Datasets:

---

ExponentialScience

/

DLT-Scientific-Literature

like 0

Follow

Exponential Science 4

ERROR: type should be string, got "https://doi.org/10.1007/s10460 023 10451 5\n\n# The adoption of conservation practices in the Corn Belt: the role of one formal farmer network, Practical Farmers of Iowa\n\n**[L. Asprooth[1] · M. Norton[2] · R. Galt[3,4]](http://orcid.org/0000-0002-9802-9252)**\n\nAccepted: 14 March 2023 / Published online: 3 May 2023\n© The Author(s) 2023\n\n**Abstract**\nSubstantial evidence has shown that involvement in peer-to-peer farming networks influences whether a farmer decides to\ntry a new practice. Formally organized farmer networks are emerging as a unique entity that blend the benefits of decentralized exchange of farmer knowledge within the structure of an organization providing a variety of sources of information and\nforms of engagement. We define formal farmer networks as farmer networks with a distinct membership and organizational\nstructure, leadership that includes farmers, and an emphasis on peer-to-peer learning. This study complements existing ethnographic research on the benefits of organized farmer networking by examining farmers in one longstanding formal farmer\nnetwork, Practical Farmers of Iowa. Using a nested, mixed-method research design, we analyzed survey and interview data\nto understand how participation and forms of engagement in the network are associated with the adoption of conservation\npractices. Responses from 677 farmers from a regular member survey disseminated by Practical Farmers of Iowa in 2013,\n2017, and 2020 were pooled and analyzed. GLM binomial and ordered logistic regression results indicate that greater participation in the network, particularly through in-person formats, has a strong and significant association with greater adoption\nof conservation practices. Logistic regression results show that building relationships in the network is the most important\nvariable for predicting whether a farmer reported adopting conservation practices as a result of participation in PFI. In-depth\ninterviews with 26 surveyed member farmers revealed that PFI supports farmers to adopt by providing information, resources,\nencouragement, confidence building, and reinforcement. In-person learning formats were more important to farmers relative\nto independent formats because they were able to have side conversations with other farmers, ask questions, and observe\nresults. We conclude that formal networks are a promising way to expand the use of conservation practices, particularly\nthrough targeted efforts to increase relationship building in the network through face-to-face learning opportunities.\n\n**Keywords Peer-to-peer networks · Formal farmer networks · Adoption · Agriculture · Conservation practices · Corn Belt**\n\n\n\n- L. Asprooth\nlsasprooth@ucdavis.edu\n\nM. Norton\n\nnorton@iastate.edu\n\nR. Galt\n\nregalt@ucdavis.edu\n\n1 Geography Graduate Group, University of California, 129\nHunt Hall, One Shields Avenue, Davis, CA 95616, USA\n\n2 Bioeconomy Institute, Iowa State University, 1140\nBiorenewables Research Lab, 617 Bissel Road, Ames,\nIA 50011, USA\n\n3 Department of Human Ecology, University of California,\n1309 Hart Hall, One Shields Avenue, Davis, CA 95616, USA\n\n4 Agricultural Sustainability Institute, University of California,\n143 Robbins Hall, One Shields Avenue, Davis, CA 95616,\nUSA\n\n\n**Abbreviations**\nAIC \u0007Akaike Information Criterion\nEQIP \u0007Environmental Quality Incentives Program\nGLM \u0007Generalized Linear Model\nPFI \u0007Practical Farmers of Iowa\n\nAgriculture in the Corn Belt is dominated by industrial farming practices that have resulted in serious damages to the\nhealth of farmers, the environment, and rural communities.\nRegenerative agricultural practices that increase cropping\ndiversity while using fewer chemical inputs and conserving\nsoil and water resources are proven to be equally as productive as conventional systems while delivering greater economic and environmental benefits (Cruse et al. 2010; Davis\net al. 2012; Dainese et al. 2019; Hunt et al. 2020; Tamburini\net al. 2020). Despite these benefits, conservation practices\n\nV l (0123456789)1 3\n\n\n-----\n\nare still relatively rare in the Midwest. In Iowa, the US state\nthat produces the most corn, hogs, and eggs (Thessen et al.\n2020), only 11% of cropland operations use cover crops,\nand 34% use conservation tillage, two critical conservation\npractices (USDA NASS 2017).\nA farmers’ decision to adopt a conservation practice is\ninfluenced by a complex and interconnected set of social,\neconomic, geographical, environmental, and agroecological\nfactors which have been documented in a large empirical\nliterature (Prokopy et al. 2008, 2019; Baumgart-Getz et al.\n2012; Carlisle 2016). Of these factors, how and through\nwhom information about a conservation practice reaches\na farmer can impact whether they choose to adopt. Diffusion of innovation theory (Rogers 2003) and social network\ntheory (Prell 2012) propose that an individual’s social network is central to the diffusion of new ideas and practices.\nInformation about innovations is spread through social\nnetworks and the presence of specific individuals or actors\nwithin one’s network, and the information flowing through\nit, provide additional explanation as to whether an individual\nadopts a new practice (Wellman 1983; Valente 1996; Rogers 2003).\nWithin an individual’s network, connections with peers\nare particularly consequential to whether one adopts a new\npractice. Explained through the theory of social learning,\nor transmitting knowledge from peer-to-peer through interpersonal communication, individuals learn best by observing and replicating the behavior of others (Bandura 1977).\nInformation diffused through interpersonal channels allows\nfor the individual to observe and validate results, and to ask\nquestions to decrease uncertainty about the practice (Ban\n1981; Rogers 2003). As a result, an individual is more likely\nto adopt a new practice when information about the practice\ncomes from someone they know and trust and who may have\nalready tested the practice (Foster and Rosenzweig 1995;\nRogers 2003; Granovetter 2005). These theories are supported by a large literature showing that farmers who are\nbetter connected to other adopters and grassroots organizations promoting adoption are more likely to adopt new\npractices (Prokopy et al. 2008; Baumgart–Getz et al. 2012;\nCarlisle 2016).\nIn addition to social learning, farmer networks provide\nmutual support and motivation to go against established\nnorms in agriculture (Bell 2004; Kroma 2006; Gosnell et al.\n2019). More broadly, participating in a social network can\ninstill social pressure to conform to the norms of the group\n(Ajzen 1991; Montgomery and Casterline 1996; Hogset and\nBarrett 2010). Social networks of farmers are particularly\nimportant for diffusing information on alternative agricultural practices that may not be available through traditional\nresearch and extension (Kroma 2006), and for sharing local\nknowledge about conservation practices that are more information-intensive and site-specific compared to simplified\n\n## 1 3\n\n\nindustrial practices (Hassanein and Kloppenburg 1995; Lyon\net al. 2011; Iles and Marsh 2012).\nSocial networks in agriculture can take many forms and\nmuch existing research has focused on personal communication networks of farmers through social network analysis\n(Garbach and Morgan 2017; Levy and Lubell 2018) and\nethnographies of networks (Hassanein 1999; Warner 2007;\nWypler 2019). Fewer studies have examined organized or\nformalized farmer-to-farmer networks, which we identify\nas a unique construct. In this paper, we look specifically at\nformal farmer networks to understand the significance of\nthis type of peer organization in catalyzing farmers to adopt\nconservation practices in the Corn Belt.\n\n### Formal farmer networks\n\nA handful of studies have examined the role of formal agricultural networks in the context of social change in agriculture (Bell 2004; Trauger 2009; Barbercheck et al. 2014;\nPape and Prokopy 2017; Šūmane et al. 2018). Within these\nstudies, no shared definition of what constitutes a formal\nfarmer network exists and only two differentiate between\ntypes of farmer networks. Pape and Prokopy (2017) define\nformal and informal social networks broadly where formal\nnetworks are structured organizations with defined roles\nand purposes (e.g., a club, church body, or team of coworkers) and informal networks are unstructured interrelations\nbetween individuals acting on their own terms, without centralized organization or planning (e.g., family, neighbors,\nand friends). Šūmane et al. (2018) specify formal knowledge\nnetworks in agriculture as formal institutions with a structured agenda, that receive public funding, and that circulate\nformal knowledge (e.g., universities, advisory services) as\nopposed to informal knowledge. Informal knowledge networks in agriculture are “fuzzier” relations between community members, family and personal relations that occur\nas part of a farmers’ daily routine (Šūmane et al. 2018, 237).\nAcross these definitions, organizations are considered formal, while exchanges with personal relations are considered\ninformal.\nYet, neither capture the characteristics of formalized\npeer-to-peer farmer networks, emerging as a unique entity\nthat blends the benefits of the decentralized exchange of\nfarmer knowledge within a structured organization. We\nbriefly examined farmer organizations and associations in\nthe US and found formal farmer-to-farmer networks to be\na particular type, with the following shared characteristics:\n(1) a defined membership and organizational structure (2)\nleadership (e.g., board of directors) that includes farmers\nand (3) an emphasis on peer-to-peer learning. Of the organizations we examined, nine met these criteria including:\nGrassworks, Iowa Organics Association (IOA), Nebraska\n\n\n-----\n\nSustainable Agriculture Society, Northeast Organic Farming Association (NOFA), Ohio Ecological Food and Farming Association (OEFFA), PASA Sustainable Agriculture,\nPractical Farmers of Iowa, Southeastern African American\nFarmers’ Organic Network (SAAFON), Sustainable Farming\nAssociation of Minnesota, and most producer-led watershed\nprotection groups.\nAcross all formal farmer networks we examined, membership is state or regionally based and representative of\nvarious farming enterprises. Some of the networks (IOA,\nNOFA, OEFFA, SAAFON), have current or historic ties to\norganic-specific production. All of the networks have been\nestablished for over thirty years with the exception of IOA\nand SAAFON, both founded in 2006. The longest standing organization is NOFA (1971), followed by OEFFA\n(1979), and PFI (1985). PFI has the most robust programming offered to members; however, all formal networks\noffer some combination of programming and resources that\ninclude annual conferences; field days; written educational\nresources; mentorship or apprenticeship opportunities; email\nlistservs; and/or classifieds to find or sell equipment, livestock, or post positions. Formal networks take on varying\ndegrees of policy advocacy. Some organizations are proactive in farm policy, such as IOA or OEFFA, and others\nremain apolitical, such as Grassworks. While PFI advocates\nfor policies promoting working lands conservation, other\ncore areas of programming such as farmer-to-farmer learning events, cost-share, and on-farm research receive priority\nover direct policy work (Personal communication with PFI\nMarketing and Communications Director, December 2021).\nWhile farmer cooperatives often result in the sharing of\ninformation and comradery inherent in networks, we do not\nconsider them to be formal farmer networks as we define\nthem, since cooperatives’ objectives are largely economic—\ne.g., to improve bargaining power (Schram 2010)—as\nopposed to social. While we attempt to characterize formal\nfarmer-to-farmer networks, we recognize that networks are\ndiverse and suited to the needs and resources of the community the network serves (White 2021).\nFormal farmer networks are particularly important in\nthe transition to sustainable agriculture because of their\nrole in bridging two valuable forms of knowledge: farmers’ experiential knowledge shared between peers and formal scientific knowledge shared from professionally trained\nexperts. Before the onset of public agricultural science, new\nknowledge was created through the cumulative experience of\nfarmers and exchanged among themselves. This local knowledge has been largely devalued by the scientific community\nand loose networks of farmers sharing information have\ntraditionally operated separately from agricultural extension (Kloppenburg 1991; Hassanein 1999; Kroma 2006).\nAgricultural extension primarily uses a top-down approach\nto disseminating scientific information from traditional\n\n\n“experts” to farmers (Warner 2007; Lubell et al. 2014).\nCriticisms of traditional agricultural extension have led to\nalternative research and extension models that involve the\nparticipation of farmers and other stakeholders (Chiffoleau\nand Desclaux 2006; Warner 2007; Healy and Dawson 2019)\nand the use of existing grower networks to extend agricultural science (Hoffman et al. 2015).\nStill, these methods place the scientist and the extension professional at the center of the production of and\ndissemination of knowledge in agriculture, while research\nindicates that farmers are more likely to test or adopt innovations when information about the innovation is transmitted\nthrough peers (Rogers 2003). Formal famer networks are a\npromising collaborative way to produce and share both scientific and farmer knowledge because they privilege farmers’ knowledge and learning in the process of dissemination.