1. Introduction
Foodborne disease causes 600 million illnesses and 420,000 deaths every year worldwide [
1,
2,
3]. In the U.S., about 48 million individuals suffer from food-related illnesses [
4], and nearly half (46%) of these are connected to outbreaks involving fruits, vegetables, and nuts, often called produce, which cause 23% of deaths in the U.S. [
5,
6]. Since the 1980s, the rise in fresh produce consumption has led to an increased likelihood of produce-related foodborne illnesses and outbreaks, with 190 outbreaks from 1973 to 1997 causing 16,058 illnesses and 8 deaths, and 972 outbreaks from 1998 to 2013 causing 34,674 illnesses and 72 deaths, indicating a growing trend [
7,
8,
9]. These types of foodborne illness significantly impact the economy of the U.S. and the overall agricultural sector [
10]. To tackle food safety-related issues, the U.S. government, industry, and others invest considerable resources. For instance, the U.S. Congress alone allocated more than USD 2.2 billion for federal food safety programs in 2014 [
11]. All these events and efforts show the criticality of food safety in the United States (U.S.).
As stipulated in the Food Safety Modernization Act (FSMA), the Food and Drug Administration (FDA) formulated the Produce Rule to regulate food safety among farm operations that grow and sell fruits and vegetables primarily meant for direct consumption in their raw state. However, the Produce Rule does not apply to farms that have average annual produce sales of USD 25,000 or more (as of 2011) in the past three years [
12]. Apart from that, farms are eligible for qualified exemptions from the Produce Rule if three years of annual food sales of such an enterprise amount to less than USD 500,000 (as of 2011) and sell the majority (>50%) of their produce straight to end-users located in-state or within 275 miles of the farm [
13,
14]. Approximately 34,000 farms whose three-year average annual produce sales fall below USD 25,000 may not be covered and roughly 76,000 farms with three-year annual average food sales below USD 500,000 could qualify for exemption from the Produce Rule [
15,
16]. Adoption of on-farm food safety practices (including agricultural water safety practices, the focus of this study) to ensure the safety of produce is completely voluntary for farms that are either not covered or qualified as exempt from the Produce Rule. This raises a concern about produce safety among these exempt growers, who are likely to prevail as distributors of fresh fruits and vegetables.
Scholars have both expressed their support considering the cost burden related to Produce Rule implementation and demonstrated their concern over such policy exemptions, like the increased threat to produce safety, especially among small-scale growers. For instance, regulation-related expenses are of particular concern to small producers because growers are afraid that such regulations might force them out of business [
17]. Other researchers indicated that the FSMA Produce Rule has the potential to cause small farms to go out of business, result in food shortages, hinder food innovation, and adversely affect the environment and consumer health because of the associated cost of complying with the Produce Rule [
18]. There are also chances of a decrease in the supply of fresh produce, which mainly comes from small-scale growers, and a reduction in the number of such suppliers could hinder the effectiveness of programs such as Farm to School, Special Supplemental Nutrition Program, and the USDA’s Farmers Market Promotion Program [
19]. Some researchers emphasized that small-scale producers were exempted from the produce rule without a proper scientific risk assessment, which created big gaps in food safety [
20]. For instance, about half of the farms growing fruits and vegetables on between 10 to 100 acres of land were operating without a food safety plan [
15]. In this context, where adherence to the Produce Rule is voluntary, it is important to investigate ways to encourage small growers to adopt on-farm food safety practices to remain competitive in the market, minimize the risk of produce contamination, and prevent any possible recalls or fines.
Estimations revealed that the average cost of implementing the proposed Produce Rule for a small domestic farm is USD 4697 per year [
21], which could exceed their profit. This amount could be far more if indirect costs were considered [
19]. The additional cost incurred while complying with Produce Rule standards will relatively increase the unit price of agricultural produce from small farms compared to larger farms [
19], which may disadvantage small farms that are not eligible for exemption [
17] and hamper their price competitiveness and ability to supply healthy or local foods [
19]. Excluding specific small-scale farms from FSMA regulations might impede their ability to market their produce to larger retail establishments because of a lack of certification or assurance that the produce is grown in compliance with the Produce Rule [
22]. Other researchers stated their curiosity about the consequences of exempting small-scale growers who sell their produce through intermediate markets and are most likely profitable [
23,
24,
25].
