PROTOCOL: Farmer Field Schools for Improving Farming Practices and Farmer Outcomes in Low‐ and Middle‐income Countries: A Systematic Review

Abstract

Agriculture has wide-ranging global impacts which extend to economic growth, poverty reduction, food security, livelihoods, rural development and the environment (Green et al., 2005). Agriculture is the main source of income for around 2.5 billion people in the developing world (FAO, 2003, p. 1). In addition, around 70 percent of the global extreme poor – or over one billion people – lives in rural areas in low and middle income countries (IFAD, 2010, p.233), most of whom rely directly or indirectly on agriculture for their livelihoods. Investment in agriculture has been shown to have beneficial impacts on agricultural growth and poverty reduction (Fan & Rao, 2003). Moreover, the poorest population quintiles benefit significantly more from agricultural growth than growth in other sectors of the economy (United Nations, 2008; World Bank, 2007). The modernisation of farming practices in the 1960s and 70s during the ‘Green Revolution’ improved agricultural yields substantially in those areas it reached and raised national production and food security (IFAD, 2001). However, two key challenges emerged (van den Berg & Jiggins, 2007). The first problem was that poor farmers were being left behind, particularly in sub-Saharan Africa where many were not reached by modernisation approaches. In addition, those technologies promoted were not appropriate to the challenges facing smallholders in the African context, particularly women farmers (Inter-Academy Council, 2004). Second, modernisation was also associated with adverse environmental and health consequences, relating to water pollution, declining soil quality, soil erosion, pest resistance and loss of biodiversity. A particular problem emerged around environmental and health consequences of chemical pesticides use. Chemical pesticides have been heavily promoted and publicly subsidised under the modernisation agenda to such an extent that their overuse led to insect pests becoming resistant and causing major outbreaks of insect pests in rice crops in Asia in the 1970s and 80s. In addition, prolonged exposure to pesticides was associated with chronic and acute health problems among rural residents (Pingali & Roger, 1995). Use of broad-spectrum insecticides in agriculture has even been linked to mosquito vectors of malaria developing resistance to insecticides used in malaria control programs (Diabate et al., 2002; cited in van den Berg & Jiggins, 2007). Ensuring that farmers have appropriate knowledge and skills to deal with changing environments, and that they have access to technology and management practices appropriate to their needs, are key components of sustainable agricultural development strategies. Agricultural extension and advisory services (hereafter extension services) are used by policy makers to achieve this aim, and comprise “the entire set of organisations that support and facilitate people engaged in agricultural production to solve problems and to obtain information, skills and technologies to improve their livelihoods” (Anderson, 2007, p.6). Extension was traditionally viewed as a means of transferring technologies developed in research stations and farm management practices to farmers, and used ‘top-down’ institutions of delivery, as characterised for example by the World Bank's Training and Visit System (Gautam & Anderson, 2000). These traditional extension approaches were criticised for providing a ‘one size fits all’ approach (Birner et al., 2006) which failed to factor in the diverse socio-economic and institutional environments faced by farmers, for failing to involve farmers in the development of technology and practices appropriate to their contexts, or to empower them more generally as problem solving decision makers. Ultimately extension has failed to achieve its main objective of farm productivity improvements, particularly in Africa (Anderson, 2007; Birkhaeuser et al., 1991). In addition, more intensive approaches were considered necessary to disseminate complex messages, such as on sustainable pest management. Since the 1980s, the approach to extension service delivery has drawn increasingly on more participatory methods, of which farmer field schools have become prominent (van den Berg & Jiggins, 2007).2 Participatory approaches to extension are based on the idea that they create spaces for farmer ‘self-learning’ and sharing and also allow the agents and agricultural researchers to learn from the farmers (Birner et al., 2006). Farmer field schools (FFSs) originated in Asia as a means of achieving several objectives, of which key was to deliver training on ‘integrated pest management’ as an alternative to intensive pesticide spraying, which was severely damaging farm production, the environment and farmers' health. Integrated pest management (IPM) was developed in the 1960s and 70s (Kelly, 2005) and aimed to