Assessing Produce Growers’ Perceptions and Adoption of Agricultural Water Safety Practices in the United States

Abstract

The Produce Rule regulates food safety among produce growers in the United States. Policy exemptions (e.g., for farms with average produce sales under USD 30,000 over three years) pose a threat to local food safety, particularly given exempt growers’ role in fresh produce production. Understanding exempt growers’ perceptions of food safety practices and the factors influencing their voluntary adoption is crucial. A cross-sectional quantitative survey was conducted to assess growers’ current engagement, their perceptions of agricultural water safety practices, and their influence on adoption decisions. The findings indicate inadequate engagement with agricultural water safety practices, with production water-related safety practices showing higher engagement compared to post-harvest water practices. Overall, growers reported favorable perceptions toward agricultural water safety practices. The perceived compatibility, relative advantage, and observability predicted adoption decisions among exempt growers, with the perceived compatibility being the strongest predictor. Policymakers and education professionals should design strategies and tailored educational interventions and messaging for exempt growers to emphasize the compatibility, relative advantage, and observability of agricultural water safety practices. Furthermore, it is recommended to explore policies and other mechanisms to increase the perceived relative advantage (immediacy of rewards and direct benefits) for exempt growers to promote voluntary adoption of these on-farm food safety practices.

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.

3. Materials and Methods

3.1. Research Design

A non-experimental and cross-sectional quantitative survey research design was used for this study [70]. The non-experimental approach was appropriate because there was no need for manipulation of the predictor variables [71] and we sought to identify population characteristics and explore their relationships [72]. Further, measurement of the predictor and response variables at a particular point of time made it a cross-sectional study [73]. The quantitative survey research design was deemed appropriate to achieve the study’s research objectives because it “provides a quantitative or numeric description of trends, attitudes, or opinions of a population by studying a sample of that population” [70] (p. 49). Prior to the start of the data collection, the study was approved by the University of Florida Institutional Review Board (IRB) as exempt (Protocol # 18155).

3.2. Target Population and Sampling Procedures

Small-scale fruit and vegetable farms in the U.S. were the target population. Due to the lack of a sampling frame, a nonprobability sampling was deemed appropriate [74,75]. Although there was an estimate provided by the FDA about not covered and qualified exempt farms [16], we found that there was no specific database of farms that are exempt from the Produce Rule. Therefore, we were not able to calculate the representative sample size for the population, and for this reason, probabilistic sampling was determined to be inappropriate for our study. A multi-sampling approach consisting of two distinct non-probabilistic sampling techniques, convenience sampling and an opt-in online panel design, was used to recruit the participants. Initially, convenience sampling was adopted to recruit participants, primarily based on ease of accessibility and the availability and willingness of participants [75,76]. To conduct convenience sampling, we utilized our connections, including extension and food safety experts working across the nation, to distribute surveys among fruit and vegetable growers through their networks, like listservs and newsletters, among others. We connected to agricultural producers’ associations (Florida Fruit and Vegetable Association, Red Hills Small Farm Alliance, Florida Organic Growers, Southwest Florida Small Farmers Network, and Feeding Florida) and extension professionals for support in survey distribution to fruit and vegetable growers in their networks. In addition, we sent emails asking for voluntary participation using publicly available contact lists of agricultural producers from government websites and agriculture producers’ associations. Two follow-up emails were sent, each after seven days, as a reminder to participate. We also visited a farmers’ market in North Carolina and asked agricultural producers to take part in the survey. The first phase of our data collection using convenience sampling started on 23 May 2023 and continued till 23 July 2023.

Due to the low numbers of responses obtained through convenience sampling, we added a second recruitment phase consisting of an opt-in online panel design to recruit growers during early July 2023. For the opt-in online panel, Centiment, a sampling firm based in Denver, Colorado, USA, was contracted to recruit an additional 74 participants who met the study’s criteria. Participants who were recruited by Centiment and completed the survey received incentives. The second phase of data collection using an opt-in online panel design began on 5 July and continued till 14 July 2023.

3.3. Sample Size

Because our study aimed to understand voluntary compliance with agricultural water safety practices among small-scale farms, we adopted the Produce Rule criteria in an attempt to identify farms that were either not covered by the rule or likely to be exempt. According to [77], if for three consecutive years, the annual sale of produce is less than USD 25,000 on average, then such farms are not covered by the Produce Rule. Taking inflation into consideration, we calculated the current (2023) value of USD 25,000 (2011), which was USD 33,380, and screened participants based on their farm’s annual sales for the last three years, with options of USD 33,830 or less, and USD 33,831 or more. Only participants who checked USD 33,830 or less were allowed to take the full survey. Based on this study’s dual sampling method, we received 76 responses from convenience sampling and 74 responses from the opt-in online panel. After excluding incomplete responses, we had 105 complete responses. Among those complete responses, 83 participants met our screening criteria (average annual income in the last three years equal to USD 33,830 or less), which were used for formal analysis.