A recent national study showed that exempt farms had the highest direct sales compared to very small, small, and medium/large farms and relatively lesser compliance with several food safety practices, such as water testing, sanitation tools, and equipment, toilets, and facilities for handwashing, disposal of trash or sewage appropriately, among others [
12]. Several other studies also highlighted the inadequate engagement of small-scale growers in agricultural water safety practices that help to minimize the risk of produce contamination [
12,
26,
27,
28]. Researchers found a lack of engagement in on-farm food safety practices among growers who were knowledgeable about safety practices in the U.S. [
29,
30]. Researchers have also reported that the cost burden and lack of knowledge associated with the Produce Rule are the primary barriers to implementing safety practices [
31]. Studies also reported low intention to engage in on-farm food safety practices among growers [
32,
33]. Inadequate compliance increases concern among consumers about the safety of produce when the demand for local produce as well as sales from direct-to-consumer and farmer markets are increasing [
34,
35].
Most importantly, farm operations that operate through direct-to-consumer options (e.g., farmers’ markets, community-supported agriculture, roadside stands, U-pick farms, and online direct-to-consumer platforms) and sell their produce locally with their own food safety mechanisms are exempt [
22]. A study conducted in Georgia, South Carolina, and Virginia by [
27] found that over 42% of the farmers’ markets had no food safety requirements in place, which simply showcased the vulnerability of local food supply chains. Similarly, another study conducted in California also reported that most of the farmers’ market managers did not ask about vendors’ on-farm food safety practices [
36].
Since small farms often sell their produce directly to consumers in the local market, an important question is whether or not the proposed regulations should apply to them. With the scale of farm operations, the impact and scale of any outbreak caused varies and these differences should be reflected in the rule. For instance, the scale of an outbreak caused by a small family farm that only sells their produce through a specific farmers’ market or roadside stand could be limited to a certain area compared to other farms that sell their produce to wholesalers or national retailers’ chains, which may cause outbreaks in multi-state or multiple locations. Produce sold through small-scale farms is still able to cause foodborne illness at the community level and cause illness and fatalities. For instance, in Oregon, USA, strawberries that were sold at the roadside and farmers’ market caused an outbreak linked to
E. coli O157:H7, which led to one death and 15 falling sick [
37]. So, researchers suggested investigating food safety threats in relation to the scale of farm operations [
38] and emphasized the voluntary adoption of the on-farm Produce Rule [
12]. Considering this inevitable risk associated with produce, it is important to design food safety policies and agricultural extension programs focused on exempted growers, which could promote their compliance with on-farm food safety practices.
Past studies have focused on understanding the role of extension, subjective and objective knowledge, intention, attitude, perceptions, and barriers to the adoption of on-farm food safety practices [
29,
30,
32,
33,
36,
39,
40,
41,
42,
43,
44]. However, how growers perceive on-farm food safety practices and whether such perceived traits of on-farm food safety practices significantly determine the likelihood of adoption has not been researched. It is highly compelling to explore growers’ perception as the driving force behind behavior change pertaining to on-farm food safety practices compliance, unlike mandatory regulation. Therefore, our study focused on understanding the adoption of on-farm food safety practices pertaining to agricultural water. Although there are many types of agricultural safety practices, agricultural water-related safety practices were selected as past studies revealed that agricultural water safety practices are the least understood and least engaged with among agricultural producers [
45,
46]. The specific objectives of the study were the following:
Assess the current engagement in agricultural water safety practices among small farms.
Quantify perceptions of small farms about agricultural water safety practices.
