AwardsUntapped motivation for agricultural conservation behavior

There are many examples of nonmonetary awards which can serve as proxies for social recognition of good agricultural stewardship and conservation behavior. However, the degree to which these awards motivate implementation and sustained use of conservation practices (such as cover cropping) has not been adequately examined. In this study, we used a serious game approach to explore the effect of nonmonetary conservation awards on participants’ agricultural management decisions in an online experiment. Our results show that study participants were highly motivated to implement cover crops on a year-by-year basis by the fictional Ecobadge award, particularly when award thresholds were set at low levels.There was no difference between participants with prior agricultural experience and those without. Although participants who were not motivated to seek the Ecobadge achieved higher mean financial returns, they also had a wider variation in their financial performance as a group. Those who attained the Ecobadge were less risk-tolerant than those who did not. Achievement of the Ecobadge decayed over several rounds of game play, except among participants who planted cover crops on a high percentage ( 50%) of their land, suggesting these participants possessed high intrinsic motivation. This exploration suggests that nonmonetary awards have high potential to serve as motivational tools to increase adoption of cover crops and potentially other agricultural conservation practices, likely as part of a suite of motivational strategies. We suggest that organizations reconsider how they issue these awards. Better integration of awards with opportunities for peer-to-peer recognition among farmers is a promising approach to expand implementation of conservation practices.


Introduction
Conservation agriculture has been defined as that which minimizes soil disturbance, integrates permanent soil coverage with crop rotations, and facilitates plant species diversity (Hobbs et al., 2008;Food and Agriculture Organization, 2016). Cover crops are often highlighted as an important approach within conservation agriculture, along with reduced tillage, mulching, and crop rotation. As an agricultural management practice, cover crops have both monetary and nonmonetary costs and benefits, both for individual farms and society. Many of these are difficult to quantify and/or assign economic value to, though this does not diminish their importance. For example, cover crops can support soil health (Basche et al., 2016), drought mitigation (Myers and Watts, 2015), carbon sequestration (Lal, 2004(Lal, , 2015, weed control (Lemessa and Wakjira, 2015), pest control (Murrell, 2017), and biodiversity (Altieri, 1991). A subgroup of cover crops, specifically legumes such as hairy vetch and field peas, fix atmospheric nitrogen in the soil. This reduces the need for farmers to purchase fertilizers. Although the cost of growing cover crops does not lead to a significant financial savings over the application of synthetic nitrogen (in regions where synthetic nitrogen is accessible and affordable; Weil and Kremen, 2007;Plastina et al., 2018), the off-farm benefits of reducing synthetic nitrogen use are considerable. These include reduced greenhouse gas (GHG) emissions associated with synthetic nitrogen production and the reduction of nutrient runoff into public waterways (Hepperly et al., 2009). These are the issues of increasing importance as GHG emissions drive climate change and climate change accelerates eutrophication processes in many parts of the world (Sharpley et al., 2013).
In addition to the public goods they deliver, cover crops are associated with multiple on-farm benefits. These include reduced erosion and improved physical, chemical, and biological properties of soil (Fageria et al., 2005). Specifically, grass cover crops can reduce nutrient leaching (by accumulating inorganic nitrogen and holding it in its organic form), legume cover crops can fix atmospheric nitrogen in the soil (through a symbiotic relationship with nitrogen-fixing bacteria), and consistent soil coverage can enhance water infiltration and reduce soil water loss through evaporation (Unger and Vigil, 1998). Additionally, organic matter supplied by decomposing cover crops contributes to the stabilization of soil aggregates, which by extension support soil water holding capacity and aeration (Kaspar and Singer, 2015;see Fageria et al., 2005, andSinger, 2015, for comprehensive summaries of the evidence-based benefits of cover crops). Despite these and other benefits, farmers are sometimes challenged to utilize cover crops due to (a) the need to plant cash crops on available acres and (b) a growing season that is too short to plant both a cash and cover crop (Stivers-Young and Tucker, 1999). This is especially true in regions with short growing seasons, such as the Northeastern United States. Despite the many advantages of cover crops, adoption on the agricultural landscape remains low in the United States.
Efforts to increase the rate of cover crop adoption include incentive programs at the state level, for example, through those in California, Maryland, New York, Pennsylvania, and Vermont. Technical and financial support for cover cropping has also been provided by the U.S. Department of Agriculture (USDA) Natural Resources Conservation Service (NRCS) through Conservation Innovation Grants, the Regional Conservation Partnership Program, and the Environmental Quality Incentives Program (EQIP). Recent analysis of the soil health practices supported by NRCS through EQIP shows that the program funds 10% of the 15 million acres of cover crop annually planted in the United States (Basche et al., 2020). Additionally, many outreach and education programs are designed to increase farmer knowledge and understanding about cover crops, originating from university-based extension programs, nonprofit organizations (e.g., the Rodale Institute), and for-profit consultancies (e.g., the Soil Health Academy). Funding for state and multistate collaborations to expand cover cropping have been awarded by the USDA National Institute for Food and Agriculture through regional Sustainable Agriculture Research and Education programs, in addition to several other program areas (Groff, 2015). In the United States, four cover crop councils (collaborations between research, industry, and agricultural stakeholders that focus in the Midwest, Northeast, Southern, and Western regions of the country) provide information, promote research, convene conferences, and provide decision support tools on cover crops (Seehaver et al., 2019).