Respondents reported residing in 27 states, where the highest number of respondents were from California (15.66%), followed by Florida (12.05%), Iowa (12.05%), Kentucky (7.23%), Texas (6.02%), and others (Figure 2). About 42% of respondents had graduated from some college, followed by a bachelor’s degree (27.7%), high school (19.3%), and so on (

Table A1. Sociodemographic characteristics of the participants.

Characteristics	n	%
Gender
Male	39	46.99
Female	38	45.78
Non-binary	4	4.82
Race/Ethnicity
American Indian or Alaska Native	10	12.05
Black or African-American	12	14.46
Asian or Pacific Islander	6	7.23
White	55	66.27
Hispanic/Latino(a)/Chicano(a)	11	13.25
Other	1	1.20
Education
Less than high school	1	1.20
High school/GED	16	19.28
Some college	35	42.17
Bachelor’s degree	23	27.71
Graduate or professional degree (i.e., MS, Ph.D., JD, MD)	8	9.64
Family income in 2022
USD 24,999 or less	7	8.43
USD 25,000 to USD 49,999	28	33.73
USD 50,000 to USD 74,999	21	25.30
USD 75,000 to USD 99,999	11	13.25
USD 100,000 or more	12	14.46
I do not wish to answer	4	4.82
Years in farming
0–5 years	29	34.94
6–10 years	22	26.51
11–15 years	13	15.66
16–20 years	4	4.82
More than 20 years	15	18.07
Acreage of land with fruit and vegetable cultivation
1–10	44	53.01
11–20	10	12.05
21–30	6	7.23
31–40	6	7.23
41–50	3	3.61
51–60	3	3.61
61–70	1	1.20
71–80	1	1.20
91–100+	5	6.02
More than 100	4	4.82

Note. N = 83. Participants were, on average, 43.51 years old (SD = 17.22).

). A relatively large segment of participants (33.7%) had a family income between USD 25,000 and USD 49,999, and 25.3% (n = 21) of respondents had an income range between USD 50,000 and USD $74,999. Regarding the gender of respondents, 46.9% (n = 39) were male and 45.8% (n = 38) were female. The average age of the surveyed growers was 43.51 years (SD = 17.22).

3.4. Measures

The measures included current engagement, likelihood of adoption, and perceptions about on-farm food safety practices. Current engagement in agricultural water safety practices was assessed to understand how frequently growers engage with safety practices related to agricultural water. The research team adapted seven agricultural water safety practices (five pre-harvest and two post-harvest) from the “Produce Rule” [77] and “Guide to Minimize Microbial Food Safety Hazards in Fresh Fruits and Vegetables” [78] (see

Table 1. Measures of current engagement in agricultural water safety practices.

On-Farm Food Safety Practices Pertaining to Agricultural Water

Source

Properly maintaining the water distribution system (e.g., evaluation of unusual activities, minimizing contamination risks, i.e., runoff, animal intrusion, maintenance of wellhead, etc.). Minimizing contact between irrigation water and the harvestable portion of crops. Using only safe water sources (e.g., treated groundwater and municipal water). Increasing the time between the last irrigation water use and harvest, and treating unsafe pre-harvest water. Monitoring recirculated post-harvest water for pathogens. Using disinfectants to ensure the safety of recirculated post-harvest water.

Adapted from [77,78]

).

The perceptions of the growers about agricultural water safety practices were evaluated using five traits of the innovation proposed by [47]. These include the relative advantage, compatibility, observability, trialability, and complexity. Items that were used to measure each of these five traits were adapted from [47,64,67]. Similarly, the researchers developed three items that measure the intent of the growers to adopt agricultural water safety practices.

3.5. Instrumentation

To ensure the survey’s credibility, instruments were developed by employing the best practices in survey design, which include professional design and layout (consistent spacing, placing the instruction where needed, etc.), pretesting of the survey instrument, clear privacy assurances, multiple contact attempts, and clear and concise instructions (e.g., conveying the affiliated institution/university identity) integrated throughout the design process to encourage survey responses and ensure data quality [79]. Additionally, the survey length was kept short, clear and unambiguous language was used, personal questions were limited, the survey was formatted for both mobile and laptop devices, confidentiality and security of information were maintained, participation of growers and their opinion were highlighted as important, and appreciation for participation was expressed [79].