Examine the relationship between the characteristics of agricultural water safety practices and the likelihood of adoption among small farms.
2. Theoretical Framework
The theory of the Diffusion of Innovations (DOI) [
47] has been widely used in social science to understand the adoption of innovation and spread. Specifically in agriculture, the DOI has been applied in various contexts, including assessing the knowledge and implementation of improved agricultural practices, understanding growers’ perception of agricultural innovations, and assessing the impact of growers’ perception on the rate of adoption. Some of the applications of the DOI include exploration of agricultural producers’ perception of agricultural best management practices [
48], System of Rice Intensification (SRI) [
49], sustainable agricultural practices [
50], integrated cover cropping [
51] and water- and input-saving technology [
52] to examine the influence such perception has on the adoption of respective technology. Apart from that, the DOI was used to assess the differences between adopters and non-adopters of tilapia aquaculture in the Solomon Islands in terms of innovation attributes [
53]. Similarly, researchers in Canada also utilized the DOI to understand the uptake of precision agriculture technology among agricultural producers [
54]. Likewise, in Vietnam, the DOI was used to explore the compatibility of alternative wetting and drying (AWD) irrigation with local farming systems [
55]. The DOI was used to study the adoption of Bt cotton and integrated pest management in India [
56,
57]. Other researchers have used the DOI to explore the adoption of agricultural technologies by farmers [
58,
59], propose strategies to promote the adoption of agroforestry [
60] uncover factors influencing the adoption of organic agriculture [
61], among others.
Despite its strength and high applicability, the DOI has not been used to explore the adoption of on-farm food safety practices. Our study used the DOI to understand growers’ perception and its influence on the likelihood of adoption of on-farm food safety practices pertaining to agricultural water. On-farm food safety practices are not just a single innovation; rather, they are a cluster of practices similar to a cluster of innovations, as described by [
47]. We conceptualized on-farm food safety practices as preventative innovation [
47] for this study because such safety practices are performed to prevent the risk of unwanted future consequences like outbreaks or recalls. According to Rogers [
47], “an idea, practice, or project that is perceived as new by an individual or other unit of adoption” (p. 12) within a social system is known as an innovation. With ever-advancing food safety science, it is not uncommon that these safety practices may be new to agricultural producers or farm workers. In our context, the social system comprises individuals or networks of growers involved in the production of fruits and vegetables that aim to market their produce. Adoption of an innovation is defined as a “decision of full use of an innovation as the best course of action available, and rejection is a decision “not to adopt an innovation” [
47] (p. 177). The speed of adoption is reliant upon the perceived characteristics of the innovation (like on-farm food safety practices), the type of innovation decision, the channels of communication, the characteristics of the social system, and the degree of promotional efforts by change agents [
47]. Our study focuses on examining the perceived characteristics of the on-farm food safety practices pertaining to agricultural water and their influence of growers’ intention to adopt safety practices.
In regard to the adoption of innovations, there may be uncertainty surrounding an innovation’s outcome, and such uncertainty is regarded as a crucial obstacle to the adoption process [
47]. Extension plays an important role in communicating agricultural innovations or technologies to growers and reducing the associated uncertainties through two-way dialogue, field demonstration, or similar other approaches. The innovation decision process includes stages through which a produce grower progresses over time to form an opinion and decide whether or not to incorporate the new idea or practice into ongoing activities. The five main stages in an innovation adoption process are knowledge, persuasion, decision, implementation, and confirmation (see
Figure 1).
At the knowledge stage, the grower becomes aware of the existence of agricultural water safety practices and learns about them. Actively seeking information about an innovation is likely to help progression in the innovation adoption process [
47]. However, it is important to understand that individuals tend to attend to ideas or information that align well with their existing attitudes and beliefs (i.e., selective perception) [
47]. Since growers who are exempted from the Produce Rule are not required to comply, adoption is up to their interests and attitudes. Based on the concept of selective perception, we can assume that growers having positive attitudes and beliefs toward on-farm food safety practices will mostly intend to learn more about such food safety practices.
