Selecting methods of agricultural extension to support diverse adoption pathways: a review and case studies

A typology of extension methods

Typologies of extension methods or approaches have been put forward by various authors over the past four decades. The food and agriculture organisation (FAO) defined eight approaches reflecting different contexts for delivering extension (i.e. general or transfer of technology approach; commodity specialised approach, training and visit approach, Participatory approach, project approach, farming system approach, cost-sharing approach, and educational institution approach; Axinn 1988). This typology was added by Bell (2015) in describing Farmer Field Schools and the Land Grant (USA) approach. Other authors have defined categories of extension methods based on the source of the approach, such as ‘push’ or extension-centred (supply driven), ‘pull’, farmer-centred or market-oriented (demand driven), and ‘participatory’ (Kareem and Phand 2018) or the client focus being the individual, the group, education or mass media (Gonsalves et al. 2005). Others distinguish the governance and organisational arrangements of extension (i.e. government, private, non-government; Davis 2008) or profit/non-profit, free/cost-recovery, multi-/single-purpose, technology-driven/need-oriented (Nagel 1997). More recently, the ethics, outcomes and means of agricultural extension has catalysed consideration of different theoretical lenses to consider the interests of which are being served in the extension process (i.e. producer-focused, interaction-focused or system-focused approaches; Rickards 2018). In Australia, a typology of extension methods developed by Coutts et al. (2017) focuses on nine different mechanisms that can encourage learning and change (Table 1). It is this typology that aligned most closely with the purpose of the review in that the mechanisms for learning and change need to be considered by all actors in the innovation system, irrespective of the institutional arrangements or philosophy underpinning extension interventions, although important. This framework was used to select and compare studies of extension methods and their effectiveness.

A structured literature review and meta-analysis of published studies and reports was conducted, informed by a PRISMA process for structured literature reviews (Page et al. 2021). Data bases searched included SCOPUS, the Web of Science Core Collection citation index, Science direct, and JSTOR. Direct searches of specialist extension journals [Journal of Agricultural Education and Extension; Rural Extension and Innovation systems (REIS) and the Journal of Extension (JoE)] were also conducted. Queries were used to retrieve the articles with the following protocols: published between 2000 and 2022; developed country focus (Australia/New Zealand; North America; UK and Europe); articles in English. Word search queries of titles, abstract and key words were conducted using the following search terms: (agricult* extension): AND (technology* adoption; adoption; demonstration) AND (adult learning and workshop) OR (training OR education) AND (nudge/nudge theory) OR (social marketing) AND (podcast; YouTube; social media; webinar; online; e-learn*; e-extension) AND (farm consult*; mentor; coach; fee-for-service) AND (boundary span*) OR (intermediary) AND (participa* action*) OR (action learn*) AND (best management practice; BMP) OR (on-line; virtual; augmented reality).

The decision to select articles for further analysis was based on their relevance to the aim of the paper, namely, to examine the effect and impact of the application of extension methods and the process by which the effect was achieved. Articles were excluded if they were general studies of factors affecting adoption rather than with a specific focus on evaluating methods, related to early versions of decision support tools (pre-2005; to focus on internet-based decision support), or books and book chapters. Additional limits were added if many results were achieved including LIMIT-TO: farm*.

In total, 2625 articles were sourced and, following a review of the abstract for relevance, a total of 96 articles was chosen for full review (Table 2). To achieve the aim of the study, analysis of the 96 articles involved identifying the extension method(s) according to the typology of methods (Table 1), noting the type of evidence collected in evaluating the extension methods, noting the adoption impacts or changes attributed to the methods, noting interrelationships between methods, and compiling critical success factors in the delivery of the method or other insights.

In addition to the review, two case studies of the application of combinations of extension approaches in the Australian dairy industry were examined. The purpose of the case studies is instrumental (Stake 2003, p. 137); that is, to provide practical insights and deepen the understanding of the review findings in the contemporary context of Australian dairy-sector RD&E. Chosen by the co-authors as examples of effective applications of methods where scientists are actively involved, the case studies include a description of the problem context and adoption challenge, the selection and delivery of extension methods, the role of scientists and the reported impacts. Critical success factors in the implementation of the combination of methods in the cases were drawn from reflections of the co-authors in light of the literature review findings and published studies.