The survey questionnaire began with the informed consent statement, which provided the study and IRB details. Next, there were screening and quality control questions to examine whether the respondent was eligible for the study and was committed to the quality of their responses. To confirm whether the respondents were engaged in the decision-making regarding on-farm food safety practices, we asked them to select their role or position on the farm from among the given five options: Owner of a farm, Operator/Manager of the farm, Other Employee who provides input or makes decisions regarding on-farm food safety practices, An employee who does not provide input or make decisions regarding on-farm food safety practices, and Other. Respondents who did not agree to participate in the study, did not commit to providing quality responses, indicated they did not provide input or make decisions regarding on-farm food safety practices, did not live in the U.S., or who were less than 18 years of age (according to their year of birth) were directed to the end of the survey and not included in the study.

To examine current engagement, respondents were asked to rate how frequently they performed the seven different on-farm agricultural water safety practices on a five-point Likert-type scale: 1 = Never, 2 = Sometimes, 3 = About half the time, 4 = Most of the time, and 5 = Always. The likelihood of adoption of these seven agricultural water safety practices was measured using a five-point Likert-type scale: −2 = Extremely Unlikely, −1 = Somewhat Unlikely, 0 = Neither Unlikely nor Likely, 1 = Somewhat Likely, and 2 = Extremely Likely. Growers’ perceptions of the agricultural water safety practices were captured using the five traits of innovations (relative advantage, compatibility, complexity, observability, and trialability). The responses were assessed by employing a five-point Likert scale, where participants reported their degree of agreement along the following sequence of value statements: Strongly Disagree (−2), Disagree (−1), Neither Agree nor Disagree (0), Agree (1), and Strongly Agree (2). The relative advantage, compatibility, complexity, observability, and trialability were captured using a series of four, five, six, four, and four statements, respectively (see

Table 2. Major constructs and internal consistency measurement.

Question Stem and Individual Items	Cronbach’s α a
Relative advantage Index	0.820
Please indicate your level of agreement or disagreement with the following statements: Using agricultural (production and post-harvest) water safety practices…
could help me in getting a higher price for produce.
would improve upon how I previously grew produce.
improves the safety of produce.
helps me avoid recall or penalties.
Compatibility Index	0.929
Again, please indicate your level of agreement or disagreement with the following statements: Using agricultural (production and post-harvest) water safety practices…
is compatible with the way I grow produce.
fits well with my lifestyle.
would fit with my customers’ expectations.
fits with my needs of producing quality, safe produce.
is a way for me to be socially responsible.
Complexity Index	0.786
Again, please indicate your level of agreement or disagreement with the following statements: Using agricultural (production and post-harvest) water safety practices…
is easy for me.
is difficult for me to perform b.
is straightforward.
is complicated b.
takes a lot of time for me to get right b.
is simple for me to understand.
Observability Index	0.851
Again, please indicate your level of agreement or disagreement with the following statements:
I have noticed others having good results using agricultural and post-harvest water safety practices.
I have seen high-quality produce grown by others using agricultural and post-harvest water safety practices.
The results of using agricultural and post-harvest water safety practices are apparent to me.
I have seen customers prefer to buy from growers who follow agricultural and post-harvest water safety practices.
Trialability Index	0.868
Again, please indicate your level of agreement or disagreement with the following statements: Using agricultural (production and post-harvest) water safety practices…
is something I can try before I make a decision about doing so permanently.
is something I can test before I commit to changing my routine.
is something I can do on a trial basis.
is something I can experiment with temporarily.
Likelihood of Adoption Index	0.880
Next, we would like to ask about on-farm practices you may use in the future. Please indicate your likelihood of following on-farm food safety practices:
Properly maintaining the water distribution system (e.g., evaluation of unusual activities, minimizing contamination risks, i.e., runoff, animal intrusion, maintenance of wellhead, etc.).
Minimizing contact between irrigation water and the harvestable portion of crops.
Using only safe water sources (e.g., treated groundwater and municipal water).
Increasing the time between the last irrigation water use and harvest.
Treating unsafe pre-harvest water.
Monitoring recirculated post-harvest water for pathogens.
Using disinfectants to ensure the safety of recirculated post-harvest water.

Note. All items except the likelihood of adoption were measured using a five-point Likert scale where −2 = Strongly Disagree, −1 Somewhat Disagree, 0 = Neither Agree nor Disagree, 1 = Somewhat Agree, and 2 = Strongly Agree. Likelihood of adoption items were measured using a five-point Likert-type scale where −2 = Extremely unlikely, −1 = Somewhat unlikely, 0 = Neither likely nor unlikely, 1 = Somewhat likely, and 2 = Extremely likely. ^a post-hoc reliability reported. ^b indicates reverse-coded responses, which were reverted prior to data analysis.

).

3.6. Measurement Reliability and Validity

Internal consistency was ensured with the repetition of items on the same topic in the questionnaire [80]. Measurement reliability was established using the internal consistency method using Cronbach’s alpha. Items that reduced the reliability below the acceptable threshold value were removed, and after that, all the major constructs exceeded the minimum threshold of 0.7 (see Table 2), ensuring the internal consistency of the items [81].