Although research and outreach on cover crops has been ongoing for several decades, adoption of this practice is still limited in many major agricultural regions. In the United States, the latest USDA Census of Agriculture shows that only 4% of cropland acres in corn-producing states and only 3% of cropland acres in wheat-producing states have cover crops planted on them (USDA-National Agricultural Statistics Service, 2019). Analysis of the challenges associated with cover crop adoption has often focused on farm-scale barriers and opportunities to overcome those barriers (see, e.g., Arbuckle and Roesch-McNally, 2015;Myers and Watts, 2015). Additional research conducted in the U.S. Corn Belt shows that structural factors (such as the demands of a national or global commodity production system) serve as significant barriers to farmers who may otherwise be inclined to integrate cover crops into intensive agricultural systems (Roesch-McNally et al., 2018). Additionally, the role of social motivation has been understudied in the context of cover crops (and conservation agriculture in general), specifically how a farmers' personal values, and the degree to which they respond to intrinsic or extrinsic motivation leads to different farm management decisions.
To address this gap in our understanding about cover crop adoption and by extension our understanding about adoption of conservation agriculture practices in general, we focus on conservation awards. These are awards given to farmers who exemplify a conservation ethic through their use of cover cropping and other ecologically sustainable practices. Such awards can be monetary or nonmonetary, awarded to a single individual or a collection of individuals. They are often designed to highlight farmers or other land managers who are addressing regional or local conservation challenges, such as soil erosion or water pollution. To explore the influence of conservation awards on planting decisions and cover crop adoption, we employed a serious game approach designed to simulate diversified vegetable farms in the Northeast region of the United States. Specifically, we attempted to enhance our understanding of how nonmonetary awards may serve as a motivation for future cover crop adoption through exploring different thresholds of adoption, the effect of adoption on financial performance, and the persistence of adoption over time. We base our approach on the assumption that nonmonetary awards serve as proxies for social recognition that motivates future behavior (Kosfeld and Neckermann, 2011), a phenomenon studied in nonagricultural contexts but is yet unexplored in the context of cover crop adoption and conservation agriculture.

Background
Conservation awards in U.S. agriculture In addition to the studies that look specifically at the challenges farmers face when implementing cover crops (Arbuckle and Roesch-McNally, 2015;Myers and Watts, 2015), there is a large body of research on what facilitates adoption of conservation practices more broadly. Historically, it was thought that farmers operate in a "profit-maximization" framework when making management decisions (Chouinard et al., 2008). Counter to this, recent research shows that profit is only one of many motivations that are applied to agricultural decision making . Recent systematic reviews of qualitative research show that social norms play a larger role than previously accounted for, as do perceptions of conservation programs  and farmers' views on the balance between production and land stewardship (Raymond et al., 2016). With these assessments in mind, we posit that conservation awards can serve as multifunctional social tools, as rewards for delivering public service, Art. 9(1) page 2 of 17 Schattman et al: Awards: Untapped motivation for agricultural conservation behavior motivators for conservation behavior, vehicles for education and outreach, and fuel for peer-to-peer learning networks (such as those described by Wick et al., 2019). Examples of conservation awards in the agricultural sphere include those led by conservation organizations, industry groups, and states. An illustration of an award issued by a conservation organization is the Leopold Conservation Award by the Sand County Foundation, which includes a monetary gift (US$10,000) and a crystal award to acknowledge "the landowners' achievements among their peers" (Rieber, n.d.). Past awardees include a diversity of land managers, including farmers, ranchers, and foresters. According to the Foundation's website, the award has been issued in partnership with other local and national organizations in 21 states, including those in New England, California, and across the Northern and Southern Central regions. The criteria for the award include responsible management, sustainable revenue, leadership, overall land health, innovation, and outreach. As one farmer states in a promotional video for the Leopold Conservation Award: "In California, the Leopold Award is highly coveted by the people who know about it. It's almost like the Nobel Prize for agriculture" (The Sand Country Foundation, 2016).
Industry-led awards, such as the American Soybean Association's Conservation Legacy Award, also highlight the efforts made by farmers toward conservation goals. The yearly recipient of this award is nominated by conservation and industry professionals who work regularly with the U.S. soybean sector. Past recipients have been chosen based on the efforts they've made toward goals, such as reducing erosion, while staying financially viable. Examples of conservation activities used by awardees include using cover crops, installing water management infrastructure such as gradient terraces and sediment control basins, and the difficult to assess outcome of "leaving things better than they found it" (The American Soybean Association, 2020). State-issued awards such as the Iowa Farm Environmental Leader Awards showcase the "exemplary voluntary efforts of Iowa's farmers as environmental leaders committed to healthy soils and improved water quality" (Iowa Department of Agriculture and Land Stewardship, 2020). Nominees for the award are evaluated based on their use of best management practices (e.g., terraces, grassed waterways, prairie restoration) and their motivation for using the practices. The Iowa program, which began issuing awards in 2012, is in part conceived as a way to address Iowa's significant challenges with agricultural nonpoint source pollution (Schilling et al., 2007, p. 200) and the associated negative impacts on public waterways.
