is seen as a critical condition for considering and asking questions such as: ‘What combination of inclusion and exclusion are desired and legitimate to achieve social transformation?’ or ‘To what extent might inclusion even import, reproduce or reinforce, social inequalities?’;

2. from an institutional perspective – Action research and its qualities, such as inclusion, are conceptualised as a way of legitimizing development projects towards diverse audiences and organising the project activities in an effective manner;

3. from a learning perspective – When we understand research as a way of fostering social change and pursuing a political agenda, action research is found to be a suitable methodological approach to gain and distribute insights and social change.

I have further taken the following methodological key lessons through my engagement in the process. Studying action research takes time and requires careful observation of the activities and participants over months. This holds especially true in complex and culturally different social settings. Agricultural systems are characterised by their local specificities (cf. Berg and Jiggins 2007), which demands careful conduct of both EAR and the study of the same in the field of agriculture.

I conclude that action research in development allows for organisational legitimacy, adult learning, and to integrate and recombine institutions at the micro-level, and can thus be interpreted as a source of, at least temporal, learning, social transformation and change. While rationalisation processes such as action research projects in development lead to change, adaptation, and learning, it remains to be seen what is ‘unlearned’ (e.g. fiction, intuition, traditional practices and beliefs related to nature).

Geolocation information

The data have been collected in the following districts in Nepal: Kaski and Bardiya. The data was analysed in Innsbruck and Vienna, Austria.

Acknowledgments

The data were collected as part of a thesis at the University of Natural Resources and Life Science, Vienna, at the Institute for Development Research. I thank the supervisors for their support, and the interview partners for sharing their experiences.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. cf. also (2021). Journal information (online) https://www.tandfonline.com/action/journalInformation?journalCode=reac20.

2. Action for Food Production (AFPRO), Sam Higginbottom Institute of Agriculture, Technology & Sciences (SHIATS), Bangladesh Rice Research Institute (BRRI) and Local Initiatives for Biodiversity, Research, and Development (LI-BIRD) (‘As of July 7, 2020, SAF-BIN stated it on its website http://www.safbin.org/project-overview#click’).

3. Elected members at the district level responsible for formulating policies.