We confirmed the face validity (a form of content validity) of the survey instrument through the expert panel review. A panel of six experts from the College of Agriculture and Life Sciences and the Institute of Food and Agricultural Sciences Extension at the University of Florida reviewed the survey instrument. The experts recruited for the review were social scientists, extension professionals, and food science and safety experts. Based on the comments and feedback received, the questionnaire was revised. In this way, the face validity of the survey instrument helped to ensure that the questions were relevant to the participants, which increased the chance of completion of the questionnaire by the participants [80].

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.

5. Results

5.1. Objective One: To Assess the Current Engagement in Agricultural Water Safety Practices among Small Farms

Growers’ mean engagement in agricultural water safety practices ranged from 3.07 (SD = 1.35) to 3.92 (SD = 1.67), meaning that on average, the frequency of engagement lies between about half of the time and most of the time (

Table 3. Agricultural water safety practices and current engagement among agricultural producers.

On-Farm Food Safety Practices	Response Scale					M	SD
	Never	Sometimes	About Half the Time	Most of the Time	Always
	% (n)	% (n)	% (n)	% (n)	% (n)
Properly maintaining the water distribution system (e.g., evaluation of unusual activities, minimizing contamination risks, i.e., runoff, animal intrusion, maintenance of wellhead, etc.)	8.43 (7)	12.05 (10)	8.43 (7)	21.69 (18)	49.40 (41)	3.92	1.35
Minimizing contact between irrigation water and the harvestable portion of crops	9.64 (8)	12.05 (10)	12.05 (10)	22.89 (19)	43.37 (36)	3.78	1.37
Using only safe water sources (e.g., treated ground water and municipal water)	20.48 (17)	12.05 (10)	8.43 (7)	14.46 (12)	44.58 (37)	3.51	1.63
Increasing the time between the last irrigation water use and harvest	15.66 (13)	16.87 (14)	16.87 (14)	28.92 (24)	21.69 (18)	3.24	1.38
Treating unsafe pre-harvest water	25.30 (21)	13.25 (11)	12.05 (10)	19.28 (16)	30.12 (25)	3.16	1.60
Monitoring recirculated post-harvest water for pathogens	25.30 (21)	13.25 (11)	13.25 (11)	21.69 (18)	26.51 (22)	3.11	1.56
Using disinfectants to ensure the safety of recirculated post-harvest water	32.53 (27)	7.23 (6)	9.64 (8)	21.69 (18)	28.92 (24)	3.07	1.67

Note. Responses were measured using a five-point Likert-type scale, where 1 = Never, 2 = Sometimes, 3 = About half of the time, 4 = Most of the time, and 5 = Always. M stands for the mean value or average score, which is the measure of central tendency. SD stands for standard deviation, which is the measure of dispersion indicating the extent to which a data point deviates from the mean value.

). The most and least frequently engaged agricultural water safety practices were properly maintaining the water distribution system (M = 3.07, SD = 1.35) and using disinfectants to ensure the safety of recirculated post-harvest waters (M = 3.92, SD = 1.67), respectively.

When examining specific practices, 32.53% (n = 27) of growers never utilized disinfectants for ensuring the safety of recirculated post-harvest waters, 25.30% (n = 21) never implemented monitoring for pathogens in such waters or the treatment of unsafe pre-harvest water, 25.30% (n = 21) never treated unsafe pre-harvest water, 20.48% (n = 17) never engaged in the exclusive use of safe water sources like treated groundwater and municipal water, 15.66% (n = 13) never engaged in increasing the time between the last irrigation water use and harvest, 9.64% (n = 8) never minimized contact between irrigation water and the harvestable portion of crops, and 8.43% (n = 7) never engaged in properly maintaining their water distribution system. Conversely, nearly half of the growers (49.40%) always adhered to the proper maintenance of their water distribution system, 44.58% (n = 37) always utilized only safe water sources, 43.37% (n = 36) always minimized contact between irrigation water and the harvestable portion of crops, 30.12% (n = 25) always treated unsafe pre-harvest water, 28.92% (n = 24) always employed disinfectants to ensure the safety of recirculated post-harvest waters, 26.51% (n = 22) always monitored recirculated post-harvest waters for pathogens, and 21.69% (n = 18) always increased the time interval between the last irrigation water use and harvest.

Out of the surveyed growers, only two indicated they never implemented any of the seven agricultural water safety practices, while two others implemented all seven practices most of the time, and an additional eight always followed these practices. The rest of the participants showed variability in the frequency of engagement across different on-farm food safety practices, as shown in Table 3.