Programs that recognize good agricultural stewardship are sometimes called certification programs rather than awards. For example, in the Chesapeake Bay watershed, the Maryland Association of Soil Conservation Districts (2016) runs a voluntary certification and assessment program designed to recognize the efforts of farmers who meet state nutrient management plan requirements plus additional conservation measures. However, there are two potential differences between traditional awards and farm certification programs. The first, which is relevant to conservation-focused certification programs, is the voluntary audit process by which agricultural practices are verified. Criteria for certification are explicitly stated in these programs. In many traditional award programs, recipients are selected by committees and criteria can be vague or opaque in nature. Second, many certification programs do not exclusively reward ecological stewardship, though this may be one of several programmatic goals. For example, the USDA National Organic Program facilitates a market advantage (i.e., the price premium) gained by certifying agricultural products. These types of market-focused programs certify a product rather than a farm operation, again departing from traditional conservation awards. However, it is sometimes difficult to tell the difference between award and certification programs, as the distinction is often ambiguous.
It has been well established that farmers balance a diversity of motivations when making management decisions and that some of these motivations are nonmonetary in nature (Carlisle, 2016). Compared to financial factors for adopting cover crops (e.g., profitability, incentives), nonmonetary motivators have been relatively underexplored. Awards and social recognition in particular are topics that have not been evaluated for their potential to extend conservation management, despite the wide array of awards issued from a diversity of conservation-and production-focused organizations. To explore the nuance of how social recognition, specifically awards, helps expand and sustain adoption of practice such as cover crops, we draw upon two theories of behavior: the theory of planned behavior (TPB; Ajzen, 1991) and self-determination theory (SDT; Ryan, 1982;Ryan and Deci, 2000).

Conceptual framework
The TPB has been frequently applied to studies of farmers' decision making and behavior. One of the important contributions of this theory that is relevant to this study is its conceptualization of the effect of three types of beliefs: (1) Behavioral beliefs are related to the consequences or outcomes of an action, (2) normative beliefs are related to the social acceptability of a practice, and (3) control beliefs are related to self-efficacy (Ajzen, 1991). These three types of beliefs have been shown in multiple studies to influence farmers' beliefs about the practice as a whole and their intentions to adopt or sustain a practice (Price and Leviston, 2014;Arbuckle and Roesch-McNally, 2015), as well as the degree to which farmers access and utilize external information to make decisions (Hu et al., 2006). Although the theory also suggests that these factors influence actual adoption of agricultural practices, recent research finds that different drivers influence intention to adopt a practice as well as actual adoption (Niles et al., 2016). The theory also states that beliefs are driven, at least in part, by contextual factors such as demographics, the type of farm operation, participation in knowledge networks, market conditions, and so on (Baumgart-Getz et al., 2012).
SDT has been used in the past to describe how different types of behavioral drivers (intrinsic and extrinsic) Schattman et al: Awards: Untapped motivation for agricultural conservation behavior Art. 9(1) page 3 of 17 influence individual decision making and behavior change over time (Ryan and Deci, 2000). The spectrum of behaviors outlined in SDT ranges from non-self-determined behavior (characterized by amotivation, or nonregulation, and lack of control) to self-determined behavior (characterized by intrinsic motivation and regulation, and internal control). Between the two extremes is extrinsic motivation, characterized by a range of regulation types including external regulation (driven by external pressures or desire to attain a reward) and integrated regulation (behavior driven by congruence, awareness, and synthesis with self; Herath, 2010). Different external contexts and stimuli can evoke behavioral responses along this spectrum. In some situations, an individual may exhibit a strong sense of self-determination and intrinsic motivation (e.g., when they are motivated by their personally held beliefs and values), while in others, the same individual may exhibit behavior demonstrating extrinsic motivation (e.g., when they are complying with externally determined rules or regulations or responding to incentives). Whether behavior is regulated by external and integrated factors has implications for behavioral persistence over time (Triste et al., 2018). Specifically, behavior motivated by integrated controls has been shown to be more resistant to decay. This is potentially due to the degree to which behavior governed by internal controls is valued for the experience itself, as opposed to any anticipated outcomes (Ryan and Deci, 2000;Vansteenkiste et al., 2010). Meta-analyses of studies using SDT have shown that tangible and expected awards undermine intrinsic motivation (Deci et al., 1999), though many prior studies neglect the multifaceted nature of behavioral motivation. The framing of an award could stimulate integrated controls by signifying social recognition of an individuals' competence, which has been shown to offset behavioral decay driven by external regulation (Elliot et al., 2000).