minimise pesticide use through use of more ‘natural’ techniques of pest management. Integrated pest management methods promoted in FFSs typically range from more simple practices, such as not applying pesticides in the first 30 days after planning (‘no early spray‘), to more complex ones that require in-depth agro-ecological and crop management knowledge, such as being able to differentiate beneficial from harmful insects, and creating a conducive environment for pest predators (Ricker-Gilbert et al., 2008). FFSs are participatory, learner-centred and characterized by heavy reliance on learning-by-doing (Pontius et al., 2002). Farmer field schools use intensive ‘discovery learning’ techniques to provide farmers with the skills and confidence to adopt different growing techniques and change the mix of inputs used on their farms. In the case of IPM-FFSs, field school participants are instructed on how to move away from pesticides to more natural techniques of pest management. Objectives of the schools include increasing farm productivity, reducing negative environmental impacts and promoting farmer empowerment. The FFS approach draws on participatory methods, both in terms of its bottom-up focus, with curricula drawing on priorities identified by farmers, and in terms of the focus on farmer experimentation and building problem solving capabilities, which empowers farmers “to handle their own on-farm decisions, using experiential learning techniques developed for non-formal adult education purposes” (Khisa, 2004). Thus, FFSs aim to provide farmers with skills which enable them to solve problems for themselves and the group activities empower farmers both within their own communities and outside. Indeed, the approach is far more intensive than other extension delivery mechanisms and broader in its objective to provide skills and empower individuals, and is therefore considered an adult education intervention by some practitioners and analysts (Braun et al., 2006; van den Berg & Jiggins, 2007).3 Originally developed for rice crops in Indonesia in the 1980s by the United Nations Food and Agriculture Organisation (FAO), FFSs had been implemented in 87 countries worldwide and produced 10-20 million field school graduates by 2008 (Braun & Duveskog, 2008). While all FFSs should be based on the same process, the approach can be adopted to suit particular needs, crops or contexts (Pontius et al., 2002). Thus, as FFSs have been promoted around the world, the IPM curriculum has been modified depending on the context, and applied to other food staples, vegetables and cotton (Braun & Duveskog, 2008). In Africa, Integrated Production and Pest Management (IPPM) has been promoted, which reflects a more ‘holistic’ approach to improving production, in which pests and pesticide use are not necessarily the main production problems (Stathers et al., 2005).4 Other variants include integrated disease management (IDM), integrated crop management (ICM), integrated plant nutrient management (IPNM), and integrated water and soil management (IWSM)5 In addition, the farmer field school curriculum has also been broadened to tackle populations in particular contexts, such as Junior Farmer Field and Life Schools (JFFLS) which have been implemented among youths across Africa and include HIV-risk reduction in addition to agriculture components more standard to FFSs (Braun & Duveskog, 2008). A typical FFS includes a field-based season-long training program delivered by a FFS facilitator, with weekly meetings nearby the plots of participating farmers (Pontius et al., 2002). Each FFS typically has from 25 to 30 participants, with farmers working together in groups of five. Facilitators can be either extension agents or selected graduates from previous FFSs, who undergo a training-of-trainers course tailored to equip them to facilitate field schools (Braun & Duveskog, 2008). The facilitators use experiential, participatory and learner-centred educational methods, including use of demonstration plots comparing farmers' existing practices with new practices promoted in the FFS to enable farmers to observe benefits (Pontius et al., 2002). Field school activities include agro-ecosystem analysis, farmer experimentation, activities to improve group dynamics, field days focusing on specific local problems and participant presentation of course material and the results of their studies (ibid). Due to the externalities associated with pesticides use (social costs exceed private costs), it may be important that integrated pest management practices are adopted throughout a community for the approach to be sustainable, particularly in extreme situations of pesticide misuse. As Feder et al. (2004a) note, lack of adoption of IPM practices by neighbouring farmers might curtail the effectiveness of the intervention, as pests from their fields may re-infest the fields of adopters, eventually leading to disadoption of IPM by FFS participants. To promote community-wide uptake of IPM, FFS