At the persuasion stage, produce growers develop a favorable or unfavorable perception or attitude toward an innovation, i.e., on-farm food safety practice. The process is affective (or a feeling), and at this stage, a grower is engaged more psychologically. After becoming aware of innovation at the knowledge stage, in pursuit of learning more, the grower assesses the reliability of the information, analyzes the information they receive, may conduct vicarious trials (thinking hypothetically with a projection into the future like thinking how implementing on-farm food safety practices would improve their current production or may be increase their production cost) and forward planning in the persuasion stage [
47]. A potential adopter judges an innovation based on the characteristics of the innovation; thus, the perception of an individual about an innovation is the result of five attributes, which include the relative advantage, compatibility, complexity, trialability, and observability [
47]. Based on past diffusion studies, about 49–87% of the change in the adoption rate of innovation is explained by these five attributes or characteristics [
47]. In our study, we contextualized these five attributes for agricultural water safety practices.
2.1. Relative Advantage
Relative advantage is the perception that an individual finds a new practice superior to an existing one through both tangible and non-tangible benefits [
47,
62]. Relative advantages might include higher yields and higher prices (e.g., in the case of adoption of rice intensification among Indonesian rice farmers; [
49]), reduced inputs, time savings, and environmental benefits (e.g., in the case of adoption of agricultural best management practices among American agricultural producers; [
48]). On the other hand, low or invisible relative advantages, such as a lower profit margin and elevated cost inputs, can impede adoption (e.g., of cover crops among agricultural producers in the U.S. [
51,
63]. When the adoption of an innovation is made mandatory, then such regulation becomes an incentive for an individual to adopt innovation [
47]. In the context of on-farm food safety practices, a relative advantage for a grower could be an opportunity to market their produce by helping them secure competitive prices, or improve the safety of their produce to reduce unwanted consequences like recall, or ensure social prestige among other fellow growers and consumers. In the meantime, the relative advantages of on-farm food safety practices may include a reduction the chances of undesirable consequences in the future [
47].
2.2. Compatibility
Compatibility is the extent to which a new idea or practice is perceived as matching with existing values, needs, past experiences, and current practices [
47,
62]. Studies from various countries, including Vietnam, the U.S., Canada, Malaysia, and India, consistently showed that the perceived compatibility of agricultural practices with existing farm systems significantly influences their adoption, with positive perceptions driving the uptake of practices like alternate wetting and drying irrigation technology, agricultural best management practices, fertilizer best management practices, sustainable agricultural practices like conservation tillage, Bt cotton, integrated pest management practices, and precision agriculture [
48,
50,
54,
55,
56,
57,
64]. Conversely, perceived incompatibility, such as issues with moisture availability and crop rotation, as found in context of cover cropping [
51], can hinder adoption. In the context of on-farm food safety practices, adoption will increase as an innovation is perceived to be compatible with a grower’s needs, values, beliefs, farming practices, and lifestyle [
47,
63].
2.3. Complexity
Complexity is the extent to which a new idea or practice is comparatively perceived as difficult to comprehend and use [
47,
62]. If an innovation has greater complexity, it might become an obstacle to adoption [
63]. Researchers identified that complexity in understanding and implementing standards like GlobalGAP hindered their adoption by Canadian growers, while factors such as the need for self-prepared compost and understanding compost making impeded the adoption of the System of Rice Intensification among Indonesian growers [
49,
65]. Complexity, identified as a significant predictor, negatively influenced the adoption of various innovations, such as landscape water conservation among U.S. residents, water and input-saving technology in the Jordan Valley, and Integrated Pest Management among agricultural producers in California and India [
52,
57,
66,
67]. In contrast, [
56] reported that a perceived lack of complexity was one of the reasons for the growing adoption of Bt (
Bacillus thuringiensis) cotton in India. In our context, if the on-farm food safety practices are perceived as simple to understand and easy to execute in the field, then their adoption is more likely.