Both TPB and SDT offer important insights to those seeking to expand conservation practice adoption among farmers. Agricultural conservation awards specifically may influence farmers' perceptions about the social acceptability of a conservation practice. If the criteria for the award are known ahead of time, it is possible that the opportunity to achieve the award may influence farmers' beliefs around self-efficacy. Additionally, the role of an award as an extrinsic motivator aligned with personal values may serve as a driver of integrated regulation, thereby increasing intrinsic motivation to apply conservation practices over time. However, the alternative may also be true: Behavior motivated by extrinsic factors may decay over time if integrated regulation does not take effect. Considering the limited degree to which conservation awards have been examined in the adoption literature, we must first explore the capacity of awards (proxies for social recognition) to serve as behavioral motivators. In this study, we begin to explore these themes by comparing the planting decisions of individuals who attain a fictional Ecobadge (signifying recognition for ecological stewardship) with those who do not. Specifically, we ask the following questions: 1. Does social recognition (i.e., a nonmonetary award) influence the ratio of cover crops to cash crops planted by game players? 2. What is the relative influence of financial gain and social recognition in game-based decisions? 3. Does the influence of a nonmonetary award decay over time? 4. Do players' self-perceived levels of risk tolerance influence the degree to which they are motivated by extrinsic rewards (e.g., financial gain, social recognition)?

Experimental design
Our experiment was developed to simulate a diversified farming system in the Northeastern United States. Diversified agriculture (specifically in vegetable cropping systems) is more prevalent in the Northeast than in many other regions of the country (Aguilar et al., 2015). In this region and in this agricultural context, using cover crops in diversified vegetable operations presents management challenges along with the many benefits previously cited.
In part due to a relatively short growing season, these challenges include the degree to which cover crops interrupt spring field preparation and fall harvest, as well as the difficulty of incorporating residue (Stivers-Young and Tucker, 1999;Sarrantonio and Gallandt, 2003). Additionally, the economic benefits of cover crops can be difficult to assess. Despite these challenges, cover crops are recognized among farmers as a useful tool, and their use is promoted by agricultural advisors (Grubinger, 1999). Previous studies on conservation behavior have used game-based and immersive platforms to test conservation decision making and pro-environmental behavior (Ahn et al., 2014). The agriculture simulation used within this experiment was built using the Unity Development Platform (Unity Technologies, Version 2017.3.1f1) and deployed online using WebGL (Parisi, 2012). Each participant acted as a farm manager and was provided information in the form of treatments that differed by combination of experimental variables: the risk of flood, the risk of drought, and the amount of cover crops needed to receive an ecological stewardship award, called the Ecobadge. Cover crops were selected as the conservation practice in this experiment because they are well-known because the costs of using them has been welldocumented.
Participants were introduced to the rules of gameplay, including the significance of the Ecobadge, through an introductory slideshow. In this slideshow, the Ecobadge was presented as a visible recognition of ecosystem service generation (i.e., reduced erosion, soil organic matter accumulation, disruption of pest/disease cycles, improved soil water availability, improved nutrient cycling, and longterm soil health). Additionally, information was provided if participants hovered over the Ecobadge icon during the Art. 9(1) page 4 of 17 Schattman et al: Awards: Untapped motivation for agricultural conservation behavior game. With the provided treatment information, participants decided whether or not to invest in management practices and how to split planting between cash and cover crops. Specifically, there was a forced trade-off as participants allocated each acre of their farmland to either cash or cover crop. This design choice was guided by the decision to contextualize the simulation in a Northeastern U.S. climate, where a relatively short growing season precludes planting both cash and full-season cover crops in the same year. Due to difficulties associated with assessing the economic benefits of cover crops, the financial implications for planting cover crops in the game were limited to (a) avoided vegetable planting costs and (b) avoided vegetable crop losses due to droughts or floods. Participants also completed a short survey before and after the main portion of the experiment. The simulation took place over a series of five rounds, with each round simulating 5 years. The first five rounds were framed as practice rounds, which allowed the participants to gain familiarity with the game. To act as a farm manager in the simulation, the participant used the computer mouse or touch screen. The farm was described as having 10 acres of land in an upland zone and 10 acres of land in a riparian zone. After the initial practice round, in which decisions did not affect how much money participants would earn, individuals participated in four rounds, for a total of five rounds. At the beginning of each round, the farm was allotted 150,000 experimental dollars. Each year was divided into two decision periods (spring and fall), with the first year of each round beginning in the fall. In the fall, individuals could choose to install one or more management practices including a riparian buffer (on the riparian zone only) and drip irrigation (on the upland zone, riparian zone, or both). Financial assumptions for each practice are described in the Supplemental Materials (Text S1), including cost to implement and any associated foregone income. Upon selection of any of the three investment options (upland drip irrigation, riparian drip irrigation, and riparian buffer strip), the investment costs and bank amount were updated. Participants were able to change their mind and select or deselect any of the options before advancing. After the fall decision was finalized by the participant, they were informed of their current year investment costs and updated bank amount.
Following the fall decision period, participants split allocation of cover crops and cash crops in the spring decision period. The planting decision was separated by field, with 10 acres of land in the riparian zone and 10 acres of land in the upland zone. Participants had the option to allocate 0-10 acres of cash crops on each of their fields, with any acres not allocated to cash crops devoted to cover crops. Financial assumptions for cash and cover crop planting net income, including gross income and expenses, are detailed in the Supplemental Materials (Text S1). When the participant changed allocation of cash versus cover crops using a slider bar, their planting costs and expected profit (if no drought or flood occurred) were updated. It should be noted that planting a greater proportion of available land in vegetables exposed participants to greater financial risk. This was due to the relative difference in planting costs associated with each type of crop (US$1,020 for 100% cover crop vs. US$69,600 for 100% vegetables). Players who opted to plant a greater proportion of cash crops stood to lose more if they experienced a drought or flood. Conversely, they stood to gain more from successfully harvest acres in the absence of these events.