participants are encouraged to share their knowledge with non-participants as a way of promoting farmer-to-farmer diffusion (Feder et al., 2004b). The extent to which it is assumed this will occur informally through FFS graduates' social networks or needs to be encouraged through community initiatives including formal training-of-trainers programmes for alumni, seems to vary from programme to programme. However, FAO guidelines on ‘community-IPM’ which focuses on institutionalising IPM at the local level, indicate formal approaches involving FFS alumni are necessary: ‘without post-FFS educational opportunities, there will be no community movement’ (Pontius et al., 2002). A wide range of factors is likely to influence effectiveness of farmer field schools. Figure 1 presents a stylized causal chain linking farmer field school delivery inputs with final outcomes, via intermediate outcomes in terms of capacity building and technological adoption. Intermediate outcomes are shown for both field school participants and for ‘exposed’ farmers who benefit via farmer-to-farmer diffusion knowledge spillovers, and are expected to be those living in close geographical proximity to field school participants or involved in their social networks. FFS theory of change Source: authors Underlying each link in the causal model are a number of assumptions which determine the extent of behaviour change and therefore the extent to which impacts materialize in practice. For instance, facilitators are a key input: they may be ‘traditional’ extension agents who have received training in the FFS approach, meaning they are required to move away from the top-down approaches to which they are familiar, and adopt a more participatory, learner-centred approach (Feder et al., 2004a). The extent to which they are able to do this may have implications for programme effectiveness. Similarly, the relevance of the FFS curriculum to farmers, to the extent that the new practices are actually appropriate, will influence farmers' attitudes and behaviour change. Farmers also need to be trained adequately, in that they have attended sufficient meetings over the planting season. Observability of improvements based on the comparison of experimental FFS with ‘control’ plots which use standard approaches to pest management, also appear to be a key component. Finally, the price of pesticide relative to (opportunity costs of) labour is also likely to determine adoption of IPM, where it involves substantial increases in demands on farmers' time. Characteristics of the intervention are important moderators of both the causal chain framework presented here and the likely impacts. As noted in Davis et al. (2010), not all FFSs are the same. Earlier FFS programmes implemented in Asia usually focused on rice IPM, while later ones focused on different crops and livestock. However, the extent to which FFSs adapted to different contexts are based on the same process, both in the design and implementation of the programmes, is likely to vary. For instance, different FFSs use different approaches, with some following a transfer-of-technology approach, while others focus to a greater extent on education or empowerment (Davis et al., 2010). These are important contextual differences which influence the way in which the concept of FFS is operationalised on the ground and may therefore also have implications for intervention effectiveness. Characteristics of local communities, such as heterogeneity in terms of land- and asset-holdings, ethnicity, education, gender roles and the degree of social cohesion, will determine the ability of the schools to reach appropriate beneficiaries, including disadvantaged farmers such as women. In the absence of formalised community building and training-of-trainers programmes for alumni, characteristics of local communities may be important determinants of the degree of diffusion of knowledge and practices from participants to non-participants. The assumption that there will be some diffusion between farmers may not be an unreasonable one in principle for simple practices. However, evidence also suggests that, in the case of technically complex issues or costly technologies, farmers prefer first-hand knowledge or advice from specialized information sources such as experts (Feder et al. 2004a). As FFSs are used to disseminate complex information and aim to improve participants' decision-making skills, FFS graduates may be limited in their ability to transmit all but the simplest of messages effectively to other farmers through informal means. Whether the diffusion mechanism is informal or formalised will therefore have implications for beneficiary targeting (Feder & Savastano, 2006). Without formal mechanisms, participants would ideally be selected if they have characteristics which will enhance diffusion, such as those respected in their communities and those with strong social networks. This may conflict with other objectives of FFS, such as targeting women