The research concludes with frameworks developed from a synthesis of findings from the literature review and case studies, which identifies the context conditions that are suited to different extension methods and strategies for combining methods. These frameworks can be used to assist scientists and others in the selection of advisory and extension methods.

Facilitated groups/farmer-led groups

Facilitated farmer groups, including farmer-led groups and demonstration/focus farms, draw on methods of peer-to-peer and participatory action learning (PAL) (Knook et al. 2018). Studies of the impact of farmer groups on improving adoption, farm performance and other outcomes are largely positive. While many of the reviewed studies did not control for self-selection bias (i.e. groups may involve farmers more motivated to implement change), where this was considered, group involvement was linked to increased profitability and likelihood of adoption (Hennessy and Heanue 2012; Läpple et al. 2013). Other studies reported increased profitability and financial returns (Bell and Allan 2000; Hansen 2015) and increased farmer empowerment (Garforth et al. 2003; Morgans et al. 2021). When combined with one-to-one advice, group involvement has been found to influence changes in farmers’ attitudes, knowledge, and practices (Hansen 2015), and can support more rapid change than do other methods (Bewsell and Brenton-Rule 2019; Patchett et al. 2020) or does adoption among non-participants (Hansen 2015). In comparing impacts from different types of group involvement, Prager and Creaney (2017) found the levels of learning and adoption were higher for discussion groups than monitor farms, even though individual monitor farmers had greater practice change. In an investigation of demonstration farms as transformational learning spaces, Cooreman et al. (2021) found that facilitated dialogue as part of farm demonstrations was more likely to promote transformational learning than did those without facilitation.

The mechanisms by which group involvement influences adoption is suggested to be the development of high social capital and trust among participants and, through this, peer support and increased pro-activity to change, and better problem solving (Hansen 2015). However, it has been noted that the social and informal interactions in groups are often devalued by those investing in group-based extension (King et al. 2001). While the more open and flexible the approach in groups is suggested to create an environment for observing, experimenting, learning, and encouraging peers (Nettle et al. 2006), it was found that such environments rely on farmers having a mindset of being active knowledge creators rather than consumers of information (Prager and Creaney 2017) and the scientists involved in such groups also need to position themselves as learners and part of the shared inquiry (Sewell et al. 2014). The outcomes from these open group processes were noted as more difficult to measure (Sewell et al. 2017). In addition, the effectiveness of action learning has been found to be dependent on the presence of a variety of activities that were aligned to desired outcomes of farmers (Sewell et al. 2014) and the quality of facilitation with an emphasis on the mobilisation of knowledge and the reflection stage of the learning process (Dooley 2020; Cooreman et al. 2021; Morgans et al. 2021).

There were exceptions to the positive outcomes reported from group-learning. Poorer outcomes were reported where groups were formed by external bodies with imposed or expert-chosen objectives and topics or where there was a lack of involvement of landholders in project design (Robertson and Vitousek 2009) or the use of groups as part of socio-political agendas (Cook et al. 2021). Prescriptive programs struggled to simultaneously encourage farmer-led processes (Prager and Creaney 2017). The need to consider group extension approaches as embedded in the wider advisory landscape was emphasised by Inman et al. (2018) who noted that some farmers believe that group-learning cannot be tailored to their individual farm-specific circumstances.

While scientists may consider farmer groups to be largely outside their sphere of professional work, the evidence of the importance of group-learning in the adoption process suggests that this assumption needs to be reconsidered, and time spent. In studying a participatory research project involving scientists and farmers, Lyon et al. (2010), observed that initially ‘farmers weren’t sure participation was worth their time, while scientists, didn’t think it would serve their careers’ (p. 555). Yet this and other studies found that farmers’ learning was promoted when they participated in a learning community with agricultural scientists (Sewell et al. 2014) and enabled the interests of both researchers and farmers to be met (Lyon et al. 2010). This suggests that scientist are missing out on being part of the adoption story if they are not engaged in these processes, including the influence to participating farmers’ networks, where Wood et al. (2014) found that farmers exchanged new scientific knowledge from interactions with scientists in facilitated groups, within their wider networks. Scientists’ involvement in groups therefore provides an opportunity for farmers to engage with the science, and interactions with scientists helps span the worlds of science and farm decision-making (Sewell et al. 2014, 2017). Further, facilitated groups and farmer-led groups are also a key foundation for other methods, including technological development/multi-actor approach (Method 2) and co-innovation (Method 7).