In addition to information about planting costs and expected profit, a separate panel was included in the spring decision period with an Ecobadge that the participant could obtain in each simulated year. There was no financial reward associated with the Ecobadge. Participants were informed that in order to receive the Ecobadge in any given game year, they must plant a certain threshold of cover crops as a percentage of their total planting allocation. The threshold to receive the Ecobadge was set at one of the following levels, which were held consistent across all rounds for each individual player: 10%, 20%, 30%, 40%, or 50% cover crop. If participants planted cover crop below the threshold, a prominent "X" appeared in front of the Ecobadge icon. After the spring decision was finalized by the participant in each game year, they were informed if flood or drought occurred, the effects of their investments (irrigation or buffer strip), and their net profit for the simulated year. By applying the Ecobadge, we employed informational aspects, which are defined as events that provided participants with information relevant to a particular decision, but which do not penalize the outcomes of that decision (Ryan, 1982). This external award signified competence if it was achieved (i.e., participants achieved a desired level of conservation) but did not signify a lack of competence if not achieved.
A cumulative sum of the total experimental dollars earned was displayed between each simulated year and round. At the end of the five experimental rounds, participants received US$1 for each 400,000 experimental dollars, plus a base pay of US$3. In addition to participant decisions related to planting and practice implementation, earnings were affected by flood and drought occurrence. Flood and drought occurrence was calculated using a pseudorandom number generator. If a flood occurred, riparian zone yield was reduced by 75%. Incurring a drought reduced crop yields by 10%-70%. Participants were informed there was a constant 30% risk of flood and 30% risk of drought, each round.

Deployment
Participants (n ¼ 1,200) were recruited using the online workplace Amazon Mechanical Turk (AMT) and informed that their payout in real dollars would be based on performance during the experiment. AMT has been identified as a representative sample for the U.S. population (Paolacci, 2010) and a viable source of data relative to traditional data collection methods (Buhrmester et al., 2011). In an experimental game looking at animal biosecurity, differences between risk behavior between pork industry professionals and AMT participants were not detectable (Clark et al., 2020), possibly due to a wide variety of risk reduction strategies employed by humans in general. Before the experiment commenced, an informational slideshow was displayed explaining the purpose of the study and mechanics of the simulation. This was followed by a screen allowing the recruit to choose between proceeding to play the game or declining to participate, and an introductory survey followed directly by the experimental gaming simulation and concluded with an exit survey. The exit survey included a question designed to show whether participants had real-world agricultural experience prior to playing the game: 11% of our participants reported they did have agricultural experience in this exit survey. An additional demographic survey of 1,500 AMT recruits conducted post hoc found 164 (10.93%) with prior agricultural experience (see Table 1 for results and Text S1 for survey questions), which shows this trend to be consistent with the AMT population at large, as well as the proportion of the U.S. workforce engaged in agricultural industries (USDA-Economic Research Service, 2020). Institutional review board approval for human subject research was obtained through the University of Vermont (IRB #STUDY00000001). The introductory slides and all three survey instruments are included in the Supplemental Materials (Slides S2 and Text S3).

Award thresholds
Our results show that participants are motivated by social recognition (i.e., a nonmonetary award) as demonstrated by the ratio of cover crops to cash crops planted on a simulated year-to-year basis. This was evident in the clustering of players planting decisions around the Ecobadge threshold, which includes both players who chose to plant cover crops and those that did not in all five treatments ( Figure  1). For example, when the threshold was set at 10%, players were required to plant at least 10% cover crop to attain the Ecobadge. Few players chose to plant more cover crops than the amount required to achieve the Ecobadge, indicating that the award was a strong motivator. Players who chose not to obtain the Ecobadge were still influenced by the presence of the Ecobadge threshold level, indicated by the shifting distribution of planting decisions shown in Figure 2 and the collective foregone income shown in Table 2.
Participants stood to make the most money in all scenarios when they planted 100% cash crops (maximum potential earnings ¼ US$75,050 experimental dollars). Even accounting for simulated years with high drought and flood incidence, this strategy yielded the greatest financial return both in game dollars and actual participant payout (see Table 3). We determined that participants who sought the Ecobadge did so at the expense of financial gain, with no participants who sought the Ecobadge attaining full payout in any threshold group. Because participants were paid in real dollars, achieving the Ecobadge reduced the amount of real financial gain study participants earned. Figure 3 shows the distributions of yearly profit between those participants who attained the Ecobadge and those who did not within each of the five treatments. Table 3 describes the difference in earning between the two groups at each treatment using the Wilcoxon Rank Sum Test, all of which were significant (P < 0.001).