farmers. Finally, the effectiveness of the schools in fostering adoption of the new technology and improved agricultural outcomes will depend on contextual factors. External factors, notably weather conditions, soil fertility, plant disease and climate trends, determine production and yields. Market prices and market access, both to purchase inputs and sell produce, determine the value of production and therefore farmer income. It is also possible that the technologies promoted by FFSs do not act to change yields (the amount of crop produced per unit of land area), but still act to improve income and net revenues (value of production less input costs) by reducing pesticide costs, provided these are not offset by any net increases in costs of labour in applying the new technologies. Moreover, in contexts where pesticides and pest management are not necessarily the key constraints to production, improvements in productivity may not necessarily arise from reduced pesticides use but as a result of adoption of the other practices being promoted, such as soil management. Since the 1980s there has been a decline or stagnation in public expenditure on agriculture in most developing countries (Akroyd & Smith, 2007). Likewise, the proportion of official development assistance (ODA) going to agriculture is estimated to have declined from around 18 per cent in 1979 to 3.5 per cent in 2004 (World Bank, 2007, p. 41). However, as noted in the World Development Report on Agriculture, “extension services, after a period of neglect, are now back on the development agenda… [but] More evaluation, learning, and knowledge sharing are required to capitalize on this renewed momentum” (World Bank, 2007, p. 175). Poverty reduction strategies in 24 African countries also listed extension as a top agricultural priority (InterAcademy Council, 2004; cited in Davis, 2006). However, age old questions remain including how to raise yields and farmer incomes, how to do so in an environmentally sustainable manner, and how to bring extension services to the poorest people. A large literature exists measuring the impacts of farmer field schools, as summarised in a number of literature reviews (Davis, 2006; Feder et al., 2010; Quizon et al., 2001; Tripp et al., 2005; van den Berg, 2004; van den Berg & Jiggins, 2007). The reviews provide conflicting conclusions about the effectiveness of farmer field schools. Further, none appears to draw on a systematic search for all available literature, applying standard inclusion criteria, and critically appraising and synthesizing literature. Moreover, most of them draw on studies that appear suspect in terms of causal validity, due to problems of confounding, selection bias and spillovers, and thus are liable to high risks of bias in attributing outcomes to the intervention.6 Van den Berg (2004) provides the most authoritative review of IPM-FFS to-date, synthesising 25 evaluation studies. Most studies focused on rice and measured immediate impact of the FFSs in terms of reduced pesticide use and changes in yields, reporting considerable reductions in pesticide use, with some studies also showing an increase in yields. The review concludes: “Studies reported substantial and consistent reductions in pesticide use attributable to the effect of training. In a number of cases, there was also a convincing increase in yield due to training.... Results demonstrated remarkable, widespread and lasting developmental impacts” (p.3). Building on the latter, Van den Berg and Jiggins (2007) argue that FFSs have had additional benefits to that of IPM including facilitating collective action, leadership, organisation and improved problem-solving skills. However, the methodology of the studies reviewed varies, and a large number are based on pretest-posttest evaluation design with no non-intervention comparison group. Moreover, the conclusions of these reviews are based on significance-based vote counting, rather than sample weighted meta-analysis of effects. Tripp et al. (2005) note that despite the sizeable investments in FFS in Asia, there is a lack of rigorous evidence on the effectiveness of the approach. Their review includes seven studies on the effects of FFS on insecticide use in addition to their own study from Sri Lanka, and all but one of these studies report that FFS participants reduced their pesticide use. However, all of these studies suffer from methodological weaknesses, due to lack of non-FFS comparison groups and high risk of selection bias. The authors conclude that while “the FFS approach has undoubtedly succeeded in lowering insecticide use in a number of Asian rice examples, judgments on its overall impact await further study” (p. 1711). The authors also found little evidence to suggest effective diffusion of knowledge between FFS participants and non-participants, nor sufficient evidence to conclude that FFS groups continue on their own. FFS is a particularly intensive intervention, with high costs in terms of both facilitation and