Technological development (multi-actor approach)

Greater interaction among scientists, technology developers, farm advisors and farmers, or ‘multi-actor’ approaches, have been recommended to co-develop decision support tools with farmers and to support the implementation challenges of precision and digital agriculture (Douthwaite et al. 2001; Hochman and Carberry 2011; Eastwood et al. 2019). The close cooperation of practitioners with developers in co-design processes has been found to be crucial to ensuring that new agricultural innovations are successfully developed with farmers’ needs, values, knowledge, and experiences in mind (Kenny and Regan 2021). Further, interactions among farmers and between farmers and scientists has been shown to generate new ideas in farming, and hence is key to the innovation process (Sewell et al. 2014). In smart-farming contexts, the absence of effective interaction and knowledge flows among these groups has been associated with poorer outcomes, yet such interactions have proved difficult to support (Knierim et al. 2019). Farmers’ variable interest in being engaged in the research and development process has been identified as an issue (Dawson and Goldberger 2008; Lyon et al. 2010); however, multi-actor approaches, when combined with training, were found to improve farmers’ ability to access and use precision-agriculture and climate-forecasting tools (Halbleib and Jepson 2015; Stitzlein et al. 2020; Harden et al. 2021) and farmers perceived that direct links with researchers were of greater benefit than was access to decision support models alone (Carberry et al. 2002). In this study, farm advisers are suggested to have an important role in leading ‘what-if’ discussions with farmers and in helping in the interpretation of data and outputs of models (Carberry et al. 2002).

Another reported advantage of multi-actor approaches to technological development is in building relationships, trust and social capital among end-users, scientists, and developers (Sewell et al. 2014; Calliera et al. 2021; Brown et al. 2022), contributing to greater knowledge and skills among all participants and generation of new ideas (Sewell et al. 2014). Scientists are often called on to lead or contribute to these approaches (Ingram et al. 2020), expanding their professional profile into roles as coordinators and facilitators. With increasing expectations that scientists will engage more with pathways to have an impact (Röling 2009), scientists will need to be more engaged in the technological development process. How scientists can be supported, along with farmers and advisers, to play this role is an area needing to be addressed in the R&D policy. Greater collaborative actions between all stakeholders in multi-agency partnerships, where learning is mutual and co-constructed, is a recommended response (Sewell et al. 2014).

Training

The impacts of training on practice change and adoption indicate mixed results. In the field of farm safety-training interventions, some studies have reported limited effects or increases in occupational fatalities following the completion of programs. This was attributed to the short-term nature of the training, a lack of on-going support and that stakeholders and the media devoted less attention to the topic following the program (De Roo and Rautiainen 2000; Rautiainen et al. 2008; Cryer et al. 2014; Svennefelt and Lundqvist 2020). There was some evidence that accompanying training with financial incentives (in this context, reduced insurance premiums) influenced training participation (Rautiainen et al. 2005). In other contexts, training was found to have a positive effect on attitudes and commitment towards change in pasture improvement (Lloyd et al. 2009), increased awareness and understanding in nitrogen management (Lawrence et al. 2000), and was a precursor to further participation in learning among resource-limited farmers (Akobundu et al. 2004). Farmers who undertook agricultural training on mastitis management were 10 times more likely to monitor milk quality through milk recording than were those who had not (Dillon et al. 2016). Additionally, those farmers in contact with an extension service and who participated in a dairy discussion group were seven times more likely to practice milk recording (Dillon et al. 2016), and the combination of education, extension and milk recording reduced somatic cell count (SCC) by 25% for the average herd (Dillon et al. 2016).

The importance of follow-up to training was a common theme. Follow-up included individualised farm visits and on-going engagement or multiple contacts with participants after training. Follow-up to training was found to improve the timing of farm operations and increase active effort to solve on-farm problems and improved planning (Akobundu et al. 2004). The use of demonstrations alongside training and follow-up was shown to reduce pesticide use (Goodhue et al. 2010). Learning design was also found to be a factor in the success of training efforts. Specifically, to ensure that the practical and procedural (or action-based) knowledge of farmers was the focus in training efforts rather than a conceptual orientation, which is more common among scientists (Thompson and Reeve 2011). This was found to be particularly relevant in the field of pest management (Halbleib and Jepson 2015). The choice of trainer was also found to be important, with formal personal sources of information and farmers’ associations and consultants suggested as being important to involve (Hochman and Carberry 2011). One study reported the importance of inclusion of advisers and researchers with farmers in training (McKenzie et al. 2018). Involving others in the family or farm team in training was also found to be important in increasing knowledge and retention of knowledge in farm family health (Brumby et al. 2009).