5.2. Objective Two: Quantify Perceptions of the Small Farms about Agricultural Water Safety Practices

All the composite mean scores of the perceived traits (relative advantage, compatibility, complexity, observability and trialability) of agricultural water safety practices were positive (see

Table 4. Descriptive analysis of the major constructs and the items constructing them.

	Response Scale					M	SD
Construct and Measures	SD	SwD	NA nor DA	SwA	SA
	% (n)	% (n)	% (n)	% (n)	% (n)
Relative advantage index						0.53	0.96
Please indicate your level of agreement or disagreement with the following statements: Using agricultural (production and post-harvest) water safety practices…
…could help me in getting a higher price for produce.	14.46 (12)	12.05 (10)	30.12 (25)	26.51 (22)	16.87 (14)	0.19	1.27
…would improve upon how I previously grew produce.	7.23 (6)	20.48 (17)	20.48 (17)	30.12 (25)	21.69 (18)	0.39	1.24
…improves the safety of produce.	3.61 (3)	6.02 (5)	25.30 (21)	27.71 (23)	37.35 (31)	0.89	1.09
helps me avoid recall or penalties.	4.82 (4)	10.84 (9)	26.51 (22)	30.12 (25)	27.71 (23)	0.65	1.14
Compatibility index						0.81	1.03
Again, please indicate your level of agreement or disagreement with the following statements: Using agricultural (production and post-harvest) water safety practices…
…is compatible with the way I grow produce.	4.82 (4)	9.64 (8)	21.69 (18)	28.92 (24)	34.94 (29)	0.80	1.17
…fits well with my lifestyle.	7.23 (6)	7.23 (6)	22.89 (19)	31.33 (26)	31.33 (26)	0.72	1.19
…would fit with my customers’ expectations.	3.61 (3)	4.82 (4)	27.71 (23)	31.33 (26)	32.53 (27)	0.84	1.05
…fits with my needs of producing quality, safe produce.	6.02 (5)	8.43 (7)	15.66 (13)	26.51 (22)	43.37 (36)	0.93	1.22
…is a way for me to be socially responsible.	6.02 (5)	8.43 (7)	22.89 (19)	30.12 (25)	32.53 (27)	0.75	1.18
Complexity index						0.21	0.87
Again, please indicate your level of agreement or disagreement with the following statements: Using agricultural (production and post-harvest) water safety practices…
…is easy for me.	12.05 (10)	10.84 (9)	31.33 (26)	24.10 (20)	21.69 (18)	0.33	1.27
…is difficult for me to perform b.	15.66 (13)	22.89 (19)	25.30 (21)	22.89 (19)	13.25 (11)	0.05	1.28
…is straightforward.	6.02 (5)	15.66 (13)	27.71 (23)	21.69 (18)	28.92 (24)	0.52	1.23
…is complicated b.	19.28 (16)	16.87 (14)	26.51 (22)	22.89 (19)	14.46 (12)	0.04	1.33
…takes a lot of time for me to get right b.	8.43 (7)	21.69 (18)	25.30 (21)	32.53 (27)	12.05 (10)	−0.18	1.16
…is simple for me to understand.	7.23 (6)	15.66 (13)	19.28 (16)	31.33 (26)	26.51 (22)	0.54	1.24
Observability index						0.36	1.01
Again, please indicate your level of agreement or disagreement with the following statements:
I have noticed others having good results using agricultural and post-harvest water safety practices.	7.23 (6)	14.46 (12)	36.14 (30)	22.89 (19)	19.28 (16)	0.33	1.16
I have seen high-quality produce grown by others using agricultural and post-harvest water safety practices.	7.23 (6)	9.64 (8)	32.53 (27)	28.92 (24)	21.69 (18)	0.48	1.15
The results of using agricultural and post-harvest water safety practices are apparent to me.	9.64 (8)	10.84 (9)	30.12 (25)	28.92 (24)	20.48 (17)	0.40	1.21
I have seen customers prefer to buy from growers who follow agricultural and post-harvest water safety practices.	13.25 (11)	13.25 (11)	32.53 (27)	16.87 (14)	24.10 (20)	0.25	1.32
Trialability index						0.65	0.95
Again, please indicate your level of agreement or disagreement with the following statements: Using agricultural (production and post-harvest) water safety practices…
…is something I can try before I make a decision about doing so permanently.	4.82 (4)	13.25 (11)	16.87 (14)	37.35 (31)	27.71 (23)	0.70	1.16
…is something I can test before I commit to changing my routine.	4.82 (4)	7.23 (6)	24.10 (20)	38.55 (32)	25.30 (21)	0.72	1.07
…is something I can do on a trial basis.	3.61 (3)	12.05 (10)	20.48 (17)	39.76 (33)	24.10 (20)	0.69	1.08
…is something I can experiment with temporarily.	4.82 (4)	18.07 (15)	20.48 (17)	36.14 (30)	20.48 (17)	0.49	1.15

Note. SD = Strongly Disagree, SwD = Somewhat Disagree, NA nor DA = Neither Agree nor Disagree, SwA = Somewhat Agree, and SA = Strongly Agree. ^b indicates reverse-coded responses, which were reverted prior to data analysis. M stands for the mean value or average score, which is the measure of central tendency. SD stands for standard deviation, which is the measure of dispersion indicating the extent to which a data point deviates from the mean value.