\nIn formal farmer networks, membership fees, grants, and/\nor donations help support network programming. Leadership facilitates the sharing of information by linking farmers\nwith each other and by linking farmers with outside actors\nand institutions. Farmers connect through newsletters, email\ndiscussion lists and social media, and through social learning activities such as on-farm research trials and field days.\nIn many of these formats, farmers are in the teaching role\nand farmer-generated knowledge is shared between participants. Leadership ties in university researchers and extension staff, conservation professionals, and other agriculture\nspecialists to the network through events such as webinars\nand workshops, often conducted alongside farmer presenters. Network leadership may provide technical assistance,\nnavigate cost share opportunities, work with farmers to trial\nnew practices, analyze data from field trials, produce reports,\nand distribute informational material members can access\nindependently. The result is a diversity of both sources\nof expertise and ways of learning within the network that\nincludes farmer and “expert” generated knowledge, and inperson and independent learning pathways.\nUsing Rogers’ (2003) explanation of diffusion, formal\nnetworks can be considered a hybrid of centralized and\ndecentralized information exchange. In a centralized diffusion system, the decision to begin to diffuse an innovation is made by formal experts and spread downwards. In\na decentralized system, information, often developed from\npractical experience, is spread horizontally among potential\nadopters. Formal farmer networks form a synergy between\nthe benefits of peer exchange and access to information produced in institutions that may be missing from informal peer\nnetworks. With ties to both institutional actors and a network\nof farmers, formal farmer networks bridge social and institutional knowledge, playing a “boundary spanning” role (Levy\nand Lubell 2018).\nIn addition to experts sharing information with farmers, and farmers sharing information with each other,\n\n## 1 3\n\n\n-----\n\nformal farmer networks provide opportunities for farmers\nto share feedback and information to experts, completing\na feedback loop. In some formal farmer networks, farmers\nshare their knowledge with agricultural researchers and\nprovide feedback to agricultural scientists and extension\nagents on the performance of different conservation and\nproduction practices at events, on email discussion lists,\nand through the dissemination of research reports. Farmers\nalso provide feedback to network leadership which drives\nprogramming decisions and signals to how to better assist\nmembers.\nDespite the unique attributes of formal farming networks, few studies have quantitatively examined the role\nof participation in a formally organized peer-to-peer network in the adoption of conservation practices in the US.\nWithin the body of literature predicting the adoption of\nbest management or conservation practices, some studies\ninclude the degree of participation in community organizations (Belknap and Saupe 1988) and in agricultural or\nconservation organizations broadly (Korsching et al. 1983;\nBates and Arbuckle 2017), which are found to be positively associated with adoption. Others examine the degree\nof farmer-to-farmer networking (Wilson et al. 2014), perceived effectiveness of learning from others including\nneighbors (Dunn et al. 2016), and importance placed on\nneighbors as an information source (Thomas et al. 1990),\nwith varying results.\nThe two studies that consider a form of formal peer\nnetworks in quantitative studies of adoption are by Barbercheck et al. (2014) and Pape and Prokopy (2017). Barbercheck et al. (2014) looked at participation in various\nagricultural organizations among female farmers in the\nNortheastern US and the use of conservation practices.\nThey found participation in an organized women’s farming\nnetwork to be positively associated with the adoption of\none of 11 measured cropping system conservation practices—manure incorporation after application (Barbercheck et al. 2014). Pape and Prokopy (2017) found that\nfarmers in two formal networks—defined as structured\norganizations with defined roles and purposes—in Indiana use a range of conservation practices more than nonnetwork farmers, most of the differences significant at the\n0.1% level. They also found a positive and significant relationship between the length of time a farmer participated\nin a network and self-reported improvements to their nutrient management practices (Pape and Prokopy 2017). Our\nresearch adds to the limited body of evaluative evidence\nof the role of farmer networks in the US in two key ways:\n1) we focus on formally organized farmer networks, complementing existing ethnographic research on their role in\nfarmer decision making and 2) we measure the intensity\nof participation for a more nuanced understanding of the\nrole of the network.\n\n## 1 3\n\n\n### Study context\n\nWe studied farmers within one longstanding formal farming network in the Midwest, Practical Farmers of Iowa\n(PFI), to understand how participation in the network\nrelates to the adoption of conservation practices. PFI\nuses “farmer-led investigation and information sharing to\nhelp farmers practice an agriculture that benefits both the\nland and people” (PFI 2022a). The group has a history in\nthe region of facilitating open dialogue between farmers\nto develop solutions to make farming more sustainable\n(Bell 2004). Established in 1985, a time when farmers\nwere under great economic pressure, PFI was formed\nby a group of like-minded farmers who came together\nto increase diversity on their farms while also reducing\ninput costs. Members learned from one another and organized to conduct randomized, replicated on-farm research\nto improve their profitability, efficiency, and stewardship\n(PFI 2022a). PFI has grown to an organization with over\n6000 members and 30 staff with program areas that extend\nbeyond the initial row crop and livestock enterprises to\ninclude horticulture, beginning farmers, habitat restoration, and policy. From the summer of 2019 to the spring of\n2020, PFI’s events and workshops had 2750 total attendees\n(Personal communication with PFI Membership Manager,\nJuly 2021). Its members are primarily located in Iowa but\ninclude members across the country and internationally.\nPFI member farmers, including both row crop and other\ntypes of member farmers, are different from the broader\nfarming population in the region in that, on average, they\nare younger, have larger farms, and have lower rates of\nland ownership (USDA NASS 2017; PFI 2020). Member\nfarmers have greater-than-average adoption of conservation practices compared to other farmers in the region.\nFor example, only 11% of farmers across the Corn Belt\nreported using cover crops according to the most recent\nUSDA agricultural census (USDA NASS 2017). Yet, in\n2020, 67% of all PFI member farmers report using cover\ncrops (PFI 2020). Similarly, PFI farmers report a higher\nuse of reduced or no-till practices and enrollment in government conservation programs compared to the larger\nfarming population in the region (USDA NASS 2017;\nPFI 2020).\n\n### Research questions and hypotheses\n\nUsing data from the three most recent PFI member surveys in 2013, 2017, and 2020, and interviews with member farmers, we ask the following questions: (1) What is\nthe role of the PFI farming network in the adoption of\n\n\n-----\n\nconservation practices amongst its members? (2) What\ndrives successful participation in the network? and (3) Do\nthe various ways member farmers participate and learn\nwithin the network matter for the adoption of conservation practices?\nKnowing that the frequency and depth to which an individual interacts with a network increases the likelihood of\nthe diffusion of information among participants (Cheng\n2021), we measure participation for our first question by\nthe degree of involvement in PFI events and workshops and\nengagement with PFI materials. The second question measures successful participation in the network by whether a\nfarmer reported adopting a conservation practice as a result\nof participation in PFI. For our third question, we test\nwhether in-person ways of participating traditionally associated with social learning in agriculture such as field days,\nsocials, and workshops have a stronger relationship with\nadoption compared to independent ways of participation\nwithin the network more unique to formal farmer networks\nsuch as email discussion lists, participatory research reports,\nwebinars, and podcasts.\nGiven the unique attributes of formal farmer networks\nto produce and share both scientific and farmer knowledge,\nand to provide opportunities for both social and independent\nlearning, we hypothesize that participation in the network\nwill be positively associated with the adoption of conservation practices. We theorize that engaging in a formal farmer\nnetwork provides impactful opportunities to learn about\nconservation practices from other farmers and agricultural\nprofessionals while receiving support and validation, thereby\nincreasing the likelihood one will adopt.\nWe also hypothesize that in-person and independent ways\nof engaging in the network do not have the same relationship\nwith adoption. PFI offers a range of in-person events and\nways of engaging remotely and accessing materials independently. The learning pathways a farmer can access independently within the PFI network can still be considered a blend\nof social and independent learning since they are created by\nor use data or information from other farmers. Nonetheless,\nwe make a distinction between in-person participation and\nindependent participation, hypothesizing that face-to-face\ninteractions are more important for adoption than independent ways of learning and engaging within the network.\nExisting theory of social learning and the diffusion of\ninnovation holds that face-to-face, interpersonal channels of\ncommunication are more likely to influence an individual’s\ndecision making compared to impersonal sources due to the\nability to observe and ask questions about practices (Bandura 1977; Ban 1981; Rogers 2003). This type of learning\nis especially important for conservation practices that tend\nto be more complex and knowledge-intensive compared to\nconventional practices (Carlisle et al. 2019; Laforge and\nLevkoe 2021). Moreover, empirical literature in agriculture\n\n\nshows that farmers place more value on social learning compared to institutional or top-down learning (Hoffman et al.\n2015; Laforge and McLachlan 2018) and that social learning\n(Garbach and Morgan 2017; Singh et al. 2018), and particularly face-to-face ways of receiving information (Bates and\nArbuckle 2017), have been shown to be a strong predictor\nof the adoption of conservation practices.\nResults from this study can provide insights to other\norganized farmer networks in the region, inform the programming and allocation of resources amongst agriculture\nprofessionals and policy makers, and inform academics and\nextension professionals of the role of organized networks\namong the many factors determining whether a farmer\nchooses to use a conservation practice.\n\n### Methods and data\n\n#### Farmer survey\n\nThe quantitative analysis used data collected by PFI in\n2013, 2017, and 2020 through their member survey, a survey distributed to all members every three to four years.\nAccess to this data was given to the researchers by the staff\nof PFI, with identifying information removed. The member\nsurveys were disseminated via email and collected information about members’ farms, farming practices, goals,\nfeedback about past PFI programming, and future priorities\nfor PFI. Members were encouraged to complete the survey\nthrough an online form, but were also given the option to\ncomplete a paper copy. The 2020 survey was disseminated\nin December of 2019 and responses were collected until\nMarch, 2020; the 2017 survey was disseminated February\nof 2017 and responses collected until July; and the 2013 survey was disseminated in September of 2013 and responses\nwere collected until January of 2014. Over the course of\neach campaign, members were sent between one to four\nemail reminders and staff posted general reminders in the\ne-newsletter and on the email discussion list. Incentives were\noffered to take the survey in the form of drawings for PFI\npromotional materials.\nIn 2020, 785 of 1941 member households responded for a\nresponse rate of 40%, in 2017, 650 of 1431 member households responded for a response rate of 45%; and in 2013, 660\nof 1286 member households responded for a response rate of\n51%. Ninety-five percent of surveys in 2020 were collected\nbefore the onset of the Covid-19 pandemic as of March 11,\n2020. Thus, the 2020 survey does not likely reflect changes\nin participation due to the pandemic. Cross sectional data\nfrom the three different survey periods were pooled and the\nmost recent survey was used for individuals that completed\nthe survey multiple years.\n\n## 1 3\n\n\n-----\n\nOnly members who reported that they were currently\nfarming were included in the analysis. We did not include\nother types of PFI members, such as farmland owners who\ndo not operate their land, aspiring farmers, or friends of\nfarmers. Farmers cultivating corn and/or soybeans, among\nother crops, are the largest membership contingency within\nPFI and constitute 53% of the farmers who responded to the\nmembership survey. The other 47% of respondents cultivated\na range of vegetables, fruits, flowers, hay, and livestock. We\nexcluded farmers who had only been a PFI member for one\nyear or less at the time of the survey, since we assume that it\nwas less likely that PFI would have been the driver of their\nuse of conservation practices. We judge a time lag of one\nyear or greater to be a sufficient amount of time to adopt the\nmajority of the practices measured.