While scientists may not be directly involved in training design, ensuring that the information conveyed from research is tailored to meet the practical needs of farmers is an important consideration for scientists (Sewell et al. 2014). Whereas scientists tend to focus on knowledge related to the ‘what and why’ of scientific principles, farmers seek ‘know how’ or practice knowledge (Klerkx and Proctor 2013), an area where scientists can place more effort.

Information provision

In a study of the efficacy of newsletters and brochures to facilitate adoption of herbicide-resistance strategies, Widderick et al. (2006) found that farmers who received a newsletter had greater knowledge of the situations causing herbicide resistance and expressed a greater concern about it. In another study, best management practice guides on management of invasive species were found to be highly valued sources of information, with detailed information on current best practices in one authoritative document of value for time-poor farmers (Coleman et al. 2017). However, the decision of whether to invest in a comprehensive best management practice guide was dependent on the urgency and risk from future invasion (Coleman et al. 2017). Offering tailored information with other methods such as financial incentives and intensive support was found to be effective; however, more specific tailoring to different segments of the population was needed with varying rates of adoption arising from different farm sizes, the future goals of farms such as succession intention and whether farmers were members of interest groups (Rollins et al. 2018). Further, the relevance, accuracy and format of information was noted as vital for farmers in the use of weather and climate services (Vedeld et al. 2020), pasture mixes (Sewell et al. 2017), alternative grazing regimes (Lyon et al. 2010), nitrogen management (Stuart et al. 2018) and carbon farming (Ingram et al. 2016).

There is an increasing use of digital tools to create and share information (Bamka et al. 2020) and in the context of sharing information about the perceived usefulness of smart-farming technologies, a positive indirect association was found between farmers’ exposure to formal personal sources of information such as farmers’ associations or training and the intention to adopt (Caffaro et al. 2020). The curation of information is therefore an important role in science communication (Llewellyn 2007) to which scientists can place greater effort, alongside their engagement with different networks where trusted sources of information are used by farmers.

Consulting (one-to-one advice, mentoring, coaching)

The provision of one-to-one advice to farmers has been associated with increased adoption in a range of studies. Farmers’ use of paid advisors was found to be a factor in the partial or full adoption of no-till cropping methods in Australia; however, farmers using multiple till practices were those less likely to employ a paid advisor (McRoberts and Rickards 2010). In a study of farmers’ contacts with advisory services relating to environmental concerns in the EU, advisory contact was positively related to the adoption of innovations, the number of information sources used and the adoption of farm risk-management measures (Herrera et al. 2019). The association between one-to-one advice and practice change has increased the interest of government and industry bodies in Australia to explore ways to support or subsidise farmers access and use of one-to-one advisers without farmers paying directly for this service (Klerkx and Jansen 2010; Nettle et al. 2021). However, the effectiveness of one-to-one advice has been found to be dependent on the extent of trust, credibility and empathy between the farmer and adviser within the advisory relationship (Kuehne et al. 2019). These attributes were shown to affect the quality of knowledge exchange, particularly in the context of facilitating farmers’ transformation to more sustainable best management practices (Ingram 2008) and where trust in government is low (Fisher 2013). Further, it has been found that farmers and advisers hold different opinions about what is effective in the advisory service provided. In one study, farmers considered that advisers were useful in providing information about grants and held relevant knowledge; however, their level of trust in the advice, the continuity and clarity in advice and how this compared to local evidence were also important attributes in advisory services valued by farmers; yet, these were areas advisers did not always acknowledge or appreciate (Vrain and Lovett 2020).