2.4. Trialability
The amount to which an innovation can be experimented or tested, and possibly reversed or adopted, in stages is known as trialability [
47,
61]. Many researchers have validated the relationship of trialability with the rate of adoption of an innovation. For example, researchers found trialability to be a significant predictor of the adoption of biogas in Pakistan [
68] and the adoption of autonomous vehicles in China [
69]. In contrast, researchers in the U.S. revealed that low trialability, which was related to the complexity of experimentation and inflexible insurance policies of crops, hindered the adoption of cover crops [
51]. In the context of our study, if growers perceive on-farm food safety practices as triable or something that they can try on a temporary basis before implementing permanently, it will help with adoption.
2.5. Observability
Observability, according to [
47], is the extent to which the adoption of an innovation is noticeable to others and has been validated by many scholars in an agricultural context. For instance, researchers in the U.S. found that observability significantly determined adoption of agricultural best management practices among producers [
48]. Other researchers in the U.S. reported that the reluctance to integrate cover crops into farmers’ current agricultural management systems stemmed from the perceived lack of visible results, despite the beneficial impacts below ground [
51]. A study conducted in the Solomon Islands reported that the low observability of tilapia aquaculture and its benefits could have hindered adoption [
53]. When the results of the adoption of innovation are distant or not readily noticeable (e.g., preventative innovation), then it may retard the rate of adoption. For instance, if growers receive an opportunity to observe the results of the adoption of on-farm food safety practices in their community or peer network, it could have positive results in terms of their perception and reinforce the adoption decision. However, if growers do not see any noticeable differences in produce grown following on-farm food safety practices, then that might inhibit adoption.
Utilizing the theory of the DOI, specially the innovation diffusion processes, our study aims to quantify growers’ perceptions and estimate the extent to which those perceptions influence their likelihood of adopting on-farm food safety practices pertaining to agricultural water.
4. Data Analysis
To perform the data-cleaning and analysis activities, IBM SPSS Statistics (version 27.1; IBM Corp., Armonk, NY, USA) was used. As part of the data-cleaning process before the actual data analysis, screening was conducted as described above and incomplete or missing responses were identified and excluded. Next, we performed data transformations to convert the raw data into an appropriate format suitable for analysis (e.g., coding ordinal and continuous variables with a standard scale as appropriate and reverse coding negative statements, etc.). The indices were calculated by averaging the scores of each respective item used to measure those constructs and could range from −2 to 2, with 2 indicating the highest possible relative advantage, compatibility, observability, and trialability and lowest possible complexity (i.e., most supportive of behavior change).
We performed descriptive analyses (means, standard deviations, percentages, and frequencies) for objectives one and two because our aim was to understand the specific characteristics of the population sample [
72]. For objective three, multiple linear regression was used to determine the relationship between the predictor variables (relative advantage, compatibility, complexity, observability, and trialability) and the outcome variable (likelihood of adoption). Regression analyses are useful for understanding the relationship between the independent and dependent variables [
82]. Before running the multiple regression model, the major assumptions, such as the linearity of predictors, normality of residuals, homoscedasticity, and multicollinearity and singularity [
83], were tested.