\nAfter exclusions, the final sample size of the pooled\ncross-sectional data was 677 member farmers including 118\nfarmers in 2013, 149 in 2017, and 410 in 2020. The network\nattracts members throughout the US, although most live in\nIowa and surrounding Corn Belt states. The states represented by farmers in the sample include Alabama (1), Arizona (1), California (1), Colorado (1), Iowa (590), Illinois\n(15), Indiana (1), Kansas (2), Kentucky (1), Michigan (1),\nMinnesota (23), Missouri (4), Montana (1), Nebraska (11),\nNew Jersey (1), Ohio (6), Ontario Canada (1), Pennsylvania\n(1), South Dakota (5), and Wisconsin (10). Given that the\nvast majority of farmers in the sample reside in Corn Belt\nstates, Iowa in particular, we use this region to compare our\nsample.\n\n#### Regression variables\n\nThe first dependent variable, the adoption of conservation\n_practices index, is a composite index ranging from 0 to 6_\nbased on the sum of farmer-indicated use of the following\npractices (each are binary variables coded as yes = 1 and\nno = 0): buffer or filter strips (riparian, grass, prairie), conservation tillage (no till, strip-till, ridge-till, or reduced till),\ncover crops, flame or electric weeding, low-use of synthetic\ninputs (organic farming or no use of synthetic inputs), and\nsustainable grazing (rotational grazing or high-density/adaptive multi-paddock grazing). A score of 6 indicates use of all\nconservation practices listed and a score of 0 indicates no\nuse of the practices listed. The distribution of the adoption\n_of conservation practices index is displayed in Fig. 1._\nThese practices were derived from the member survey\nand are key practices elevated by PFI, clearly improving\nof conservation on farms, and applicable to all farms in the\nsample. While conservation practices have different levels of environmental and agronomic benefit (Reimer et al.\n2014), we do not weight each practice within the adoption\nindex. A lack of comparable data on the relative environmental impact across each practice, and the varying impact\n\n## 1 3\n\n\n**Fig. 1 Distribution of the dependent variable** _adoption of conserva-_\n_tion practices index using the pooled member survey data from 2013,_\n2017, 2020 (n = 677)\n\ndepending upon the biophysical conditions of each farm\nwhich are spread across a heterogenous geographical area,\nwould make it difficult to do so. While some studies have\nattempted to differentially weight certain practices (Napier\nand Tucker 2001; Napier and Bridges 2002), other studies\nhave used composite indexes similar to ours when analyzing adoption of best management practices in agriculture\n(Thomas et al. 1990; Park and Lohr 2005; Filson et al. 2009;\nCampbell et al. 2011).\nThe independent variable general participation index\nmeasures a farmer’s participation in a range of events and\nresources offered through the network. We created an index\naccording to the number of events attended and engagement\nwith resources reported by the farmer completed in the past\n12 months. In addition, we test the association of ways of\nparticipating and learning in the network and the adoption\nof conservation practices. We do so by defining two types\nof participating and learning—in-person and independent.\nIn-person ways involve face-to-face social learning and independent ways allow the farmer to learn on their own through\nPFI materials.\nWe disaggregated the general participation index into\ntwo discrete indexes of in-person and independent ways of\nparticipation (Table 1). The in-person participation index\nincludes the following categories: attended 1–2 events,\nattended 3 or more events, and acted as a farmer-leader.\nActing as a farmer leader includes at least one of the following: conducted on-farm research, hosted a field day,\ntalked to the media, served on a committee, or served as\na mentor. This category is weighted doubly to account for\ngreater integration into the network. The independent par_ticipation index includes: visited the PFI website, read the_\nPFI weekly e-newsletter, and used the PFI e-mail discussion\nlist. Each category that was checked is given a 1 (except for\n\n\n-----\n\n**Table 1 Main independent variable composition**\n\nIncluded Variables General participation index\n(0–7)\n\n\nSurvey question\n\n\nIn-person participation index\n(0–4)\n\n\nIndependent\nparticipation Index\n(0–3)\n\n\nAttended 1–2 events Y Y The following are events\n\nAttended 3 or more events Y Y and resources offered\n\nthrough PFI. Please select\n\nActed as a farmer-leader (doubly Y Y\n\nany you have done in the\n\nweighted)\n\npast 12 months\n\nVisited the PFI website Y Y\nRead the PFI weekly e-newsletter Y Y\nUsed the PFI e-mail discussion list Y Y\n\n\nfarmer-leader, which is weighted doubly as noted above)\nand summed for a final score, ranging from 0 to 7 for the\ncombined index, 0–4 for the in-person index, and 0–3 for\nthe independent index.\nEach model includes additional covariates also shown to\nbe associated with adoption in review studies (BaumgartGetz et al. 2012; Carlisle 2016; Prokopy et al. 2019) and\nnarrowed their inclusion to the best model fit according to\nthose that increased the R[2] value. Of the predictor variables\ngleaned from the member survey, age of primary member\nfarmer, total farm size in acres, participation in EQIP (Environmental Quality Incentives Program), participation in a\n_watershed group, proportion of acres owned, and years in_\n_PFI, were included in the model._\n_Age of primary member is the age of the person who_\ncompleted the survey. Farm size is the combination of acres\nowned and rented that are operated by the member farmer.\n_Participation in EQIP measures present or past participation_\nin EQIP. Participation in a watershed group measures current or past involvement in a watershed project. Watershed\nprojects are farmer-led, governmentally funded cost-share\nprograms with the aim of improving water quality through\nthe use of conservation practices. Proportion of acres owned\nmeasures the proportion of acres owned and operated to\ntotal acres operated (rented and owned) and ranges from 0\nto 1. The number of years of membership in PFI (years in\n_PFI) was included to control for increased adoption due to_\nlength of time in the network and to isolate the association\nof increased participation and the adoption of conservation\npractices. A dummy variable for year in which the survey\nwas distributed was included as a fixed effect to control for\nvariation across years such as fluctuations in government\nconservation program supports, commodity prices, and\nweather which may have impacted adoption. The state in\nwhich the farmer is located (state) was included to control\nfor distance to in-person events, as over 95% of PFI events\nare held in the state of Iowa (Personal communication with\nPFI Membership Manager, July 2021). Including state fixed\neffects also help to control for time-invariant environmental characteristics at the state-level that may influence the\n\n\nadoption of conservation practices such as climate and soil\ntype.\nThe second dependent variable is adopted a conservation\n_practice(s) as a result of participation in PFI.[1] We selected_\nindependent variables available through survey data to predict whether a farmer adopted conservation practice(s) as\na result of participation in PFI, including age of primary\n_member, whether the farmer reported having formed rela-_\ntionships through the network (formed relationships),\nwhether the farmer reported having felt a sense of community through the network (felt a sense of community), years\n_in PFI, state, and year fixed effects._\n\n#### Regression analysis\n\nUsing R statistical software (Version 4.2.1), we created a\nseries of regression models to understand the relationships\nbetween the adoption of conservation practices index and\nthe three key independent variables described in Table 1:\ngeneral, in-person, and independent participation in PFI.\nThe three main independent variables were included in\nseparate models due to theoretical and observed issues of\nmulticollinearity. For general participation, we ran bivariate\nregressions between the adoption of conservation practices\n_index and the general participation index, a multivariate_\nregression with covariates and fixed effects, and a multivariate regression with covariates and fixed effects including an interaction between the general participation index\nand the number of years in PFI. The interaction was chosen\ndue to the likelihood that general participation in PFI had a\n\n1 In 2020 and 2017 the survey question read: “Have you changed any\nof the following as a result of your participation with PFI: conservation practices (reduce inputs, diversifying rotations, cover crops etc.”\nIn 2013 the survey question read: “My participation with PFI helped\nme improve stewardship on my farm” and noted “Stewardship can be\nanything you do to leave your farm in better shape for future generations, such as improvements that affect energy use, soil, water, and\nbiodiversity or field conservation.” While the questions are different,\nwe believe that the intent is similar enough to be pooled in one variable for analysis.\n\n## 1 3\n\n\n-----\n\ndifferent relationship with adoption depending on the number of years the member had participated in the network. For\nin-person and independent modes of participation, we ran\nmultivariate regressions with covariates and fixed effects. In\naddition, we ran a multivariate regression to understand the\nfactors that predict whether a farmer adopted conservation\npractice(s) as a result of participation in PFI. We standardized the independent variables in all models to compare the\nrelative importance of each while holding all other variables\nin each model constant.\nWe estimated the association with the adoption of conser_vation practices index and the independent variables in the_\nmodel using two model types: a Generalized Linear Model\n(GLM) based on the binomial probability distribution and\nan ordered logistic model. The binomial GLM model best\nfits data with a dependent variable that is a discrete count\ncomposed of positive integers with an upper bound (McCullagh 2019). Unlike other count models, the binomial GLM\nmodel predicts the sum of the number of “successes” of a\npractice being adopted out of a set number of independent “trials” and will therefore not predict values outside the\nrange of the adoption index. Because this data may have\nminor violations of independence of trials from the same\nindividual, we conducted an ordered logistic regression to\ntest the robustness of the binomial GLM results. For the\nordinal logistic regression, we treated the dependent variable, adoption of conservation practices index, as ordered\nand discrete, allowing for the distances between intervals\nin the dependent variable to vary (Pampel 2021). A binary\nlogistic regression model is used to predict the factors associated with the variable adopted conservation practice(s) as\n_a result of participation in PFI, appropriate for models with_\na binary dependent variable.\nResults for each model are expressed as the likelihood of\nbeing in a higher level of adoption with a unit increase in\neach independent variable (Models 1–6 and 8–11), or the\nlikelihood of having adopted conservation practice(s) as a\nresult of participation in PFI (Model 7) controlling for all\nother independent variables. We calculated a McFadden’s\npseudo R[2] to measure the predictive power of each model,\nan appropriate measure for the generalized linear, ordered\nlogistic, and binary logistic regression with no R[2] value. To\nverify that our models did not violate assumptions and to\nimprove confidence in our model selection, we performed\na series of examinations for each model type. A study of\ncorrelation coefficients showed no signs of multicollinearity among the independent variables in any of the models.\nResidual versus fitted scatter plots indicated that the relationships between the independent and dependent variables\nwere linear and the variance of the error terms constant.\nFor the ordinal logistic regressions, a test of parallel lines\n(Brant 1990) showed no indication of a varied effect of the\nindependent variables across levels of the adoption index,\n\n## 1 3\n\n\nexcept for the variables farm size and year = 2013. Given\nthat proportional odds held for the key independent variables\nacross the fitted models, we believe that the dependent variables are more appropriately treated as ordinal compared to\nnominal. No evidence of lack of model fit was found using\na Wald Chi-squared test across all models, nor for the ordinal logistic models using a Lipsitz, Hosmer–Lemeshow and\nPulkstenis-Robinson tests (Fagerland and Hosmer 2016).\n\n#### In‑depth interviews\n\nIn-depth, semi-structured interviews with a subset of the 410\nfarmers who filled out a PFI member survey in 2019–2020\ntook place from July to December of 2020. Nesting our\nresearch design, we sampled farmers to interview from the\nmember survey using a stratified, maximum variation sampling technique (Patton 2002) to understand perspectives\nfrom a variety of member farmers. First, we limited the sample to those growing corn and or soybeans to focus on the\ngroup of farmers with the greatest environmental impact on\nthe Midwestern landscape and representative of a majority\nof PFI’s membership. Next, we divided surveyed farmers\ninto four groups based on two axes, (1) their score on the\nadoption of conservation practices index and (2) their score\non the general participation index, and randomly sampled\nwithin each of these four groups. Finally, after finding our\nsample to be composed of primarily men, we recruited and\ninterviewed two additional women farmers through snowball\nsampling. Thirty-five farmers were contacted to participate\nin an interview; 71% agreed and were interviewed. In total,\nwe interviewed 22[2] current members and 4 previous member\nfarmers who had not renewed their membership. The farmers we interviewed represent 1.3% of member households\nin the PFI network.