). Among all the perceived traits, growers showed relatively higher agreement on compatibility (M = 0.81, SD = 1.03), trialability (M = 0.65, SD = 0.95), and relative advantage (M = 0.53, SD = 0.96) when compared to observability (M = 0.36, SD = 1.01) and complexity (M = 0.21, SD = 0.87). We looked at each of these perceived characteristics of agricultural water safety individually and the items constructing them. The average score of four items that were used to capture perceptions about the relative advantage of agricultural water safety practices were positive (see Table 4). On average, growers reported somewhat agreement that safety practices help improve the safety of the produce and help avoid recall or penalties. However, the average scores for getting a higher price or opportunity to improve the way they used to grow were close to neither agree nor disagree (see Table 4). Likewise, the average scores for all five items that were used to capture compatibility were positive and close to somewhat agreement. It can be further stated that on average, growers reported approaching somewhat agreement that agricultural water safety practices are compatible with their production methods, align with their lifestyle, meet customer expectations, support produce quality and safety, and demonstrate social responsibility (see Table 4).

Similarly, the average score of the six items that were used to measure the complexity trait of agricultural water safety practices were positive except takes a lot of time for me to get right (see Table 4). For observability, the average scores of all four items were positive but were close to neither agree nor disagree (see Table 4). Finally, the average scores of all four items that were used to capture perceptions about the trialability of agricultural water safety practices were positive. Among those four items, I can try before I make a decision about doing so permanently, I can test before I commit to changing my routine and I can do on a trial basis had average scores close to somewhat agreement (see Table 4).

5.3. Objective Three: Examine the Relationship between Characteristics of Agricultural Water Safety Practices and the Likelihood of Adoption among Small Farms

The overall model was significant, F (5, 77) = 25.55, p < 0.001, and predicted 59.9% (adjusted R square) of the variance in the likelihood of adoption of on-farm agricultural water safety practices when all the other variables were held constant. Among all five independent variables, compatibility (B = 0.413, t (77) = 3,79, p = < 0.001), relative advantage (B = 0.272, t (77) = 2.630, p = 0.10), and observability (B = 0.245, t (77) = 2.47, p = 0.16) were significant at p ≤ 0.001, p = 0.05, and p = 0.05, respectively. The standardized coefficients or partial regression coefficients (β) for the significant predictors ranged from 0.243 to 0.417 (see

Table 5. Multiple linear regression analysis: likelihood of adoption and characteristics of an innovation.

	B	SE	95% CI		β	t	p
			LL	UL
Constant ***	3.159	0.095				33.422	<0.001
Relative Advantage (X1) **	0.272	0.103	0.07	0.48	0.256	2.630	0.010
Compatibility (X2) ***	0.413	0.109	0.20	0.63	0.417	3.794	<0.001
Complexity (X3)	−0.162	0.104	−0.37	0.05	−0.139	−1.557	0.124
Observability (X4) **	0.245	0.099	0.05	0.44	0.243	2.471	0.016
Trialability (X5)	0.114	0.095	−0.08	0.30	0.106	1.203	0.233

Note. *** refers to significant at p ≤ 0.001 and ** means significant at p = 0.05. Reported B, SE, β are unstandardized regression coefficients, standard error, and standardized regression coefficients. R² = 0.624, R² adjusted = 0.599, F (5, 77) = 25.55, p < 0.001. CI = confidence interval. LL = lower limit, UL = upper limit.

). A one-unit increase in the relative advantage, compatibility, and observability increases the likelihood of adoption of on-farm food safety practices related to agricultural water by 0.272, 0.413, and 0.245 units, respectively, when the other variables were held constant. Among the significant independent variables, compatibility had the highest partial regression coefficient (β = 0.417). Complexity and trialability were not significant.