Financial risk versus award attainment
Our analysis shows that participants who did not seek the Ecobadge took greater risks than participants who did obtain the Ecobadge. Participants who planted more vegetables in lieu of attaining the Ecobadge were exposed to greater risk of loss due to drought and floods in the game. This was due to the higher costs Participants could choose both rangeland and/or agricultural management experience; therefore, the total of these options may be greater than the total number of participants who indicated general agricultural experience.
Art. 9(1) page 6 of 17 Schattman et al: Awards: Untapped motivation for agricultural conservation behavior associated with planting vegetables compared to cover crops (i.e., those who planted cover crops had more to lose). This is demonstrated by the significant difference in mean earnings between those participants in simulated years where the Ecobadge was attained versus years that it was not attained (see Figure 3). It should be noted that there were not statistically significant differences in participants' application of riparian buffers or irrigation installation. Although the Ecobadge was a strong motivator, we observed that not all treatment groups achieved a similar proportion of participants on a year-by-year basis who achieved the award. Figure 3 shows that the greatest proportion of participants to achieve the Ecobadge on a yearly basis were found in the 10% threshold group (the group who only had to plant 10% of their land in cover crops to achieve the Ecobadge). The smallest proportion of participants to achieve the Ecobadge on a yearly basis were found in the 50% threshold group (the group who had to plant 50% of their land in cover crops to achieve the award). We also observed decay in player pursuit of the award over the course of the game, meaning that players were more likely to seek the Ecobadge in early simulated years and rounds and less likely to pursue it in later years and rounds. This is described in Figure 4, which shows that fewer participants over time achieved the Ecobadge in all threshold groups. Decay was statistically significant in all threshold treatments with the exception of the 50% threshold group, suggesting that participants who sought the Ecobadge in this treatment were less likely to be motivated by financial incentives than other participants and may have been more motivated by intrinsic factors.

The effect of risk tolerance
Participants' self-perceived risk tolerance was associated with their attainment of the Ecobadge. In the pregame survey, participants were asked to self-identify their level of risk tolerance (see survey questions results summarized  in Table 4). Using a series of Mann-Whitney U tests, we found that those who self-identified as being risk averse (n ¼ 399) were more likely to attain the Ecobadge than the participants who identified as being risk tolerant (n ¼ 307; U ¼ 44,042, P < 0.001). As stated earlier, planting a higher proportion of cash crops represented a more risk tolerant strategy, due to the relatively higher cost of planting vegetables and also the higher potential return on investment. Additionally, participants who self-identified as being risk neutral (n ¼ 153) attainted the Ecobadge significantly less often than those who identified as risk averse (U ¼ 23,223, P < .03). We did not find a significant difference in Ecobadge attainment between those who self-identified as risk neutral and risk tolerant individuals (U ¼ 22,115.5, P ¼ 0.148). It should be noted that different survey methods for assessing risk tolerance among agricultural producers have been shown to yield inconsistent results (Fausti and Gillespie, 2006). This finding likely extends to risk surveys targeted toward general populations as well. However, the alignment between self-reported risk tolerance (in the surveys) and Ecobadge attainment (in gameplay) indicates that comparisons of these two sets of responses are meaningful.
The association between reported award-motivation, risk tolerance, and game play is demonstrated through a closer examination of the players who achieved the Ecobadge at the 30% threshold (see Figure 5). At this threshold level, Ecobadge obtainment is positively correlated with Ecobadge motivation as self-reported in the postgame survey. Those participants who reported that they were highly motivated by the Ecobadge did, in fact, achieve it more often than not. Self-reported Ecobadge attainment was also negatively correlated with financial performance, indicated by the low number of experimental dollars earned by participants who reported that they were highly motivated by the Ecobadge. Additionally, there was a weak but significant positive correlation between self-reported risk aversion and self-reported Ecobadge motivation (also at the 30% threshold level, n ¼ 800), indicating that those who perceived themselves as being more risk averse were more likely to be motivated by the Ecobadge (r ¼ .148; P < 0.001). Similar tests were conducted using data from other threshold treatments, though results were insignificant (likely due to the smaller sample sizes in these treatments).

The effect of agricultural experience
The decisions simulated in this experiment are a simplified representation of real-world management decision made by farmers, effectively capturing trade-offs between ecosystem services and economic viability. We recognize, however, that a large portion of our participant sample was not likely to have prior experience with agricultural management. This is reflective of the U.S. population at large (USDA-ERS, 2020), and the proportion of the national population engaged in agricultural production is closely reflected both in our sample and the population of AMT players (see Table 1). To explore the potential difference between participants in this study with prior agricultural Table 3. Earnings comparison. DOI: https://doi.org/10.1525/elementa.2021.00120.t3 experience and those without, we asked participants in the pregame survey whether or not they had agricultural experience. This allowed us to conduct a series of Kruskal-Wallis H-tests to determine whether the results of players with agricultural experience differed from those without.