\nFarmers were told in the beginning of the interview\nthat PFI was financially supporting the study, but that the\ninterviewers were independent and no information from the\ninterviews would be published or relayed back to PFI staff\nwithout all individual identifying information removed.\nInterviews were conducted via Zoom or in person depending on location and farmer preference, and lasted between\n45 and 90 min. Questions focused on the factors that allowed\nthem to transition to conservation practices, preferred learning and information sources, and experience with and opinions on the PFI network.[3] Interview data were analyzed\nusing NVivo software and coded according to an inductive\napproach that identified reoccurring themes.\n\n2 Two farmers were interviewed together, and their responses are\ncounted as one in the findings because they took turns answering\nquestions.\n\n3 Interview protocol is available on request from the corresponding\nauthor.\n\n\n-----\n\n### Results\n\n#### PFI and the adoption of conservation practices\n\n**Regression results**\n\nThe dependent variable, adoption of conservation practices\n_index, is displayed disaggregated by practice and by the_\nextent of participation in the PFI network in Fig. 2. Each\npractice that was included in the index is displayed by the\nproportion of member farmers that reported use. Low use\nof synthetic inputs and use of cover crops had the highest\nadoption among PFI farmers, at 75% and 61%, respectively.\nTo understand how the adoption of each practice varied by\nthe extent of participation in the network, we divided farmers into two groups, high and low participators, where high\n(n = 292) participators engaged in 4–7 of the activities listed\nand low participators (n = 385) engaged in 0–3. For all conservation practices, high participators are also those with\nhigher rates of adoption, on average (Fig. 2).\nFigure 3 shows our main independent variable, general\n_participation index, disaggregated by each type of activity_\nor resource and by level of adoption of conservation practices. The most common way PFI member farmers participated in the network was through the PFI website. Over\n80% of farmers said that they visited the website in the past\n12 months. The PFI website includes resources organized\nby program area, PFI publications including the quarterly\nmagazine, annual report, the Cooperators’ Program on-farm\nresearch reports, a member portal, and information about\nupcoming events. Farmers also frequently reported reading\nthe PFI weekly e-newsletter (75%), attending 1–2 events per\nyear (66%), and using the PFI email discussion list (53%).\nThe newsletter shares information about upcoming events,\n\n**Fig. 2 Dependent variable** _adoption of conservation practices index_\ndisaggregated by practice and by proportion that selected yes using\nthe pooled member survey data from 2013, 2017, 2020 (n = 677) and\n\n\nmember and organizational achievements or features in\nthe media, publications, announcements, and highlights of\ntimely resources. PFI offers in-person events including field\ndays, conferences, workshops, regional meet-ups, socials,\nmulti-day bus trips, and mentorship opportunities. Discussion lists are used by members and staff to ask and answer\nquestions, post items for sale or needed, send announcements, and disseminate information on other resources and\nevents outside of PFI. Specific lists exist concerning general\nannouncements, field crops, livestock, and horticulture, and\na perspectives list that provides space for members to engage\nin open-minded debate and discussion related to agriculture\n(PFI 2022b). Fewer farmers attended 3 or more events per\nyear, or served as a farmer leader (i.e., conducted on-farm\nresearch, hosted a field day, talked to the media, served on\na committee, or served as a mentor). Figure 3 also displays\nparticipation in the network by level of adoption of conservation practices where high adopters (n = 258) used between\n4 and 6 conservation practices and low adopters (n = 419)\nused between 0 and 3. High adopters participated slightly\nmore in each avenue listed besides attending 3 or more\nevents, which was almost equal across adopter groups.\nTable 2 summarizes the results of the regressions measuring the relationship between the adoption of conserva_tion practices index and the general participation index._\nThe results show that farmers’ level of participation in PFI\nis statistically significant in explaining the level of adoption of conservation practices. The higher a farmer's level\nof participation in PFI, the more likely it is that they will\nhave a higher level of adoption of conservation practices,\non average. Results are robust across both the bivariate and\nmultivariate GLM binomial (Models 1 & 2) and ordinal\nlogistic (Models 4 & 5) models. General participation in\nPFI has the second most statistically significant relationship\n\nfurther disaggregated by level of participation in PFI (High: n = 292;\nLow: n = 385)\n\n## 1 3\n\n\n-----\n\n**Fig. 3 Main independent variable** _general participation index dis-_\naggregated by activity/resource and by proportion that selected\nyes using the pooled member survey data from 2013, 2017, 2020\n\nwith adoption, following participation in EQIP, across all\ncovariates (excluding Year 2013 fixed effects) included in\nthe GLM binomial Model 2 (p = 0.0001) and ordinal logistic Model 4 (p = 0.0003). These findings support existing\nscholarship showing that the degree of participation in conservation organizations (Korsching et al. 1983; Belknap\nand Saupe 1988; Bates and Arbuckle 2017) and the degree\nintegration into or centrality within social networks broadly\n(Phelps et al. 2012; Cheng 2021) drive the diffusion of\ntechnology.\nWhile years in PFI is not significant in the multivariate Models 2 and 5, when interacted with the general par_ticipation index, the two combined have a strong effect on_\nthe _adoption of conservation practices index (Model 3:_\np = 0.001, Model 6: p = 0.003). Figure 4 visually displays\nthe moderating role of the length of time in the network on\nthe relationship between participation and the adoption of\nconservation practices. For farmers who have been in the\nnetwork for 11 years (long dotted green line), the association between their participation in PFI and their adoption\nof conservation practices is the strongest, indicated by the\nsteepest slope. For farmers who have spent less than three\nyears in the network (solid red line), the association between\ntheir participation in PFI and their adoption of conservation\npractices is the weakest, shown by the flattest slope. Thus,\nthe interaction term in Models 3 and 6 and the relationships\ndisplayed in Fig. 4 show that the effect of greater participation in the network is the strongest when a member has\nbeen in the network for a longer period of time. Some existing research points to the importance of length of time in a\nfarmer network to the adoption of conservation practices;\nPape and Prokopy (2017) found that the length of time spent\nin formal networks in Indiana was positively associated with\n\n## 1 3\n\n\n(n = 677) and further disaggregated by level of adoption of conservation practices (High: n = 258; Low n = 419)\n\nthe use of better nutrient management practices. However,\nours is the first research to date that has examined the combined effects of length of time spent in a farmer network and\nthe degree of participation on the adoption of conservation\npractices.\nAcross the covariates, a farmers’ participation in EQIP\nand proportion of acres owned are associated with adoption\nat the 1% level of significance in all multivariate models\n(Models 2–3 & 5–6). The association between participation\nin EQIP and higher adoption is in line with several review\nstudies that find conservation program participation has an\noverall positive relationship with adoption of conservation\npractices (Knowler and Bradshaw 2007; Prokopy et al. 2008;\nCarlisle 2016). The more acres of their total operation a PFI\nmember farmer owns, as opposed to rents, the less likely\nthey are to adopt a greater number of conservation practices\nwas a surprising finding given that existing theory points to\nthe advantages of owning land for receiving the longer-term\nbenefits of conservation practices (Belknap and Saupe 1988;\nSoule et al. 2000). However, renting more acres may allow\nfarmers to expand their farm size and larger farms are shown\nto enable farmers to absorb risk and adopt more due to economies of scale (Khanna et al. 1999; Ulrich-Schad et al. 2017;\nDurant et al. 2023) which may explain why farm size is also\nsignificant at the 5% level in the GLM multivariate models\n(Models 2 & 3). Within the fixed effects, we note that 2013\nis strongly associated with adoption compared to 2017 or\n2020. In the years following 2013, there was a greater effort\nplaced on membership recruitment and a broader group of\nboth conservation and conventional-minded farmers joined\nPFI (Personal communication with PFI Senior Programs\nand Member Engagement Director, November 2021). For\n\n\n-----\n\n**Table 2 GLM binomial and ordinal logistic regression models predicting adoption of conservation practices using the pooled member survey**\ndata from 2013, 2017, 2020\n\nDependent variable:\n\nAdoption of conservation practices index\n\nGLM logistic Ordered logistic\n\nModel (1) Model (2) Model (3) Model (4) Model (5) Model (6)\n\nGeneral participation index **0.121[***]** **0.148[***]** **0.123[***]** **0.220[***]** **0.298[***]** **0.251[***]**\n\n(0.032) (0.038) (0.039) (0.068) (0.081) (0.083)\np = 0.0002 p = 0.0001 p = 0.002 p = 0.002 p = 0.0003 p = 0.003\nAge of primary member 0.015 0.007 0.011 − 0.007\n(0.042) (0.043) (0.092) (0.092)\np = 0.731 p = 0.873 p = 0.904 p = 0.942\nFarm size **0.087[**]** **0.084[**]** **0.150[*]** 0.146\n(0.042) (0.042) (0.091) (0.092)\np = 0.038 p = 0.047 p = 0.098 p = 0.111\nParticipation in EQIP (yes = 1) **0.166[***]** **0.170[***]** **0.354[***]** **0.359[***]**\n\n(0.039) (0.039) (0.084) (0.085)\np = 0.00002 p = 0.00002 p = 0.00003 p = 0.00003\nParticipation in a watershed project 0.055 0.045 0.109 0.094\n(yes = 1) (0.036) (0.037) (0.075) (0.076)\np = 0.129 p = 0.218 p = 0.149 p = 0.217\nProportion of acres owned − 0.120[***] − 0.114[***] − 0.246[***] − 0.238[***]\n\n(0.040) (0.040) (0.087) (0.087)\np = 0.003 p = 0.005 p = 0.005 p = 0.007\nYears in PFI − 0.001 − 0.015 0.012 − 0.011\n(0.040) (0.041) (0.086) (0.087)\np = 0.981 p = 0.709 p = 0.893 p = 0.901\nState 0.073[**] 0.070[**] 0.170[**] 0.162[**]\n\n(0.035) (0.035) (0.074) (0.074)\np = 0.034 p = 0.043 p = 0.022 p = 0.029\nYear 2017 0.001 0.009 0.005 0.021\n(0.042) (0.042) (0.091) (0.092)\np = 0.986 p = 0.832 p = 0.961 p = 0.818\nYear 2013 **0.199[***]** **0.195[***]** **0.477[***]** **0.466[***]**\n\n(0.038) (0.038) (0.085) (0.085)\np = 0.00000 p = 0.00000 p = 0.00000 p = 0.00000\nCombined participation index* **0.123[***]** **0.226[***]**\n\nYears in PFI (0.036) (0.075)\np = 0.001 p = 0.003\nConstant 0.015 **0.071[*]** 0.056\n(0.031) (0.037) (0.037)\np = 0.637 p = 0.056 p = 0.133\nObservations 677 516 516 677 516 516\nLog Likelihood − 1,158.620 -835.720 − 829.512\nPseudo ­R[2] (McFadden’s) 0.02 0.13 0.15 0.005 0.28 0.28\nAkaike Inf. Crit 2,321.24 1,693.44 1,683.02 2,304.03 1,697.25 1,689.96\nWald Chi-squared 0.22 76.40*** 82.90*** 10.31** 80.56*** 88.98***\nHosmer and Lemeshow test p = 0.95 p = 0.73 p = 0.59\nLipsitz test p = 0.58 p = 0.90 p = 0.49\nPulkstenis-Robinson test p = 0.62 p = 0.58 p = 0.79\n\n## 1 3\n\n\n-----\n\n**Table 2 (continued)**\n\nAsterisks and bolding note statistical significance at or below the 10% level; [⋆] is < 0.1; [⋆⋆] is < 0.05, and [⋆⋆⋆] is < 0.01\n\nStandard errors are listed in parentheses. All coefficients have been standardized\n\n_PFI, and the state in which they reside are not significant._\nTaken together, the independent variables in the model are\nstrong predictors of whether a farmer adopted one or more\nconservation practices as a result of PFI indicated by a high\npseudo ­R[2] (0.55).\n\n**Farmer‑reported results**\n\n\n**Fig.4 Interaction plot between the percentage of practices adopted**\nwithin conservation practices index and the general participation\nindex by the number of years in PFI (n = 677). The colors represent\nvarying levels of years in PFI. Years in PFI are shown in the lower,\nmedian, and upper quartiles. The width of the bands of each group of\nyears in PFI represent the upper and lower values of the 95% confidence intervals\n\nthis reason, it is likely that in 2013, adoption rates were\nhigher due to the composition of the membership that tended\ntowards higher adopters.\nKnowing that participation in PFI is important for the\nadoption of conservation practices among members, it is\nuseful to understand more precisely what drives successful participation in the network. Table 3 shows the logistic\nregression model (Model 7) with the dependent variable\n_adopted conservation practice(s) as a result of PFI and sev-_\neral independent variables from the survey data that might\npredict the change. We find that, on average, those who\nsaid they formed relationships through the network were\nthe most likely to report adopted conservation practice(s)\n_as a result of PFI in terms of significance (p < 0.000), and_\ncoefficient size (1.789) holding all other variables in the\nmodel constant. This is in line with theories of social learning suggesting that farmers who are friends or in intentional\nrelationships share more information, increasing the likelihood of adoption (Liverpool-Tasie and Winter-Nelson 2012;\nMekonnen et al. 2022). Additionally, those who participated\nmore (general participation index, p = 0.014) and those\nwho reported feeling a sense of community (felt a sense of\n_community through the network, p = 0.026) were also more_\nlikely to report adopting a conservation practice as a result\nof PFI. The age of primary member, the number of years in\n\n## 1 3\n\n\nFarmers reported their own experience with PFI and the use\nof conservation practices through the survey and through\nin-depth interviews. Using the pooled survey data from\n2013, 2017, and 2020 we find that 71% of farmers said they\nadopted conservation practice(s) as a result of their participation in PFI (Table 4). Similar to the survey, when asked\nabout the role of PFI in their transition to sustainable practices during in-depth interviews, 68% of farmers told us that\ntheir engagement in PFI helped. Farmers said that the PFI\nnetwork provided information, resources, encouragement,\nand confidence building. Below we provide two illustrative examples from the 18 farmers who noted this. Farmer\n2 explained:\n\nI feel like PFI has been a huge source of information and encouragement…Probably, because of PFI's\nopportunities for learning and information sharing, it\nhas given us maybe confidence to try something we\nmight've been thinking about.