opportunity costs of beneficiaries' time. Leading authors from the literature have therefore noted that FFS is unlikely to be a solution to extension delivery, and only likely to be scalable under certain circumstances, and to face particular challenges (Braun et al., 2006; Davis, 2006). Braun and Duveskog (2008) argue that relative cost-effectiveness of FFS should be put in the context of rural adult education rather than extension “when FFS are regarded as a form of public investment in farmer education to tackle rural poverty – and hence as a tool for achieving the Millennium Development Goals” (p.19). Van den Berg and Jiggins (2007) also note that discussions on the fiscal sustainability of FFSs should include considerations of who will pay for the externalities of pesticide use; they conclude that the evidence gathered in their review suggests that FFSs can be a sustainable way of increasing farmers' skills and thus contributing towards escaping poverty. However, Quizon et al. (2001) note that lack of fiscal sustainability is a generic problem affecting many large-scale public extension services, concluding that FFSs face the same issues as other approaches. The cost per farmer is likely to be high compared to less intensive extension approaches and the evidence from Indonesia suggests there is a low rate of informal diffusion from direct beneficiaries of the schools to non-beneficiaries based locally.7 They suggest that as the situation for farmers, in terms of political power, governance systems and day-to-day interactions among farmers, is quite similar in many other developing countries in Asia and Africa, the results are relevant for discussions of similar extension activities in these areas. They warn that while pilot projects might indicate the viability of the FFS approach in certain circumstances, the issue of fiscal sustainability is particularly relevant when scaling up. The existing reviews provide some suggestive evidence of the effects of FFS, but come to widely different conclusions in a hotly debated, policy important field. Our systematic review aims to shed light on this debate by aiming to provide a systematic and exhaustive literature search, together with a comprehensive and unbiased synthesis of the existing evidence. Our review will contribute to the discussion, and systematic review methods development more generally, by synthesising both quantitative and qualitative literatures. To ensure the review is adequately oriented towards both reporting effects and explaining the reasons for them, we will synthesise effects along the causal chain, together with qualitative evidence. For quantitative synthesis, where studies are judged sufficiently similar to do so, we will use meta-analysis to pool study effects. An important benefit of meta-analysis is that it aims to account for the statistical power of studies. The alternative to meta-analysis is ‘vote counting’, which simply reports the numbers of positive, negative and insignificant findings. However, using this approach, one is more likely to come to incorrect conclusions about studies that find insignificant effects on outcomes, when this could simply be the result of low statistical power (small sample size), and therefore about the body of literature as a whole (Borenstein et al., 2009). One of the main benefits of meta-analysis is that by pooling across a larger sample, it takes into account both magnitude and precision of effects, allowing the researcher to correct for possibly under-powered studies. The primary objective of the review is to synthesise evidence on the effectiveness of farmer field school interventions in disseminating information on integrated pest management (IPM). The review aims to answer the following questions: Review question (1): What is the impact of farmer field schools on their objectives in terms of ‘endpoint’ outcomes such as increased yields, net revenues and farmer empowerment, and intermediate outcomes such as capacity building and adoption of improved practices (e.g. reduced use of pesticides) in low and middle income countries? Review question (2): Under which circumstances and why: what are the facilitators and barriers to FFS effectiveness and sustainability? The review will follow Campbell and Cochrane Collaboration approaches to systematic reviewing (Becker et al., n.d.; Hammerstrøm et al., 2010; Higgins & Green, 2011; Shadish & Myers, 2004; Shemilt et al., 2008). The review is also informed by theory-based impact evaluation (White, 2009) using the theory of change (Figure 1) as the framework for the review, to guide the types of studies included, data collection and analysis. The review will systematically collect and synthesize quantitative evidence from high quality impact evaluations of farmer field school interventions using meta-analysis, to answer review question (1). Outcomes will be synthesised along the causal chain, from intermediate outcomes such as capacity building, technological adoption and diffusion