How to access smart agriculture empirically
To explore the ways in which information and communication technologies have informed both visions and practices of agricultural development in Eastern Africa, we used three different empirical approaches. Combining document analysis, event ethnography on agricultural fairs and conferences, and participant observation with NGOs and farmers at the implementation sites of diverse agricultural development projects, we were able to follow the idea of a digital transformation from the reports of major organizations to conference presentations to the ways in which it informs concrete project designs.
Qualitative document analysis
Qualitative document analysis has recently been used to understand shifts in development policy in general (Berndt 2015) and for digital agriculture in Africa in particular (Abdulai 2022).
Here, development policy and related documents are seen to act as hegemonic instruments and mechanisms to create order. Their paradigmatic shifts are being expressed in working papers and reports of major development organizations. In this text, such documents are examined in terms of the visions and future agendas they entail, as these contribute to a legitimizing discourse for influential development actors and technology stakeholders (see Korf 2004).
Given the durability of documents as archives of visions, a qualitative document analysis allows the tracing of discursive patterns and shifts over time. At the same time, when combined with other modes of qualitative inquiry, it may serve as verification of results gained through other methods (Bowen 2009: 28–31).
For our analysis, over 200 documents dealing with the digitization of agriculture in the Global South were collected, covering the time period from 2003, when the first World Summit on the Information Society (WSIS) took place in Geneva, until today (May 2024). These documents included annual reports, conference proceedings, technical briefs, strategy papers, case studies, and press releases, among others. In light of principles of relevance, completeness, balance, and evidence regarding the purpose of the study (see e.g. Bowen 2009: 33, Karpinnen and Moe 2019: 257–9), 35 key publications were identified as relevant in that they represent a comprehensive, well-balanced body of documents dealing with the digital transformation of African agriculture explicitly and stemming from a wide range of actors ranging from non-governmental and international organizations to development agencies, and the private sector.
However, the frameworks and policies formulated in these documents are not to be taken as a given. They arise in the sense of order-creating measures within organizations to clear up internal ambiguities, to be translated into best practices and, in the end, to be adapted and reinterpreted according to successful or failed project outcomes (Korf 2004). Nevertheless, policies stabilize and maintain a rational logic in corporate negotiations on funds and projects and may thus serve as an interpretative pattern for the norms, administration, and analytics of development (Lewis and Mosse 2006). They describe an ideal world as envisioned by their authors and editors which can rarely be directly translated into practice. They stem from a rather homogeneous group of experts, characterized by similar values and lifestyles and often closely connected through employment circles, digital communication, and certain events (Mosse 2014). It is such events that we incorporated into our research as a complementary qualitative approach.
Event ethnography
In addition to the document analysis, we conducted so-called event ethnographies in two major conferences and one online course dedicated to agricultural development through digital technologies. Comprising presentations, competitive pitches, Q&A sessions, and informal interactions, these events can be seen as settings in which participants ‘embody their professional identities: performing for peers, conveying information, and telling stories’ (Monahan and Fisher 2015: 714). Investigating these settings ethnographically makes it possible to understand how ‘knowledge of practice’ emerges in interactions among professional elites and ‘authorized speakers’ (Baird 2017: 189). Here, we focused on the situated discursive and performative processes that shape a field of practice as well as determine who participates in these processes (Duffy 2014). This allowed us to gain insights into which types of knowledge are treated as legitimate or not, and which ones are ignored altogether (Campbell et al. 2014).
In this contribution we focus on two conferences that brought together practitioners and decision-makers from research institutions, government and non-government agencies and the private sector: ‘ICTforAg’, a one-day conference held in Washington, DC, in June 2018 and ‘Decoding the Data Ecosystem: Big Data in Agriculture Convention’, which took place over three days in October 2018 in Nairobi. In addition, we also participated in a four-week online course on agricultural development and technological innovation held by experienced experts and successful entrepreneurs.
At these events, we listened to talks, took part in Q&A sessions, and interacted with other participants, e.g. in group work sessions and breaks. The material we generated from this therefore mainly consists of detailed notes from presentations and podium discussions, as well as notes on forum entries, which made up the discursive space of the online course. Furthermore, we generated field notes with general observations, e.g. on the structure of conferences and how people (inter)act. To complement this, we also used materials such as product brochures, advertised reading materials, and slides and presentation videos.
Given the staged character of presentation and product pitches on such events, it must be noted that these insights are never complete or fully comprehensive. Nonetheless, it is possible to draw conclusions about a professional culture because participants ‘reveal more than they intend, especially at the level of the deep background assumptions that shape what they say and what they do not say, as well as the body displays and interaction rituals they perform’ (Ortner 2010: 220). Furthermore, the experts for ICT4Ag or D4Ag bring their ideas, views and expertise to the places of implementation.
Participant Observation at Project Sites
At events it becomes clear that technical knowledge and information gained through digital technologies is considered more true, objective, and neutral than, and therefore rather superior to, local forms of knowledge. However, the idea that these technologies can help overcome farmers’ ‘knowledge gaps’ often turns out to be more complex than expected. The encounters between development practitioners who bring their visions to the fields and farmers tend to lead to complex processes of negotiation or even neglect (Mosse 2014). Against this backdrop, participant observation of development practice lends itself as an insightful approach, characterised by intensive and prolonged participation (Korf 2004).
Accordingly, we engaged with selected projects in Kenya and Tanzania in which the development and use of digital technologies was a central concern. One of these projects concentrated on an agricultural information service providing the farmers with weather forecasts and agricultural advice in Western Kenya, another one comprised plant clinics for the identification and treatment of pests and diseases. In the rural South of Tanzania, the projects studied range from participatory land-use mapping, to IT innovations to improve agricultural productivity. Most of these projects are still in their pilot stage, which – as we will show – plays a crucial role when examining the character of digital interventions.
These five different project settings all required multi-layered roles on the part of the researcher. While we were collecting observations and participating in the ways in which digital transformations were supposed to be brought about, we sometimes found ourselves in the role of an ‘antagonistic observer’ who, as Lewis poignantly described, is ‘characterised by critical distance and a basic hostility towards both the ideas of development and the motives of those who seek to promote it’ (Lewis 2005: 472). In other situations, we could rather be considered as ‘reluctant observers’ (Lewis 2005, Lewis and Mosse 2006), for example, when uncomfortably sitting next to the development professional who introduces himself as chief at the annual evaluation meeting and poses the question: ‘Will intensified agriculture lead you out of the poverty trap one day?’ following which farmers were encouraged to answer ‘yes’, ‘no’, or ‘maybe’. However, since it was common to be asked for advice either by the implementing organization or by farmers, each seeking support on how best to deal with the situation, the research also involved more activist and engaged moments.
Of particular interest to us was to experience the technology in action. Living in the villages, interning at the NGOs involved, joining focus groups and workshops, helping farmers on their fields and assisting them in their daily work, we especially focused on the ways in which different actors interact with the technologies, how those technologies are adapted to the specific contexts, and what role they play not just on the projects’ agendas but also in actual doings (Adams and Thompson 2016: 17, 18). Furthermore, we were able to gain insights into the ways in which the digitization of agricultural development was perceived by farmers. As a participant observer in these contexts it becomes irrevocably evident that ‘[m]eanings are contested and organizing space is negotiated at multiple levels, blurring the range of insider/outsider, indigenous/scientific or formal/informal dualisms that are common in both development theory and practice’ (Lewis and Mosse 2006: 10).
As our results show, such a reflexive social scientific approach is crucial to better understand the implications of African agriculture going digital.