\n\nFarmer 3 echoed the regression findings on the importance of the length of time spent in the network for the\nimpact of participation to be fully realized, which they\nattribute to the ability to observe practices used successfully\nby other members over longer periods of time:\n\nI probably wouldn’t have used cover crops quite as\nmuch if I hadn’t seen PFI members doing it for a long\nperiod of time. That gave me confidence to try. Even\nthough I probably philosophically knew it was the\nright thing to do, it was probably a little bit scary for\nme to do it at the beginning…seeing people do that…\nthat were farmers, allows me to, not be quite as scary\nto try something like that.\n\nTwo other farmers noted that PFI helped to reinforce their\ndecisions to adopt a new practice:\n\nFarmer 16: I’m not sure I’ve changed a lot of practices\nwith PFI…But they have solidified some things that, yes,\nI think I'm going down the right track. Or maybe I’ve\nlearned something that’s tweaked the practice a little bit.\n\n\n-----\n\n**Table 3 Logistic regression**\nmodel predicting the change\nof conservation practices as a\nresult of participation in PFI\nusing the pooled member survey\ndata from 2013, 2017, 2020\n\n\nDependent variable:\n\nAdopted conservation practice(s) as a result of participation in PFI\n\nModel (7)\n\nGeneral participation index **0.491[**]**\n\n(0.198)\np = 0.014\nAge of primary member 0.012\n(0.173)\np = 0.945\nFormed relationships (yes = 1) **1.789[***]**\n\n(0.150)\np = 0.000\nFelt a sense of community (yes = 1) **0.335[**]**\n\n(0.150)\np = 0.026\nYears in PFI 0.179\n(0.191)\np = 0.348\nState − 0.085\n(0.146)\np = 0.562\nYear 2017 − 0.328[**]\n\n(0.147)\np = 0.027\nYear 2013 − 0.583[*]\n\n(0.314)\np = 0.064\nConstant **1.386[***]**\n\n(0.202)\np = 0.000\n\nObservations 548\nLog Likelihood − 145.904\nPseudo ­R[2] (McFadden’s) 0.55\nAkaike Inf. Crit 309.808\nWald Chi-squared 47.30***\n\nAsterisks and bolding note statistical significance at or below the 10% level; [⋆] is < 0.1; [⋆⋆] is < 0.05, and [⋆⋆⋆]\nis < 0.01\n\nStandard errors are listed in parentheses. All coefficients have been standardized\n\n\nFarmer 19: The role of PFI in my life was to reinforce\nthe decisions that I was making. To feel validated, I\nguess, about the decisions that I was making…So,\nI was already headed down the road toward a more\ndiversified production, longer rotations, soil conservation, protecting the quality of the water. Heading\ndown that road. But PFI just kind of gave me the validation and reinforcement that I felt like I needed to\nkeep doing that. Which is an important role. I mean,\nyou know?\n\n\n**Table 4 Famer reported experience of the PFI network using the**\npooled member survey data from 2013, 2017, 2020\n\nPct SD _n_\n\nAdopted conservation practices as result of partici- 71% 0.45 667\npation in PFI (1 = Yes)\n\nFormed relationships through the network (1 = Yes) 73% 0.44 567\nFelt a sense of community through the network 78% 0.41 570\n(1 = Yes)\n\n## 1 3\n\n\n-----\n\nFarmers also reported on the social impacts of their\nparticipation in the member survey. Seventy-three percent of farmers said that they formed friendships, business relationships, or relationships through the network\nand 78% percent reported feeling a sense of community\nthrough the network (Table 4). In an open-ended survey\nquestion about the impact of their PFI relationships, the\nmost common responses included receiving new information and ideas, learning from the experience of others, and\nfeeling social support from like-minded farmers. These\nfarmer-reported impacts of network participation and the\nresults of the regression in Table 3 show that the personal\nrelationships formed through the network, together with\ngreater participation (as measured by the general par_ticipation index) and a widespread sense of community,_\nare important to explaining successful participation in\nthe network.\n\n#### Ways of participating and learning in the network\n\n**Regression results**\n\nWhen considering ways of participating in the network\nseparately, engagement through both in-person (Models 8\n& 9) and independent (Models 10 & 11) pathways have a\npositive and significant relationship with increased adoption of conservation practices (Table 5). Thus, participating more, regardless of the format, is associated with\ngreater adoption. However, in-person participation has a\nstronger relationship with adoption compared to independent participation in terms of statistical significance and\nthe standardized coefficient size across both GLM binomial and ordinal logistic models. We use the McFadden’s\nPseudo R[2] and Akaike Information Criterion) (AIC) values\nto compare model fit across the in-person and independent\nparticipation models. The best model is signified by the\nhighest Pseudo R[2] value and lowest AIC value. A difference of 2 between the AIC values is considered substantial (Burnham 2002). The in-person participation models\n(Models 8 & 9) have higher Pseudo R[2] values and AIC\nvalues are 10 units lower than the independent participation models (Models 10 & 11), indicating that in-person\nparticipation is significantly better than independent participation at predicting the level of adoption of conservation practices.\nOther studies have reached similar conclusions regarding the greater effectiveness of social learning in agriculture compared to formal learning (Hoffman et al. 2015) and\nindependent or institutional learning (Laforge and McLachlan 2018). Social learning has been shown to be a predictor of the adoption of conservation practices in the form of\ninteraction with innovative neighbors (Garbach and Morgan\n2017) and attendance at demonstration sites and field days\n\n## 1 3\n\n\n(Singh et al. 2018). Our results mirror findings from Bates\nand Arbuckle (2017) who looked at farmer preferences for\nreceiving information through face-to-face formats (field\ndays, meetings, or workshops) and non-face-to-face formats\n(online videos or downloaded publications). The authors\nfound that those who preferred to receive information about\nnutrient management through face-to-face formats reported\nusing more diverse nitrogen management practices that help\nreduce nutrient loss (Bates and Arbuckle 2017).\n\n**Farmer‑reported results**\n\nIn terms of preferred ways of learning and participating in\nthe network, farmers reported that they value a range of ways\nof engaging, including both in-person and independent formats.[4] In the 2020 member survey, farmers were asked to\nrate the importance of learning formats offered through the\nnetwork. The most highly-ranked ways of participating were\nattending field days followed by reading research reports,\nreading E-newsletters, watching videos, participating in\nfarminars, and participating in full-day workshops (Fig. 5).\nTo improve our confidence in the results of the in-person\nversus independent participation regressions (Models 8–11)\nand farmer-reported learning preferences in the member survey (Fig. 5), as well as to understand the reasoning behind\nwhy farmers prefer one learning format to another, we asked\nfarmers about their learning and participation preferences\nduring in-depth interviews. While farmers reported that they\nvalue a combination of in-person and independent learning\nformats in the survey, during in-depth interviews the majority (68%) of farmers told us that they preferred in-person\nformats. Farmers said that they use independent sources\nlike research reports and videos for their ease of access,\nand to stay up to date and connected. However, several key\nthemes emerged around the relatively greater importance\nof in-person formats that align with existing literature on\nthe benefits of face-to-face interaction for social learning\n(Bandura 1977; Ban 1981; Rogers 2003): the ability to have\nside conversations, ask follow-up questions, and to observe\nresults for themselves.\n_Side conversations: Eight of the farmers interviewed dis-_\ncussed the importance of being able to talk with and learn\nfrom other farmers during unstructured portions of network\nevents, allowing them to ask follow-up questions and gain a\nmore detailed understanding of the practice. Three illustrative quotes from the interviews are below:\n\nFarmer 5: …some of the things that I do miss since\nCovid is like, we would go to three, four, five field days\n\n4 Figure 5 includes only 2020 member survey data. Data was not\navailable for this question for the 2013 and 2017 surveys.\n\n\n-----\n\n**Table 5 GLM binomial and**\nOrdinal logistic regression\nmodels comparing in-person\nand independent participation\nindices and the level of adoption\nof conservation practices using\nthe pooled member survey data\nfrom 2013, 2017, 2020\n\n\nDependent variable\n\nAdoption of conservation practices index\n\nGLM logistic Ordered logistic GLM logistic Ordered logistic\n\nModel (8) Model (9) Model (10) Model (11)\n\nIn-person participation index **0.124[***]** **0.248[***]**\n\n(0.032) (0.066)\n\np = 0.0001 p = 0.0002\n\nIndependent participation index **0.087[**]** **0.176[**]**\n\n(0.041) (0.088)\n\np = 0.032 p = 0.046\n\nAge of primary member 0.005 − 0.003 − 0.001 − 0.023\n\n(0.042) (0.091) (0.042) (0.091)\n\np = 0.911 p = 0.976 p = 0.981 p = 0.804\n\nFarm size **0.083[**]** 0.143 **0.080[*]** 0.136\n\n(0.041) (0.088) (0.042) (0.092)\n\np = 0.045 p = 0.104 p = 0.056 p = 0.137\n\nParticipation in EQIP (yes = 1) **0.168[***]** **0.359[***]** **0.177[***]** **0.370[***]**\n\n(0.038) (0.084) (0.038) (0.084)\n\np = 0.00002 p = 0.00002 p = 0.00001 p = 0.00002\n\nParticipation in a watershed project 0.046 0.089 0.060 0.114\n(yes = 1)\n\n(0.036) (0.075) (0.037) (0.076)\n\np = 0.203 p = 0.235 p = 0.101 p = 0.137\n\nProportion of acres owned − 0.117[***] − 0.242[***] − 0.126[***] − 0.252[***]\n\n(0.040) (0.087) (0.040) (0.087)\n\np = 0.004 p = 0.006 p = 0.002 p = 0.004\n\nYears in PFI 0.007 0.025 0.015 0.038\n\n(0.040) (0.086) (0.040) (0.086)\n\np = 0.859 p = 0.773 p = 0.712 p = 0.662\n\nState **0.077[**]** **0.174[**]** **0.059[*]** **0.138[*]**\n\n(0.035) (0.074) (0.034) (0.073)\n\np = 0.027 p = 0.019 p = 0.087 p = 0.058\n\nYear 2017 0.008 0.020 -0.001 0.011\n\n(0.042) (0.091) (0.042) (0.091)\n\np = 0.849 p = 0.830 p = 0.978 p = 0.906\n\nYear 2013 **0.201[***]** **0.483[***]** **0.182[***]** **0.447[***]**\n\n(0.038) (0.085) (0.037) (0.085)\n\np = 0.00000 p = 0.00000 p = 0.00000 p = 0.00000\n\nConstant − 0.085 − 0.107\n\n(0.055) (0.093)\n\np = 0.127 p = 0.251\n\nObservations 516 516 516 516\n\nLog Likelihood − 835.653 − 841.043\n\nPseudo ­R[2] (McFadden’s) 0.13 0.28 0.11 0.27\n\nAkaike Inf. Crit 1,693.31 1,697.00 1,704.09 1,707.04\n\nWald Chi-squared 8.78*** 81.10*** 7.77*** 70.80***\n\nHosmer and Lemeshow test p = 0.39 p = 0.40\n\nLipsitz test p = 0.34 p = 0.84\n\nPulkstenis-Robinson test p = 0.55 p = 0.26\n\nAsterisks and bolding note statistical significance at or below the 10% level; [⋆] is < 0.1; [⋆⋆] is < 0.05, and [⋆⋆⋆]\nis < 0.01\n\nStandard errors are listed in parentheses. All coefficients have been standardized\n\n## 1 3\n\n\n-----\n\n**Fig. 5 Average ranked importance of PFI learning formats using the member survey data from ­2020[a] (n = 410) by type of learning format [a]Mem-**\nber survey data was not available for this question for the 2013 and 2017 surveys\n\n\nin a summer, and the side conversations that we had\nwere just as valuable as the content of the field day.\nFarmer 16: …where do you really learn something?\nIt's the breaks or having a beer at the bar afterwards.\nThat’s where you really learn something, in my opinion. [Regarding reading a blog or research report,] I\nlearn something, I learn a concept. But not the details.\nFarmer 21: Attending [the field day] live you get some\nof that community because you can ask questions…\nwhat often would happen is you might ask a question\nand the person who's running the field day will answer\nit and then you end up having a discussion with the\nfarmer next to you about their answer. You get two\nfarmer's perspectives, not just one...\n\n_Seeing is believing: A core tenant of farmer-to-farmer_\nlearning, as emphasized by Rosset et al. (2011) is “seeing\nis believing,” which highlights the importance of seeing\nresults with one’s own eyes for the successful diffusion of\ninnovation. This sentiment was echoed by six PFI farmers\nwhen explaining the importance of in-person learning. Two\nof them described this succinctly as:\n\nFarmer 17: The physical presentation of the evidence\nand getting to see it is just always more impactful than\nreading about it in a textual format.\nFarmer 19: What helps the most is seeing other farmers adopt any new practices and that's where I think\nthat PFI excels is in their field days and showing people how it can be done and making public what their\nneighbors are doing.