to final outcomes such as agricultural yields, household income and other indicators of household wellbeing. Farmer field schools are complex interventions implemented in a range of different contexts, making the limitations of a systematic review focusing solely on effectiveness particularly apparent. For the review to be more useful for policy-makers and practitioners, we will extend the review of effectiveness by including quantitative and qualitative studies to address review question (2), focusing on underlying factors that determine or hinder the effectiveness of FFSs. We will conduct the two syntheses in parallel, before integrating them in a final synthesis. Our methodology is also informed by chapter 20 in the Cochrane Handbook (Higgins & Green, 2011), the additional guidance developed by the Cochrane Qualitative Methods Group (Noyes et al., 2011) and the increasing number of examples of systematic reviews in international development based on or incorporating qualitative evidence (e.g., Munro et al., 2007; Williamson et al., 2009). Studies will be included in the review if they meet the following selection criteria. The review includes arable farmers, living in developing (low- or middle-income) countries, as defined by the World Bank, at the time the intervention was carried out. For studies to be included, they need to collect and report on data at the farm or household level. Many of the included populations are by definition disadvantaged, but studies focusing on particular disadvantaged groups, or conducting analysis across disadvantaged groups, will be included in the review. The review excludes livestock farmers, who receive different types of training than arable farmers, and farmers based in high-income countries where the challenges facing farmers in terms of poverty, land size, crops, and agro-ecological and environmental contexts are usually very different. Studies must report specific FFS-IPM interventions, or similar. Interventions are identified as farmer field schools that contain all the following components: If farmers in low and middle income countries do access agricultural extension services at all, it is usually through visits from public extension agents, through observation of public demonstration plots, or through extension provided by the private sector. Public extension may take the form of centralised or more decentralised systems (Birner et al., 2006). We will include studies which compare farmers receiving FFS education to those who receive no, or other types of extension, including where they receive IPM (or equivalent) training from another source. We will collect relevant information on the intervention received by control/comparison groups. Due to the possibilities of spillovers, we will assess whether comparisons are geographically separated from intervention groups, which usually means they are living in a different village to the farmer field school participants.8 The review primarily looks at economic outcomes, including agricultural yields (production per unit of land), profits (revenues minus costs), household income/ expenditure/ poverty status, and empowerment outcomes such as feelings of self-esteem. The review is interested in effects on two groups of beneficiaries: those participating directly in the field school and those living or working in close proximity to participants (so-called FFS ‘exposed‘). We will collect information from studies assessing outcomes for either type of beneficiary, and will compare outcomes for both to the non-FFS comparison group. Intermediate outcomes include farmer knowledge and capacity, adoption of new approaches (including reduced pesticides use) and diffusion of new approaches to ‘exposed’ farmers who may live in the same communities as field school graduates, or interact with them at market. We will also collect data on other final outcomes measured including health and environmental outcomes. These include, for example, self-reporting of health conditions such as respiratory infections or eye irritation, or indices of environmental impact based on assessment of active ingredients in pesticides (Kovach et al., 1992). To answer review question (2) we will include data on barriers and facilitators to FFS effectiveness and sustainability. This will include factual information on participation rates and follow up activities and measures of beneficiaries' attitudes and experiences with FFS. It will also include process and implementation information. Assessing the measured impacts of FFSs requires an appropriate evaluation methodology. However, designing impact evaluations of agricultural programmes is complicated by the wide range of additional (confounding) factors that influence agricultural outcomes and by biases caused by self-selection of individuals and communities into programmes, meaning that differences in outcomes between participants and non-participants might result from pre-existing differences rather than the programme under eval