The type of advisory model used was found to be an important factor in the effectiveness of one-to-one services. In a study of the application of coaching models to support livestock farmers’ understanding of GHG emissions and improving profit, 79% of participants either intended to, or had already made changes because of participating, which was a 9% higher level of intention to change than reported in traditional adoption programs in the industry. Other changes included a 43% increase in skills post-coaching and an overall reduction of 19% in GHG emissions intensity (Sobotta et al. 2016). Coaching models work with farmers’ own knowledge and local context while supporting their autonomy, and coaches can operate as intermediaries between farmers and researchers (Cobon et al. 2021). One study reported that advisors and advisory organisations often fail to consider these functions (Herrera et al. 2019) and hence the advisory context (the organisation and the interests of advisers) can be an important factor in the effectiveness of the one-to-one approach (Nettle et al. 2018). For instance, in one study, crop advisers were most likely to recommend practices associated with soil health that were directly related to crop production (e.g. higher yields and lower inputs) rather than to other benefits to the environment (Eanes et al. 2019). It is argued that if advisers do not buy-in to the importance of sustainability, then the link between the provision of advisory services and sustainability outcomes is weakened (Sutherland et al. 2013).

Overall, a positive, long-term interactions between an adviser and a farmer were found to be more important in engendering trust than whether the advisory service provider was from the public, private or a charitable service (Sutherland et al. 2013). In the Australian context, all advisers then should be considered part of the knowledge system (Nettle et al. 2021) and, given the importance of advisers in adoption, scientists need to understand the advisory landscape and make efforts to communicate and collaborate with the range of trusted advisory services in adoption efforts.

E-extension

There is increasing farmer acceptance of the use of information and communication technologies (ICT) for accessing relevant knowledge and engaging in extension programs. In one study, a large proportion of organic producers that accessed an e-extension platform with webinars, videos, information/e-newsletters, and broadcasts, supported changes in practices in some instances (Jenkins et al. 2019), and in another, podcasts and videos were found to have far greater acceptance and use since the COVID-19 pandemic (Chivers et al. 2021). Farmers favour a wide range of information delivery channels; however, the development and introduction of ICTs should be planned and evaluated with the users, implying that local knowledge systems, a focus on experiential knowledge and working in conjunction with local organisations and advisers is needed with e-extension methods (Anastasios et al. 2010; Chivers et al. 2021).

One application of e-extension is in on-line or E-learning. In a study of the use of a feed-management learning package for meat and wool producers, E-learning based on videoconferencing was found to be effective with small-group, interactive sessions with practical ‘homework’ tasks rather than large group, face-to-face workshops (Brown and Bewsell 2009). The shorter, repeated sessions increased reflection and decisions to try the suggested actions. Having access to an ’expert’ was also an important aspect of success. Similarly, Hesse et al. (2019) found that micro-learning courses were popular and effective in colostrum management on dairy farms, creating feelings of confidence and accuracy in work performance and in establishing standing operating procedures. Little difference was found between the efficacy of online training, hands-on training, or a combined approach in training dairy farmers to successfully administer pain control (Winder et al. 2018) with the on-line trained groups rating their technical knowledge slightly lower and taking longer to complete the course, yet on-line training was considered most useful for motivated producers who lacked access to hands-on training.

Another area of application of e-extension is in the development of Apps, virtual experiences and video demonstrations and podcasts (Chivers et al. 2021). In a study examining the use of smart phone Apps in dairy herd management, Michels et al. (2019) concluded that the benefits of App use should be clearly visible and the handling of Apps as simple as possible to ensure Apps are attractive for farmers, irrespective of farmers’ educational background and knowledge. In a pest-management application, webinar conversion to YouTube videos were the most frequently watched videos from all videos produced (Wright et al. 2018). In a cropping application, sugar cane growers and agronomists were receptive to the use of the web-based virtual world Second Life™ platform for modelling virtual conversations between characters (avatars) which integrated relevant climate information and industry-recommended management practices in practical farm decision-making scenarios that were used as ‘discussion support’ tools to improve farm climate-risk decision-making (Reardon-Smith et al. 2015). Other studies noted increased engagement of farmer audiences through digital gamification for improving awareness of Natural Resource Management (NRM) issues (Leitão et al. 2022) and the use of Instagram (Stock 2020) and Facebook, WhatsApp, and Twitter in extension programming (Mueller et al. 2018). Scientists can engage in these platforms to increase the awareness of their work and work with e-extension providers to develop content and channels relevant to farmer needs.