\n\n## 1 3\n\n\nAs interviews took place during the first year of the\nCovid-19 pandemic in the US (July–December 2020), we\nasked farmers whether their learning and participation preferences were changing due to the newly offered virtual and\nhybrid formats. Seven farmers expressed appreciation for\nthe flexibility and accessibility of virtual options that they\ncould view on their own time and forgo the travel time and\ncost, explained by two of them as:\n\nFarmer 2: (Remote field days are) not as good as being\nin person. But in reality, with taking on the cropping\nmore this year, there is no way I could have taken time\nto actually drive to all the ones that I’d be interested\nin. And as much as I really liked doing it in person, the\nfew field days I did attend virtually, it really just took\nliterally the hour that it lasted or whatever. And I could\ngo right back out and work again, and not take the time\nto drive and come back and whatnot.\nFarmer 7: With the Covid this year it was kind of nice\nbecause there was a couple field days that I probably\nwouldn’t have attended due to distance. But because\nthey were on video, I watched them…I could watch\nit on my time versus when it was happening. I didn’t\nhave to take off of work. It kind of fit my schedule a\nlittle better.\n\nOne farmer noted the benefits of hybrid participation\noptions that allow for live feedback and questions:\n\nFarmer 15: I think after Covid I’m going to want to go\nto live events, I’m going to want to go and meet people\n\n\n-----\n\nface-to-face. But I really like the approachability of the\nFacebook Live in addition to face-to-face, both being\nheld at the same time. I mean, we have the ability to\ndo all that so it's nice to have options.\n\nWhile nearly all farmers we interviewed appreciated having both in-person and remote options available to them,\npreferences for face-to-face interaction persisted during\nthe early stages of the Covid-19 pandemic and 72% said\nthat they planned to return to in-person participation when\npossible. This general sentiment was encompassed by one\nfarmer as:\n\nFarmer 18: I can’t wait to be able to have the in-person\nstuff again because you still need that interaction with\npeople that listserv and Farminars don’t provide.\n\nThese quotes re-iterate the power of in-person learning\nand help explain the results of the regressions in Table 5\nthat attest to the greater importance of in-person formats for\nthe adoption of conservation practices. The benefits of and\npreference for in-person learning likely feeds into the sense\nof community and relationship building occurring through\nthe network as reported by farmers in the member survey.\n\n### Discussion, limitations, and areas for future research\n\nBecause few farmer networks are as large and as long standing as PFI and even fewer collect systematic data from members, we use these unique data to gain greater insight into\nhow the degree of participation in a network is associated\nwith levels of adoption of conservation practices among\nnetwork participants. We then add to and triangulate the\nsurvey data with our qualitative data from in-depth interviews with a subset of member farmers. Our findings support\nour hypothesis that greater engagement in the PFI network\nleads to greater use of conservation practices amongst member farmers. Quantitative results show a strong and positive\nassociation between participation in the PFI network and\nthe use of conservation practices among members. These\nresults are robust across GLM binomial and ordinal logistic models using pooled data from three years of member\nsurveys. Qualitative interviews reinforce this finding: 76%\nof farmers told us that participation in the network helped\nthem to transition to sustainable practices, or reinforced\ntheir decision to do so, by providing information, resources,\nencouragement, and confidence building. The effects of participation on adoption are enhanced by participating in the\nnetwork for a longer period of time, as shown by the positive\nsign and significance of the interaction between the general\nparticipation index and the adoption of conservation practices index. Interviews with farmers suggest this could be\n\n\ndue to feedback loops involved with seeing practices used\nby peer farmers successfully over longer periods of time.\nFor farmers’ participation in the network to be successful,\nmeasured as adopting one or more conservation practices as\na result of participation, social ties within the network are\nimportant, particularly forming relationships. These results\nadd to the literature quantifying the role of formal farmer\nnetworks in the adoption of conservation practices and add\na nuanced analysis of the role of social networks by measuring the extent of engagement in one well known network in\nthe Midwest.\nThis research also tests theories of social learning and\ndiffusion of innovation that suggest that in-person, interpersonal channels of communication are a more effective means\nof changing an individual’s behavior compared to information from impersonal sources. Findings from both our quantitative and qualitative data affirm existing theories of the\nrelatively greater effectiveness of in-person learning, but\nalso point to the complementary nature of the two modes.\nWhile PFI farmers appreciated a variety of both in-person\nand independent ways of engaging in the network, regression results show a stronger relationship between adoption\nand in-person participation compared to independent participation. In-depth interviews improved our confidence in\nthese regression results by showing that more of the farmers\nwe interviewed (68%) preferred in-person learning formats\nto support them in their transition to sustainable agriculture. Interviews suggest that this is because PFI farmers are\nmore likely to adopt when they can observe other farmers\nusing a practice, hear about other farmers’ experiences,\nand ask questions through activities such as field days and\nworkshops.\nAs researchers continue to search for quantitative models\nthat explain the use of conservation practices in agriculture,\nour results show that involvement in a formal social network\nis an important variable for predictive models determining\nadoption. Our results also show that the type of participation matters, and that when farmers engage in person, they\nbenefit from the direct interaction with peers these formats\nafford.\nHowever, due to the cross-sectional research design of\nthis study, which it shares with many studies in this literature, we cannot quantitatively prove a causal relationship\nbetween the degree of participation in PFI and the adoption\nof conservation practice because of two main limitations.\nFirst, those who join PFI may have certain characteristics\nthat lead them to both participate in a sustainable farming\ngroup like PFI and to adopt conservation practices. As we\nsaw from the differences between PFI farmers and the general farming population (see Study context), PFI farmers\ncan be considered early adopters and motivated by deeply\nheld values like altruism or other intrinsic motivations.\nWe include covariates in our models that are shown to be\n\n## 1 3\n\n\n-----\n\nassociated with adoption in other studies, but due to the limitations of the survey data set, we cannot control for other\npotential confounding factors that may influence the intensity of both adoption and participation such as environmental ethic, labor availability, free time, and financial resources.\nSecond, farmers may have adopted a practice before\njoining PFI or greater adoption may be causing greater participation as opposed to the other way around. While we\nremoved observations that had not been a PFI member for\nmore than one year since taking the survey to ensure that\nadoption is measured at least one subsequent year since\nparticipating in the network, we do not have sufficient data\non the same farmers over multiple years to compare each\nfarmers’ change in participation and adoption over time.\nOnce a farmer adopts a conservation practice, the need for\ninformation may increase, which is possible given the more\nknowledge-intensive nature of conservation practices. In\nthis case, adoption may be driving participation as farmers\nseek additional resources and support. Even if it is the case,\nthis scenario, along with data from farmer interviews, still\nspeaks to the importance of the network as a resource and\none that may spur greater or sustained adoption.\nLongitudinal studies tracking adoption of a set of farmers before and after network participation, or better yet,\nexperimental studies that randomly facilitate membership\nin a network and measure any changes in adoption, can help\nto gain a more precise understanding of whether formal\nfarmer networks are in fact driving the adoption of conservation practices. Still, the effects of network participation are\nnuanced, with impacts causal studies may not capture. Interview results show that in some cases PFI is a driver of adoption while in others it helps sustain adoption for those that\nhave already started on this path. Both effects are important\nas the disadoption of conservation practices among farmers\nis common and can erase their accumulated environmental\nbenefits (Sawadgo and Plastina 2022).\nLongitudinal and experimental studies, however, are often\nmore costly, time consuming, and can ask more of already\ntime-constrained farmers. This study uses existing data collected by PFI to contribute to the limited literature aimed at\nunderstanding the role of formal farmer-to-farmer networks.\nWhile we cannot say from regression results whether farmers who choose to participate in a network may be those\nalready likely to adopt conservation practices, our mixedmethods design allows us to understand causality through\nmultiple modes (Sayer 1992). Across surveys in 2013, 2017,\nand 2020, 71% of member farmers self-reported that they\nhad adopted conservation practice(s) as a result of participation and during interviews, 76% of a sample of member\nfarmers told us that PFI played a direct role in their transition\nto sustainable practices, or supported their continued use\nof these practices. In addition, other research conducted on\nPFI supports a causal relationship between involvement in\n\n## 1 3\n\n\nPFI and the use of sustainable practices (Bell 2004; Warner\n2007; Blesh and Wolf 2014; Carlisle 2016; Blesh and Galt\n2017). In a study of Iowa grain farmers by Blesh and Wolf\n(2014), almost all listed PFI as an important resource for\ntransitioning towards sustainability. The in-depth ethnographic research on PFI from Bell (2004) suggests that the\ndialogic culture that is a crucial element of PFI is a key\nway its members were able to transition to more sustainable\npractices.\nSimilar to Bell’s (2004) findings, our research shows\nthat building relationships within the network is an important component to adoption as farmers learn and vet ideas\nthrough these friendships and feel a sense of community and\nsocial support to try different practices. Our interview data\nshow that PFI acts as much more than a channel for the diffusion of information. The network also provides an important source of encouragement and problem solving from\nlike-minded farmers and network leadership, and a means\nthrough which to access resources. In the context of formal\nfarmer networks broadly, more research would be useful to\nunderstand the ways in which networks encourage farmers to\ntransition to conservation practices and the characteristics of\nfarmer networks that are useful and desirable from farmers’\nperspectives. To gain the greatest societal benefit from formal farmer networks, more research is also needed to understand how to encourage greater and sustained participation,\nhow to increase network reach to farmers on the periphery\nwho may benefit most from information exchanged in networks (Granovetter 1973), and how to make networks work\nbetter for different types of farmers.\n\n### Conclusion\n\nWhile our findings are not directly generalizable beyond PFI\nmember farmers, several implications arise for strategies and\npolicies to encourage greater use of conservation practices\nthat may be applicable to farmers in other formal farmerto-farmer networks in the US and to early adopters of conservation practices broadly. First, formal farmer-to-farmer\nnetworks may be a promising way to build and enhance\nfarmers’ technical capacity and provide the peer support\nnecessary to create substantial progress toward sustainable\nagriculture at a grassroots level. The fact that engagement\nin PFI (both participation and length of time in the network)\nand adoption of conservation practices are strongly associated in our analyses, and that member farmers reported\nduring in-depth interviews that PFI helped them to adopt\nor sustain adoption, suggests that there is a strong synergy\nbetween the two, regardless of causality. That early adopters\nparticipate frequently in this network means that it is a ripe\navenue to support their transition to sustainable agriculture.\n\n\n-----\n\nSupporting these early adopters as role models is likely to\ncatalyze wider-scale change.\nSecond, to create meaningful change in agriculture, it\nwill be important to understand the most impactful ways\nof engaging farmers in learning about new practices. This\nis especially timely to consider given the growing use of\nremote learning formats brought on by the Covid-19 pandemic. This study finds that while network farmers value\nboth independent and in-person ways of interacting in the\nnetwork, regression and interview results point to the greater\nimportance of in-person formats in adopting conservation\npractices. This finding is important as it contrasts other\nwork elevating the importance of self-learning within early\nadopters compared to later adopters with whom face-toface exchange is of greater importance (Rogers 2003; Dunn\net al. 2016). Thus, networks and other farming organizations\nmay benefit from prioritizing face-to-face programming\nthat encourages relationship-building amongst members.\nFarmers in this study appreciated the benefits of remote and\nhybrid ways of learning through the Covid-19 pandemic,\nyet the strong desire to return to in-person learning formats\nsuggests that remote options are beneficial, when possible,\nbut returning to in-person offerings should be a priority.