Among participants in the 30% threshold group (the largest sample in our experiment), we found that there was no significant difference in yearly Ecobadge attainment between players with agricultural experience and those without (H ¼ 1.631, P ¼ 0.201). In this sample, 77 participants reported having agricultural experience versus 716 participants without. Distributions of planting behaviors were significantly different between groups (H ¼ 14.319, P < 0.001). Using a Mann-Whitney U test, we found those with prior agricultural experience planted significantly more cover crops than vegetables on average (0.34 average cover crop proportion), compared to participants without experience (0.28 average): (U ¼ 20,344, P < 0.001). This may be an indication that participants with prior industry knowledge were more risk averse in their simulated planting behaviors. Although some planting differences occurred between participants with agricultural experience and those without, we did not detect a difference in willingness to acquire the Ecobadge. It should be noted that Ecobadge attainment was binary (the award was either achieved or not based on the different thresholds in each treatment), while planting decisions were continuous. This suggests that Ecobadge attainment has value for both those with agricultural experience and those without and thus merits further research for conservation applications.

Discussion
Several studies on farmers' motivations to adopt conservation practices (Baumgart-Getz et al., 2012; Prokopy et al.,   Ranjan et al., 2019) show that farmer motivations are complex and multifaceted. Evidence suggests that both farmers and agricultural advisors perceive specific management practices as having multifaceted tradeoffs associated with them (Schattman et al., 2017). Awards as proxies for social recognition have thus far not been included in many of these assessments. The results of this experiment highlight three novel findings that demonstrate the potential for better integration and leveraging of awards to motivate conservation practices.
First, when decision makers know about awards and associated achievement criteria ahead of time, the potential to achieve the award can influence behavior. This was demonstrated by the high proportion of study participants that made cover crop planting decisions at or above each Ecobadge threshold. Our results suggest that the criteria for achieving an award influence the number of individuals' who adopt a conservation practice (and to what degree they adopt it), with more people choosing to adopt the practice when the threshold for achieving the award is low. This suggests that awards have the potential to promote wider adoption of conservation practices, should they be more fully integrated into conservation programs. To secure the full benefit of awards at motivating factors, however, changes to how awards are currently made should be considered. For example, rather than recognizing a small number of farmers on a yearly basis based on opaque criteria (as is the case with many awards), all eligible conservation program enrollees who meet specified standards could be issued the award. Clearly communicating what conservation agriculture is (both the practice and how the practice is applied) is necessary. Findlater et al. (2019) show that farmers sometimes believe they are practicing conservation agriculture but tend to overestimate the degree to which they are actually doing so. Additionally, awards could be leveraged to increase additional adoption throughout farmer networks if awardees were called upon to share their successes and challenges. Promoting peer-to-peer learning has been shown to enhance farmer willingness to try new management approaches (Franz, 2007;Franz and Westbrook, 2010;Wick et al., 2019), and our findings suggest that integration of conservation awards could further enhance some of these proven social network approaches. Further research focused on the interplay between social networks and award motivation is therefore needed.
Second, our results suggest that awards may be particularly useful for motivating a segment of farmers who consider themselves to be risk-averse. In our experiment, participants had a high degree of control over their attainment of the Ecobadge but comparatively low control over their financial performance (because of simulated droughts and floods). The relative certainty with which participants could predict successful attainment of the Ecobadge may explain why a large proportion opted for the award, as a means to project control in a situation where other outcomes were uncertain. Here, behavioral control in Ecobadge attainment was strong, whereas control of economic return was weak. Both TPB and SDT suggest that options that represent a higher locus of control, such as attainment of the Ecobadge, are more attractive (Ajzen, 2002;Vansteenkiste et al., 2010). Indeed, a segment of our participants strongly preferred attainment of the Ecobadge over financial reward, perhaps because participants felt more in control over the Ecobadge. Locus of control and perceived behavioral control (Ajzen, 1991) have previously been shown to be important predictors of farmer management decisions (Price and Leviston, 2014;Perry and Davenport, 2020). Despite both long historic and recent applications of this concept in research on decision making in natural resource fields (Huebner and Lipsey, 1981;McNairn and Mitchell, 1992), financial incentives have had much greater attention from scholars as motivators of farmer behavior. What is missing from many of these studies is a comparative assessment of the financial motivation and social motivation, and how each of these motivations shift a farmers' locus of control, and by extension their intentions and their management decisions. It should also be noted that the relatively low real-world payout for our study participants may have increased the salience of the Ecobadge. Repeating our work with larger financial incentives would clarify the relationship between nonmonetary and monetary incentives in this context. Third, we show that, among individuals who are motivated to achieve conservation awards, persistence in conservation behavior may decay over time. This was demonstrated by the proportion of participants who obtained the Ecobadge in early game rounds and then adopted a more financially rewarding strategy in later game rounds. Decay in award attainment may have been driven by a progressive discounting of the value of the award (i.e., participants realized that social recognition was less valuable over time), as predicted by several theories of behavioral change (Kwasnicka et al., 2016). This confirms that decay of behavior over time is likely when behavioral motivation is extrinsic in nature (Deci et al., 1999). There have been few empirical studies that document persistence of conservation management following incentives or awards. The literature that does exist shows that individuals who adopt various practices are either motivated by an extrinsic motivator (e.g., an award) or not and will persist in using the practice after the extrinsic motivator has been removed if practice aligns with their values, financial constraints, and planning horizons (Claassen and Ribaudo, 2016; Dayer et al., 2018).