Linearity was confirmed through scatter plots (see
Figure A1,
Figure A2,
Figure A3,
Figure A4,
Figure A5 and
Figure A6) and normality was confirmed by creating a histogram of the residuals (see
Figure A7) [
84]. The standard regression residuals were approximately normally distributed in the population [
83]. In addition, the Kolmogorov–Smirnov test [
85,
86] was conducted and it was found to be significant (
p = 0.200), meaning the standard regression residuals follow the normal distribution. Homoscedasticity is upheld when the variance of the errors is persistent across all the levels of the predictor variables [
87]. The assumption of homoscedasticity was checked by plotting the residuals against the predicted values in a scatter plot [
83]. Through visual inspection, we confirmed that as we moved toward higher values on the
X-axis, there was no specific pattern found and neither point consistently narrowed nor widened as it progressed toward the
X-axis. Also, we did not find any distinct pattern such as a funnel-shape. In addition, the Breusch–Pagan test was conducted and did not find evidence of heteroscedasticity (
p = 0.119), suggesting that the variance of the errors is likely constant (homoscedastic) across the predictors. Multicollinearity and singularity exist when two or more independent variables have a high correlation or low correlation, respectively [
88]. We tested the multicollinearity and singularity by using the Pearson correlation because our predictor variables were continuous. The results showed that the correlation coefficient did not exceed the threshold (
r = 0.9) for multicollinearity or singularity among the independent variables. In addition, the variance inflation factor values were less than 10 for our data, thus proving the non-existence of multicollinearity [
86]. After completion of the assumption testing, the predictors (relative advantage, compatibility, complexity, observability, and trialability) were regressed with the likelihood of adoption using a multiple linear regression model.
8. Recommendations for Policymakers and Practitioners
The development of favorable growers’ perceptions toward on-farm food safety practices, including agricultural water safety practices, is crucial in translating knowledge into action. Based on our findings, to bridge the knowledge and action gap concerning the adoption of agricultural water safety practices, stakeholders should focus on developing policies and education that enhance the perceived compatibility, relative advantage, and observability of on-farm food safety practices among exempt growers. On a broader level, these findings should be taken into consideration by policymakers in various regions as they consider whether or how to offer exemptions from these types of regulations. If perceptions among the affected audience are positive, policymakers may be able to expect voluntary compliance. However, if they are negative, exemptions might need to be reconsidered or policies might need to make provisions to improve perceptions.
Extension professionals and other educators should focus on the segment of growers who are not consistently performing on-farm food safety practices. It is suggested to educate growers about the importance of each on-farm food safety practice and their role in combating the risk of produce contamination as well as to create awareness that only the consistent application of such on-farm food safety measures will ensure the safety of produce. However, considering the current exemption of small-scale produce growers, it could be a challenging task to ensure full engagement with on-farm food safety practices. With the help of audience research, extension professionals must explore strategies or areas to tap into to encourage growers and increase the uptake of on-farm food safety practices voluntarily. To motivate the growers toward full compliance, initiatives like offering incentives, facilitating certification of agricultural commodities, or rewarding farms through promotion of their business through county extension social media could be beneficial. Policies should contain provisions to recognize exempt or small-scale growers for their voluntary compliance with the Produce Rule or similar produce safety standards.
9. Recommendations for Researchers
At the global level, food safety practices are unregulated in numerous countries, and it is important in such contexts to understand the human dimension of food safety innovations to overcome the challenges associated with, and leverage the existing opportunities for diffusion and adoption of, such practices. There is an opportunity to conduct a study aimed at understanding growers’ knowledge, engagement, and perception in different domains of on-farm food safety practices, which will provide a broader understanding of where growers are highly engaged and where they lack adequate compliance. Further, such a study could be used to understand their knowledge and perception of food safety regulations like the Produce Rule and how they impact their voluntary adoption of the on-farm food safety practices. Additionally, the diffusion of innovation could be used to understand non-exempt growers’ perceptions to understand how they perceive the adoption of food safety practices, how their perceptions compare to exempt growers, and even to assess their knowledge and implementation under mandatory adoption condition.
It is recommended to investigate factors that impede growers’ consistent engagement in on-farm food safety practices, which would provide important insights for designing future policies and interventions to promote the adoption of safety measures among small-scale growers. Also, it is important to explore why there is comparatively less engagement for post-harvest water safety practices compared to production water-related safety practices. Recognizing this gap in adoption, researchers could investigate how adopters differ from non-adopters, which will provide insight into understanding the factors underlying the adoption and non-adoption of on-farm food safety practices among growers. To dive deeply into such phenomena and rooted intricacies, researchers could utilize a qualitative research approach to explore growers’ responses in this regard.