\nThe challenge remains that formal farmer networks do\nnot exist everywhere and are limited by a lack of public and\nprivate support, and not all farmers have adequate information about networks or the ability to participate. Support\nis needed to help establish more formal farmer networks\nfocused on conservation and to increase the reach of existing networks to conventional farmers, those in social circles\nwho have not heard about them, and those who may not\nhave the means to participate. We believe that government,\nland-grant university, and philanthropic investments in,\nas well as private sector partnerships with, formal farmer\nnetworks are a promising way to help farmers transition\nto more regenerative and resilient practices. Specialized\nfarmer knowledge diffused through social networks is vital\nfor developing farmers’ adaptive capacity to respond to a\nchanging climate, fluctuating markets, and socio-economic\ninstability (Petersen–Rockney et al. 2021). Likewise, Cooperative Extension, embedded in university research, is vital\nfor bringing knowledge from formal institutions to farmers.\nCreating and disseminating both farmer and formal scientific\nknowledge, a model unique to formal networks, likely creates synergies far greater than when the two remain separate.\n\n**Acknowledgements This research was supported by the Walton Family**\nFoundation. We thank Dr. Jess Rudder for her thoughts and feedback\non the quantitative analysis, and to Dr. Linda Prokopy and her lab\nmembers at Purdue University for their review of the manuscript. We\nare grateful for the staff and member farmers at Practical Farmers of\nIowa for their thoughts and feedback throughout the study.\n\n**Funding Walton Family Foundation, 00101930, Maggie Norton.**\n\n\n**Open Access This article is licensed under a Creative Commons Attri-**\nbution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long\nas you give appropriate credit to the original author(s) and the source,\nprovide a link to the Creative Commons licence, and indicate if changes\nwere made. The images or other third party material in this article are\nincluded in the article's Creative Commons licence, unless indicated\notherwise in a credit line to the material. If material is not included in\nthe article's Creative Commons licence and your intended use is not\npermitted by statutory regulation or exceeds the permitted use, you will\nneed to obtain permission directly from the copyright holder. To view a\n[copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/.](http://creativecommons.org/licenses/by/4.0/)\n\n### References\n\nAjzen, I. 1991. The theory of planned behavior. Organizational Behav_[ior and Human Decision Processes 50: 179–211. https://​doi.​org/​](https://doi.org/10.1016/0749-5978(91)90020-T)_\n[10.​1016/​0749-​5978(91)​90020-T.](https://doi.org/10.1016/0749-5978(91)90020-T)\nBan, A. W. van den. 1981. Interpersonal communication and the\ndiffusion of innovations. In Extension Education and Rural\nDevelopment.\nBandura, Albert. 1977. Social learning theory. Prentice-Hall Series\nin Social Learning Theory. Englewood Cliffs, N.J: Prentice Hall.\nBarbercheck, M., K. Brasier, N.E. Kiernan, C. Sachs, and A. Trauger.\n2014. Use of conservation practices by women farmers in the\nNortheastern United States. Renewable Agriculture and Food Sys_[tems 29 (1): 65–82. https://​doi.​org/​10.​1017/​S1742​17051​20003​48.](https://doi.org/10.1017/S1742170512000348)_\nBates, H., and J.G. Arbuckle. 2017. Understanding predictors of nutrient management practice diversity in Midwestern agriculture. The\n_Journal of Extension 55 (6): 48._\nBaumgart-Getz, A., L.S. Prokopy, and K. Floress. 2012. Why farmers\nadopt best management practice in the United States: A metaanalysis of the adoption literature. Journal of Environmental\n_[Management 96 (1): 17–25. https://​doi.​org/​10.​1016/j.​jenvm​an.​](https://doi.org/10.1016/j.jenvman.2011.10.006)_\n[2011.​10.​006.](https://doi.org/10.1016/j.jenvman.2011.10.006)\nBelknap, J., and W.E. Saupe. 1988. Farm family resources and the\nadoption of no-plow tillage in Southwestern Wisconsin. Review\n_of Agricultural Economics 10 (1): 13–23._\nBell, M.M. 2004. Farming for us all: Practical agriculture and the\n_cultivation of sustainability. Pennsylvania: Pennsylvania Univer-_\nsity Park Press.\nBlesh, J., and R. Galt. 2017. Transitions to agroecological nutrient\n_management practices in the USA Corn Belt: Principles, appli-_\n_cations, and making the transition Agroecological Practices for_\n_Sustainable Agriculture, 85–126. London: World Scientific Pub-_\n[lishing. https://​doi.​org/​10.​1142/​97817​86343​062_​0004.](https://doi.org/10.1142/9781786343062_0004)\nBlesh, J., and S.A. Wolf. 2014. Transitions to agroecological farming systems in the Mississippi River Basin: Toward an integrated\nsocioecological analysis. Agriculture and Human Values 31 (4):\n[621–635. https://​doi.​org/​10.​1007/​s10460-​014-​9517-3.](https://doi.org/10.1007/s10460-014-9517-3)\nBrant, R. 1990. Assessing proportionality in the proportional odds\nmodel for ordinal logistic regression. Biometrics 46 (4): 1171–\n[1178. https://​doi.​org/​10.​2307/​25324​57.](https://doi.org/10.2307/2532457)\nBurnham, K.P. 2002. Model selection and multimodel inference: A\n_practical information-theoretic approach, 2nd ed. New York, NY:_\nSpringer.\nCampbell, J.T., T.M. Koontz, and J.E. Bonnell. 2011. Does collaboration promote grass-roots behavior change? Farmer adoption of\nbest management practices in two watersheds. Society and Natural\n_[Resources 24 (11): 1127–1141. https://​doi.​org/​10.​1080/​08941​920.​](https://doi.org/10.1080/08941920.2010.512358)_\n[2010.​512358.](https://doi.org/10.1080/08941920.2010.512358)\n\n## 1 3\n\n\n-----\n\nCarlisle, L. 2016. Factors influencing farmer adoption of soil health\npractices in the United States: A narrative review. Agroecology\n_[and Sustainable Food Systems 40 (6): 583–613. https://​doi.​org/​](https://doi.org/10.1080/21683565.2016.1156596)_\n[10.​1080/​21683​565.​2016.​11565​96.](https://doi.org/10.1080/21683565.2016.1156596)\nCarlisle, L., M. Montenegro de Wit, M.S. DeLonge, A. Iles, A.\nCalo, C. Getz, J. Ory, K. Munden-Dixon, R. Galt, B. Melone,\nR. Knox, and D. Press. 2019. Transitioning to sustainable agriculture requires growing and sustaining an ecologically skilled\nworkforce. Frontiers in Sustainable Food Systems 3 (2019): 96.\n[https://​doi.​org/​10.​3389/​fsufs.​2019.​00096.](https://doi.org/10.3389/fsufs.2019.00096)\nCheng, H.W.J. 2021. Factors affecting technological diffusion\nthrough social networks: A review of the empirical evidence.\n_[The World Bank Research Observer. https://​doi.​org/​10.​1093/​](https://doi.org/10.1093/wbro/lkab010)_\n[wbro/​lkab0​10.](https://doi.org/10.1093/wbro/lkab010)\nChiffoleau, Y., and D. Desclaux. 2006. Participatory plant breeding:\nThe best way to breed for sustainable agriculture? International\n_[Journal of Agricultural Sustainability 4 (2): 119–130. https://​](https://doi.org/10.1080/14735903.2006.9684795)_\n[doi.​org/​10.​1080/​14735​903.​2006.​96847​95.](https://doi.org/10.1080/14735903.2006.9684795)\nCruse, M.J., M. Liebman, D.R. Raman, and M.H. Wiedenhoeft.\n2010. Fossil energy use in conventional and low-external-input\n[cropping systems. Agronomy Journal 102 (3): 934–941. https://​](https://doi.org/10.2134/agronj2009.0457)\n[doi.​org/​10.​2134/​agron​j2009.​0457.](https://doi.org/10.2134/agronj2009.0457)\nDainese, M., E.A. Martin, M.A. Aizen, M. Albrecht, I. Bartomeus,\nR. Bommarco, L.G. Carvalheiro, et al. 2019. A global synthesis reveals biodiversity-mediated benefits for crop production.\n_[Science Advances 5 (10): eaax0121. https://​doi.​org/​10.​1126/​](https://doi.org/10.1126/sciadv.aax012)_\n[sciadv.​aax012.](https://doi.org/10.1126/sciadv.aax012)\nDavis, A.S., J.D. Hill, C.A. Chase, A.M. Johanns, and M. Liebman.\n2012. Increasing cropping system diversity balances productivity, profitability and environmental health. PLoS ONE 7 (10):\n[e47149. https://​doi.​org/​10.​1371/​journ​al.​pone.​00471​49.](https://doi.org/10.1371/journal.pone.0047149)\nDunn, M., J.D. Ulrich-Schad, L.S. Prokopy, R.L. Myers, C.R. Watts,\nand K. Scanlon. 2016. Perceptions and use of cover crops\namong early adopters: Findings from a national survey. Journal\n_[of Soil and Water Conservation 71 (1): 29–40. https://​doi.​org/​](https://doi.org/10.2489/jswc.71.1.29)_\n[10.​2489/​jswc.​71.1.​29.](https://doi.org/10.2489/jswc.71.1.29)\nDurant, J.L., L. Asprooth, R.E. Galt, S.P. Schmulevich, G.M. Manser,\nand N. Pinzón. 2023. Farm resilience during the COVID-19\npandemic: The case of California direct market farmers. Agri_[cultural Systems 204: 103532. https://​doi.​org/​10.​1016/j.​agsy.​](https://doi.org/10.1016/j.agsy.2022.103532)_\n[2022.​103532.](https://doi.org/10.1016/j.agsy.2022.103532)\nFagerland, M.W., and D.W. Hosmer. 2016. Tests for goodness of fit in\nordinal logistic regression models. Journal of Statistical Com_[putation and Simulation 86 (17): 3398–3418. https://​doi.​org/​10.​](https://doi.org/10.1080/00949655.2016.1156682)_\n[1080/​00949​655.​2016.​11566​82.](https://doi.org/10.1080/00949655.2016.1156682)\nFilson, G.C., S. Sethuratnam, B. Adekunle, and P. Lamba. 2009. Beneficial management practice adoption in five Southern Ontario\nwatersheds. Journal of Sustainable Agriculture 33 (2): 229–252.\n[https://​doi.​org/​10.​1080/​10440​04080​25874​21.](https://doi.org/10.1080/10440040802587421)\nFoster, A.D., and M.R. Rosenzweig. 1995. Learning by doing and\nlearning from others: Human capital and technical change in\nagriculture. Journal of Political Economy 103 (6): 1176–1209.\n[https://​doi.​org/​10.​1086/​601447.](https://doi.org/10.1086/601447)\nGarbach, K., and G.P. Morgan. 2017. Grower networks support adoption of innovations in pollination management: The roles of social\nlearning, technical learning, and personal experience. Journal of\n_[Environmental Management 204 (2017): 39–49. https://​doi.​org/​](https://doi.org/10.1016/j.jenvman.2017.07.077)_\n[10.​1016/j.​jenvm​an.​2017.​07.​077.](https://doi.org/10.1016/j.jenvman.2017.07.077)\nGosnell, H., N. Gill, and M. Voyer. 2019. Transformational adaptation\non the farm: Processes of change and persistence in transitions to\n‘climate-smart’ regenerative agriculture. Global Environmental\n_[Change 59 (2019): 101965. https://​doi.​org/​10.​1016/j.​gloen​vcha.​](https://doi.org/10.1016/j.gloenvcha.2019.101965)_\n[2019.​101965.](https://doi.org/10.1016/j.gloenvcha.2019.101965)\nGranovetter, M.S. 1973. The strength of weak ties. American Journal\n_of Sociology 78 (6): 1360–1380._\n\n## 1 3\n\n\nGranovetter, M. 2005. The impact of social structure on economic out[comes. Journal of Economic Perspectives 19 (1): 33–50. https://​](https://doi.org/10.1257/0895330053147958)\n[doi.​org/​10.​1257/​08953​30053​147958.](https://doi.org/10.1257/0895330053147958)\nHassanein, N. 1999. Changing the way America farms: Knowledge and\n_community in the sustainable agriculture movement. Lincoln: U_\nof Nebraska Press.\nHassanein, N., and J.R. Kloppenburg. 1995. Where the grass grows\nagain: Knowledge exchange in the sustainable agriculture move[ment. Rural Sociology 60 (4): 721–740. https://​doi.​org/​10.​1111/j.​](https://doi.org/10.1111/j.1549-0831.1995.tb00603.x)\n[1549-​0831.​1995.​tb006​03.x.](https://doi.org/10.1111/j.1549-0831.1995.tb00603.x)\nHealy, G.K., and J.C. Dawson. 2019. Participatory plant breeding and\nsocial change in the Midwestern United States: Perspectives from\nthe Seed to Kitchen Collaborative. Agriculture and Human Values\n[36 (4): 879–889. https://​doi.​org/​10.​1007/​s10460-​019-​09973-8.](https://doi.org/10.1007/s10460-019-09973-8)\nHoffman, M., M. Lubell, and V. Hillis. 2015. Network-smart extension could catalyze social learning. California Agriculture 69 (2):\n[113–122. https://​doi.​org/​10.​3733/​ca.E.​v069n​02p113.](https://doi.org/10.3733/ca.E.v069n02p113)\nHogset, H., and C.B. Barrett. 2010. Social learning, social influence,\nand projection bias: A caution on inferences based on proxy\nreporting of peer behavior. Economic Development and Cultural\n_[Change 58 (3): 563–589. https://​doi.​org/​10.​1086/​650424.](https://doi.org/10.1086/650424)_\nHunt, N.D., M. Liebman, S.K. Thakrar, and J.D. Hill. 2020. Fossil\nenergy use, climate change impacts, and air quality-related human\nhealth damages of conventional and diversified cropping systems\nin Iowa, USA. Environmental Science & Technology 54 (18):\n[11002–11014. https://​doi.​org/​10.​1021/​acs.​est.​9b069​29.](https://doi.org/10.1021/acs.est.9b06929)\nIles, A., and R. Marsh. 2012. Nurturing diversified farming systems\nin industrialized countries: How public policy can contribute. _Ecology and Society 17 (4): 42._ [https://​doi.​org/​10.​5751/​](https://doi.org/10.5751/ES-05041-170442)\n[ES-​05041-​170442.](https://doi.org/10.5751/ES-05041-170442)\nKhanna, M., O.F. Epouhe, and R. Hornbaker. 1999. Site-specific crop\nmanagement: Adoption patterns and incentives. Review of Agri_[cultural Economics 21 (2): 455–472. https://​doi.​org/​10.​2307/​](https://doi.org/10.2307/1349891)_\n[13498​91.](https://doi.org/10.2307/1349891)\nKloppenburg, J. 1991. Social theory and the de/reconstruction of agricultural science: Local knowledge for an alternative agriculture.\n_[Rural Sociology 56 (4): 519–548. https://​doi.​org/​10.​1111/j.​1549-​](https://doi.org/10.1111/j.1549-0831.1991.tb00445.x)_\n[0831.​1991.​tb004​45.x.](https://doi.org/10.1111/j.1549-0831.1991.tb00445.x)\nKnowler, D., and B. Bradshaw. 2007. Farmers’ adoption of conservation agriculture: A review and synthesis of recent research. Food\n_[Policy 32 (1): 25–48. https://​doi.​org/​10.​1016/j.​foodp​ol.​2006.​01.​](https://doi.org/10.1016/j.foodpol.2006.01.003)_\n[003.](https://doi.org/10.1016/j.foodpol.2006.01.003)\nKorsching, P.F., C.W. Stofferahn, P.J. Nowak, and D.J. Wagener. 1983.