Co-innovation

Co-innovation approaches in agriculture involve shared problem solving across organisations and/or among farmers, commercial parties, government, advisers, researchers and often consumers or other stakeholders (Fielke et al. 2018). This requires formal or informal collaboration and iterative processes of coordination, co-design, and co-creation of solutions. The outputs of co-innovation can either be new business models or value chains, new products and services or changes in practices. There are many similarities in the features of co-innovation methods and facilitated groups (Method 1) and technological development or multi-actor methods (Method 2). Whereas facilitated groups are often farmer-to-farmer oriented and technological development often farmer-to-scientist oriented, the co-innovation method is often directed to pre-commercial stages of the innovation cycle, where defining the problem and deciding on options or pathways for addressing problems is less well defined and where a wider range of stakeholders outside farmers, advisers and agricultural scientists are involved (such as other science disciplines, commercial companies, and consumers). However, it is recognised that there is often little difference among the three methods in the underlying learning processes involved. One of the challenges in reviewing the studies of the impacts from co-innovation approaches is that multiple terms have been used so that co-innovation is described under titles of ‘participatory extension programs’ (Knook et al. 2018) or ‘participatory (action) research (PAR)’ or ‘PAL’ and involving facilitated groups (Method 1) and multi-actor approaches (Method 2).

Focusing on the evidence around co-innovation involving a wide range of stakeholders, participatory approaches broadly have been found to provide formal opportunities for researchers to positively interact with farmers in a transparent and effective way (Burbi et al. 2016). For instance, in climate-change adaptation, a climate learning network involving farmers, agricultural extension specialists, researchers, and climate scientists found that dialogue among researchers and practitioners can enhance agricultural adaptation to climate change (Bartels et al. 2013) and higher rates of adoption of climate-change mitigation practices (Knook et al. 2020). Similarly, Titterton et al. (2011) found that farmers’ knowledge and intuitive understanding of the need to adapt to changing weather patterns were significant when part of learning projects with scientists, and Del Corso et al. (2015) found that group deliberation and the exchange of arguments in learning processes helped disseminate new technical knowledge, enlarge the range of individual experiences, and change the normative frameworks for action in water use. In another study, innovation platforms with producers allowed farmers to contribute directly to extension-program design (Brown et al. 2022). However, in these studies, while the processes were found to influence the questions of scientists, changing the roles of scientists and farmers was more difficult, whereby farmers were often passive listeners to information and scientists as expert presenters, rather than involved in joint exploration and inquiry (Sewell et al. 2014; Berthet and Hickey 2018).

Some authors note that co-innovation is more recently being applied to promote policy objectives, such as in public-good and environmental schemes; however, the application to policy goals can be contradictory in principle and problematic in practice (Bruges and Smith 2008; Knook et al. 2018). Successful co-innovation approaches are described as having built high levels of trust and confidence among participants (Smithers and Furman 2003) and directly engaging with socio-cultural dimensions and institutional arrangements affecting participation (Knook and Turner 2020; Paschen et al. 2021). For instance, farmer participation has been found to be reliant on the institutional arrangements of programs such as the source of funds, trust in, and goal-orientation of government funding agencies; cost-sharing arrangements; the degree of social recognition involved, the role of local and regional level intermediaries, and economic relationships between growers and their processors, contractors and suppliers (Taylor and Van Grieken 2015). Farmer participation was also found to be conditioned by perceptions of the risks of that participation to their local institutions of economic and social cooperation, and what it means to be a ‘good farmer’. It has been suggested that agricultural extension design could be significantly improved by more directly and explicitly engaging with these risks through dialogue and mutual inquiry (Sewell et al. 2017). Co-innovation approaches are also considered to require a conducive policy and institutional environment to produce a more transformative ambition (Turner et al. 2016) and successful co-innovation approaches have been found to hinge on deliberate institutional design (Vedeld et al. 2020). While involvement in co-innovation processes may appear distant from the core work of agricultural scientists and farm advisers, Ingram et al. (2020) noted that facilitation of multi-actor groups and innovation brokering are increasingly considered as core competencies in agricultural innovation, and involvement of scientists in these processes, essential.