6. Discussion

6.1. Growers’ Engagement

Growers’ engagement in agricultural water safety practices revealed room for improvement in achieving full engagement among surveyed growers. Our finding is similar to past reports of agricultural producers’ inadequate engagement in preventative measures like the use of safe water, addition of sanitizer to wash water, and water testing, among others [26,29,30,89]. In addition, growers’ engagement in post-harvest water safety practices was relatively less compared to production water safety practices. Although a separate study is needed to understand the exact barriers to full compliance and discrepancies in engagement across agricultural water safety practices, some inferences can be drawn from past studies. For example, the variability in growers’ engagement across different agricultural water safety practices could be linked to a lack of recognition that contamination can occur at the farm level [29] or a lack of awareness of how distinct on-farm food safety practices minimize the contamination risk [30]. This finding emphasizes the importance of how-to knowledge and principle knowledge, in addition to awareness knowledge, for the adoption of a practice, as stated by [47]. All of these possibilities demonstrate the strong need for on-farm food safety educational programs and training related for small-scale growers, which should be given priority by the federal, regional, and local food safety policies and outreach programs, similar to what was suggested by other scholars [15,27]. Finally, the majority of surveyed growers performed such safety practices ranging from sometimes to most of the time, implying that engaging in these behaviors is not a binary decision (e.g., complete adoption or non-adoption), but rather, it happens in a continuum that could range from never to always, especially for behaviors that must be repeated over and over.

6.2. Growers’ Perceptions and Significant Predictors of Adoption Intention

The average scores for relative advantage, compatibility, complexity (which is reverse coded, thus measuring simplicity), observability and trialability of agricultural water safety practices were positive. Growers exhibited comparatively higher levels of agreement on the compatibility of agricultural water safety practices and less agreement on the simplicity of these practices. In addition, compatibility (compatible with their values, goals, needs, and lifestyle), relative advantage (how advantageous are safety practices over their current practices), and observability (extent to which results of adoption are noticeable) were significant predictors of the likelihood of adoption of agricultural water safety practices, all with a positive relationship. This finding supports the theoretical assumptions that the rate of adoption of agricultural water safety practices increases with increases in the perceived compatibility, relative advantage, and observability of those safety measures [47]. Our findings also align with [48], who showed that relative advantage, compatibility, and observability were the most important characteristics of agricultural best management practices determining adoption among agricultural producers in the U.S. Complexity and trialability were not significant, meaning that a grower’s perceptions about the complexity or simplicity of agricultural water safety practices and the extent to which these practices can be experimented do not influence their likelihood of adoption when other factors were considered. However, other researchers have found that complexity was significant regarding the adoption of integrated pest management practices [50] and water- and input-saving technologies [52]. Similarly, low trialability, which was related to the complexity of experimentation, was one of the reasons behind the inadequate adoption of cover crops [51]. Whether these influences significantly predict adoption often depends on the nature of a practice/behavior and growers’ perceptions. Food safety policies and outreach programs should be built on these significant perceived characteristics of on-farm food safety practices, which drive growers’ intention to adopt.

6.3. Compatibility: The Strongest Predictor of the Likelihood of Adoption

Among the significant predictors, compatibility had the greatest strength in predicting the likelihood of adoption of on-farm agricultural water safety practices, implying that the likelihood of adoption of agricultural water safety practices is greatest when there is a strong realization among growers that agricultural water safety practices fit with their values and beliefs, needs, and farming practices. Past studies have shown that small-scale farms need to make more changes to meet the Produce Rule standards than large-scale farms and they spend more dollars per acre on food safety practices [90]. Applied to our context, it is challenging to encourage exempt growers to voluntarily adopt produce rules or similar good agricultural practices unless they fit well, cost less, and require minimum changes to their current farming operations. A similar problem existed with the diffusion of integrated pest management among small agricultural producers in the U.S. due to costs [47]. The perceived incompatibility of cover crops with current agriculture management system such as moisture availability and current crop rotation system among U.S. growers was one of the factors negatively influencing the adoption of cover crops [51]. Similarly, ref. [57] reported the adoption of resistant transgenic crop varieties was hindered by incompatibility with local tastes, leading farmers to continue growing traditional varieties. Past studies revealed that rather than the scale of the farm operation, the market channel they use to sell produce influences growers’ decision to adopt food safety practices [91]. Due to increasing consumer demand for safe food, small-scale growers’ need to produce safe and quality fruits and vegetables might have increased, leading them to adopt on-farm food safety practices to meet demand and remain competitive in the market. Rather than just offering an exemption from food safety rules, policymakers need to find ways to address the concerns about food safety among these exempt or qualified exempt growers. Unless food safety policies and outreach programs are designed to enhance perceived compatibility, the adoption on-farm food safety practices will be less likely among exempt growers, which can compromise food safety at a local level.