\nAdopter characteristics and adoption patterns of minimum tillage: Implications for soil conservation programs. Journal of Soil\n_[and Water Conservation 38 (5): 428–431. https://​doi.​org/​10.​1016/​](https://doi.org/10.1016/0167-1987(85)90030-3)_\n[0167-​1987(85)​90030-3.](https://doi.org/10.1016/0167-1987(85)90030-3)\nKroma, M. 2006. Organic farmer networks: Facilitating learning and\ninnovation for sustainable agriculture. Journal of Sustainable\n_[Agriculture 28 (4): 5–28. https://​doi.​org/​10.​1300/​J064v​28n04_​03.](https://doi.org/10.1300/J064v28n04_03)_\nLaforge, J.M.L., and C.Z. Levkoe. 2021. The impacts of agroecology training within and beyond farmer training. In Handbook\n_on the human impact of agriculture. Cheltenham: Edward Elgar_\nPublishing.\nLaforge, J.M.L., and S.M. McLachlan. 2018. Learning communities\nand new farmer knowledge in Canada. Geoforum 96 (2018): 256–\n[267. https://​doi.​org/​10.​1016/j.​geofo​rum.​2018.​07.​022.](https://doi.org/10.1016/j.geoforum.2018.07.022)\nLevy, M.A., and M.N. Lubell. 2018. Innovation, cooperation, and the\nstructure of three regional sustainable agriculture networks in\nCalifornia. Regional Environmental Change 18 (4): 1235–1246.\n[https://​doi.​org/​10.​1007/​s10113-​017-​1258-6.](https://doi.org/10.1007/s10113-017-1258-6)\nLiverpool-Tasie, L.S.O., and A. Winter-Nelson. 2012. Social learning\nand farm technology in Ethiopia: Impacts by technology, network\ntype, and poverty status. The Journal of Development Studies 48\n[(10): 1505–1521. https://​doi.​org/​10.​1080/​00220​388.​2012.​693167.](https://doi.org/10.1080/00220388.2012.693167)\n\n\n-----\n\nLubell, M., M. Niles, and M. Hoffman. 2014. Extension 3.0: Managing agricultural knowledge systems in the network age. Society &\n_[Natural Resources 27 (10): 1089–1103. https://​doi.​org/​10.​1080/​](https://doi.org/10.1080/08941920.2014.933496)_\n[08941​920.​2014.​933496.](https://doi.org/10.1080/08941920.2014.933496)\nLyon, A., M.M. Bell, C. Gratton, and R. Jackson. 2011. Farming without a recipe: Wisconsin graziers and new directions for agricul[tural science. Journal of Rural Studies 27 (4): 384–393. https://​](https://doi.org/10.1016/j.jrurstud.2011.04.002)\n[doi.​org/​10.​1016/j.​jrurs​tud.​2011.​04.​002.](https://doi.org/10.1016/j.jrurstud.2011.04.002)\nMcCullagh, P. 2019. Generalized linear models. Routledge.\nMekonnen, D.K., S. Yimam, T. Arega, K. Matheswaran, and P.M.V.\nSchmitter. 2022. Relatives, neighbors, or friends: Information\nexchanges among irrigators on new on-farm water management\n[tools. Agricultural Systems 203 (2022): 103492. https://​doi.​org/​](https://doi.org/10.1016/j.agsy.2022.103492)\n[10.​1016/j.​agsy.​2022.​103492.](https://doi.org/10.1016/j.agsy.2022.103492)\nMontgomery, M.R., and J.B. Casterline. 1996. Social learning, social\ninfluence, and new models of fertility. Population and Develop_[ment Review 22 (1996): 151–175. https://​doi.​org/​10.​2307/​28080​](https://doi.org/10.2307/2808010)_\n[10.](https://doi.org/10.2307/2808010)\nNapier, T.L., and T. Bridges. 2002. Adoption of conservation production systems in two Ohio watersheds: A comparative study. Jour_nal of Soil and Water Conservation 57 (4): 229–235._\nNapier, T.L., and M. Tucker. 2001. Use of soil and water protection\npractices among farmers in three Midwest watersheds. Environ_[mental Management 27 (2): 269–279. https://​doi.​org/​10.​1007/​](https://doi.org/10.1007/s002670010148)_\n[s0026​70010​148.](https://doi.org/10.1007/s002670010148)\nNass, U.S.D.A. 2017. 2017 Census of Agriculture. Washington, DC:\nUSDA National Agricultural Statistics Service.\nPampel, F.C. 2021. Logistic regression: A primer, 2nd ed. Thousand\nOaks: SAGE Publications Inc.\nPape, A., and L.S. Prokopy. 2017. Delivering on the potential of formal\nfarmer networks: Insights from Indiana. Journal of Soil and Water\n_[Conservation 72 (5): 463–470. https://​doi.​org/​10.​2489/​jswc.​72.5.​](https://doi.org/10.2489/jswc.72.5.463)_\n[463.](https://doi.org/10.2489/jswc.72.5.463)\nPark, T.A., and L. Lohr. 2005. Organic pest management decisions: A\nsystems approach to technology adoption. Agricultural Econom_[ics 33 (s3): 467–478. https://​doi.​org/​10.​1111/j.​1574-​0864.​2005.​](https://doi.org/10.1111/j.1574-0864.2005.00400.x)_\n[00400.x.](https://doi.org/10.1111/j.1574-0864.2005.00400.x)\nPatton, M.Q. 2002. Qualitative research and evaluation methods, 3rd\ned. Thousand Oaks: Sage Publications.\nPetersen-Rockney, M., P. Baur, A. Guzman, S.F. Bender, A. Calo, F.\nCastillo, K. De Master, et al. 2021. Narrow and brittle or broad\nand nimble? Comparing adaptive capacity in simplifying and\ndiversifying farming systems. Frontiers in Sustainable Food Sys_[tems 5 (2021): 56. https://​doi.​org/​10.​3389/​fsufs.​2021.​564900.](https://doi.org/10.3389/fsufs.2021.564900)_\nPFI. 2020. Practical Farmers of Iowa member survey.\n[PFI. 2022a. History. Practical Farmers of Iowa. https://​pract​icalf​armers.​](https://practicalfarmers.org/about/history/)\n\n[org/​about/​histo​ry/. Accessed 17 April 2022a.](https://practicalfarmers.org/about/history/)\nPFI. 2022b. E-Mail Discussion Groups. Practical Farmers of Iowa.\n\n[https://​pract​icalf​armers.​org/​get-​invol​ved/​conne​ct-​with-​pfi/e-​mail-​](https://practicalfarmers.org/get-involved/connect-with-pfi/e-mail-discussion-groups/)\n[discu​ssion-​groups/. Accessed 17 April 2022b.](https://practicalfarmers.org/get-involved/connect-with-pfi/e-mail-discussion-groups/)\nPhelps, C., R. Heidl, and A. Wadhwa. 2012. Knowledge, networks, and\nknowledge networks: A review and research agenda. Journal of\n_[Management 38 (4): 1115–1166. https://​doi.​org/​10.​1177/​01492​](https://doi.org/10.1177/0149206311432640)_\n[06311​432640.](https://doi.org/10.1177/0149206311432640)\nPrell, C. 2012. Social network analysis: History, theory & methodol_ogy. Los Angeles: SAGE._\nProkopy, L.S., K. Floress, J.G. Arbuckle, S.P. Church, F.R. Eanes, Y.\nGao, B.M. Gramig, P. Ranjan, and A.S. Singh. 2019. Adoption of\nagricultural conservation practices in the United States: Evidence\nfrom 35 years of quantitative literature. Journal of Soil and Water\n_[Conservation 74 (5): 520–534. https://​doi.​org/​10.​2489/​jswc.​74.5.​](https://doi.org/10.2489/jswc.74.5.520)_\n[520.](https://doi.org/10.2489/jswc.74.5.520)\nProkopy, L.S., K. Floress, D. Klotthor-Weinkauf, and A. BaumgartGetz. 2008. Determinants of agricultural best management\n\n\npractice adoption: Evidence from the literature. Journal of Soil\n_[and Water Conservation 63: 300–311. https://​doi.​org/​10.​2489/​](https://doi.org/10.2489/jswc.63.5.300)_\n[jswc.​63.5.​300.](https://doi.org/10.2489/jswc.63.5.300)\nReimer, A., A. Thompson, L.S. Prokopy, J.G. Arbuckle, K. Genskow,\nD. Jackson-Smith, G. Lynne, L. McCann, L.W. Morton, and P.\nNowak. 2014. People, place, behavior, and context: A research\nagenda for expanding our understanding of what motivates farmers’ conservation behaviors. Journal of Soil and Water Conser_[vation 69 (2): 57A-61A. https://​doi.​org/​10.​2489/​jswc.​69.2.​57A.](https://doi.org/10.2489/jswc.69.2.57A)_\nRogers, E.M. 2003. Diffusion of innovations, 5th ed. New York: Free\nPress.\nRosset, P.M., B. Machín Sosa, A.M. Roque Jaime, and D.R. Ávila\nLozano. 2011. The Campesino-to-Campesino agroecology movement of ANAP in Cuba: Social process methodology in the construction of sustainable peasant agriculture and food sovereignty.\n_[The Journal of Peasant Studies 38 (1): 161–191. https://​doi.​org/​](https://doi.org/10.1080/03066150.2010.538584)_\n[10.​1080/​03066​150.​2010.​538584.](https://doi.org/10.1080/03066150.2010.538584)\nSawadgo, W., and A. Plastina. 2022. The Invisible Elephant: Disadoption of Conservation Practices in the United States. Choices.\nQuarter 1.\nSayer, A. 1992. Problems of explanation and the aims of social science.\n_In Method in social science: a realistic approach, 2nd ed., 211–_\n[258. London: Routledge. https://​doi.​org/​10.​4324/​97802​03163​603.](https://doi.org/10.4324/9780203163603)\nSchram, S. G. 2010. Cooperatives: Pathways to Economic, Democratic and Social Development in the Global Economy. The U.S.\nOverseas Cooperative Development Council. The U.S. Overseas\nCooperative Development Council.\nSingh, A., B. MacGowan, M. O’Donnell, B. Overstreet, J. UlrichSchad, M. Dunn, H. Klotz, and L. Prokopy. 2018. The influence\nof demonstration sites and field days on adoption of conservation practices. Journal of Soil and Water Conservation 73 (3):\n[276–283. https://​doi.​org/​10.​2489/​jswc.​73.3.​276.](https://doi.org/10.2489/jswc.73.3.276)\nSoule, M.J., A. Tegene, and K.D. Wiebe. 2000. Land tenure and the\nadoption of conservation practices. American Journal of Agri_[cultural Economics 82 (4): 993–1005. https://​doi.​org/​10.​1111/​](https://doi.org/10.1111/0002-9092.00097)_\n[0002-​9092.​00097.](https://doi.org/10.1111/0002-9092.00097)\nŠūmane, S., I. Kunda, K. Knickel, A. Strauss, T. Tisenkopfs, I. de\nRios, M. Rivera, T. Chebach, and A. Ashkenazy. 2018. Local and\nfarmers’ knowledge matters! How integrating informal and formal\nknowledge enhances sustainable and resilient agriculture. Journal\n_[of Rural Studies 59 (2018): 232–241. https://​doi.​org/​10.​1016/j.​](https://doi.org/10.1016/j.jrurstud.2017.01.020)_\n[jrurs​tud.​2017.​01.​020.](https://doi.org/10.1016/j.jrurstud.2017.01.020)\nTamburini, G., R. Bommarco, T. Wanger, C. Kremen, M. Heijden, M.\nLiebman, and S. Hallin. 2020. Agricultural diversification promotes multiple ecosystem services without compromising yield.\n_[Science Advances 6 (45): 1715. https://​doi.​org/​10.​1126/​sciadv.​](https://doi.org/10.1126/sciadv.aba1715)_\n[aba17​15.](https://doi.org/10.1126/sciadv.aba1715)\nThessen, G., C. Adamson, and B. Parks. 2020. 2020 Iowa Agricultural\n_Statistics. USDA, National Agricultural Statistics Service Upper_\nMidwest Regional Office.\nThomas, J.K., H. Ladewig, and W.A. McIntosh. 1990. The adoption of\nintegrated pest management practices among Texas cotton grow[ers. Rural Sociology 55 (3): 395–410. https://​doi.​org/​10.​1111/j.​](https://doi.org/10.1111/j.1549-0831.1990.tb00690.x)\n[1549-​0831.​1990.​tb006​90.x.](https://doi.org/10.1111/j.1549-0831.1990.tb00690.x)\nTrauger, A. 2009. Social agency and networked spatial relations in\n[sustainable agriculture. Area 41 (2): 117–128. https://​doi.​org/​10.​](https://doi.org/10.1111/j.1475-4762.2008.00866.x)\n[1111/j.​1475-​4762.​2008.​00866.x.](https://doi.org/10.1111/j.1475-4762.2008.00866.x)\nUlrich-Schad, J.D., S. García de Jalón, N. Babin, A. Pape, and L.S.\nProkopy. 2017. Measuring and understanding agricultural producers’ adoption of nutrient best management practices. Journal of\n_[Soil and Water Conservation 72 (5): 506–518. https://​doi.​org/​10.​](https://doi.org/10.2489/jswc.72.5.506)_\n[2489/​jswc.​72.5.​506.](https://doi.org/10.2489/jswc.72.5.506)\nValente, T. 1996. Network models of the diffusion of innovations.\n_Computational and Mathematical Organization Theory 2 (1995):_\n163–164.\n\n## 1 3\n\n\n-----\n\nWarner, K. 2007. Agroecology in action: extending alternative agricul_ture through social networks Food, Health, and the Environment._\nCambridge: MIT Press.\nWellman, B. 1983. Network analysis: Some basic principles. Sociologi_[cal Theory 1: 155–200. https://​doi.​org/​10.​2307/​202050.](https://doi.org/10.2307/202050)_\nWhite, E. 2021. 3 Insights about farmer-to-farmer learning communities from our work on America’s conservation ag movement.\nTrust In Food. [https://​www.​trust​infood.​com/​2021/​06/​14/3-​insig​](https://www.trustinfood.com/2021/06/14/3-insights-about-farmer-to-farmer-learning-communities-from-our-work-on-americas-conservation-ag-movement/)\n[hts-​about-​farmer-​to-​farmer-​learn​ing-​commu​nities-​from-​our-​work-​](https://www.trustinfood.com/2021/06/14/3-insights-about-farmer-to-farmer-learning-communities-from-our-work-on-americas-conservation-ag-movement/)\n[on-​ameri​cas-​conse​rvati​on-​ag-​movem​ent/ Accessed 14 June 2021](https://www.trustinfood.com/2021/06/14/3-insights-about-farmer-to-farmer-learning-communities-from-our-work-on-americas-conservation-ag-movement/)\nWilson, R.S., G. Howard, and E.A. Burnett. 2014. Improving nutrient management practices in agriculture: The role of risk-based\nbeliefs in understanding farmers’ attitudes toward taking additional action. _Water Resources Research 50 (8): 6735–6746._\n[https://​doi.​org/​10.​1002/​2013W​R0152​00.](https://doi.org/10.1002/2013WR015200)\nWypler, J. 2019. Lesbian and queer sustainable farmer networks in the\n[Midwest. Society & Natural Resources 32 (8): 947–964. https://​](https://doi.org/10.1080/08941920.2019.1584834)\n[doi.​org/​10.​1080/​08941​920.​2019.​15848​34.](https://doi.org/10.1080/08941920.2019.1584834)\n\n**Publisher's Note Springer Nature remains neutral with regard to**\njurisdictional claims in published maps and institutional affiliations.\n\n## 1 3\n\n\n**L. Asprooth** is a Ph.D. candidate at the University of California, Davis.\nHer research focuses on the role of grassroots movements and top-down\npolicy mechanisms to support a transition to regenerative agriculture\nin the Upper Midwestern United States. She holds a Master of Public\nAdministration from the Evans School of Public Policy and Governance at the University of Washington\n\n**M. Norton** is a project manager for the Bioeconomy Institute at Iowa\nState University. As part of the regenerative agriculture team, she\noversees transdisciplinary projects related to perennial landcover and\nbiogas production. Formerly, she led the farmer-to farmer speaking and\noutreach program at Practical Farmers of Iowa. She holds a Master of\nScience in rural sociology and sustainable agriculture from ISU.\n\n**R. Galt** is a professor in the Department of Human Ecology and the\nDirector of the Agricultural Sustainability Institute at the University of\nCalifornia, Davis, California. He has conducted research on pesticidedependent smallholder agriculture in Costa Rica, Community Supported Agriculture in the United States and California, and fair and\nsustainable chocolate. He holds M.S. and a Ph.D. degrees in geography\nfrom the University of Wisconsin-Madison.\n\n\n-----\n\n"