Best management practices (BMPs)

Studies of the adoption of best management practices (BMPs) suggest that high-quality best management practice guides are useful to farmers (Baumgart-Getz et al. 2012) but were not, on their own, effective in practice change. However, in combination with other extension approaches, stronger results were realised. For instance, membership of advisory clubs was found to have a statistically significant positive effect for farmers adopting environmental BMPs (Tamini 2011). A study of the adoption of nutrient management plans found social pressure and perceived ease of adopting were strong predictors of intentions to adopt BMPs (Daxini et al. 2019). Because farmers were found to express differing levels of trust in different organisations, different strategies to encourage adoption of recommended management practices were suggested to differing groups of farmers and involving a variety of organisations who are able to meet with and provide advice to landholders (Emtage and Herbohn 2012a, 2012b).

The importance of farmer involvement in establishing BMPs has also been highlighted in contexts including BMPs to deal with climate risks (George et al. 2019) and water quality (Merhaut et al. 2013). Farmer involvement in developing the guidelines and how they were communicated increased trust by farmers and ensured that BMP guides contained practical inputs concerning business and operational issues along with science. Additional methods such as on-site visits and direct on-farm recommendations increased understanding of issues and significantly increased the implementation of appropriate BMPs (Merhaut et al. 2013) as did meetings with advisers, attendance at seminars (Rahelizatovo and Gillespie 2004) and involvement in demonstrations and field-scale projects with ‘model growers’ where multiple practices integrated into whole-farm management are displayed (Hopkins et al. 2007). The role of farm data in farmer learning and monitoring and tracking progress toward BMPs has also been identified as an emerging consideration in the design of BMP programs (Turner and Irvine 2017; Sumner et al. 2018).

BMP guides are a common output from agricultural research; however, researchers need to be engaged in additional processes to see BMPs effectively applied. This will include involving farmers and advisors in designing guides, aligning best-practice recommendations with the ease and level of social acceptance of practices and ensuring that interactive approaches are used to explore complex or uncertain practices in demonstrations and group-learning settings with farmers (Lacy 2011) and explore novel approaches in collecting data associated with the application and impact from BMPs.

Social marketing

Social-marketing approaches focus on addressing the psychological and behavioural dimensions of practice change and adoption mainly through the use of tools and techniques of social comparison. Commonly applied in agricultural environmental schemes to promote positive choices to desirable behaviour change, various studies have explored the strengths and weaknesses of nudges (non-financial and non-regulatory tools) to shift people’s behaviour without closing options or changing economic incentives, or budges (tools which restrict or eliminate choice; Barnes et al. 2013). Marketing approaches, including incentives and demonstrations, are also part of the suite of methods (Heiman et al. 2020). Applying the concepts can be complex. For instance, in a study comparing the voluntary adoption of water-quality management techniques, both ‘budging’ and ‘nudging’ were found to be needed but the mix of approaches may differ on an individual basis and regulations created both divisions and alliances towards attitudinal acceptance and resistance (Barnes et al. 2013). In this study (Barnes et al. 2013), influencing social norms through group sharing of information and demonstrations that raise the visibility of individual farmer practices was suggested to make acceptable behaviour more explicit and to reduce a victimisation culture from environmental schemes. While nudging was found to have a preventive effect on fertiliser use (Peth et al. 2018), social comparison was a disincentive to change, indicating that social comparisons might work for one group but not for others. In a systematic review of the evidence on green nudging (i.e. environmentally focused) interventions, Ferrari et al. (2019) found that almost all studies on farmers provide evidence that ‘green’ nudging can be effective. This is supported by other studies involving nudges to reward individual participation and conveying information on other farmers’ pro-environmental practices (Kuhfuss et al. 2016a, 2016b).

The short-term impact of nudges has been raised as an issue if nudges do not allow for farmers to internalise values and build new habits, such as in environmental management (Mills et al. 2017). The mixed evidence between the role of social comparison highlights the potential for negative effects from poorly designed nudges (Chabé-Ferret et al. 2019) and, more generally, challenges with undertaking ‘controlled’ intervention studies (Main et al. 2012).

While scientists may be hesitant in engaging with social marketing to promote the relevance of their research to different farmer groups, scientists can consider the principles and reflect on the signals that farmers would need so as to consider changing practices and the extent to which social comparisons can be beneficial.

Across the studies of extension methods, it was the application of combinations of methods that were a common feature in successful adoption outcomes. The following section reports on two case studies from the Australian dairy industry that describe the design of extension approaches using multiple methods and their effect.

Case study 1: how extension methods influenced uptake of white grain sorghum (WGS) for silage on Queensland dairy farms

Case study 2: grazing management support in Tasmanian dairy systems