6.4. Relative Advantage: The Second Strongest Predictor of the Likelihood of Adoption

Our finding that relative advantage significantly predicted the adoption of agricultural water safety practices is consistent with [47]. Other scholars also revealed relative advantages as a main trait driving the adoption of conservation tillage, organic fertilizer/compost, and crop rotation among Malaysian vegetable farmers [15]. In the meantime, a perceived lack of profitability and increased expenses limited the adoption of cover crops in the U.S. [51]. Relative advantage is often regarded as the most crucial determinant of innovation success [92]. However, in the context of the adoption of recommended on-farm food safety practices or the Produce Rule, past studies raised undeniable concerns about the implementation challenges among small-scale growers, mostly because of the associated costs [17,19,42]. For innovations like on-farm food safety practices, which are preventative in nature and often adopted in current times to lower the probability of some unwanted future consequences, like foodborne illness outbreak, recall, or penalties, the immediacy of the reward is reduced, which impedes the rate of adoption [93]. For instance, non-adopters of the System of Rice Intensification in Indonesia reported that the time gap between the application of the SRI method and the return was a reason to discontinue their adoption decision [49]. The desired direct or indirect benefits are temporally distant and so the relative advantage of on-farm food safety practices may have a delayed reward. Additional challenges are introduced pertaining to relative advantage because the results of the adoption of preventative innovations occur at an undetermined point in the future and there exists the possibility that they may not manifest at all [47,93].

More than half of growers reported some degree of agreement that the adoption of agricultural water safety practices improves produce safety, helps avoid recall or penalties, and improves upon how produce was previously grown (see Table 4). However, only 43.38% of growers reported some degree of agreement that the adoption of agricultural safety practices can help secure a higher price. Extension professionals can use such information to tailor their educational programs to frame messages that help convince growers about the potential benefits of adoption of safety practices. Similarly, the diffusion of innovation increases when individuals have the anticipation of more benefits from the adoption of certain new practices than what they currently do [94]. If the adoption of on-farm food safety practices could be tied to direct or indirect benefits such as a premium price or prestige/reputation of farm, then it would help perceive the adoption of such practices as more advantageous. In addition, for preventative innovations like on-farm food safety practices, some sort of mandatory provisions like minimum produce safety standards or local food safety rules could possibly play a role as an incentive for growers. Therefore, policymakers and extension and other educators should establish mechanisms to increase the perceived relative advantage for the adoption of on-farm food safety practices, especially among small-scale growers.

6.5. Observability: The Third Significant Predictor of the Likelihood of Adoption

Observability was significant and had a positive relationship with the likelihood of adoption, meaning that with increasing perception of observability, there is an increasing tendency to adopt agricultural water safety practices. Our findings corroborate the results of other studies, such as the readily visible benefits of BMPs (observability) being one of the significant predictors of adoption among U.S. agricultural producers [48]. In contrast, [53] reported that the lack of observability of tilapia aquaculture’s results, such as economic benefits, led people to leave their aquaculture occupation for other economically viable activities in the rural Solomon Islands. Growers need to see other growers using an innovation and realize that it is beneficial and safe to use [47]. However, in the case of preventative innovations like on-farm food safety practices, whose effects are not immediately received or readily noticeable, educators need to help them realize or observe how the adoption of these preventative innovations can help them in the long run because these kinds of innovations, which are less observable, have a slower rate of adoption.

7. Implications

There is limited understanding of how food safety innovations diffuse among growers, possibly in part due to the mandatory nature regulated by governments or retail and wholesale buyers. However, this is not always the case, because many countries around the world do not have such mandatory food safety regulations in place. Food safety policy should not only dictate mandatory food safety regulations but also speak explicitly about how it plans to promote voluntary adoption among stakeholders across the food chain. In the context of the U.S., the Produce Rule requires growers to adhere to on-farm food safety practices while producing fruits, vegetables and edible nuts that are often eaten raw. However, the Produce Rule offers exemptions based on certain criteria, which have left a number of farms, especially small-scale farms, unregulated by the Produce Rule standards. We strongly believe that whether food safety practices are mandatory or voluntary in nature, it is critical to understand the human dimension of such practices among growers to strengthen the diffusion and adoption of food safety practices. Understanding the key predictors and causal relationships in growers’ adoption of on-farm food safety practices could be valuable for effective policy development. In this context, our study adds to the existing body of literature on understanding the adoption of on-farm food safety practices among small-scale growers who are likely to be exempt.

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.

10. Limitations

Despite our best efforts (e.g., expert panel review of the survey instrument, clarity in questions, etc.), measurement errors could have existed because of individual interpretations of the instrument (e.g., Likert-type scales), inaccurate or incorrect participant responses, and social desirability biases where participants respond in such a way that their responses match with what society expects or sets as the right thing to do [71,79,95,96]. Because sample statistics differ from population parameters, a sampling error could exist [82,97]. Coverage errors occurred because growers out of internet reach were not sampled [79,82]. Since the sample is not representative of the U.S. small growers’ population, the results are not generalizable beyond the study participants [82]. Finally, measurement of the perceived characteristics of on-farm food safety practices’ cross-sectionally at one point in time provides only a partial picture of the relationship between such characteristics and the rate of adoption of on-farm food safety practices [47].