Our findings suggest that some participants in our study who attained the Ecobadge in early years and rounds but opted to forgo the award in later years and rounds were at first motivated by the award but later were motivated by financial performance. Other participants continued to pursue the award consistently throughout the experiment. The literature suggests that persistent behavior could be explained in several ways. First, it is possible that participants who sought the Ecobadge at high thresholds experienced greater intrinsic motivation than other participants. Ryan and Deci (2000) posit that intrinsic motivation is more persistent than extrinsic motivation, which would therefore make the decision to plant cover crops more resistant to decay. Second, it is possible that participants developed a habit of obtaining the Ecobage. Habits have been found to be a factor in behavioral maintenance in situations where relevant behavioral cues are offered on a regular basis (as was the case in our experimental design) and that these cues can ultimately lead to intrinsic motivations (Verplanken et al., 2008;Kwasnicka et al., 2016). Third, it is possible that the design of our experiment, which gave participants a high degree of autonomy over their choice to achieve the Ecobadge, facilitated integration of conservation goals into individuals' intrinsic motivations. Prior research shows that individuals who make autonomous decisions are more likely to report higher rates of internalized proenvironmental goals (Osbaldiston and Sheldon, 2003) and consistently demonstrate pro-environmental behavior over time (Villacorta et al., 2016). This does not diminish the role that social networks, iterative or one-time feedback, and behavioral reinforcement play in decision making (Biel and Thøgersen, 2007;Grilli and Curtis, 2019). Considering the range of possibilities, further assessment in the persistence of conservation behavior over time (what Frey and Rogers, 2014, call persistence pathways) and how awards do or do not moderate persistence is needed.
Finally, we note that participants in lab experiments likely behave differently than actors in real-world settings. As previously stated, it was necessary to simplify simulated decisions in order to detect behavior differences between participants. Most notably, participants in this experiment contended with fewer risks than actual farmers face and did not accumulate the multiple, long-term and contextspecific benefits associated with conservation agriculture. Levitt and List (2007) additionally identify several factors that lead people to act differently in the real world than they do in experiments, including (1) ethical considerations, (2) the experience of being observed by others, (3) the decision-making context, (4) self-selection of participants, and (5) what is at stake. Still, our results show that the influence of the Ecobadge on participant behavior in this experimental context was clear, justifying future fieldbased experiments to assess the strength of this approach in a real-world context. By utilizing both experimental and field-based approaches to understand award influence, we can better assess both its internal and external validity in conservation decision making (Roe and Just, 2009).

Conclusion
Nonmonetary awards, which can serve as proxies for social recognition, have the potential to serve as extrinsic motivators of good agricultural stewardship and conservation behavior. In this experiment, we used a serious game approach to explore the degree to which awards motivate implementation and sustained use of cover cropping. We show that experiment participants were highly motivated by a fictional Ecobadge award, particularly when award thresholds were set at low levels. Many participants sought and obtained the Ecobadge at the expense of financial gain, both in experimental dollars and real participant stipends. When participants chose not to obtain the Ecobadge, they achieved higher mean financial returns and accepted a greater level of risk, while those who attained the Ecobadge demonstrated (and self-reported) lower risk tolerance. Attainment of the Ecobadge decayed over the course of the experiment for some players, with the exception of those players who sought the Ecobadge at the highest rate (50%), suggesting these participants possessed intrinsic motivation to plant cover crops. This exploration suggests that nonmonetary awards have high potential to serve as motivational tools to increase adoption of cover crops and potentially other agricultural Art. 9(1) page 12 of 17 Schattman et al: Awards: Untapped motivation for agricultural conservation behavior conservation practices. However, it should be noted that real-world application of awards and/or conservation certification programs vary widely. Future research that parses programs based on factors such as level of transparency in selection criteria, application of award (for social recognition vs. market advantage), and the function of awards within social networks would inform best practices for leveraging awards in conservation programing. Although this study utilized cover crops as its experimental context, we anticipate that our findings are applicable to other conservation agriculture practices or systems of stacked practices. Our results suggest that awards may be better integrated into conservation programs both to boost adoption of conservation practices and to specifically target those farmers who consider themselves to be risk averse. To achieve this, award criteria should be transparent and well publicized, and awards should be given to all qualifying farmers. Additionally, the opportunity for award recipients to share their expertise within peer networks could enhance the impact of award programs. Further research is needed on the persistence of conservation practices after the award is given, and specifically on pathways of persistence, or the relationship between extrinsic motivators and more persistent, intrinsic motivators. Despite these gaps in our understanding, this exploration shows that nonmonetary awards have high potential to serve as motivational tools to increase adoption of cover crops and potentially other agricultural conservation practices as well.

Data accessibility statement
Our complete survey instrument and game description is available as Supplementary Material to this article. Deidentified raw data will be shared upon request, pending IRB approval. .

Supplemental files
The supplemental files for this article can be found as follows: Text S1. Financial assumptions for the experiment; introductory slides shown to experiment participants; pregame survey, postgame survey, and post hoc survey instruments.